Skip to main content

Full text of "Encyclopedia of Caves"

See other formats



ie oe » Edited by 



Alwertets Uigreeroey 


The Bhangra Hee Olererrsiey 


Amutenicey Pouter Alrtain Meng oxox ina: Kart Chola! 
fin Saw Ove Sau Greiner Liner perr Eyal ebpa 

The chapter entitled Nitrate Contamination in Karst Groundwater by Brian G. Katz is in the public domain and doesn’t carry 
the Elsevier copyright. 

Alexander Klimchouk owns the copyright for his three chapters entitled Gypsum Caves, Krubera (Voronja) Cave, and 
Ukranian Giant Gypsum Caves. Copyright © 2005 by Alexander Klimchouk. 

Elsevier Academic Press 

200 Wheeler Road, 6th Floor, Burlington, MA 01803, USA 

525 B Street, Suite 1900, San Diego, California 92101-4495, USA 
84 Theobald’s Road, London WC1X 8RR, UK 

This book is printed on acid-free paper. 

Copyright © 2005, Elsevier Inc. All rights reserved. 

No part of this publication may be reproduced or transmitted in any form or by any 
means, electronic or mechanical, including photocopy, recording, or any information 
storage and retrieval system, without permission in writing from the publisher. 

Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: 

(+44) 1865 843830, fax: (+44) 1865 853333, e-mail: You may also complete your request 
on-line via the Elsevier homepage (, by selecting “Customer Support” and then “Obtaining Permissions.” 

Library of Congress Cataloging-in-Publication Data 
Application submitted 

British Library Cataloguing in Publication Data 
A catalogue record for this book is available from the British Library 

ISBN: 0-12-198651-9 

For all information on all Elsevier Academic Press publications 
visit our Web site at 

Contents by Subject Area ix 
Contributors xi 

Guide to the Encyclopedia xv 
Foreword xvii 

Preface xix 

Adaptation to Darkness 1 
Elke Aden 

Adaptation to Low Food 4 
Kathrin Hiippop 

Adaptation to Low Oxygen 10 
Frédéric Hervant and Florian Malard 

Adaptive Shifts 17 
Francis G. Howarth and Hannelore Hoch 

Anchialine Caves, Biodiversity 
in 24 
Thomas M. Iliffe 

Anchialine Caves 30 
Boris Sket 

Bats 39 
Susan W. Murray and Thomas H. Kunz 

Beetles 45 
Oana Teodora Moldovan 

Behavioral Adaptations 51 
Jakob Parzefall 

Breakdown 56 
Elizabeth L. White 

Burnsville Cove, Virginia 60 
Gregg S. Clemmer 

Camps 73 
Gregg S. Clemmer 

Castleguard Cave, Canada 77 
Derek Ford 

Cave, Definition of 81 
William B. White and David C. Culver 

Cave Dwellers in the Middle 

East 85 
Paul Goldberg and Ofer Bar-Yosef 

Chemoautotrophy 90 

Annette Summers Engel 

Clastic Sediments in Caves 102 
Gregory S. Springer 

Closed Depressions 108 
Ugo Sauro 

Coastal Caves 122 
John E. Mylroie 

Contamination of Cave Waters by 

Heavy Metals 127 
Dorothy J. Vesper 

Contamination of Cave Waters by 

Nonaqueous Phase Liquids 131 
Caroline M. Loop 

Cosmogenic Isotope 

Dating 137 
Darryl E. Granger and Derek Fabel 

Crustacea 141 
Horton H. Hobbs IIT 

Databases 155 
Keith D. Wheeland 

Dinaric Karst, Diversity in 158 
Boris Sket 

Diversity Patterns in the 
Tropics 166 

Louis Deharveng 

Diversity Patterns in the United 

States 170 
Horton H. Hobbs IIT 

Diversity Patterns in 

Australia 183 
William FE. Humphreys 

Diversity Patterns in 

Europe 196 
Janine Gibert and David C. Culver 

Early Humans in the Mammoth 

Cave Area 203 
Patty Jo Watson 

vi Contents 

Ecotones 206 
David C. Culver 

Entranceless Caves, Discovery 

of 208 
Nevin W. Davis 

Entranceless Caves, Geophysics 

of 210 
William B. White 

Entrances 215 
William B. White 

Epikarst 220 
Michel Bakalowicz 

Epikarstic Communities 223 
Anton Brancelj and David C. Culver 

Evolution of Lineages 230 

Eleonora Trajano 

Exploration and Light 

Sources 234 
William B. White 

Fish 241 
Horst Wilkens 

Flooding 251 
Chris Groves and Joe Meiman 

Food Sources 255 
Thomas L. Poulson 

Friars Hole Cave System, West 
Virginia 264 

Stephen R. H. Worthington and Douglas 

M. Medville 

Glacier Caves 271 

Andrew G. Fountain 

Guano Communities 276 
Pedro Gnaspini 

Gypsum Caves 283 
Alexander Klimchouk 

Gypsum Flowers and Related 

Speleothems 288 
William B. White 

Hydrogeology of Karst 

Aquifers 293 
William B. White 

Hydrothermal Caves 300 
Yuri Dublyansky 

Invasion, Active versus 

Passive 305 
Dan L. Danielopol and Raymond Rouch 

Jewel Cave, South Dakota 311 
Mike E. Wiles 

Karren 315 
Joyce Lundberg 

Karst Water Tracing 321 
William K. Jones 

Kazumura Cave, Hawaii 330 
Kevin Allred 

Krubera (Voronja) Cave 335 
Alexander Klimchouk 

Lechuguilla Cave, 
New Mexico 339 

Patricia Kambesis 

Life History, Evolution 346 
David C. Culver 

Mammoth Cave System 351 
Roger W. Brucker 

Mapping Subterranean 

Biodiversity 355 
Mary C. Christman 

Marine Regressions 361 
Claude Boutin and Nicole Coineau 

Maya Caves 366 
Andrea Stone and James E. Brady 

Microbes 369 
David C. Culver 

Minerals 371 
Bogdan P Onac 

Modeling Karst Aquifers 378 
Carol M. Wicks 

Molluscs 382 
David C. Culver 

Morphological Adaptations 386 

Kenneth Christiansen 

Multilevel Caves and Landscape 

Evolution 397 
Darlene M. Anthony 

Mulu Caves, Malaysia 400 
Joel Despain 

Myriapods 404 
David C. Culver 

Myth and Legend, Caves in 406 
Paul Jay Steward 

Natural Selection 409 
Thomas C. Kane and Robert C. Richardson 

Neutral Mutation 411 
Horst Wilkens 

Nitrate Contamination in Karst 

Groundwater 415 
Brian G. Katz 

Nullarbor Caves, Australia 418 
Julia M. James, Annalisa K. Contos, and 
Craig M. Barnes 

Paleomagnetic Record in Cave 

Sediments 427 
Tra D. Sasowsky 

Paleontology of Caves: 

Pleistocene Mammals 431 
Kazimierz Kowalski 

Passages 436 
George Veni 

Passage Growth and 

Development 440 
Arthur N. Palmer 

Pits and Shafts 444 
John W. Hess 

Population Structure 447 
Valerio Sbordoni, Giuliana Allegrucci, and 
Donatella Cesaroni 

Postojna-Planinska Cave System, 

Slovenia 456 
Stanka Sebela 

Protecting Caves and Cave 

Life 458 
William R. Elliott 

Recreational Caving 469 
John M. Wilson 

Rescues 475 
John C. Hempel 

Root Communities in Lava 
Tubes 477 

Fred D. Stone, Francis G. Howarth, Hannelore 
Hoch, and Manfred Asche 

Salamanders 485 
Jacques Pierre Durand 

Saltpetre Mining 492 
David A. Hubbard, Jr. 

Show Caves 495 
Arrigo A. Cigna 

Siebenhengste Cave System, 

Switzerland 500 
Pierre-Yves Jeannin and Philipp Héauselmann 

Sinking Streams and Losing 

Systems 509 
Joseph A. Ray 

Sistema Huautla, Mexico 514 
C. William Steele and James H. Smith, Jr. 

Soil Piping and Sinkhole 

Failures 521 
Barry Beck 

Solution Caves in Regions of 

Moderate Relief 527 
Arthur N. Palmer 

Solutional Sculpturing 536 
Phillip J. Murphy 

Species Interactions 539 
David C. Culver 

Speleothem Deposition 543 
Wolfgang Dreybrodt 

Speleothems: Helictites and 

Related Forms 549 
Donald G. Davis 

Spiders and Related 

Groups 554 
James R. Reddell 

Springs 565 
William B. White 

Stalactites and Stalagmites 570 

Silvia Frisia 

Sulfuric Acid Caves 573 
Arthur N. Palmer and Carol A. Hill 

Contents Vii 

Ukrainian Giant Gypsum 

Caves 583 
Alexander Klimchouk 

Vertebrate Visitors—Birds and 

Mammals 589 
Nikoa Turtkovieé 

Vicariance and Dispersalist 
Biogeography 591 
John R. Holsinger 

Volcanic Caves 599 
William B. White 

Wakulla Spring Underwater Cave 
System, Florida 603 

Barbara Anne am Ende 

Water Chemistry in Caves 609 

Janet S. Herman 

Worms 614 

Elzbieta Dumnicka 

Glossary 619 
Index 631 


Anchialine Caves 

Cave, Definition of 

Coastal Caves 

Glacier Caves 

Gypsum Caves 

Hydrothermal Caves 

Solution Caves in Regions of 
Moderate Relief 

Sulfuric Acid Caves 

Volcanic Caves 



Pits and Shafts 


Closed Depressions 


Hydrogeology of Karst Aquifers 

Karst Waters Tracing 

Modeling Karst Aquifers 

Passages Growth and 

Sinking Caves and Cave 

Solutional Sculpturing 

Water Chemistry in Caves 


Clastic Sediments in Caves 

Gypsum Flowers and Related 



Speleothem Deposition 


Stalactites and Stalagmites 


Cosmogenic Isotope Dating 
Multilevel Caves and Landscape 


Paleomagnetic Record in Cave 
Paleontology of Caves: Pleistocene 



Burnsville Cove, Virginia 
Castleguard Cave, Canada 
Friars Hole Cave System, West 
Jewel Cave, South Dakota 
Kazumura Cave, Hawaii 
Krubera (Voronja) Cave 
Lechuguilla Cave, 
New Mexico 
Mammoth Cave System 
Mulu Caves, Malaysia 
Nullarbor Caves, Australia 
Postojna-Planinska Cave System, 
Siebenhengste Cave System, 
Sistema Huautla, Mexico 
Ukranian Caves 
Wakulla Spring Underwater 
Cave System, Florida 

x Contents by Subject Area 






Guano Communities 





Spiders and Related Groups 

Vertebrate Visitors—Birds and 




Epikarstic Communities 

Food Sources 

Population Structure 

Root Communities in 
Lava Tubes 

Species Interactions 


Invasion, Active versus Passive 

Marine Regressions 

Vicariance and Dispersalist 


Anchialine Caves, Biodiversity in 
Dinaric Karst, Diversity in 
Diversity in the Tropics 
Diversity in the United States 
Diversity Patterns in Australia 
Diversity Patterns in Europe 
Mapping Subterranean 


Adaptation to Darkness 
Adaptation to Low Food 
Adaptation to Low Oxygen 
Adaptive Shifts 

Behavioral Adaptations 
Evolution of Lineages 

Life History Evolution 
Morphological Adaptations 
Natural Selection 

Neutral Mutation 




Entranceless Caves, Discovery of 
Entranceless Caves, Geophysics of 
Exploration and Light Sources 


Protecting Caves and Cave Life 
Recreational Caving 
Show Caves 



Cave Dwellers in the Middle East 

Early Humans in the Mammoth 
Cave Area 

Maya Caves 

Myths and Legends, Caves in 

Salpetre Mining 






Contamination of Cave Waters by 
Heavy Metals 

Contamination of Cave Waters by 
Nonaqueous Phase Liquids 


Nitrate Contamination in Karst 

Soil Piping and Sinkhole Failures 


Zoological Institute and Zoological Museum, 
University of Hamburg, Germany 
Adaptation to Darkness 

Tor Vergata University, Italy 
Population Structure 


Hawaii Speleological Survey 
Kazumura Cave, Hawaii 

Wakulla Spring Underwater Cave System, 

Deep Caves Consulting 

Purdue University 

Multilevel Caves and Landscape 


Museum fur Naturkunde 

Root Communities in Lava Tubes 



University of Sydney, Australia 
Nullarbor Caves, Australia 

Harvard University 

Cave Dwellers in the Middle East 


PE. LaMoreaux & Associates, Inc. 
Soil Piping and Sinkhole Failures 

Université Paul Sabatier, France 
Marine Regressions 

California State University, Los Angeles 
Maya Caves 

National Institute of Biology, Slovenia 
Epikarstic Communities 

Mammoth Cave System 

Tor Vergata University, Italy 
Population Structure 

Grinnell College 
Morphological Adaptations 

University of Maryland 
Mapping Subterranean Biodiversity 


International Show Caves Association (Union 
Internationale de Spéléologie), Italy 

Show Caves 

Butler Cave Conservation Society, Inc. 
Burnsville Cove, Virginia 

Laboratoire Arago, France 
Marine Regressions 
University of Sydney, Australia 
Nullarbor Caves, Australia 
American University 

Cave, Definition of 
Diversity Patterns in Europe 


Epikarstic Communities 

Life History Evolution 



Species Interactions 

Austrian Academy of Sciences, Austria 
Invasion, Active versus Passive 

National Speleological Society 
Speleothems: Helictites and Related Forms 

Butler Cave Conservation Society, Inc. 
Entranceless Caves, Discovery of 


Museum National d'Histoire Naturelle de Paris, 

Diversity in the Tropics 


Gunung Buda Project and Sequoia and Kings 
Canyon National Park 

Mulu Caves, Malaysia 


University of Bremen, Germany 

Speleothem Deposition 


Institute of Mineralogy and Petrography, Russia 
Hydrothermal Caves 


Institute of Freshwater Biology, Polish Academy of 
Sciences, Poland 


Laboratoire Souterrain, France 

xii Contributors 

Missouri Department of Conservation 
Protecting Caves and Cave Life 

The University of Texas at Austin 

The Australian National University 
Cosmogenic Isotope Dating 

McMaster University 
Castleguard Cave, Canada 

Portland State University 
Glacier Caves 

Museo Tridentino di Scienze Naturali, Italy 
Stalactites and Stalagmites 

Université Lyon, France 
Diversity Patterns in Europe 

University of Sao Paulo, Argentina 
Guano Communities 

Boston University 

Cave Dwellers in the Middle East 

Purdue University 
Cosmogenic Isotope Dating 

Western Kentucky University 

Hohlenforschergemeinschaft Region Hohgant 
(HRH) and Swiss Institute for Speleology and 
Karst Studies (SISKA), Switzerland 
Siebenhengste Cave System, Switzerland 

EEI Geophysical 

University of Virginia 
Water Chemistry in Caves 

Université Lyon, France 
Adaptation to Low Oxygen 


Geological Society of America 
Pits and Shafts 

University of New Mexico 
Sulfuric Acid Caves 


Wittenberg University 


Diversity Patterns in the United States 

Museum fur Naturkunde, Germany 
Adaptive Shifts 

Root Communities in Lava Tubes 

Old Dominion University 
Vicariance and Dispersalist Biogeography 

Bernice P. Bishop Museum 
Adaptive Shifts 

Root Communities in Lava Tubes 


Virginia Speleological Survey and Virginia 
Department of Mines, Minerals and Energy 
Saltpetre Mining 

Western Australian Museum, Australia 
Diversity Patterns in Australia 

Institute of Avian Research, Germany 
Adaptation to Low Food 

Texas A&M University at Galveston 
Anchialine Caves, Biodiversity 

University of Sydney, Australia 
Nullarbor Caves, Australia 

Hohlenforschergemeinschaft Region Hohgant 
(HRH) and Swiss Institute for Speleology and 
Karst Studies (SISKA), Switzerland 
Siebenhengste Cave System, Switzerland 

Karst Waters Institute 

Karst Water Tracing 

Hoffman Environmental Research Institute 

Lechuguilla Cave, New Mexico 

University of Cincinnati 
Natural Selection 


U.S. Geological Survey 

Nitrate Contamination in Karst 

Institute of Geological Sciences, Ukraine 
Gypsum Caves 

Krubera (Voronja) Cave 
Ukrainian Giant Gypsum Caves 


Polish Academy of Sciences, Poland 
Paleontology of Caves: Pleistocene 

Boston University 


The Pennsylvania State University 
Contamination of Cave Waters by 
Nonaqueous Phase Liquids 

Carleton University, Ottawa 

Université Lyon, France 

Adaptation to Low Oxygen 

West Virginia Speleological Survey 
Friars Hole Cave System, West Virginia 

Mammoth Cave National Park 

Emil Racovitza Speleological Institute, Romania 

University of Leeds, United Kingdom 
Solutional Sculpturing 

Boston University 

Mississippi State University 
Coastal Caves 


Babes-Bolyai University and Emil Racovita 
Institute of Speleology, Romania 



State University of New York 

Passage Growth and Development 
Solution Caves in Regions of Moderate 

Sulfuric Acid Caves 


Zoologisches Institut und Zoologisches der 
Universitit Hamburg, Germany 
Behavioral Adaptations 


University of Illinois at Chicago 

Food Sources 


Kentucky Division of Water 

Sinking Streams and Loosing Streams 

Laboratoire Souterrain, France (Retired) 
Invasion, Active versus Passive 

The University of Texas at Austin 
Spiders and Related Groups 
University of Cincinnati 

Natural Selection 


University of Akron 
Paleomagnetic Record in Cave Sediments 

University of Padova, Italy 

Closed Depressions 


Tor Vergata University, Italy 
Population Structure 

Karst Research Institute, Slovenia 
Postojna—Planinska Cave System, Slovenia 

Univerza v Ljublijani, Slovenia 
Anchialine Caves 

Dinaric Karst, Diversity in 
Environmental Protection Agency 
Sistema Huautla, Mexico 
Ohio University 

Clastic Sediments in Caves 


Boy Scouts of America 

Sistema Huautla, Mexico 

Cave Research Foundation 

Myth and Legend, Caves in 

University of Wisconsin-Milwaukee 
Maya Caves 

Hawaii Community College 
Root Communities in Lava Tubes 

Universidade de Sao Paulo, Argentina 
Evolution of Lineages 


Croatian Natural History Museum, Croatia 
Vertebrate Visitors—Birds and Mammals 

George Veni & Associates 


West Virginia University 

Contamination of Cave Waters by Heavy 

Washington University, St. Louis 
Early Humans in the Mammoth Cave Area 

The Pennsylvania State University, Retired 

Contributors Xili 

The Pennsylvania State University 


The Pennsylvania State University 

Cave, Definition of 

Entranceless Caves, Geophysics of 

Exploration and Light Sources 

Gypsum Flowers and Related Speleothems 
Hydrogeology of Karst Aquifers 


Volcanic Caves 

University of Missouri—Columbia 
Modeling Karst Aquifers 

Jewel Cave National Monument 

Jewel Cave, South Dakota 

University of Hamburg, Germany 

Neutral Mutation 


Marks Products, Inc. 
Recreational Caving 

Worthington Groundwater 
Friars Hole Cave System, West Virginia 

The Encyclopedia of Caves is a complete source of information 
on the subject of caves and life in caves, contained within a 
single volume. Each article in the Encyclopedia provides an 
overview of the selected topic to inform a broad spectrum of 
readers, from biologists and geologists conducting research in 
related areas, to students and the interested general public. 

In order that you, the reader, will derive the maximum 
benefit from the Encyclopedia of Caves, we have provided this 
Guide. It explains how the book is organized and how the 
information within its pages can be located. 


The Encyclopedia of Caves presents 107 separate articles 
on the entire range of speleological study. Articles in the 
Encyclopedia fall within 15 general subject areas, as follows: 

¢ ‘Types of Caves 

° Cave Features 

e Hydrology and Hydrogeology 

¢ Speleothems and Other Cave Deposits 

* Cave Ages and Paleoclimate 

e Exceptional Caves 

¢ Biology of Particular Organisms in Caves 

* Ecology 

¢ Cave Invasion 

¢ Biogeography and Diversity 

¢ Evolution and Adaptation in Caves 

e Exploration of Caves 

¢ Contemporary Use of Caves 

¢ Historical Use of Caves 

¢ Ground Water Contamination and Land Use Hazards 
in Cave Regions 


The Encyclopedia of Caves is organized to provide the 
maximum ease of use for its readers. All of the articles are 
arranged in a single alphabetical sequence by title. An 
alphabetical Table of Contents for the articles can be found 
beginning on page v of this introductory section. 

So they can be more easily identified, article titles begin 
with the key word or phrase indicating the topic, with any 
descriptive terms following this. For example, “Invasion, 
Active versus Passive” is the title assigned to this article, 
rather than “Active versus Passive Invasion,” because the 
specific term Invasion is the key word. 

You can use this alphabetical Table of Contents by itself to 
locate a topic, or you can first identify the topic in the 
Contents by Subject Area on page x and then go to the 
alphabetical Table to find the page location. 


Each article in the Encyclopedia begins with introductory text 
that defines the topic being discussed and indicates its 
significance. For example, the article “Behavioral 
Adaptations” begins as follows: 

Animals living in darkness have to compete for food, 
mates, and space for undisturbed reproduction just as their 
epigean conspecifics do in the epigean habitats, but there is 
one striking difference: In light, animals can use visual 
signals. Thus, important aspects of their behavior driven by 
visual signals cannot apply in darkness. The question arises, 
then, of how cave dwellers compensate for this disadvantage 
in complete darkness. This article uses several examples to 
compare various behavior patterns among cave dwelling 

populations with epigean ancestors. 

xvi Guide to the Encyclopedia 

Major headings highlight important subtopics that are 
discussed in the article. For example, the article “Beetles” 
includes these topics: Adaptations, Colonization and 
Geographical Distribution, Systematics of Cave Beetles, 
Ecology, Importance and Protection. 


Cross-references appear within the Encyclopedia as 
indications of related topics at the end of a particular article. 
As an example, a cross-reference at the end of an article can 
be found in the entry “Camps.” This article concludes with 

the statement: 

See Also the Following Articles 
Recreational Caving ° Exploration of Light Sources 

This reference indicates that these related articles all 
provide some additional information about Camps. 


The Bibliography section appears as the last element of the 
article. This section lists recent secondary sources that will 
aid the reader in locating more detailed or technical 
information on the topic at hand. Review articles and 
research papers that are important to a more detailed 
understanding of the topic are also listed here. The 
Bibliography entries in this Encyclopedia are for the benefit 
of the reader to provide references for further reading or 
additional research on the given topic. Thus they typically 
consist of a limited number of entries. They are not intended 
to represent a complete listing of all the materials consulted 
by the author or authors in preparing the article. The 
Bibliography is in effect an extension of the article itself, and 

it represents the author’ choice as to the best sources 
available for additional information. 


The Encyclopedia of Caves presents an additional resource for 
the reader, following the A-Z text. A comprehensive glossary 
provides definitions for more than 450 specialized terms used 
in the articles in this Encyclopedia. The terms were identified 
by the contributors as helpful to the understanding of their 
entries, and they have been defined by these authors 
according to their use in the actual articles. 


The Subject Index for the Encyclopedia of Caves contains 
more than 4600 entries. Within the entry for a given topic, 
references to general coverage of the topic appear first, such 
as a complete article on the subject. References to more 
specific aspects of the topic then appear below this in an 
indented list. 


The Encyclopedia of Caves maintains its own Web page on the 
Internet at: 

This site provides information about the Encyclopedia and 
features links to related sites that provide information on 
subjects covered in the Encyclopedia. It also hosts sample 
material, published reviews, and the opportunity to purchase 
additional copies of the Encyclopedia on a secure Web site. 
The site will continue to evolve as more information 
becomes available. 

ew things capture man’s imagination, as do caves. Mark 

Twain recognized this in writing 7om Sawyer. Tom, Huck 
Finn, Becky Thatcher, and Injun Joe are well-known 
characters of interest to children and adults alike. Their cave 
exploration, for example, with candles flickering, currents of 
fresh air causing flames to flutter, and the discovery of Injun 
Joe in the dim light of a lantern with a hidden treasure was a 
masterpiece of literature. Twain captured it all. 

Caves, stalactites, stalagmites, albino sightless fish, and 
underground rivers have been around for millions of years. 
The earliest “cave art” dates to 15,000 B.P. in the caves of 
Altimara in Spain and in Lascaux in France, which contain 
spectacular drawings of animals. Cave artifacts are the earliest 
evidence of cave occupation at the time of Australopithecus 
(3.5 m.y.b.p). Caves provided evidence that Homo (2.4 
m.y.b.p.), used caves for shelters, a water source, and safety. 
Caves have provided information about the evolution of 
mankind through their artifacts, weapons, fire, and art. 
Shanedar Cave, in Iraq, as an example, was the burial place 
for nine Neanderthal skeletons that provided knowledge 
about care of the sick and elderly and the ritualistic burial of 
their dead over 34,000 years b.p. 

Approximately one-fifth of the earth’s surface is underlain 
by carbonate rocks of a complex physical character that 
produced a diverse topographic expression by weathering 
under varied climatic conditions. Carbonate terranes in 
some areas are underlain by broad, rolling plains, whereas in 
others they are characterized by steep bluffs, canyons, sinks, 
and valleys. Owing to the variability of the solubility of 
limestone, man’s inhabitation and development in limestone 
areas has sometimes been difficult. There are areas of 
limestone covered by fertile soils, whereas in others, soils are 
missing. In the United States, a large area in the Midwest is 
underlain by limestone and covered by a very rich soil that 

produces large quantities of food. This area is called, literally, 
“the breadbasket of a nation.” 

Carbonate rocks are a source of abundant water supplies, 
minerals and oil, and gas. Even though there are many 
blessings associated with carbonate terranes, there are also 
many problems related to developing an adequate water 
supply, assuring proper drainage, providing stable foundation 
conditions, and preventing serious pollution problems. 
Because of this complexity, the evolution of concepts related 
to the movement and occurrence of ground water in karst, 
methods of exploration and development of water, safe 
engineering practices in construction of all kinds, and 
adequate environmental safety precautions cannot be based 
on one set of uniform rules. 

Caves Karst areas are dynamic and environmentally sensi- 
tive. The geologic structure, solubility of the rocks involved, 
and the climatic conditions determine to a great degree how 
rapid these changes can take place. Therefore, investigations 
must consider the dynamic nature of karst. It is necessary to 
recognize the synergistic relation between circulation of 
water and the solution of the rock. The greater and more 
rapid the solution of the rock can lead to changes in and 
progressive lowering of water tables—base levels and cave 
enlargement, changes that can take place in a relatively brief 
period of time and can impact the hydrogeologic history of 
and area, and bring about major environmental problems. 

During the past few years many outstanding publications 
illustrated by fine graphics have described caves and the 
science of karst. The literature on the subject is voluminous 
with over 50 new textbooks, approximately 1000 technical 
manuscripts, and over 300 field trips associated with 
meetings or congresses. 

Because of this proliferation of scientific material and the 
diversification of disciplines involved—botany, biology, 

XVili Foreword 

chemistry, archaeology, geology, engineering, speleology, 
conservation and planning, history and resource exploration, 
and development—it is not easy to maintain a grasp of the 
status of karst research. The Encyclopedia of Caves, with 107 
articles by world-wide experts on caves, is quite unique with 
contributions from multi-disciplines and a great variety of 
subject matter, academic, as well as practical, with case 
histories describing all types of cave-associated problems such 

as: drainage, water supply, construction, environmental 

exploration, development, and management of water 

supplies in karst. This volume is a must for karst researchers, 

cave enthusiasts, teachers, and developers. 

Philip LaMoreaux 
June, 2004 

hroughout history, caves have always been of at least 

some interest to almost everyone. During the past few 
centuries caves have been a passionate interest to at least a 
few people. The number of those with a passionate interest 
has been continuously growing. The core of the cave enthu- 
siasts are, of course, the cave explorers. However, scientists of 
various sorts, mainly geologists and biologists, have also 
found caves useful and fascinating subjects for scientific 

There have always been cave explorers. Some, such as E. A. 
Martel in France in the late 1800s, achieved amazing feats of 
exploration of deep alpine caves. In the United States, the 
number of individuals seriously interested in the exploration 
of caves has grown continuously since the 1940s. Cave 
exploration takes many forms. Some cavers are interested in 
caving simply as a recreational experience, not intrinsically 
different from hiking, rock climbing, or mountain biking. 
But many pursue genuine exploration. Their objective is the 
discovery of new cave passages never before seen by humans. 
As the more obvious entrances and the more accessible caves 
have been explored, cave exploration in the true sense of the 
word, has become more elaborate and more difficult. To meet 
the challenge of larger, more obscure and more difficult 
caves, cavers have responded with the invention of new 
techniques, new equipment, and the training required to use 
it. To meet the challenge of long and difficult caves, cavers 
have been willing to accept the discipline of project and 
expedition caving and to accept the arduous tasks of 
surveying caves as they are explored. The result has been the 
accumulation of a tremendous wealth of information about 
caves that has been invaluable to those studying caves from a 
scientific point of view. 

In the early years of the twentieth century, a few geologists 
became interested in the processes that allowed caves to form. 
Biologists were interested in the unique habitats and the 

specialized organisms that evolved there. In both sciences and 
in both Europe and the United States, the interest was in the 
caves themselves. The study of caves was focused inward and 
some proposed the study of caves to be a separate science 
called speleology. In the latter decades of the twentieth 
century, there was a gradual change in perspective. The study 
of caves came to be seen as important for its illumination of 
other realms of science. 

In the past few decades, the geological study of caves has 
undergone a tremendous expansion in point of view. The 
caves themselves are no longer seen as simply geological 
oddities that need to be explained. Caves are repositories and 
are part of something larger. As repositories, the clastic 
sediments in caves and the speleothems in caves have been 
found to be records of past climatic and hydrologic 
conditions. Cave passages themselves are recognized as 
fragments of conduit systems that are or were an intrinsic 
part of the groundwater system. Active caves give direct 
insight into the hydrology and dry caves are records that tell 
something of how drainage systems have evolved. Techniques 
for the dating of cave deposits have locked down events 
much more accurately in the caves than on the land surface 
above. Caves then become an important marker for 
interpreting the evolution of the landscape above. Even the 
original, rather prosaic, problem of explaining the origin and 
development of caves has required delving into the chemistry 
of groundwater interactions with carbonate rocks and on the 
fluid mechanics of groundwater flow. 

Cave biology has likewise evolved from an exercise in 
taxonomy—discovering, describing, and classifying organ- 
isms from caves—to the use of caves as natural laboratories 
for ecology and evolutionary studies. The central question 
that has occupied the attention of biologists at least since the 
time of Lamarck is how did animals come to lose their eyes 
and pigment. The question gets answered each generation 

xx Preface 

using the scientific tools available and its most contemporary 
form is a question of the fate of eye genes themselves. Cave 
animals have also served as models for the study of 
adaptation because of their ability to survive in the harsh 
environments of caves. There are also interesting biological 
questions about the evolutionary history of cave animals that 
are being unraveled using a variety of contemporary 
techniques. Finally, there is increasing concern about the 
conservation of cave animals. Nearly all have very restricted 
ranges and many are found in only a single cave. The past 
two decades have seen a phenomenal growth in the 
understanding of how to manage cave and karst areas to 
protect the species that depend on them. 

One should not suppose that caves are of interest only to 
geologists and biologists. Caves are repositories of archaeo- 
logical and paleontological resources. Ancient art has been 
preserved in caves. Caves appear in folk tales, legends, 
mythology, and in the religions of many peoples. Caves 
appear frequently in literature, either as an interesting setting 
for the story or as a metaphor. The latter has a history 
extending at least to Plato. 

In planning the content of the Encyclopedia of Caves, the 
editors were faced with this great variety of “clients” with 
their highly diverse interests in caves. Several decisions were 
made. One was that we would address the interests of as 
many “clients” as possible given the limitations of space. 
Thus, the Encyclopedia, in addition to the expected articles 
on biology and geology, also contains articles on exploration 
techniques, archaeology, and folklore. A second decision was 
to allow authors a reasonable page space so they could discuss 

their assigned subject in some depth. As a result of this 
decision the Encyclopedia contains a smaller number of 
articles and thus a smaller number of subjects than might be 
expected. The object was to provide a good cross-section of 
contemporary knowledge of caves rather than attempt an 
entry for every possible subject. 

The level of presentation was intended to be at the college 
level. In this way the articles would have sufficient technical 
depth to be useful to specialists but would still be accessible 
to the general reader. Some of the subjects are intrinsically 
more technical than others but we have attempted to keep to 
a minimum the specialist jargon and in particular the 
obnoxious acronyms that turn many technical subjects into a 
secret code known only to insiders. 

The selection of authors was made by the editors. We 
attempted to select contributors who we knew were expert in 
the subject being requested of them. For many subjects there 
was certainly a choice of potential experts and our selection 
was to some degree arbitrary. We sincerely hope that no 
one is offended that some other person was selected rather 
than them. 

We take this opportunity to thank the authors for their 
hard work. The Encyclopedia is a collective effort of many 
peoples in many disciplines. We are particularly appreciative 
of everyone’s efforts to communicate with cave enthusiasts 
outside of their particular discipline. 

David C. Culver 
William B. White 
September 2004 

Adaptation to Darkness 

Elke Aden 
Zoological Institute and Zoological Museum, University of Hamburg, 


i ight is the origin of all life, and primary production is the 
source of biodiversity; nevertheless, life also exists in 
darkness. Animals living in this “unreal” world have their 
origin in the light and have to be adapted to this environment. 
Probably no animals have a more intimate environmental 
adaptation than those inhabiting caves. What does adapta- 
tion mean and how dark is darkness? 


The most obvious characteristic of caves is the darkness 
beyond the twilight zone that extends a short distance in 
from the entrance (Fig. 1). An experiment to characterize 
cave darkness as absolute darkness illustrates this point well. 
A photographic film developed after being exposed for a 
week in the depths of a cave was completely blank (Moore 
and Sullivan, 1978). This should be the definition for dark, 
and it is darker than the darkest night. This level of darkness 
is comparable with deep-sea ocean depths, because below 
1000 to 1200 meters there is no penetration of sunlight, but 
sometimes bioluminescence does occur. Normally, no bio- 
luminescence is observed in caves; only caves in New Zealand 
and Australia are known to have light-producing insect 
larvae. There are some parallels in adaptations to darkness 
between the deep sea and caves. The large domain on earth 
where life exists without light is termed allobiosphere, after 
Hutchinson (Danielopol et a/., 1996). Perpetual darkness is a 
characteristic of most rock void habitats anyway, so let us 
choose to define a cave as a habitat entirely without natural 

illumination (Chapman, 1993). For another definition of a 
cave, see Culver and White (Cave, Definition of). 


To explain the origin and maintenance of patterns of organic 
diversity, biologists normally use two principles: phylogeny 
and adaptation. The term adaptation is applied to several 
different biological phenomena. Generally, three different 
types of responses of organisms to their environment have 
been termed adaptive: 

Physiological adaptation is the ability of organisms to 
adjust phenotypically to short-term changes in the 
environment. This is somatic or phenotypic plasticity, 
and it is not heritable. For example, adaptation of the 
eye is sensitivity adjustment effected after considerable 
exposure to light (light adapted) or darkness (dark 

Adaptation in behavior means fatigue in responses to a 
repeated, uniform stimulus. 

Evolutionary adaptation is the long-term, hereditary 
change that occurs in species in response to a 
particular set of environmental factors. 

Putting all these things together, let us say that adaptation 
means any morphological, physiological, or behavioral 
characteristics that fit an organism to the conditions under 
which it lives. The term adaptation arises from the Latin ad 
+ aptus (“towards a fit”). Adaptive traits are those that are 
correlated with an aspect of the environment and are postu- 
lated to have arisen and been subsequently maintained by 
the same selective pressure (Northcutt, 1988). Culver et al. 
(1995) stress the predominant importance of the genetic 
component: Adaptation is a progress of genetic change 
resulting in improvement of a character with reference to a 
specific function or a feature of a selective advantage that has 
become prevalent in a population. Because most traits have 

2 Adaptation to Darkness 

multiple functions, it is necessary to determine how a trait is 
actually needed by an organism in the real world—its 
biological role—when trying to evaluate its adaptiveness. 


Animals that live permanently in the dark zone and are found 
exclusively in caves are termed troglobites, from troglos 
(“cave”) and bios (“life”). Typical troglobites are pale and 
blind. These genetically fixed characters evolved in all 
troglobites of many different systematic groups subsequent 
to exposure to darkness for a sufficiently long period. Usually 
pigmentation is necessary as protection against sunlight, 
especially the ultraviolet wavelengths. In caves, this protec- 
tion is obsolete. The eye is a sense organ to detect light, so in 
caves even the most highly developed eyes are completely 
useless. When we recall the definition of adaptation we can 
say that the loss of eyes is not an adaptation to fit an animal 
for the dark environment, but it is a restraining condition 
that occurs only in the darkness as a matter of course. 
Nevertheless, in the field of regressive evolution, the loss of 
the eye has received strong attention. Another view is that the 
regression of ocular structures is surely an adaptive feature, 
because it is submitted to natural selection in order to save 
energy (for discussion, see Culver and Wilkens, 2000). 


The term troglomorph refers to any morphological, physio- 
logical, or behavioral feature that characterizes cave animals 
(Christiansen, 1992). The troglomorphic suite in cave 
animals exemplifies evolutionary convergence resulting 
from life in similar environments. Troglomorphy comprises 
constructive as well as regressive characteristics of the cave 
animals. In consequence, not every troglomorphic trait is 
adaptive, but on the other hand all cave adaptations result in 
troglomorphy. Not every cave animal displays the complete 
set of troglomorphic traits. Their expression depends on the 

characteristics of the epigean ancestor. A prerequisite for 
constructive traits is their genetic availability in epigean 
forms. For example a fish that lacks barbels probably will not 
evolve them as a cave fish. To see the degree of adaptation, 
one has to search in each individual species to determine 
what trait could be improved. The reason for the similarities 
of troglomorphies in different troglobites is the process 
of evolution, with mechanisms underlying parallelism and 
convergence. The term troglomorphy is restricted to cave 
animals; the same traits lead to the more general concept of 
darkness syndrome, which is comprised of morphological 
and biological changes of animals inhabiting the aquatic 
allobiosphere (Danielopol et a/., 1996). 

The most conspicuous troglomorphic feature in troglo- 
bites is regression of their eyes. In cave fishes, the degree of 
eye reduction is commonly considered as a reflection of the 
period of cavernicolous evolution and therefore indicates the 
relative phylogenetic age of related cave fish species. Within 
the ontogeny, the eye is built up to a species-specific degree 
of development followed by a more or less pronounced 
regression (Langecker, 2000). Either a rudiment or even a 
total loss of any eye tissue can be found. Loss of the eyes is 
linked to learning how to navigate blindly. This problem is 
solved by cave animals in different ways. To find food and to 
reproduce are essential. Animals able to find food and a 
partner without the use of the eye are good candidates for 
generating a cave form. Without a need to improve the 
efficiency of the other nonvisual senses, there will not be a 
constructive evolution. Alternatively , when the nonvisual 
senses are not good enough for blind navigation, there will be 
a constructive evolution of other structures or features. 


To begin adaptation to an environment such as a cave, a 
population has to pass through stages of subterranean 
evolution. For a critical review of the relevant theories of the 
evolution of subterranean animals, see Culver and Wilkens 

Entrance Zone 

Twilight Zone 

Dark Zone 

FIGURE 1 The darkness beyond the twilight zone that extends a short distance in from the cave entrances. 

(2000). All known cave species originate from surface- 
dwelling species that have incidentally invaded caves. When 
the ancestral form has remained extant, a direct comparison 
can be made between the ancestral and derived cave forms. 
In a few cases among isopods, amphipods, and fish, the cave 
and epigean form are regarded as being conspecific, but for 
most cave animals direct epigean ancestors are unknown. The 
surface-dwelling ancestors of many cave forms have become 
extinct, leaving their underground derivatives as the sole 
representatives of the taxon. 

Comparing the epigean and hypogean forms of one 
species, we can identify what may be an obvious adaptation 
to darkness; however, both lineages are susceptible to evolu- 
tionary change, even the epigean one, so it may be difficult 
to recapitulate the common ancestral state. The interpreta- 
tion might be complicated by whether a structure or feature 
is adapted or not. Additionally, it is necessary to evaluate if 
a supposed troglomorphic trait is, in fact, apomorphic or 
merely a plesiomorphy of the group or lineage. 


Even though the majority of troglobiotic species live in 
terrestrial cave environments, some species (e.g., fish) evolve 
in aquatic caves. Fish, in terms of numbers, are by far the 
most important of all vertebrate classes, probably accounting 
for over half of all recognized vertebrate species. In many 
ways, their visual systems and specifically their eyes are 
similar to our own. The retina is widely used as a model 
system for the study of the central nervous system (Douglas 
and Djamgoz, 1990). Much can be learned about the visual 
system by examining how it is reduced under such strong 
environmental conditions as total darkness. For further 
discussion of eye reduction in cave fishes, see Wilkens (Fish). 
Troglobitic fishes can be found all over the world. On the 
other hand, a single species is often endemic; that is, it is only 
found in a single cave or a small karst area. About 81 species 
of hypogean fishes are known in 18 families (Weber, 2000; 
updated by G. Proudlove, pers. comm.). All of them are able 
to navigate in complete darkness, and several exhibit a 
compensatory development of extraoptic sensorial organs. 

Silurids are characterized by long tactile and gustatory 
barbels and do not depend on sight in their feeding activities. 
This is, of course, a favorable requisite for successful cave 
colonization. Indeed, about a third of the cave fishes belong 
to the catfishes. 

Fishes may lose their eyes, but never their auditory and 
lateral line systems. Some of them are more sensitive to vibra- 
tion in the surrounding water than their seeing relatives—for 
example, by virtue of a better lateral line system. To improve 
orientation, the tactile sense, the olfactory sense, taste, and 
hearing can be advanced. This can be achieved by increasing 
the number of receptors, or an elongation of the neuromast 
cupulae may occur. A dispersion of structures can be seen in 
some cases. Often, elongated antennae or different appen- 

Adaptation to Darkness 3 

dages in invertebrates or elongated barbels in cave fish have 
evolved to improve orientation. These modifications in sensory 
structures have converged in different cave animals and/or in 
different populations. No general way to compensate the 
visual system exists in all cave animals. For more information 
on sensory compensation in specific cave fish species, see 
Wilkens (Fish). 

Due to compensatory improvement of extraocular senses, 
the signal processing structures are altered correspondingly. 
For example, an enhancement in taste bud number is 
matched by an increase in the size of the forebrain, which 
contains the teleost gustatory center (Jeffery, 2001). 
Similarly, eye regression leads to a reduced tectum opticum, 
which is the main center of vision processing. 

In summary, adaptation to darkness means managing to 
live blindly, if necessary, via an improvement of nonvisual 
senses that may enhance the fitness of the animal to life in a 
dark environment like a cave. Each cave is characterized by 
constraints that depend on its mode of formation, location, 
age, size, temperature, humidity, and food supply. The 
adaptations of cave animals depend not only on darkness but 
also on the remaining cave conditions. 

See Also the Following Articles 
Adaptive Shifts 


Chapman, P. (1993) Caves and Cave Life, Harper Collins, London. 

Christiansen, K.A. (1992) Biological processes in space and time: cave life in 
the light of modern evolutionary theory, in The Natural History of 
Biospeleology, Camacho, A.I., Ed., Museo Nacional de Ciencias 
Naturales, Madrid, pp. 453-478. 

Culver, D.C., T.C., Kane, and D.W. Fong, (1995) Adaptation and Natural 
Selection in Caves, Harvard University Press, Cambridge, MA. 

Culver, D.C. and H. Wilkens (2000) Critical review of the relevant theories 
of the evolution of subterranean animals. In Ecosystems of the World, 
Vol. 30, Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.E, 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 381-397. 

Danielopol, D.L., A. Baltanas, and G. Bonaduce (1996) The darkness 
syndrome of subsurface-shallow and deep-sea dwelling Ostracoda. In 
Deep-Sea and Extreme Shallow-Water Habitats: Affinities and Adaptation 
(E. Uiblein, J. Ott, and M. Stachowitsch, Eds.). Austrian Academy of 
Sciences, Vienna, pp. 123-143. 

Douglas, R.H. and M.B.A. Djamgoz (1990) The Visual System of Fish, 
Chapman & Hall, London. 

Jeffery, W.R. (2001) Cavefish as a model system in evolutionary 
developmental biology. Dev. Biol., 231: 1-12. 

Langecker, T.G. (2000) The effect of continuous darkness on cave ecology 
and cavernicolous evolution. In Ecosystems of the World, Vol. 30, 
Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.E, 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 135-157. 

Moore, G.W. and G.N. Sullivan (1978) Speleology: The Study of Caves, 
Zephyrtus Press, Teaneck, NJ. 

Northcutt, R.G. (1988) Sensory and other neural trails and the adaptationist 
program: mackerels of San Marco? In: Sensory Biology of Aquatic Animals 
(J. Atema, R.R. Fay, A.N. Popper, and W.N. Tavolga, Eds.). Springer- 
Verlag, New York, pp. 870-883. 

Weber, A. (2000) Fish and amphibia. In: Ecosystems of the World, Vol. 30, 
Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.F, 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 109-132. 

4 Adaptation to Low Food 

Adaptation to Low Food 

Kathrin Hiippop 

ave animals share numerous adaptations to the relative 

food scarcity in their habitat. A refined sensory orienta- 
tion due to elongated appendages combined with enlarged or 
multiplied sensory areas and changes in foraging behavior 
improve food-finding ability in hypogean environments 
which are generally and/or patchily scarce in food. A reduced 
energy demand, realizable through reduced metabolic rates 
and/or changes of life history toward more K-selection, is a 
method to cope with the general food scarcity in caves. 
Higher fat accumulation ability additionally aids survival 
during starvation periods in caves periodically low in food. 
All of these and several other factors concerning adaptation 
to food scarcity in subterranean habitats can be illustrated in 
a causal network. 


Subterranean environments are characterized not only by 
continuous darkness but also by a reduced variability in 
the number of specific abiotic conditions such as moisture, 
temperature, and water chemistry, as well as by isolation 
and restriction in space. Additionally, hypogean systems 
are relatively energy limited compared to photosynthetically 
based epigean systems. As a response, many cave animals share 
numerous adaptations to the food scarcity of their environ- 
ment. They not only show morphological and behavioral 
adaptations but also have evolved several special physiological 
characters. Especially energy economy, which is a reduction in 
energy consumption, has a high selective advantage in cave 
animals and has been observed in numerous species in a 
variety of phyla (Poulson, 1963; Culver, 1982; Hiippop, 
2000). All factors concerning adaptation to food scarcity in 
caves can be illustrated in the causal network shown in Fig. 1 

On one hand, the high environmental stability in caves, 
including darkness and sometimes predator scarcity, allows 
the evolution of characters; on the other hand, it requires 
character changes. In fact, most characteristics of adaptation 
to food scarcity can only be realized in ecologically stable 
and, above all, predator-poor caves (Fig. 1). Food scarcity 
acting as a selective force in caves requires adaptations. 
Possible adaptations of cave animals to survival in caves low 
in food are an improved food-finding ability, an improved 
starvation resistance, a reduced energy demand by reduced 
metabolic rates, and life history characters changed toward 
more K-selected features (Hiippop, 2000). Further, feeding 
generalism and dietary shift may be realized. Many of these 
characteristics have evolved coincidentally, depending on 
the kind of food scarcity. As a consequence, most real cave 

animals show no or only minor signs of malnutrition despite 
the low food availability in their environment. 


Not only the intensity but also the quality of the food 
scarcity and the duration of this selective force determine the 
degree of adaptation. Food scarcity in caves can have three 
facets: general food scarcity, periodic food supply, and patchy 
food scarcity. General food scarcity holds for nearly all caves 
and occurs especially in caves with no or low but continuous 
food input. Additionally, many caves are not stable through- 
out the year. Periodic food supply characterizes caves that are 
flooded periodically (normally several times during the rainy 
season) or caves with periodic food input by visiting animals. 
Seasonally flooded caves are subject to severe changes regard- 
ing food input, water quality, oxygen content, temperature, 
and competitors or predators. During the wet season, food 
supply can be very high and even abundant for some weeks 
or months. After exhaustion of these food reserves, animals 
in such caves suffer food scarcity like animals in generally 
food poor caves. Some cave animals have to cope with patchy 
food scarcity. This means that food is not necessarily limited 
but is difficult to find and exploit. Under such conditions, 
cave organisms can be observed aggregated at patchy food 

Food Input 

The basic food resource in most caves is organic matter from 
external origin. Wind, percolating surface water, flooding, 
and streams provide input of many kinds of organic matter, 
such as detritus, microorganisms, feces, and accidental or 
dead animals. Some caves are visited actively by epigean 
animals for shelter or reproduction. Such caves are much 
richer in food than are more isolated ones, because the 
visitors provide an additional food input in the form of their 
feces or their carcasses. Bat guano can present an immense 
source of food for guanobionts. Bacteria and above all micro- 
fungi decompose detritus and guano, thus building the basis 
for a food pyramid in caves. Lava tubes can be rich in food 
due to exudates from roots growing through the ceilings into 
the caves (Poulson and Lavoie, 2000). 


As the only primary producers in caves, a few species of 
chemoautotrophic bacteria may support the survival of cave 
animals, especially in caves that have no natural entrance and 
where the absence of water infiltration from the surface 
excludes the input of photosynthetic food (Sarbu, 2000). 
However, these chemoautotrophic systems are quantitatively 
important in only a few exceptional caves, the best known 
example being the Movile Cave in Romania. 

changes in 





NX life 

changes in life history * 

Wi r- to K-selection) Y 


food spectrum 


higher survival rate 
of the young 



adult size t\ Y 
reduced A L\ 
energy demand 

DX energy demand 

smaller £\ fat content 
swim-bladder /N 

|| fat-accumulation 1 

Adaptation to Low Food D, 

fright resistance 

routine MR 


changes in 
feeding behavior 
and locomotion 

(see left) 

“improved * 

multiplied and 
improved sensory 


energy demand 

standard MR 
(see above) 

during starvation 

LD reduced 
mass-specific MR 

routine MR 
(see above) 

FIGURE 1 Network of characters observed in many cave animals and their causes, consequences, and interrelations. Closed arrowheads require something; 

open arrowheads allow something. (Reprinted from Hiippop, K., in Subterranean Ecosystems, Wilkins, H. et al., Eds., Elsevier Press, Amsterdam, pp. 159-188. 

With permission.) 

Influence of Cave Type 

The amount of food supply in caves depends on the cave 
type, on surface connections, and on the geographic location. 
Generally, the food supply in tropical and subtropical caves is 
greater than in temperate ones because the biomass in the 
tropical epigeum is greater and its production is mostly un- 
interrupted (Poulson and Lavoie, 2000). As a consequence, 
selection pressure can be expected to be weaker, the evolu- 
tionary rate slower, and the appearance of troglobites not as 
fast in such caves compared to caves with low energy input 
such as temperate ones. In fact, troglobites are far more 
abundant in temperate zones than in the tropics, and species 
richness in caves is often correlated with the amount of 
available energy. 


A variety of morphological and physiological adaptations and 
changes in feeding behavior are the basis for more efficient 

foraging and increased food-finding ability of cave animals in 
darkness and under conditions of continuous or patchy food 
scarcity compared to surface relatives. Such alterations are 
only advantageous under low food conditions amd in dark- 
ness. In the case of high prey density or in light conditions, 
cave species are inferior to competing epigean relatives. 

Appendages and Sensory Equipment 

The most obvious morphological alterations in cave animals 
are longer legs, antennae, fins and barbels, or enlarged or 
flattened heads. If these body parts bear sensory organs, their 
enlarged surface can be correlated to an increased number 
of chemosensitive or mechanosensitive organs. As a conse- 
quence, an increased sensitivity to chemical and mechanical 
stimulants and changes in feeding behavior are possible. Cave 
animals can detect the food faster and at a greater distance 
from their bodies than can epigean ones and, as a side effect, 
spend less energy for food searching. All of these characters 

6 Adaptation to Low Food 

have been observed in a broad variety of taxa, from 
amphipods over crayfish, isopods, spiders, beetles, and fish to 
salamanders, and more. Cave fish have been studied most 
intensively in this respect, above all those of the famous North 
American cave fish family, the Amblyopsidae, or the Mexican 
characid fish Astyanax fasciatus and cave salamanders. 

In the Amblyopsidae, a positive trend in several of the 
specified troglomorphic features progresses from an epigean 
species over four gradually more cave-adapted species 
(Poulson, 1963). Adaptive alterations to the cave conditions 
are correlated with enlarged associated brain parts, whereas 
smaller optic lobes reflect the reduction of eyes as a conse- 
quence of darkness and uselessness. For cave salamanders, the 
most likely function of the elongated limbs is to raise the 
body and particularly the head above the cave floor to 
increase efficiency of the lateral-line system. They also permit 
the salamanders to search a larger area per unit of energy 
expended and thus increase feeding efficiency. In interstitial 
species, the evolution of appendage length is different than in 
cave species. Due to the small size of the interstitial gaps, 
they tend to have shorter appendages and a more worm-like 
appearance (Coineau, 2000). 


Changes in foraging behavior can also increase the food- 
finding ability. In the darkness of caves, a food-searching 
behavior concentrated on only the two-dimensional bottom 
or other surface areas can be much more economic in time 
and cost than a food search in a three-dimensional space, as 
exhibited by most surface animals in light and what they also 
try to do in darkness. Several cave animals have abandoned 
the shoaling or grouping behavior and adopted a continuous 
moving mode as a consequence of darkness and food scarcity 
in the cave habitat. They compensate for the optically orien- 
tated and spatially limited food searching mode of epigean 
relatives by covering a greater area using chemo- and 
mechanosensors. The amblyopsid cave fish have developed a 
different swimming behavior, referred to as glide-and-rest 
swimming. This behavior, also enabled by the larger fins, 
not only conserves energy but also results in a reduction of 
interference noise for neuromast receptors thus improving 
prey detection. 

Other Factors 

Most cave animals cope with food scarcity by taking a wide 
range of food or exhibiting a different food preference com- 
pared to surface relatives. Sometimes they show a dietary 
shift if one food source becomes scarce. A higher food utiliza- 
tion efficiency in cave animals as adaptation to the food 
scarcity is still not proven. Finally, the improvement of one 
feature sometimes may have more than one positive effect 
on the cave animals. The elaboration of the antennae in 
amphipods not only enhances food-finding ability, and thus 

survivorship, but also improves the mate-finding ability in 
populations with often low densities. Elongated bodies pre- 
sumably facilitate movement through an interstitial medium. 

Back to the Network 

An improved food-finding ability is adaptive predominantly 
in patchily food-limited cave habitats and may be realized 
through changes in foraging behavior and improved ability 
for sensory orientation. The latter includes improvements not 
only in taste and smell senses but also in spatial orientation, 
as is required in the darkness of caves. A multiplied and 
improved sensory equipment can be causally connected with 
longer appendages and, by this, with neoteny. An improved 
food-finding ability itself can reduce the general energy 
demand of cave animals or improve the ability of fat 


Besides a general food scarcity, many cave animals are faced 
with temporal periodicity of food; hence, they need an im- 
proved ability to survive long periods of starvation. Seasonality 
in caves, as already mentioned, is based on periodic flooding 
or on animals visiting the cave periodically, such as bats. 
Normally, this results in annual cycles. 

Fat Accumulation 

The main way to improve the survival capacity in periodi- 
cally food-scarce cave environments is the accumulation of 
large amounts of adipose tissue during food rich seasons. 
High lipid contents have been observed not only in cave 
animals but also in many surface species subjected to seasonal 
changes in food supply. The energy content per gram of 
lipids is roughly twice that of proteins or carbohydrates; 
therefore, fat accumulation is the best way to store energy. 
This may be achieved through excessive feeding, increased 
feeding efficiency, or improved metabolic pathways favoring 
lipid deposition. Cave animals build up fat reserves during 
the food-rich season and store them in their abdominal 
cavity, in subdermal layers, intra- or extracellularly in the 
hepatopancreas (decapods) or in the muscle mass, within the 
orbital sockets of the reduced eye, or in the more or less 
reduced swim bladder (as some fish do). They are able to 
survive starvation periods from several weeks to one year (as 
proven with fish), probably even more. Additionally, cave 
species cope better with starvation periods when metabolic 
rates are reduced, and cave animals can enable their young to 
resist starvation periods by producing eggs with more yolk. 


Amphipods, decapods, remipedes, collembola, beetles, and 
several fish species all over the world have been observed to 

accumulate fat deposits, some up to huge amounts, and to be 
able to survive starvation periods better than their epigean 
relatives. For example, the different stages of various 
morphological adaptations in the five North American 
amblyopsid cave fish species are correlated with the time the 
species have been isolated in caves. The increasing ability 
to cope with food follows the same order (Poulson, 1963), 
which supports the high adaptive value of starvation 
resistance in cave animals. 

Individuals of one hypogean variety of the Mexican 
characid fish Astyanax fasciatus fed ad libitum in the labora- 
tory were able to accumulate fat up to 71% of dry body mass 
compared to only 27% in the conspecific epigean fish variety. 
A good measure to compare the nutritional states of animals 
within one species or within closely related species is 
calculation of the condition factor (CF): 

CF = 100 wl? 

where w is the wet body mass in grams, and / is the body 
length in centimeters. Not until after a starvation period of 
almost half a year did the CF of individuals of the hypogean 
variety of A. fasciatus fall below the CF of the epigean fish in 
an unstarved condition (Table I). 

Back to the Network 

A higher fat content is highly useful in cave animals con- 
fronted with periodic food scarcity in that it increases their 
resistance to starvation. Cave animals may be able to improve 
their capacity of fat accumulation by an improved food- 
finding ability, perhaps supported by a higher food utilization 
efficiency. Lowered metabolic rates reduce energy demand 
during starvation and thus increase starvation resistance. An 
increased starvation resistance possibly also has an influence 
on the survival rate of the young. 

TABLEI Fat Content and Condition Factors of the Epigean 
and One Hypogean Variety of the Mexican Characid Astyanax 
fasciatus in Relation to the Number of Experimental 
Starvation Days 

Days of Starvation 

0 109 174 
Fat content (% wet body mass) 
Epigean fish 9 2 _— 
Hypogean fish 37 28 27 
Fat content (% dry body mass) 
Epigean fish 27 8 — 
Hypogean fish 71 63 62 
Condition factor (100 g cm™) 
Epigean fish 2.0 1.6 — 
Hypogean fish 2.9 2:5 1.9 

Source: Adapted from Hiippop, K., in Subterranean Ecosystems, Wilkins, 
H. et al, Eds., Elsevier Press, Amsterdam, pp. 159-188. 

Adaptation to Low Food 7 


A reduced energy demand is highly adaptive in the food 
scarcity of caves. It reflects a resistance not only to starvation 
during periodic food limitation or to general food scarcity 
but also to food patchiness, low oxygen content, or other 
abiotic factors in the cave environment. The energy demand 
of an animal usually is quantified by its metabolic rate. 
Meaningful information on the metabolic rate is given by the 
measurement of oxygen consumption of the entire organism 
or parts of it. In addition, indirect parameters such as respira- 
tory frequency, resistance to anoxia, ability to survive starva- 
tion periods, body composition, growth rate, gill area, or 
turnover rate of adenosine triphosphate (ATP) give informa- 
tion on the metabolic rate. All of these methods have been 
used, and in most investigations the metabolic rates of the 
hypogean species were found to be more or less lower than 
that of their epigean relatives. 

Aquatic Cave Animals 

Not only in caves but also in interstitial habitats, aquatic 
animals above all practice striking energy economy. Several 
cavernicolous amphipod, isopod, decapod, and fish species 
have been shown to live with metabolic rates much lower 
than those of surface relatives. The most detailed analysis 
of cave adaptation in fish (Poulson, 1963) demonstrates a 
decreasing trend in the metabolic rate from the epigean 
species in the Amblyopsidae over the troglophilic to 
gradually more cave-adapted ones. This trend in the 
metabolic rate is negatively correlated with starvation 
resistance. High fat reserves together with low metabolic rates 
explain the long survival time of the most troglobitic 
amblyopsid species when starved. However, high fat contents 
may lead to misinterpretations of metabolic rates. Studies in 
the Mexican cave fish A. fasciatus regarding metabolic rate, 
body composition, and starvation survival revealed lower 
metabolic rates in one cave form as compared to the epigean 
relative and very high fat contents (Hiippop, 2000) (see 
Table I). Because fat tissue is known to have a relatively 
low maintenance metabolism compared to other tissues or 
organs, lean body mass or bodies with comparable fat 
contents should be preferred as a metabolic reference to avoid 
misinterpretations of the metabolic rate. The recalculation of 
metabolic rates in A. fasciatus resulted in nearly identical 
values in both varieties of the fish species. Although 
obviously adapted to a periodically low energy environment, 
as can be seen from the high fat content, the hypogean 
A, fasciatus were not yet able to reduce their energy turnover 
rate in adaptation to a general food scarcity. 

Terrestrial Cave Animals 

Only a few investigations on metabolic rates of terrestrial 
cave animals exist. Although food scarcity generally is even 
greater in the terrestrial than in the aquatic cave environ- 

8 Adaptation to Low Food 

ment, only a few cave arthropod species were found to show 
a tendency toward energy economy. 


Every activity increases the energy consumption of animals. 
The standard metabolic rate (z¢., the lowest oxygen con- 
sumption rate that can be measured during a test) excludes 
motion activity and is a measure of the physiological adapta- 
tion of cave animals to food scarcity. However, the routine 
metabolic rate (z.e., the mean metabolic rate over 24 hours 
which includes spontaneous activity) is a more appropriate 
index of actual energy expenditures in nature; it actually may 
have the highest rank among other adaptations reducing 
energy demand in cave animals. The routine metabolic rate 
may be reduced in cave animals due to reduced motion 
activity, to changed motion patterns (temporal as well as 
morphological), to reduced or no longer practiced aggressive 
and territory behavior, or to reduced fright reactions. 
Actually, in most cave animals activity is reduced. Although 
an increase in food-finding ability in cave animals often 
seems to go along with an increase in food-searching activity, 
changed motion patterns result in a reduction of energy 
expenditure, sometimes to a fantastic extension. For example, 
in the most cave-adapted species of the amblyopsid fish in 
North America, over 90% of the total energy savings by adap- 
tations are based on the reduced activity (Poulson, 1985). 

Excitement and Aggression 

Metabolic rates definitely are elevated by an animal’s reaction 
to disturbance (excitement) and by aggressive behavior. 
Interior activity, or excitement without expression in motion 
activity, elevates the standard metabolic rate. External 
activity, including motion activity resulting from excitement 
or aggression, increases the routine metabolic rate. An 
increased resistance to disturbance has been shown to be 
important for energy economy in the amblyopsid fish 
(Poulson, 1963). The generally low standard and routine 
metabolic rates of cave amblyopsids and their resistance to 
disturbance are interpreted not only as adaptations to the 
reduced food supply, by a factor of about 100 compared to 
the surface, but also as a by-product of relatively stable cave 
conditions and a general lack of predators in the amblyopsid 
cave environment (c.f’, regressive evolution). 

The Conflicts of Body Size 

In addition to food scarcity, the interstitial habitat is 
constrained by the grain size of the substrate. Interstitial 
animals are constrained in size and shape due to the small size 
of the interstitial gaps. They often are very small and have 
shortened appendages excluding the posterior appendages, 
which tend to be elongated (Coineau, 2000). In this special 
subterranean habitat, the motion activity for food searching 

in interstitial animals can actually be increased and can result 
in an elevated routine metabolic rate compared to surface 
forms (Danielopol et a/., 1994). Often, interstitial species are 
much smaller than their surface relatives, so their higher 
mass-specific routine metabolic rate corresponds to a smaller 
routine metabolic rate per individual. The energy reserves 
of larger animals last longer and are more resistant to food 
shortage than of small animals because the metabolic rate 
of animals is not directly proportional to body mass but is 
related to mass by the following equation: 

Metabolic rate = aM? 

where a is the intercept, M is the body mass, and 6 is the 
mass exponent or slope smaller than 1 (Withers, 1992). Con- 
sequently, subterranean animals have to resolve the conflict 
between two advantages: (1) to be larger with a lower energy 
demand per unit mass but a higher one per individual, or (2) 
to be small, thus requiring less energy per animal and/or 
being able to live in crevices. Interstitial forms may have 
reduced their body size to fit better into the small crevices 
and to cope better individually with food scarcity in their 
special habitat (Danielopol et al, 1994). 

Ectothermy and Neoteny 

Troglobites are exclusively ectotherms. The generally very 
low metabolic rates of ectotherms (only 10 to 20% or even 
less that of similar sized endotherms) are the basis for their 
success in zones characterized by limited resource supplies, 
such as shortages in food, oxygen or water. Ectotherms can 
utilize energy for reproduction that endotherms are forced to 
use for thermoregulation. Finally, ectotherms are able to exploit 
a world of small body sizes unavailable to endotherms. Body 
sizes less than 2 grams are not feasible for endotherms 
because the curve relating metabolism to body mass becomes 
asymptotic to the metabolism axis at body masses lower than 
2 grams (Withers, 1992). 

Within the ectothermic vertebrates, only fish and 
amphibians evolved cave species. Because they are the largest 
animals in cave communities, they usually represent the 
highest trophic level in the cave food web and can survive in 
large populations only in relatively food-rich caves. Most 
troglobitic salamanders are aquatic and show the retention of 
larval characters, known as neoteny, which enables them to 
survive in the relatively less food-scarce aquatic cave habitat 
compared to the terrestrial cave habitat (Culver, 1982). 
Finally, suppression of the energetically expensive metamor- 
phosis in hypogean salamanders can be interpreted as an 
adaptation to general food scarcity. 

Hypoxic Conditions 

Besides food scarcity, numerous cave or interstitial species 
have to cope with temporary or permanent hypoxic 
conditions. Also, this character of some cave environments 

forces reduced metabolic rates and has been proven in 
crustaceans and fish. In contrast to surface species, several 
hypogean species have no sharp break in the oxygen uptake 
lines under depleting oxygen concentrations. This absence 
of a discontinuity in the oxygen uptake line is called oxy- 
regulation and is considered to be adaptive in environments 
characterized by variable oxygen conditions (Danielopol et 

al., 1994). 

Character Reduction 

Many features become reduced during the evolution of cave 
animals. This regressive evolution generally is described as 
the reduction of “functionless” characters in cave animals 
and not only concerns structural but also behavioral and 
physiological traits. It obviously should be an advantage for 
cave animals to use the energy saved from not building 
or maintaining useless characters when living in strongly 
food-limited cave environments by transferring it to the 
development or support of other characters or to growth, 
reproduction, or survival during starvation periods. There 
exist a few hints among beetles and spiders of this possible 
strategy of cave animals to adapt to a food-restricted cave 
environment. Nevertheless, more often the reduction of cha- 
racters in cave animals seems to be the result of accumulated 
neutral mutations. 

Back to the Network 

A reduced energy demand is adaptive mainly in those caves 
that are generally low in food. A lowered interior activity, 
meaning a reduced standard metabolic rate, can be the result 
of an increased fright resistance, a lowered aggressive 
behavior, and possible changes in biochemical mechanisms 
(e.g, ATP turnover rates). Reduced motion activity (i¢., a 
lowered routine metabolic rate) can be achieved by means 
of reduced body movement to escape (a reduced number of) 
predators or for aggression, changes in feeding behavior and 
locomotion, and an improved sensory orientation resulting 
in fewer movements for food searching. In the end, reduced 
metabolic rates result in a higher availability of energy for 
growth and/or a greater resistance to starvation. Additionally, 
character reduction, reduced growth rates, and smaller adult 
body size have the ability to reduce energy demand in cave 
animals. The reduced energy demand in cave animals can 
have two effects. Under the aspect of metabolic span, a 
reduction per time is correlated with a prolonged lifetime 
combined with iteroparity. On the other hand, the reduction 
per individual life enables higher survival rates of individuals 
or even the increase of population size. 


The extremes of the spectrum of life history adaptations are 
characterized as r- and K-selection. Whereas r-selection (7 

Adaptation to Low Food 9 

being the slope of the population growth curve) means a 
trend toward high population growth rate under temporarily 
good conditions in relatively unpredictable and changing 
habitats, K-selection (K being the carrying capacity of the 
habitat) can be realized only in more predictable and stable 
habitats, and the appropriate fitness measure is the maximum 
lifetime reproduction. K-selected species are characterized by 
low or no population growth; they have reached a maximal 
K. This situation is connected with fewer but larger and 
more nutrient-rich eggs, increased time required for 
hatching, prolonged larvae stage, generally decreased growth 
rate, delayed and perhaps infrequent reproduction, increased 
longevity, and parental care. Many cave animals show a 
couple of these characters, demonstrating a trend toward 
more K-selection in cave species. The life history of cave 
animals has been the subject of some reviews; the main inves- 
tigations were done on a variety of invertebrates, particularly 
crustaceans and arthropods, and on fish and salamanders 

(Hiippop, 2000). 

Egg Size 

The adaptive value of fewer but bigger eggs, not only for 
cave fish but for all animals living in food-poor habitats, is 
obvious. Bigger eggs with more energy-rich yolks release 
bigger larvae. These larvae have a bigger head with larger 
mouth, so they can start external feeding on a larger spec- 
trum of food particles and have a better chance to survive. 
Furthermore, bigger larvae may have a higher resistance to 
starvation and a higher mobility for food searching and for 
effective escape reactions. The reduction of the clutch size 
can finally result in a single larva per reproductive season that 
possibly never feeds, as in cave beetles. 

Growth Rate 

A reduced growth rate in cave animals is adaptive to food 
scarcity because it means a reduced energy demand per time. 
More animals can live on a defined amount of food, or 
a defined group of animals can survive longer on it. A 
reduction in energy demand per unit time through a lower 
metabolic rate, together with a reduction of absolute and 
relative costs of reproduction, can make possible an increase 
of population density and hence an increase in the number 
of females actually breeding per year. 


There is evidence that the total metabolic turnover in a 
lifetime not only of endotherms but also in ectotherms is the 
product of the energy turnover rate and the duration of 
life, called the metabolic span. Generally, lower metabolic 
rates and slower growth rates (that is, “slower living”) are 
connected with increased longevity. Because the reproductive 
success of an animal might be defined by the ability to 

10 Adaptation to Low Food 

live long enough to survive the gap between good years, 
an increased lifetime in cave animals is advantageous in a 
generally food-scarce environment or in caves where 
relatively food-rich reproductive seasons occur irregularly. 
The increased longevity of cave animals connected with a 
delay in maturity and a trend from semelparity to iteroparity 
means that the population is less likely to disappear in years 
when food supply is too low to allow females to produce 


A Case Study 

Extremely prolonged lifetimes of more than 150 years are 
known among North American cave crayfish; however, the 
amblyopsid cave fish, intensively investigated by Poulson 
(1963), are the best known example of how cave animals 
adapt their life history to food scarcity. Within this group of 
fish species it is obvious how cave animals with increasingly 
slower energy turnover rates have increasingly prolonged life 
cycles connected with many increasingly K-selected features, 
such as bigger and fewer eggs with prolonged developmental 
time, prolonged branchial incubation time (= parental care), 
bigger larvae at first external feeding, reduced growth rate, 
delayed maturity, and multiplied chances to reproduce with 
increasing cave adaptation (Table I). Population growth rate 
and population density decrease with phylogenetic age of 
the cave species, and the population structure shifts toward 

adults (Poulson, 1963). 

Back to the Network 

Life history changes toward more K-selected characters 
are correlated with a prolonged lifetime (and consequently 
iteroparity) and/or with a shift toward more adults in the 

TABLE II Some Life History Characters of the Epigean 
Amblyopsid Fish Chologaster cornuta Compared to the 
Troglobitic Amblyopsis rosae 

C. cornuta A. rosae 

Egg size (mm) 0.9 to 1.2 1.9 to 2.2 
Developmental time few weeks 5 to 6 
Larval size (mm) When hatching 3 5 
When leaving gill cavity 8 12 
Adult size (mm) 23 to 55 36 to 48 
Female body mass (g) 0.93 1.25 
Adult maturity (year) 1 3 
Maximal life span (years) i 5 to6 
Number of mature ova per female 98 23 
Average number of reproduction per lifetime 1 0.6 
Maximum number of reproduction per lifetime 1 3 

Source: Adapted from Poulson (1963, 1985) and Culver (1982). 

population. Additionally, more K-selection may include 
bigger and fewer eggs and longer incubation and brood care, 
giving the offspring a higher chance of survival. A reduced 
growth rate and a smaller adult size may save energy. A 
prolonged developmental period may result in neoteny 
which in turn can have an influence on food-finding ability 
through appendage lengthening and changes in foraging 

See Also the Following Articles 
Food Sources * Adaptive Shifts 


Coineau, N. (2000) Adaptations to interstitial groundwater life. In 
Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.E, 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 189-210. 

Culver, D.C. (1982) Cave Life: Evolution and Ecology, Harvard University 
Press, Cambridge, MA. 

Danielopol, D.L., M. Creuzé des Chatelliers, F Mésslacher, P. Pospisil, and 
R. Popa (1994) Adaptation of crustacea to interstitial habitats: a practical 
agenda for ecological studies. In Groundwater Ecology (J. Gibert, D.L. 
Danielopol, and J.A. Stanford, Eds.). Academic Press, New York. 

Hiippop, K. (2000) How do cave animals cope with the food scarcity in 
caves? In Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.F, 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 159-188. 

Poulson. T.L. (1963) Cave adaptation in amblyopsid fishes. Am. Midl. Nat., 
70, pp. 257-290. 

Poulson, T.L. (1985) Evolutionary reduction by neutral mutations: 
plausibility arguments and data from amblyopsid fishes and linyphiid 
spiders. Natl. Speleol. Soc. Bull., 47, pp. 109-117. 

Poulson, T.-L. and K.H. Lavoie (2000) The trophic basis of subsurface 
ecosystems. In Subterranean Ecosystems (H. Wilkens, D.C. Culver, and 
W.E, Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 231-249. 

Sarbu, S.M. (2000) Movile Cave: a chemoautotrophically based 
groundwater ecosystem. In Subterranean Ecosystems (H. Wilkens, D.C. 
Culver, and W.E, Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 

Withers, PC. (1992) Comparative Animal Physiology, Saunders College 
Publishing, Fort Worth, TX. 

Adaptation to Low Oxygen 

Frédéric Hervant and Florian Malard 
Université Lyon 


Before the late 1970s, ecological studies were carried out 
mainly in the unsaturated zone of karst aquifers, more 
particularly in caves. Because cave water bodies are exposed 
to the atmosphere, they are usually saturated with oxygen, 
but dissolved oxygen (DO) was not measured routinely, 
and oxygen availability in subterranean biotopes was rarely 
considered as a key ecological factor that governed the 
occurrence and spatiotemporal distribution of hypogean 
animals. Biological activity, animal density, and organic 

matter content in the groundwater were assumed to be too 
low to induce oxygen deficiency; therefore, the possibility 
that hypogean organisms may have to face hypoxic stress was 
not considered. Physiological studies essentially concerned the 
adaptive responses of hypogean animals to low food supply 
including their higher food-finding ability, strong resistance 
to starvation, and reduced metabolism (Hiippop, 2000). 

Since the 1970s, this view has been reevaluated by many 
researchers who have frequently reported low dissolved 
oxygen concentrations from shallow groundwater in un- 
consolidated sediments, more particularly in the hyporheic 
zone of streams (Boulton et a/., 1992; Strayer et al., 1997); 
the groundwater environment has been described as being 
hypoxic or weakly oxygenated. Also, several authors have 
conducted laboratory studies to test the resistance and 
adaptive strategies of animals in response to low oxygen 
(Malard and Hervant, 1999). 

The ensuing material begins by examining the oxygen 
status of different groundwater systems including deep- and 
shallow-water-table aquifers and the hyporheic zone of 
streams. Then, based on laboratory studies conducted by one 
of the authors (Hervant et al, 1996-1999; Malard and 
Hervant, 1999), we examine the behavioral, respiratory, and 
metabolic responses of groundwater organisms, especially 
crustaceans, to low oxygen concentrations. Finally, we 
suggest that the selection of organic-matter-rich habitats in 
groundwater increases the risk of facing hypoxic stress. 


Because of permanent darkness, there is no photosynthesis in 
groundwater, thus no production of oxygen; therefore, the 
oxygen status of groundwater is determined by the rate of 
oxygen transport from the surface environment and by the 
rate of oxygen consumption in the — subsurface. 
Replenishment of dissolved oxygen occurs by air diffusion 
from the unsaturated zone or by recharge with normoxic 
rainwater or river water. Fluctuations of the groundwater 
table enhance air entrapment, thereby increasing dissolved 
oxygen transfer from entrapped air. Oxygen transport within 
groundwater may occur as a result of oxygen diffusion, 
convection currents caused by heat transfer, and advection of 
water in response to hydraulic gradients. The diffusive 
movement of oxygen in water is negligible, and convective 
currents are limited in groundwater by sediment and weak 
thermal gradients (the geothermal gradient is usually about 
0.01°C m') (Malard and Hervant, 1999); therefore, oxygen 
transport in groundwater is primarily due to advective move- 
ment of water in response to hydraulic gradients. Because 
groundwater velocity is usually low (i.e, 10°°-10~% ms"), 
the available flux of DO in groundwater is much slower than 
in surface waters. On the other hand, oxygen consumption 
by microorganisms is limited in many aquifers by the avail- 
ability of biodegradable organic carbon. Consequently, DO 

Adaptation to Low Oxygen 11 

may persist at considerable distances from the recharge zone 
in deep-water-table aquifers where soil-generated dissolved 
organic carbon is completely degraded during the transit 
of infiltrating water in the unsaturated zone. In confined 
aquifers of the Ash Meadows basin (southcentral Nevada), 
Winograd and Robertson (1982) sampled ground water with 
2 mg L™ O, at a distance of 80 km from the recharge zone. 
In contrast, dissolved oxygen may be totally consumed over 
very short distances (z.e., a few meters or even centimeters) 
in shallow-water-table aquifers or in the hyporheic zone of 
rivers because of the input of soil- or river-labile dissolved 
organic carbon. Malard and Hervant (1999) reported strong 
variability among groundwater systems in the length of 
underground pathways for dissolved oxygen. Based on cross- 
system comparison from literature data, these authors 
suggested that differences among hyporheic zones reflect 
variation in the contact time of water with sediment, whereas 
differences among confined aquifers are primarily a result of 
differences in the rate of DO consumption. 


Small-scale investigations of oxygen distributions in sub- 
surface waters revealed strong variations over distances of a 
few centimeters or meters. This heterogeneity, an essential 
feature of the groundwater environment, was observed in a 
number of subsurface water habitats, including the saturated 
zone of karst aquifers, the water-table region of deep- and 
shallow-water-table porous aquifers, the halocline of anchia- 
line caves, the hyporheic zone of rivers, and the interstitial 
environment of marine and freshwater beaches (Malard and 
Hervant, 1999) (Fig. 1). Small-scale spatial heterogeneity in 
DO reflects changes in sediment composition and structure, 
subsurface water flow velocity, strength of hydrological 
exchanges with the surface environment, dissolved and 
particulate organic matter content, and activity of micro- 
organisms. Strong temporal changes in DO may also occur 
in the hyporheic zone of streams as well as in the recharge 
zone of aquifers, but these fluctuations are strongly damped 
with increasing distance from the stream and the recharge 
zone. Whereas diminished oxygen concentration is typically 
not a rule for the groundwater environment, the high 
spatial heterogeneity of DO at meso- (meter) and micro- 
(centimeter) scales is considered a peculiarity of groundwater 
habitats (Malard and Hervant, 1999). This implies that 
animals living in groundwater have to experience highly 
variable oxygen concentrations as they move through a 
mosaic of patches with contrasted DO concentrations. 


Results of field studies reveal that most animals can be found 
living in a wide range of DO, even anoxia in some cases 



30-m deep 0 
water table 



Depth below the water table (cm) 

Adaptation to Low Oxygen 


O, (mg.L!) 
0 1 2 3 4 


Depth below the soil surface (m) 


0 0.1 0.2 O, [mM] 
—— ne 
200 1 
2 Eh 
250 2; 
0 2 4 6 8 a4 
Cc O, (mg.L) < 
0 6 
SE 10 
z x) 15 -100 0 +100 +200 +300 Eh [mV] 
= oe -20 0 0.2 0.4 H,S [mM] 
ee 25 
A a0 

0 1 2 3 4 5 6 =7 

O, (mg.L-!) 

Stream length (10 m) 

: CBW] © 

(aI €) YIpia weens 


FIGURE 1 Mesoscale (meter) and microscale (centimeter) heterogeneity in dissolved oxygen concentrations in (A) a deep water-table aquifer (Ronen e¢ al., 
1987); (B) a shallow water-table aquifer (Pospisil et a/., 1994); (C) an anchialine cave (Yaeger, 1994); (D) the interstitial environment of a sandy marine beach 
(Revsbech and Jorgensen, 1986); and (E) the hyporheic zone of a backwater (Danielopol, 1989). 

(Malard and Hervant, 1999). Based on 700 faunal samples 
collected in the hyporheic zone of several desert streams in 
Arizona, Boulton et al, (1992) examined the field tolerance 
to DO of 23 common taxa. These authors showed that most 
taxa could be encountered in subsurface waters with less than 
1 mgL! O. Strayer et al. (1997) obtained similar results 
based on 167 samples of hyporheic invertebrates collected 
at 14 sites in the eastern United States; however, in both 
studies, several animal groups, particularly crustaceans and 
insects, occurred more frequently in well-oxygenated 
sediments than where oxygen was scarce. Several crustaceans 
(including the amphipod Niphargus hebereri Schellenberg, 
the thermosbaenacean Monodella argentarii Stella, and some 
species of remipedes) are known to develop dense popula- 
tions in the sulfide zone of anchihaline caves (Malard and 
Hervant, 1999). It is not yet clear, however, whether these 
crustaceans permanently live in deoxygenated waters or if 
they temporarily seek shelter in a more aerated environment. 
A diversified aquatic fauna was also found to live in the 
uppermost hypoxic layer (DO < 0.3 mg L™' O3) of sulfidic 
groundwater in Movile Cave, Romania (Malard and 
Hervant, 1999). On the other hand, the paucity of fauna in 
extensive areas of hypoxic groundwater suggests that most 
hypogean taxa are probably not able to survive a lack of 
oxygen for very long. Reports of groundwater animals among 
poorly oxygenated groundwaters are equivocal because their 
presence may be due to either tolerance to low oxygen 
concentrations by hypogean invertebrates or the existence 
of microzones of high dissolved oxygen. Recent detailed 
laboratory studies on the behavioral, respiratory, and 
metabolic responses of several subterranean crustaceans to 
anoxia have enabled us to define more precisely the degree of 
tolerance of groundwater animals to low oxygen concentra- 
tions and to elucidate some the mechanisms responsible 
for extended survival under oxygen stress (Hervant et al, 



Most data available on adaptations to low oxygen in ground- 
water organisms arise from comparisons of the adaptive 
responses of three hypogean crustaceans (the amphipods 
Niphargus virei Chevreux and Niphargus rhenorhodanensis 
Schellenberg and the isopod Stenasellus virei Dolfus) and two 
epigean crustaceans (the amphipod Gammarus fossarum 
Koch and the isopod Asellus aquaticus L.) to anoxia (Hervant 
et al., 1996-1999; Malard and Hervant, 1999). The 
responses of these five crustaceans were examined in darkness 
(at 11°C) under three different experimental conditions: 
anoxia, normoxia following an anoxic stress (postanoxic 
recovery), and declining DO concentration (from 10.3 to 
0.7 mg L" O,). Information gained from these experiments 
concerned: (1) the lethal time for 50% of the population 
(LT 50%) in anoxia; (2) the locomotory activity (number of 
periods of locomotion per minute) and the ventilatory 

Adaptation to Low Oxygen 13 

activity (number of pleopods beats per minute) of animals; 
(3) the oxygen consumption rates in normoxia and under 
declining Po, (in order to determine the critical partial pres- 
sure of oxygen, Pc); and (4) changes in the concentrations 
of key metabolites such as high energy compounds (e.g, 
adenosine triphosphate [ATP], phosphagen), anaerobic 
substrates (glucose, glycogen, amino acids), and anaerobic 
end products (lactate, alanine, succinate, malate, propionate). 
Groundwater crustaceans show survival times (LT 50%) 
of about 2 to 3 days under anoxia (Table I) and several 
months in moderate hypoxia (Danielopol, 1989; Malard and 
Hervant, 1999). They are much more resistant to oxygen 
deprivation than morphologically closed epigean species 
whose LT 50% values range from a few hours to one day. 
Hypogean fishes, crayfishes (Huppép, 2000), and urodele 
amphibians (Hervant, unpublished data) also display much 
higher survival times than their related epigean species. 


Every spontaneous or stress-induced activity increases ener- 
getic expenditures and, therefore, oxygen consumption. 
Some hypogean species show a very low activity rate when 
compared to morphologically closed epigean animals (Table 
I). This reduced activity has been interpreted to be a result 
of (1) food scarcity, stable environmental conditions, and 
general lack of predators and territorial behavior; (2) cramped 
interstitial habitats; and/or (3) low oxygen availability (Hervant 
et al., 1996; Hiippop, 2000). A low activity rate in ground- 
water animals which is often associated with an increase in 
food-finding ability is an efficient adaptation to low food 
supply because it results in a reduced energy demand. 
Epigean Gammarus fossarum and Asellus aquaticus respond 
to experimental anoxia by a marked hyperactivity (Table I) 
that, in natural conditions probably corresponds to an 
attempt to move to more oxygenated habitats (z¢, to a 
behavioral compensation). In contrast, all three hypogean 
crustaceans show a drastic reduction in locomotory activity 
under lack of oxygen (Table I). This adaptive behavior 
reduces energy expenditure during oxygen deprivation (by 
decreasing the oxygen consumption) and therefore increases 
survival time in deoxygenated groundwater. This “sit-and- 
wait” strategy, very advantageous under hypoxic (and/or 
food limited) environments, was also observed in starved 
stygobite crustaceans (Hervant and Renault, 2002) and cave 
salamanders (Hervant et a/., 2001). Danielopol et al. (1994) 
then Mosslacher and Creuzé des Chitelliers (1996) showed 
that the hypogean isopod Proasellus slavus actively explored 
its interstitial environment but slowed down exploratory 
movement under severe hypoxic conditions (i.e., 0.1 mg L™ 
O,). The lack of an escape behavior is probably not universal 
for hypogean aquatic organisms, however, because active 
animal migration in response to fluctuating DO concen- 
trations has already been documented in natural biotopes 
and in a laboratory flume (Henry and Danielopol, 1999). 

14 Adaptation to Low Oxygen 


Comparison of Locomotory and Ventilatory Activities, Oxygen Consumption (SMR), and Intermediary and Energetic 

Metabolism During Normoxia, Anoxia, and Postanoxic Recovery in Two Surface-Dwelling and Three Groundwater Crustaceans 

Epigean Hypogean 

G. fossarum A. aquaticus N. rhenorho. N. virei S. virei 
Survival during anoxic stress 
Lethal time for 50% of the population 6.3 h 19.7 h 46.7 h 52.1h 61.7h 
Number of periods of locomotion per minute (in normoxia) 19.3 6.7 12.5 1.4 4.5 
Changes during anoxia 4 \ x X x 
Number of pleopods beats per minutes (in normoxia) 60 43 44 48 38 
Changes during anoxia Ls LS \ \ \ 
Respiratory metabolism 
Normoxic O, consumption (SMR; ML )2/g dw/h) 940 1325 605 305 425 
O, critical pressure (mg.L"') 3.6 3.8 2.0 1.8 2.0 
Body energy stores (in normoxia) 
Stored glycogen (umol/g dw) 110 165 356 243 307 
Stored arginine phosphate (lumol/g dw) 70 10.5 26.5 26.5 31.5 
Post-anoxic recovery 
Glycogen re-synthesis rate (mol/g dw/min) 0.003 0.01 0.07 0.03 0.04 
% of glycogen re-synthesis 8 19 47 49 53 

Source: Adapted from Hervant et al., 1995-1998; Malard and Hervant, 1999. 


Subterranean animals generally have reduced standard 
metabolic rates (SMRs) and reduced routine metabolic rates 
(RMRs) when compared to their surface-dwelling counter- 
parts (Hiippop, 2000). For example, the cave salamander 
Proteus anguinus Laurenti and the groundwater crustaceans 
Niphargus virei, Niphargus rhenorhodanensis, and Stenasellus 
virei show SMRs in normoxia from 1.6 to 4.5 times lower 
than their epigean relatives (Table I) (Hervant et al., 1998). 
Low SMRs and RMRs among hypogean animals have long 
been interpreted only as an adaptation to low food avail- 
ability (Hiippop, 2000); however, reduced SMRs in some 
populations of subterranean animals may also reflect an 
adaptation to low oxygen concentrations in the groundwater 
(Hervant and Renault, 2002). A reduced metabolic rate (z.e., 
low energetic requirements) results in a lower oxygen removal 
rate and an increased survival time in anoxia. 

The hypogean amphipods N. virei and N. rhenorhoda- 
nensis and the hypogean isopod S. virei reduce their venti- 
latory activity during anoxic stress (Table I), thereby limiting 
their energy expenditure. Several animals contract an oxygen 
debt during oxygen deprivation that is repaid upon return 
to normoxia (Herreid, 1980). The repayment of this debt 
during recovery from anaerobic stress involves a significant 
increase in SMR. The oxygen debt of hypogean N. virei 
and LV. rhenorhodanensis is 2.2 to 5.3 times lower than that of 
epigean G. fossarum and A. aquaticus (Hervant et al, 1998). 
This lower oxygen debt indicates a reduced energetic expen- 
diture (z.e., an energy sparing) which is probably linked to 
lower locomotory and ventilatory activities during anoxia. 

The resistance to anoxia (and/or long-term fasting) 
displayed by numerous hypogean species may be explained 
by their ability to remain in a prolonged state of torpor. This 
state enables hypogean organisms to tolerate a prolonged 
reduction in oxygen (and/or food) availability by maximizing 
the time during which metabolism can be fuelled by a given 
energy reserve (or a given food ration). This supports the 
classic suggestion that difficulties in obtaining food in 
stressful environments may select for conservative energy use 
(Hervant and Renault, 2002). These adaptive responses may 
be considered for numerous subterranean organisms as an 
efficient energy-saving strategy in a harsh and unpredictable 
environment where hypoxic (and/or starvation) periods of 
variable duration alternate with normoxic periods (and/or 
sporadic feeding events). Hypoxia-tolerant (and/or food- 
limited) groundwater species appear to be good examples 
of animals representing a low-energy system (Hervant et al, 
2001; Hervant and Renault, 2002). 

Theoretically, aerobic organisms can be described as meta- 
bolic conformers or regulators if their oxygen consumption 
varies directly with or is independent of Po, respectively; 
however, these two types are merely the two ends of a large 
spectrum of respiratory responses of species. Animals never 
fully belong to one or the other type, and below a critical Po, 
(Pc) a regulator becomes a conformer (Herreid, 1980); there- 
fore, Pc is a good indicator of the tolerance or adaptation of 
an organism to low Po). Gammarus fossarum, A. aquaticus, 
N. virei, S. virei and N. rhenorhodanensis are able to maintain 
a relatively constant rate of oxygen consumption relatively 
constant that is independent of Po, and between normoxia 
and Pc (Hervant et al, 1998). However, the Po, at which 

respiratory independence is lost is significantly lower for 
hypogean species than for surface-dwelling ones (Table I). 
This implies that groundwater species are able to maintain an 
aerobic metabolism for a longer time in declining Po, 
(progressive hypoxia) instead of partly switching to a low- 
energy anaerobic metabolism. The maintenance of an aerobic 
metabolism (ze. survival at a lower energetic cost) under 
hypoxic conditions is partly due to the lower SMR of 
hypogean animals. Mésslacher and Creuzé des Chatelliers 
(1996) also found that the RMRs of an unpigmented and 
eyeless subterranean form of A. aquaticus decreased with 
decreasing environmental DO concentration. In contrast, 
Danielopol et al. (1994) showed that the hypogean isopods 
Proasellus slavus, Proasellus strouhali, and a blind population 
of A. aquaticus maintain a high ventilation activity inde- 
pendent of the external oxygen concentration (ze., from 0.1 
to 9mgL™! O,). However, the occurrence of respiratory 
regulation over such a wide range of DO concentrations had, 
until then, not been observed among epigean crustaceans 
(Malard and Hervant, 1999) and had rarely been observed in 
other invertebrates (Herreid, 1980). 

Metabolic Responses During Anoxia 

A number of biochemical adaptations that permit extended 
survival under prolonged hypoxia or anoxia have been 
identified in various well-adapted epigean groups, especially 
marine annelids and intertidal mollusks (Fields, 1983). These 
include the maintenance of high reserves of fermentable 
fuels (such as glycogen and amino acids) in some tissues 
under normoxic conditions, the use of anaerobic pathways to 
enhance ATP yield and to maintain redox balance during low 
oxygen conditions, and mechanisms for minimizing meta- 
bolic acidosis often associated with anaerobic metabolism. 
Nevertheless, epigean crustaceans have been recognized as 
being of a “modest anaerobic capacity” without special and 
efficient mechanisms of anaerobic metabolism (Zebe, 1991). 

Examination of biochemical (z.¢., intermediary and energy 
metabolism) responses of hypogean crustaceans N. virei, N. 
rhenorhodanensis, and S. virei during anoxic stress show that, 
similarly to epigean crustaceans (for review, see Zebe, 1991), 
anaerobic metabolism does not lead to a high ATP produc- 
tion rate (Hervant et al, 1996, 1997). The five crustaceans 
studied respond to severe experimental anoxia with a slight 
improvement of a classical anaerobic metabolism that is 
characterized by a decrease in ATP and phosphagen (arginine 
phosphate, representing an immediate source of ATP), a 
coupled utilization of glycogen and some amino acids 
(mainly glutamate), and the accumulation of lactate and 
alanine as end-products. The only difference is that both 
Niphargus species also accumulate a low proportion of succi- 
nate, which slightly enhances ATP yield during anaerobiosis 
(Fields, 1983). Lactate is largely excreted by all five 

Adaptation to Low Oxygen 15 

crustaceans. This excretion, which is unusual for crustaceans 
(Zebe, 1991), can be considered a simple way to fight against 
metabolic acidosis linked to anaerobic end-product 
(including H*) accumulations. 

There is a striking difference in the respective amounts of 
glycogen and phosphagen stored by epigean and hypogean 
crustaceans. Glycogen body reserves (Table I) are 1.5 to 3.2 
times higher in the three hypogean crustaceans studied than 
in the epigean G. fossarum and A. aquaticus (Hervant et al, 
1996, 1997) but are also higher than those reported for 
all surface-dwelling crustaceans, even those most tolerant 
of anoxia or hypoxia (Malard and Hervant, 1999). High 
amounts of fermentable fuels result in a more sustained 
supply for anaerobic metabolism, thereby increasing survival 
time during oxygen deprivation. Moreover, glycogen utiliza- 
tion rates and lactate production rates are significantly lower 
in hypogean crustaceans. This finding is probably linked 
to lower SMRs and to the reduction of locomotory and 
ventilatory activity during anoxia (Table I) (Hervant et al, 

1996, 1998). 

Metabolic Responses During a Postanoxic Recovery 

It is ecologically very important for organisms to recover 
quickly and completely from hypoxic or anoxic stress when 
oxygen is available once more. This recovery implies a resto- 
ration of high energy compounds (mainly ATP, phosphagen, 
and glycogen), as well as disposal of anaerobic end-products 
(mainly lactate, alanine, and succinate). End-products can be 
disposed of in three different ways during a postanoxic 
recovery phase: complete oxidation, conversion back into 
storage products such as glycogen (via the glyconeogenesis 
pathway, such as glycogen de novo synthesis from lactate, 
amino acids, and/or glycerol), and excretion into the 
medium (Hervant et a/., 1996). Excretion is an important 
mechanism for the disposal of lactate during aerobic recovery 
in the epigean G. fossarum and A. aquaticus. This is a costly 
strategy because it implies a loss of energy-rich carbon 
chains. In contrast, hypogean crustaceans preferentially use 
glyconeogenesis to convert lactate into glycogen stores. The 
existence of glyconeogenesis has already been demonstrated 
in several crustaceans (Hervant et al, 1999), although the 
organ sites for this metabolic pathway have not been 
identified clearly. Recent experiments using injections of 
labeled glucose and lactate (Hervant et al., 1999) revealed 
that the gluconeogenesis rate in NV. virei during postanoxic 
recovery was higher than any rate measured previously 
for epigean crustaceans. Glycogen reserve restoration was 
indeed 2.5 to 6.6 times greater in hypogean species than in 
G. fossarum and A. aquaticus (Table I). This ability to quickly 
resynthesize during recovery periods the body stores depleted 
during lack of oxygen allows groundwater organisms to fuel 
successfully an ensuing hypoxic or anoxic period; therefore, 
groundwater species are well adapted to live in habitats show- 

16 Adaptation to Low Oxygen 

ing frequent and unpredictable alternations of normoxic and 
hypoxic/anoxic phases. 


The occurrence of adaptive strategies in response to low 
oxygen among animals living in an oligotrophic environment 
may seem paradoxical. Notwithstanding the fact that the 
supply of organic matter is typically lower in groundwater 
than in surface water, field ecological studies carried out 
for the last 30 years have shown that several groundwater 
habitats (particularly unconsolidated sediments) have 
reduced DO concentrations. Although a few subterranean 
habitats are known to be organic-matter rich (Huppép, 
2000), low DO concentration in many groundwater habitats 
is most likely attributable to the lack of oxygen production 
and low transport rate of DO than to elevated concentrations 
of organic matter. If food availability drives habitat selection 
in groundwater, hypogean animal populations would prefer- 
entially occur in groundwater biotopes receiving higher 
fluxes of organic matter from the surface environment. These 
habitats are also more likely to exhibit reduced DO concen- 
trations because of increased respiration rates associated with 
the input of organic matter. Thus, the selection of habitats 
having increased food supply increases the probability of 
facing hypoxic stress among hypogean animals. Meantime, 
the development of behavioral, respiratory, and metabolic 
strategies to low food supplies also selects for higher 
resistance to a low oxygen supply. Clearly, the role of food 
availability and the significance of low oxygen supply in 
determining the development of adaptive strategies and 
distribution patterns of animals in groundwater are over- 
lapping aspects that can hardly be treated independently. 

Some metabolic pathways, therefore, that are specifically 
linked to the response to long-term starvation and/or adapta- 
tion to oxygen stress are associated with energy-limited 
subterranean organisms. Hervant and Renault (2002) 
demonstrated that the groundwater crustacean S. virei 
preferentially utilizes lipids during food shortage, in order to 
(1) save carbohydrates and phosphagen, the two main fuels 
metabolized during oxygen deficiency in crustaceans (Zebe, 
1991), and (2) save proteins (and therefore muscular mass) 
for as long as possible. Thus, this species can (1) successfully 
withstand a hypoxic period subsequent to (or associated 
with) an initial nutritional stress, and (2) rapidly resume 
searching for food during short-term, sporadic, nutrition 

A general adaptation model for groundwater animals 
involves their ability to withstand prolonged hypoxia and/or 
long-term starvation (Hervant et al, 2001; Hervant and 
Renault, 2002; Hiippop, 2000) and to utilize in a very 
efficient way the high-energy body stores. Because the three 
hypogean crustaceans studied lack a high-ATP-yielding 

anaerobic pathway (such as observed in permanent anaerobic 
organisms) (Fields, 1983), their higher survival time in 
anoxia is mainly due to the combination of four mechanisms: 
(1) high storage of fermentable fuels (glycogen and phos- 
phagen); (2) low SMR in normoxia; (3) further reduction in 
metabolic rate by lowering energetic expenditures linked to 
locomotion and ventilation during hypoxia; and (4) high 
ability to resynthesize the depleted body stores during subse- 
quent recovery periods. The ability to maintain and rapidly 
restore (without feeding) high amounts of fermentable 
fuels for use during lack of oxygen can be considered an 
adaptation to life in a patchy environment. Through their 
efficient exploratory behavior in a moving mosaic of patches 
of low and high DO concentration (Malard and Hervant, 
1999), numerous groundwater animals probably experience 
highly variable DO and/or food concentrations. The behav- 
ioral, physiological, and metabolic responses of numerous 
hypogean animals partly explain why they occur in ground- 
water systems with a wide range of DO (Malard and 
Hervant, 1999). 

A high resistance to lack of oxygen (and/or to food 
deprivation) is not universally found in subterranean 
organisms but is probably more related to oxygen availability 
and/or to the energetic state of each subterranean ecosystem. 
Indeed, groundwater ecosystems are far more complex 
and diverse than earlier presumed. The aquatic amphipod 
Gammarus minus showed no significant difference in behav- 
ioral, physiological, and metabolic responses to experimental 
anoxia and subsequent recovery among a spring and a cave 
populations (see Malard and Hervant, 1999). Despite a 
strong tendency toward morphological convergence, 
subterranean organisms do not form a homogeneous group 
(Malard and Hervant, 1999). Further developments in the 
study of their physiology would highlight the diversity 
of adaptive responses among hypogean animals that have 
colonized contrasted groundwater ecosystems. 

See Also the Following Articles 
Adaptive Shifts 


Boulton, A.J., H.M. Valett, and S.G. Fisher (1992) Spatial distribution and 
taxonomic composition of the hyporheos of several Sonoran Desert 
streams. Archiv fiir Hydrobiol. 125: 37-61. 

Danielopol, D.L. (1989) Groundwater fauna associated with riverine 
aquifers. J. North Am. Benthol. Soc. 8: 18-35. 

Danielopol, D.L., M. Creuzé des Chatelliers, E Mésslacher, P. Pospisil, and 
R. Popa (1994) Adaptation of Crustacea to interstitial habitats: a 
practical agenda for ecological studies. In Groundwater Ecology (J. Gibert, 
D.L. Danielopol, and J.A. Stanford, Eds.). Academic Press, San Diego, 
pp. 217-243. 

Fields, J.H.A. (1983) Alternatives to lactic acid: possible advantages. J. Exp. 
Zool. 228: 445-457. 

Henry, K.S. and D.L. Danielopol (1999) Oxygen dependent habitat 
selection in surface and hyporheic environment by Gammarus 
roeseli Gervais (Crustaceana, Amphipoda): experimental evidence. 

Hydrobiologia 390: 51-60. 

Herreid, C.F. (1980) Hypoxia in invertebrates. Comp. Biochem. Physiol. 
67A: 311-320. 

Hervant, FE. and D. Renault (2002) Long-term fasting and realimentation in 
hypogean and epigean isopods: a proposed adaptive strategy for 
groundwater organisms. J. Exp. Biol. 205: 2079-2087. 

Hervant, E, J. Mathieu, D. Garin, and A. Fréminet (1996) Behavioral, 
ventilatory, and metabolic responses of the hypogean amphipod 
Niphargus virei and the epigean isopod Asellus aquaticus to severe hypoxia 
and subsequent recovery. Physiol. Zool. 69: 1277-1300. 

Hervant, E, J. Mathieu, and G. Messana (1997) Locomotory, ventilatory 
and metabolic responses of the subterranean Stenasellus virei (Crustacea: 
Isopoda) to severe hypoxia and subsequent recovery. C.R. Acad. Sci. Paris, 
Life Sci. 320: 139-148. 

Hervant, FE, J. Mathieu, and G. Messana (1998) Oxygen consumption and 
ventilation in declining oxygen tension and posthypoxic recovery in 
epigean and hypogean aquatic crustaceans. J. Crust. Biol. 18: 717-727. 

Hervant, E, D. Garin, J. Mathieu, and A. Fréminet (1999) Lactate 
metabolism and glucose turnover in the subterranean Niphargus virei 
(Crustacea: Amphipoda) during posthypoxic recovery. /. Exp. Biol. 
202: 579-592. 

Hervant, F, J. Mathieu, and J.P. Durand (2001) Behavioural, physiological 
and metabolic responses to long-term starvation and refeeding in a blind 
cave-dwelling salamander (Proteus anguinus) and a facultative cave- 
dwelling newt (Euproctus asper). J. Exp. Biol. 204: 269-281. 

Hiippop, K. (2000) How do cave animals cope with the food scarcity in 
caves? In Ecosystems of the World, Vol. 30 (D. Wilkens, C. Culver, and 
W.E Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 159-188. 

Malard, F. and F. Hervant (1999) Oxygen supply and the adaptations of 
animals in groundwater. Freshwater Biol. 41: 1-30. 

Mésslacher, E and M. Creuzé des Chatelliers (1996) Physiological and 
behavioural adaptations of an epigean and a hypogean dwelling 
population of Asellus aquaticus (L.) (Crustacea: Isopoda). Archiv fiir 
Hydrobiol. 138: 187-198. 

Revsbech, N.P. and B.B. Jorgensen (1986) Microelectrodes: their use in 
microbial ecology. Adv. Microb. Ecol. 9: 293-352. 

Ronen, D., Magaritz, M., Almon, E., and Amiel, A.J. (1987) Anthropogenic 
anoxicication (“eutrophication”) of the water table region of a deep 
phreatic aquifer. Water Resources Res. 23, 1554-1560. 

Strayer, D.L., S.E. May, P. Nielsen, W. Wollheim, and S. Hausam (1997) 
Oxygen, organic matter, and sediment granulometry as controls on 
hyporheic animal communities. Archiv. fiir Hydrobiol. 140: 131-144. 

Winograd, I. and F. Robertson (1982) Deep oxygenated groundwater. 
Anomaly or common occurrence? Science 216: 1227-1229. 

Yaeger, J. (1994) Speleonectes gironensis: new species (Remipedia: 
Spelonectidae) from anchialine caves in Cuba, with remarks on 
biogeography and ecology. J. Crust. Biol. 14: 752-762. 

Zebe, E. (1991) Arthropods. In Metazoan Life Without Oxygen (C. Bryant, 
Ed.). Chapman & Hall, London, pp. 218-237. 

Adaptive Shifts 

Francis G. Howarth 
Bishop Museum, Honolulu 

Hannelore Hoch 
Museum fiir Naturkunde, Berlin 

daptive shift is an evolutionary phenomenon in which 
individuals from an existing population change to exploit 
a new habitat or food resource. If successful, the new 
population may diverge behaviorally, morphologically, and 

Adaptive Shifts 17 

physiologically to become a distinct population or species. 
The phenomenon is inferred from the many examples of 
adaptive radiation in which separate species within a lineage 
have adapted to occupy different habitats or food resources. 
We can only see the end results of the process, making it 
difficult to test the model except indirectly through appro- 
priate comparative studies of the closely related species. 
Adaptive shifts are most recognizable on islands where pairs 
of closely related species overlap in distribution yet are 
adapted to radically different habitats. These species retain 
characteristics that indicate a logical progression from an 
immediate common ancestor, corroborating the view that 
they diverged from one another through the process of adap- 
tive shift. Only in the last few decades have adaptive shifts 
been advanced to explain the evolution of troglobites. The 
generally accepted view of cave adaptation held that surface 
populations of facultative cave species became locally 
extirpated (for example, by changing climate), thus isolating 
the surviving populations in caves where they could evolve in 
isolation. Any interbreeding with its surface population was 
believed to be sufficient to swamp any incipient specializa- 
tion. Divergence by adaptive shifts is usually envisioned as 
sympatric or parapatric; that is, the diverging populations 
remain in contact during the split. It now appears that many 
cave-adapted animals evolved by the process of adaptive 
shifts from representatives of the local surface fauna. How 
this occurs is the focus of this chapter. 


For over a century, the evolution of obligate cave species 
(troglobites) was assumed to be restricted to continental 
regions that had been influenced by glacial events and that 
the ecological effects resulting from the severe climatic 
fluctuations provided the isolation and necessary elements to 
facilitate the evolution of troglobites. Because most troglo- 
bites appear to have no surviving ancestors, it has been 
assumed that they evolved allopatrically (z¢, as a conse- 
quence of extinction of their closely related epigean species); 
however, it is possible that many temperate-zone troglobites 
evolved parapatrically through adaptive shifts and that their 
current isolation occurred after cave adaptation. 

In the adaptive shift model, populations of epigean species 
invade subterranean habitats to exploit novel resources, and 
the drastic change in habitat is considered to be the driving 
force for genetic divergence and speciation. The presence 
of exploitable food resources provides the evolutionary 
“incentive” for adaptation. Evolutionary theory predicts that 
adaptation and new species are more likely to arise from 
large, expanding populations to exploit marginal resources at 
the edge of their habitat and to make adaptive shifts to 
exploit new environments than are small populations. 

Adaptive shifts proceed in a three-step process. First, a 
new habitat or resource becomes available for exploitation. 
This opening of an ecological niche can occur either by an 

18 Adaptive Shifts 

organism expanding its range or by geological or successional 
processes creating an opportunity for a resident population. 
Second, there is a shift in behavior to exploit the new habitat 
or food resource. Third, if the behavioral shift successfully 
establishes a new population, natural selection fosters adapta- 
tion. In a classic comparative study of cave and surface 
springtails, Christiansen (1965) demonstrated that a 
behavioral shift to exploit a new resource or survive a new 
environmental stress occurs first, and, if it is successful, 
morphological, physiological, and additional behavioral 
changes follow. Even if the populations remain in contact, 
the selection pressures imposed by the different environ- 
ments and augmented by environmental stress, if strong 
enough, can force the two populations to diverge. 

Speciation by adaptive shift is parapatric in that adaptive 
differentiation, accompanied by a reduction in gene flow, 
proceeds across a steep environmental gradient within a 
contiguous area. If the colonization is successful, the found- 
ing population reproduces and expands into underground 
habitats. Maintenance of the barrier to gene flow could be 
facilitated by selection against hybridization (e.g., decreased 
hybrid viability), assortative mating, or spread of the 
incipient cave population away from the narrow hybrid zone, 
thereby reducing the effect of introgression from epigean 
individuals. It has been hypothesized for cave species that 
selection for novel mating behaviors may be the principal 
origin of isolation. Many adaptive shifts undoubtedly fail; 
that is, the founding population dies out. Conversely, 
interbreeding and selection following a shift can produce a 
single more adaptable population capable of exploiting both 


The trigger for an adaptive shift into caves is the availability 
of suitable habitat and exploitable resources. Only a small 
percentage of epigean groups have representatives inhabiting 
caves, and several factors appear to be involved in deter- 
mining whether or not an organism can take advantage of the 
opportunity to exploit caves. These factors often act in 
concert but for clarity can be considered either intrinsic or 
extrinsic. Intrinsic factors are characteristics inherent in the 
organism that allow it to make the shift—for example, the 
ability to live in damp, dark habitats. Extrinsic factors are 
those imposed on the organism by the environment. 

Intrinsic Factors 

PREADAPTATION The role of preadaptation has been 
recognized for more than a century as one of the principal 
factors explaining which taxa have successfully adapted to 
live in caves. Most preadaptations simply result from the 
correspondence of an organism’s preferred environment with 
that found in caves. That is, organisms that have character- 

istics that allow them to live in damp, dark, wet-rock 
microhabitats on the surface have a better chance of 
surviving in caves than do organisms that do not possess 
these traits. Similarly, survival is enhanced if their normal 
food resource naturally occurs in caves. Nymphs of epigean 
cixiid planthoppers are admirably preadapted to caves as they 
feed on plant roots and therefore already have a suite of 
behavioral, morphological, and physiological adaptations 
to survive and feed underground. In the cave species, these 
preadapted nymphal characters have been retained into 
adulthood so that the adults are also able to live 

GENETIC REPERTOIRE In order to adapt to caves, a 
population must have the ability to change. For example, if 
an organism is genetically hard-wired to require light, tempe- 
rature change, or other environmental cues to complete its 
life cycle and reproduce, it is unlikely to colonize caves. In 
cixiids, salamanders, and a few other groups, the adaptive 
shift to living in caves was greatly facilitated by neoteny 
or the retention of nymphal characters into adulthood. 
Development of neoteny in salamanders is believed to 
involve relatively small changes in the genes that regulate 
development. This demonstrates that a small genetic change 
can have a large effect on phenotype. 

FLUSH AND CRASH CYCLES Because only a few indivi- 
duals from the parent population make the initial shift, only 
a subset of the genetic diversity found in the parent popula- 
tion can be carried into the founding population. Certain 
alleles may be lost, while certain rare ones may become more 
abundant. Inbreeding, response to environmental stress, and 
expression of previously rare alleles may result in coevolved 
blocks of loci becoming destabilized, thus allowing new 
combinations of genes. If the new colony expands rapidly 
into the new habitat (e.g., due to abundant food or reduced 
competition and predation), the relaxed selection pressure 
allows the survival of new mutations and unusual recom- 
binants that ordinarily would be lost due to low fitness. Thus, 
the genetic variability may be quickly reestablished in the 
founding population; however, by this time there may be 
significant genetic divergence from the parent population. 
A subsequent population crash (as when the new colony 
exceeds the carrying capacity) places extreme selection 
pressure on the population. These cycles of expansions and 
crashes in founding populations are thought to facilitate 
adaptation to a new environment. 

In parapatric divergence, occasional backcrosses with the 
parent population can increase genetic diversity and provide 
additional phenotypic variability on which natural selection 
can act. Mixing of the two gene pools by interbreeding 
becomes less likely as the new population expands its range 
beyond that of the parent population and adapts to the 
stresses of the new environment. 

RESPONSE TO STRESS Environmental stresses in a novel 
habitat may facilitate adaptive shifts. Organisms living under 
environmental stresses often experience higher mutation 
rates and display greater phenotypic and genetic variation. In 
nature, selection and the high energetic costs of stress usually 
reduce this variation; however, where sufficient exploitable 
food energy compensates for the extra costs of survival in a 
novel habitat, some of the enhanced variation may survive 
and in time result in a new population adapted to cope with 
the stresses. The higher energetic costs required to cope with 
stress augment both natural and relaxed selection pressures to 
favor the loss of unused characters, such as eyes and bodily 
pigment of troglobites. The lower metabolic and fecundity 
rates possessed by troglobites may in large part be an 
adaptation to cope with stresses found in subterranean 

tant constraint among organisms colonizing caves is the 
ability to locate mates and reproduce underground. The 
normal cues used by a species may not be present or the mate 
recognition signal system may be confusing in the unusual 
environment. For example, sex pheromones would not 
disperse in the same way in caves as they do on the surface. 
Also, an animal may have difficulty following the plume in a 
three-dimensional dark maze. The rarity of troglobites that 
use airborne sounds to locate and choose a mate attests to the 
difficulty of using sound underground and suggests that 
muteness in crickets is a preadaptation for colonizing caves. 
In addition, environmental stress can reduce sexual selection 
and disrupt mating behaviors. Founder events can exacerbate 
these effects and the resulting release of sexual selection 
pressure may allow hybridization with relatives. The main- 
tenance of limited gene flow between the diverging popu- 
lations is thought to be important in providing the additional 
genetic variability to accelerate adaptation; however, 
evolution of more appropriate mating behaviors in a newly 
established cave population may provide the critical factor in 
isolating the diverging populations. 

Extrinsic Factors 

there must be an opportunity for a surface population to 
colonize caves; that is, the surface population must live in 
contact with cavernous landforms. Geological processes 
(such as lava flows) can create new cave habitats, ecological 
succession or the immigration of a new organism can intro- 
duce suitable new food resources into caves, or a preadapted 
surface population can migrate into the cave area. The 
Hawaiian Islands formed sequentially in line, and the 
successful colonization of each island after it emerged from 
the sea had to proceed in an orderly fashion. Phytophagous 
insects had to wait for their hosts to establish and so on. 
Terrestrial obligate cave species are not likely to have crossed 

Adaptive Shifts 19 

the wide water gaps between each island. Even though 
suitable lava tubes and cave habitats were available from the 
beginning, the evolution of cave species had to wait for their 
surface ancestors to become established and for appropriate 
food resources to accumulate in caves. These circumstances 
provided tremendous advantages to the first organisms that 
could exploit a newly available resource and preempt the 
resource from subsequent colonists. 

exploitable food resource is the paramount prerequisite that 
allows individuals of a population to shift into a new habitat 
or life style. In fact, in most examples of adaptive shift, both 
the cave and surface species feed on the same food resource. 
Sparse soil and organic litter with areas of barren rock on the 
surface are characteristic of cavernous landforms, including 
those created by limestone, lava, and talus. Where an inter- 
connected system of cave and cave-like subterranean voids 
occur, organic material does not accumulate on the surface 
but is washed underground by water, falls underground by 
gravity, or is carried underground by living organisms. 
Many surface organisms will follow their food underground 
and attempt to exploit it. Even though their food may be 
abundant in caves, most organisms adapted to living in 
surface habitats may become lost in deep cave passages 
because their normal cues are reduced or absent or because 
they cannot cope with the harsh environment. Thus, much 
of this sinking material is outside the reach of most surface 
organisms; however, it becomes a rich reward for any 
organism that can adapt to exploit it. 

ENVIRONMENTAL STRESSES The environment in deep 
caves and cave-like voids is highly stressful for most surface 
animals, and relatively few surface species can survive for 
long underground. Even though food may be adequate, it is 
much more difficult to find and exploit in the complex three- 
dimensional dark maze, especially because many environ- 
mental cues (e.g., light/dark cycles, temperature changes, and 
air currents) are absent. Furthermore, the atmosphere is 
characteristically above the equilibrium humidity of body 
fluids, making respiration and water balance difficult to 
sustain. The substrate is moist, and the voids occasionally 
flood. Carbon dioxide and oxygen concentrations can reach 
stressful levels. In many cave regions, radiation from decay- 
ing heavy isotopes may exceed hazardous levels. Because of 
the mazelike space, escape to a less stressful environment is 
virtually impossible, and cave animals must either adapt to 
cope with these and other stresses i situ or die out. 

ANCESTRAL HABITATS ‘Troglobites originated from a 
variety of damp surface habitats. Specifically, the ancestral 
habitats include soil, leaf litter, mosses, and other damp 
microhabitats in forests; damp, wet-rock habitats (such as 
marine littoral, riparian, and cracks in barren lava flows); 
and guano-inhabiting species living in caves and animal 

20 Adaptive Shifts 

burrows, These source habitats corroborate che concep of 


‘Twelve taxonomic groups have independently adapred co 
caves on nwo or more islands in Hawaii, indicating thar cave 
adaptation is a general phenomenon fosered by ocological 
and evolucionary factors, Om the young fless than one 
million years} Hawaii Island, some taxa have undergone for 
arc undergeing) adaptive shifts inte both caves and new 
surface habitats ina phenomenon called saeptipe ereliagion, 

Many of these nroglobites still have extant close surface 
relarives living in nearby habirars, The better known ameeng 
these are lied in Table | and discussed below. The number 
af examples will increase as more taxa are studied, Cave 
species on the older islands tend co be celices todays however, 
they also are beliewed no have originally evolved by adaptive 

MORPHA: CIXIIDAE) ‘Cisiid planthoppers exemplify how 
adaprive shifts might occur (Fig, 1). The nymphs of nearly alll 
species in the farnily live in or close co the soil and feed on 
the xylem sap of roots, Cixiid nymphs of surface species 
molt co adules (Fig, 2), emerge above ground, and live, food, 

TARLE | Parapatric Cave anal Surhce Species Maire Occurring om Hawaii Iskond 

Khare Bjpwcien Comins Mane 

Saorface Itelatiee Ancestral Holhicad 

Alaete [titel 

OMtarce moxrboeil 
Oaner paler" 
OMtarna dernier 
Nevialolestes aus” 

Comormabiy parr! 
Aaalrhe beecartht 

iyewa dewarity 

Qi, weed bons 
Cixitd plarhopper 
Cizikl plarhepper 
Cixiiel flarhoepper 
Thread-kegeed beog, 
Rack cricket 


Wolk spicler 

fattaros ilscia tera 

OMtere dbaavrea Wewic: Koren 

CNGerses ppt Rasen heres 

CMoarses appt Cre shorts Larned 
Fitutalodcetice anbeverr Coypie rain heen hahtiass 
Caraneoraber fort" Barren brea Hw 

( swale felartme bite! 

Aneokutur maritime Marine lineal 

A, dotwertireur 

fyceed =p" 

Barren beva (iwi? 

Barren ba How 

* Represcieed by several distinc: populations or species. Polymorphic cave populasions may represent separate inmsions or could reaile from divergence of 

a dingle Lineage after cave adapiaion. 

oA asco 
with ioe 

ren: TE topiephale; teri, cregkohine. 

onené nicer 

young kee 
fish foovemous) 

FROUAE I Chagrammatic view of an adageive shily from sees bo srklerground WY Ie Le ka acioed planthegper AC, accidental; AS, aclape ine ahilt; fel, 

FIGURE 2 Epigean adult dbdkd, Chir apedics, beam Herwuii Jalan. 
(Plot graph by H. Haeh.3 

and mace on vegetation. Plants living on or near cavernous 
landforms often send their roors through the interconnected 
sverem of voids deep underground following nutrients and 
percolating groundwater. Thus, rears ane often an inypertant 
sures OF food energy in caves, especially in the cropics 
and lava muibes. Nymphs of surface species may miprace along 
rats deep into caves from their nermal shallow habitat. 
These nymphs can exploit deep roots, as the deep halbitar 
is similar co their normal ome exceps for the perpetually 
saturated anmosphere, and for chis stress chey already have 
mechanisms to excrete excess water as xylem sap is very 
dilute, When they molt, the adults may have little chance to 
find their way to the surface wo reproduce and would even- 
wally die from che stressful environment: hewever, if host 
plant roots were abundance and large numbers of nymphs 
continually made their way underground, an adult mighe 
occasionally find a mace and reproduce formuicaush, Most of 
these carly chance matings, if successful ar all, would produce 
normal surface adulrs thar would also become lost, Eventually, 
some offspring might acquire teats from random mucation 
ar genetic recombination that are passed on anal char allen 
the offspring co survive and reproduce underground more 
easily, If this incipient colony successfully establishes 
an underground population, it could mere fully exploit che 

Adaptive Shifts 21 

FIGURE 4 Mieaph af Chuan pushover Within was Glament ovceon an its 

lust pert. (Pastagrapls ty H. Hack 

deeper root resources, In fact, a whole new habitat and 
resturce would be opened tothe expanding population as it 
could move throughout the interconnected voids in the cave 
region and leave its surface relatives behind. The surface 
SVviniment on cavernous landberms provides lithe: suitable 
habirar for epipean planthoppers, so thac the subterranean 
population is often larger than surface populations over 
caves, This is especially true on volcanoes where eruprions 
can creare vast new habinars for cave populations. [In Hawaii, 
the principal pioneering plant on young lava flows is the 
endemic tree Waverider potrmenis (Myrtaceae), the races 
of which are the hosts of Offers: poligobernas sf ‘eases dates! 
nymphs (Fig, 3). Even widely scattered small ness on 100- 
year-old flows send abundant roots deep inno caves, As envi- 
sioned, adapration can be rapid in the invasive expanding 
phass, Subsequently, the newly established subrerrancan 
population will acquire the troglomerphic traits character 
istic for obligate cavernicoles: reduction of eves, wings, and 
bodily pigment. 

In Hawaii, the genus GWerey has unederpene extensive 
adaptive radiation in surtace habitats after the successful 
colonization of a single ancestral species, and about 80 
endemic species and subspecies have been described from all 
major islands, Eighteen surface species (eg. Pig, 2) are 
known from the island of Hawaii, where at leas three 
independent adaptive shifts into subterranean habicacs have 
occurred {Table 0}. The clearest case for adaptive shift among 
cave cpdids invelwes (2 nerkoilt and () dkormead, Ch aeanbatki 
has reduced wings, «ves, and bodily pigment and is known 
froma single 1400-00 3000-year-old lava cube from Hualalai 
Volcano, Its male genitalia fwhich usually provide an 
excellent means for distinguishing species in the group) are 
virtually identical co the fully winged and eyed epipean 0 
degen, OL beaved is widespread in mesic forests on Hawaii. 
Its nymphs are frequent in shallow caves and even occur near 
the ennrance in the same cave with OL menbete, Adults of 0. 
dogo may be accidental in caves, as accumulations of body 

22 Adaptive Shifts 

FIGURE 4 Adult female of Oliarus polyphemus. (Photograph by W.P. Mull.) 

parts from thousands of dead individuals have been found in 
dead-end passages in some caves. 

The highly troglomorphic O. polyphemus s.1. (Fig. 4) also 
occurs in the same cave with O. makaiki, as well as in 
virtually every suitable cave containing its host roots, 
Metrosideros polymorpha; however, each cave system appears 
to harbor a unique population of this troglobite. Each 
population differs only slightly in morphology, but the 
mating calls are highly distinctive and probably sufficient 
to reduce hybridization if the populations were to come in 
contact. The ancestor of O. polyphemus s.l. has not been 
determined, although it certainly belongs in the endemic 
Hawaiian group of species. It is also unknown whether O. 
polyphemus s.1. represents a single colonization event in caves 
with subsequent divergence underground or whether some 
populations represent separate invasions by the same or 
closely related surface species. 

Adult planthoppers use substrate-borne sounds to recog- 
nize and locate their mates. Among surface species, females 
feed on the host plant while the males fly from host to host 
and initiate a species-specific mating song. If a receptive 
female hears an appropriate song, she will answer with her 
own song. The male orients to the call, and the two alter- 
nately sing and listen until the male finds the female, who 
remains in situ. Studies have shown that Hawaiian cave 
Oliarus essentially use the same mate recognition system: 
substrate-borne vibrations, with living roots being an 
excellent transmitting medium. In laboratory studies on the 
cave cixiid Oliarus polyphemus s.1., the flightless females feed 
on the host root and occasionally call. Males are also flightless 
and wander from root to root listening. When a male hears 
a female, he answers, and if the song is appropriate the two 
sing and listen alternately until the male locates the 
stationary female. Although the behavior has not been 
confirmed in natural settings in caves, it is believed that this 
switch in behavior occurred as an adaptation to the sub- 

terranean environment. Singing is energetically costly, and 
males expend extra energy finding roots in the dark, three- 
dimensional maze. Also, predators can use these songs to 
locate prey, and males would be especially vulnerable if they 
had to sing on every root they found; therefore, males would 
be more likely to find a female if they did not have to initiate 
calling. Females conserve energy by staying in place and 
feeding, and they produce fine wax filaments that deter 

PHILOSCIIDAE) Species in the genus Littorophiloscia are 
marine littoral and inhabit tropical shorelines worldwide. 
One epigean species, L. hawaiiensis, is endemic to Hawaii 
and is so far known only from littoral habitats on the islands 
of Hawaii and Laysan. Surprisingly, an undescribed troglo- 
bitic species also lives on Hawaii Island. The cave form is 
remarkably similar in morphology to L. hawaiiensis, differing 
only in the troglomorphic characters (reduced eyes and 
pigment) displayed by the cave population. The males even 
share the same distinctive sexual characters. Results from 
a recent molecular phylogenetic study showed that ZL. 
hawaiiensis and the cave Littorophiloscia are distinct species 
that diverged from each other most likely by an adaptive 
shift. The initial adaptive shift must have been from salty 
marine littoral habitats to freshwater terrestrial habitats, 
possibly in coastal springs in young lava flows. Suitable fresh- 
water shorelines are not well developed on Hawaii Island, but 
lava tubes would offer a vast new habitat. The complex three- 
dimensional system of anastomosing voids characteristic of 
young basaltic lava flows provides an immeasurably large 
habitat for cave animals. In contrast, L. hawaiiensis is 
restricted to a narrow band along the coast where it lives under 
rocks resting on soil. Thus, an adapting cave population 
would quickly move away from contact with its ancestral 
population and greatly exceed it in size. 

A possible alternative explanation is the classical model 
of cave adaptation in which the coastal species became 
extinct stranding an incipient cave population; however, this 
model still requires that the adaptive shift to freshwater (and 
possibly to caves) had to occur before extinction, and that 
circumstance is precluded in the classical model. This 
scenario also requires a subsequent colonization by the halo- 
philic littoral species that secondarily reestablished parapatry. 

of Hawaii, the marine littoral rock cricket Caconemobius 
sandwichensis has colonized both unvegetated lava flows and 
caves, both of which are barren-rock habitats like its ancestral 
home. Several distinct populations of rock crickets inhabit 
each of these inland environments, but it is unknown 
whether these represent multiple adaptive shifts from the 
marine littoral habitat or result from divergence after 
colonization of the new habitat, or both. A hybrid zone 
between the marine littoral species and the lava flow cricket 

C. fori was found around a pool of brackish water in a deep 
crack where a young lava flow entered the sea, suggesting that 
the adaptive shift is ongoing. C. fori is a nocturnal scavenger 
on very young (z.¢., 1 month to about a century old) lava 
flows on Kilauea, disappearing when plants colonize the flow. 
It hides in deep cracks and caves during the day where it 
overlaps with several distinct cave populations or species. 
Additional recognizable populations of lava crickets, 
resembling C. fori, occur allopatrically on Mauna Loa and 
Hualalai volcanoes, also on Hawaii Island. 

Up to three morphologically, behaviorally, and physio- 
logically distinct populations of cave crickets can be found 
in larger caves, and individual caves may harbor a unique 
population of one or more of these forms. The total diversity 
of cave-adapted rock crickets on Hawaii is astounding, 
especially given the young age of the island. These multiple 
simultaneous adaptive shifts into such different habitats from 
the same ancestor further corroborate the contention that 
cave adaptation via adaptive shifts is a general phenomenon. 

A possible scenario for the origin of troglobitic rock 
crickets and wolf spiders on the island of Hawaii might 
proceed like this. Marine littoral crickets could have been 
among the earliest colonists of newly emerging islands, 
because all that they required was already present: a rocky 
shoreline and ocean-derived flotsam. Barren lava was the 
original terrestrial habitat, and sea birds would have sought 
these new islands for nesting, presenting an inviting habitat 
for the littoral crickets to shift to the land where they could 
feed on wind-borne debris. Subsequent accumulation of 
organic material underground would have allowed the 
colonization of caves by either the seacoast or lava cricket 
preempting some resources, possibly before many plants and 
potential competitors arrived on the island. 

eyed wolf spiders (family Lycosidae) live on barren landscapes 
in many parts of the world and are good long-distance 
dispersers. They would also have been among the earliest 
colonists on emerging islands in Hawaii. There are several 
barren-ground populations of an undescribed species of 
Lycosa on Hawaii Island that differ in behavior, morphology, 
and color, but surprisingly they can hybridize. They occupy 
a range of habitats from hot, dry coastal barren lava plains 
to the freezing stone deserts above 4000 meters on Mauna 
Kea and Mauna Loa. On young lava flows, they prey on 
Caconemobius \ava crickets. One member of this group 
followed its prey underground, becoming the anomalous 
small-eyed, big-eyed wolf spider (Lycosa howarthi). 

The lycosid wolf spiders are characterized by the possession 
of four huge eyes (for spiders) and four smaller eyes, and they 
are among the better-sighted spiders. The troglobitic species 
has six vestigial eyes, yet differs from the lava flow species 
only by the characters associated with cave adaptation. As 
noted, both the epigean spider and cricket were early 
colonists on Hawaii, and possibly the spider was able to 

Adaptive Shifts 23 

preempt the cave habitat because it was among the first 
predators to have the opportunity to exploit the young caves 
on the island. 

Other Islands 

Adaptive shifts are best known from islands in large part 
because of their youth and isolation. Thus, many diverging 
species pairs are still extant, and their evolutionary history 
has not been obscured by geological or ecological events. 
Examples of troglobite evolution by adaptive shift are known 
from the Galapagos, Canary, and Greater Antilles islands. 
In the Galapagos, at least ten species of troglobitic terrestrial 
invertebrates still live parapatrically with their putative epigean 
ancestors. At least seven extant parapatric pairs of troglobitic 
and epigean species are known from the Canary Islands. 
Some of these, such as the Dysdera spiders and cockroaches, 
represent multiple invasions into caves, in parallel with 


Clear examples of adaptive shifts in continental caves are 
relatively rare, possibly because the great age of these systems 
has obscured the evolutionary history of the taxa involved or 
because the phenomenon has not generally been considered; 
nevertheless, there are a few cases. Perhaps the best docu- 
mented is the detailed study (Culver et a/., 1995) done on the 
spring- and cave-inhabiting amphipod Gammarus minus, 
which occurs in freshwater basins in temperate eastern North 
America. The study demonstrated that several cave popula- 
tions were derived from independent invasions from springs 
and that some hybridization had occurred; however, the 
constraints imposed by the dynamics of water flow and other 
factors may have isolated some populations from each other. 
In spite of considerable genetic and morphological diversity, 
these populations are all considered one species. In parallel 
with some Hawaiian forms, the definition of what consti- 
tutes a species is sometimes not clear. Adaptive shifts have 
been proposed to explain the origins of cave species in North 
Queensland, Australia, but the few phylogenetic analyses 
completed to date have not supported the contention. 


In cavernous regions, organic energy is continually being 
transported into an extensive system of subterranean voids, 
as evidenced by the characteristic presence of areas with a 
barren exposed rocky surface. Epigean species can exploit 
food resources on and near the surface, but increased 
environmental stress levels and absent or inappropriate cues 
prohibit their access to this resource in the deeper voids. 
Animals following their normal food deeper may eventually 
become lost and be unable to return to the surface. Surface 
species continue to be frequent accidentals in caves and 

24 Anchialine Caves, Biodiversity in 

provide a large proportion of the food to the cave-adapted 
predators and scavengers. Over time, a few of these acci- 
dentals may eventually survive and found a new population. 
A major initial factor facing a new cave population may be 
the ability to locate a mate and reproduce underground. 
Once a population is able to reproduce underground, a large 
new habitat will be opened up to it, and the population can 
expand rapidly. Evolution of troglobites by adaptive shift 
may be a common phenomenon; however, many examples of 
adaptive shift remain unrecognized, especially because the 
phenomenon is not generally considered. Other scenarios 
(such as allopatric evolution, which appears to be supported 
for many temperate troglobites) are also conceivable. The 
adaptive shift model and the classical theory of troglobite 
evolution may not be mutually exclusive. 


Adaptive shifts offer extraordinary opportunities for research. 
Comparative studies of closely related pairs of species 
adapted to such different environments as caves and surface 
habitats should provide better understanding not only of the 
processes of cave adaptation but also of selection and species 
formation in general. Molecular systematics will clarify the 
evolutionary history of taxonomic groups as well as elucidate 
the genetic basis of adaptive shifts. Phylogenetic analyses 
can elucidate the evolutionary transformation of specific 
morphological, behavioral, physiological, and ecological 
traits that occurred during the process of adaptation to novel 
habitats. The remarkable divergence from their epigean 
relatives displayed by troglobites makes such studies espe- 
cially interesting. Comparative physiological experiments can 
potentially dissect out for study the specific mechanisms 
involved in coping with a variety of environmental stresses. 
For example, terrestrial cave species live in an air-filled 
aquatic environment, while the close relatives of some live in 
desert-like environments. These species pairs are ideal models 
for studies on the physiological mechanisms involved in 
regulating water balance. 

See Also the Following Articles 
Adaptation to Darkness « Adaptation to Low Food « 
Adaptation to Low Oxygen 


Christiansen, K. (1965) Behavior and form in the evolution of cave 
Collembola. Evolution 19: 529-537. 

Culver, D.C., T.C. Kane, and D.W. Fong. (1995) Adaptation and Natural 
Selection in Caves: The Evolution of Gammarus minus, Harvard University 
Press, Cambridge, MA. 

Hoch, H. and EG. Howarth (1993) Evolutionary dynamics of behavioral 
divergence among populations of the Hawaiian cave-dwelling 
planthopper Okarus polyphemus (Homoptera: Fulgoroidea: Cixiidae). 
Pacific Sci. 47: 303-318. 

Hoch, H. and EG. Howarth (1999) Multiple cave invasions by species of 
the planthopper genus Oliarus in Hawaii (Homoptera: Fulgoroidea: 
Cixiidae). Zool. J. Linnean Soc. 127(4): 453-475. 

Howard, D.J. and S.H. Berlocher, Eds. (1998) Endless Forms: Species and 
Speciation, Oxford University Press, New York. 

Howarth, EG. (1980) The zoogeography of specialized cave animals: a 
bioclimatic model. Evolution 34: 394—406. 

Howarth, EG. (1983) Ecology of cave arthropods. Annu. Rev. Entomol. 
28: 365-389. 

Howarth, EG. (1993) High-stress subterranean habitats and evolutionary 
change in cave-inhabiting arthropods. Am. Naturalist 142: S65-S77. 
Oromi, P., J.L. Martin, A.L. Medina, and I. Izquierdo (1991) The evolution 

of the hypogean fauna in the Canary Islands. In The Unity of Evolu- 
tionary Biology, Vol. 1 (E.C. Dudley, Ed.). Dioscorides, Portland, OR. 
Rivera, M.A.J., EG. Howarth, S., Taiti, and G.K. Roderick. (2002) 
Evolution in Hawaiian cave-adapted isopods (Oniscidea: Philosciidae): 
vicariant speciation or adaptive shift? Molec. Phylogenet. Evol. 25: 1-9 
Peck, S.B. and T. Finston (1993) Galapagos islands troglobites: the questions 
of tropical troglobites, parapatric distributions with eyed-sister-species, 
and their origin by parapatric speciation. Mém. Biospéol. 20: 19-37. 

Anchialine Caves, 
Biodiversity in 

Thomas M. Iliffe 
Texas AGM University at Galveston 


In 1973, Holthuis described several new and unusual red 
shrimps collected from land-locked saltwater pools in the 
tropical Indo-West Pacific. Because these pools shared 
common features, Holthuis coined the term anchialine (from 
the Greek meaning “near the sea”) and described this type of 
habitat as “pools with no surface connection with the sea, 
containing salt or brackish water, which fluctuates with the 
tides.” In 1966, Riedl had referred to similar habitats as 
Randhoelen (or “marginal caves”). Anchialine pools occur in 
uplifted reef limestone as well as irregular porous lava flows. 
The water in these pools varies in salinity from nearly fresh 
to fully marine. Dampened tidal fluctuations indicate the 
influence of a continuous water table extending inland from 
the open sea. Both salinity and the degree of connection to 
the sea control the nature of the biota, as pools close to the 
sea contain typical marine species, while those farther inland 
have fewer, but more unusual species. 

Explorations by cave divers have resulted in the discovery 
of extensive, entirely submerged cave systems of which the 
classical anchialine pool represents only a small portion 
(Fig. 1). Thus, a new and ecologically more refined definition 
of the term anchialine was proposed at the 1984 Interna- 
tional Symposium on the Biology of Marine Caves: 
“Anchialine habitats consist of bodies of haline waters, 
usually with a restricted exposure to open air, always 
with more or less extensive subterranean connections to the 
sea, and showing noticeable marine as well as terrestrial 

FIGURE 1 Diver swimming past submerged stalagmitic columns in Crystal 

Cave, an anchialine cave in Bermuda. (Photo by Cristian Lascu.) 

influences.” An alternate spelling of anchialine is anchihaline, 
which adds an “h” to reflect international conventions on 
classification of salinity. 


Anchialine stygobites are almost exclusively of marine origin 
and consist primarily of crustaceans (Table I). In addition, 
various freshwater stygobites including many nematodes, 
triclad turbellarians, polychaetes, copepods, mysids, thermos- 
baenaceans, isopods, amphipods, shrimp, and fish are marine 
relicts that may have invaded fresh groundwater by way of 
anchialine habitats. The reason why crustaceans make up 
80 to 90% of anchialine stygobites is unclear. Perhaps 
crustaceans are better able to adapt to such limiting factors in 
the anchialine environment as lack of light, low levels of 
dissolved oxygen, and limited food supply. 

Many new higher taxa have been discovered from 
anchialine caves, including the crustacean class Remipedia, 
the peracarid order Mictacea, and the copepod order Platy- 
copioida, plus 10 new families and more than 50 new genera. 
A diverse group of higher crustacean taxa characteristically 
inhabit anchialine caves. These include: 

Remipedia: Free-swimming stygobitic crustaceans with a 
short head and elongate trunk composed of numerous 
segments, each with a pair of biramous swimming 
appendages (Fig. 2). This class includes 12 described 
species—eight from anchialine caves in the Bahamas, 
plus one each from the Canary Islands, Yucatan 
Peninsula, Cuba, and Western Australia (Fig. 3). 
Remipedes are considered the most primitive of living 

Ostracoda: Small, free-swimming crustaceans with a 
calcareous bivalve shell. The ostracod orders 
Halocyprida, Myodocopida, Podocopida, and 
Platycopida include anchialine species. Of these, the 
halocyprids contain the most stygobitic species with 

Anchialine Caves, Biodiversity in 25 

TABLEI Approximate Number of Strictly Anchialine Species 
Sponges 4 
Turbellarians 1 
Gastropods 5 
Annelids 10 
Chaetognaths 4 
Tantulocarids 1 
Copepods 55-60 
Ostracods 40-45 
Mysids 35-40 
Remipedes 12 
Thermosbaenaceans 30-35 
Mictaceans 1 
Tanaidaceans 2 
Leptostracans 1 
Bochusaceans 2 
Isopods 35-40 
Amphipods 95-100 
Decapods 45-50 
Syncarids 2 
Water mites 3 
Pisces 10 

Source: Adapted from Pesce (2000). 

FIGURE 2 A remipede from Norman’s Pond Cave in the Exuma Cays, 

three primarily or exclusively anchialine genera. 
Danielopolina includes 13 species with 5 species from 
caves in the Bahamas, 2 from the Canary Islands, and 
1 each from Cuba, Jamaica, Yucatan, Galapagos, 
Western Australia, and deep-sea Atlantic (Fig. 4). 
Speleaoecia has 11 species, including 6 from the 
Bahamas, 2 from Cuba, and 1 each from Jamaica, 
Yucatan, and Bermuda. Deeveya has the most restricted 
distribution, with 8 species all from the Bahamas. 
Copepoda: Small crustaceans that are the most abundant 
and diverse type of animal on the planet. Copepods 
from the orders Calanoida, Cyclopoida, 
Harpacticoida, Misophrioida, and Platycopioida 


Anchialine Caves, Biodiversity in 

FIGURE 3 World map showing the distribution of the crustacean class Remipedia. 

inhabit anchialine caves. The calanoid family 
Epacticeriscidae, cyclopoid family Speleoithonidae, 
and harpacticoid families Rotundiclipeidae, 
Superornatiremidae, and Novocriniidae are almost 
exclusively anchialine. Misophrioids include both 
anchialine and deep-sea species. Several exclusively 
anchialine copepod genera are widely distributed. For 
example, seven species from the epacteriscid genus 
Enantiosis inhabit caves in Palau (two species), plus 
one species each in the Bahamas, Bermuda, Belize, 

Galapagos Islands, and Fiji (Fig. 5). 

Mysidacea: Small, shrimp-like crustaceans with stalked 

eyes; embryos carried ventrally. Nineteen genera of 
mysids include stygobiont representatives. Most 
colonized groundwater during regressions of the 
Tethys and Mediterranean Seas. 

Mictacea: Small crustaceans with seven pairs of pereopods 

(legs) (Fig. 6). The order Mictacea includes four 
species—one each from anchialine caves in Bermuda 
and the Bahamas and two from deep waters of the 
Equatorial Atlantic and the Indo-Pacific. A new order, 
Bochusacea, has been proposed to include the 
Bahamian species, as well as the two deep-sea species 
previously included within the Mictacea. 

Lsopoda: Small, sessile-eyed crustaceans inhabiting a 

variety of terrestrial, marine, and freshwater habitats. 
The isopod family Cirolanidae includes six anchialine 
genera. The anthurid genus Curassanthura consists of 
three anchialine species from Bermuda, the Canary 
Islands, and the Netherlands Antilles. The family 
Atlantasellidae includes only two species, which 
inhabit anchialine caves in Bermuda and the 
Dominican Republic. 

Amphipoda: Small crustaceans with a laterally compressed 

body. The amphipod families Bogidiellidae, 
Gammaridae, Hadziidae, Hyalidae, Ingolfiellidae, 
Liljeborgiidae, Lysianassidae, Pardaliscidae, 
Phoxocephalidae, and Sebidae include anchialine 
species. Of these families, the Hadziidae + Melitidae 
complex contains the largest number of anchialine 

Thermosbaenacea: Small crustaceans with embryos carried 

dorsally (Fig. 7). Four genera of anchialine 
thermosbaenaceans include Monodella and 
Tethysbaena from the Mediterranean region; 
Halosbaena from the Canary Islands, Caribbean, and 
Western Australia; and Tudumella from the Yucatan 
Peninsula and the Bahamas. 

CLE ate 

Anchialine Caves, Biodiversity in 27 

‘ aoa 

can yr 


FIGURE 4 World map showing the distribution of the halocyprid ostracod genus Danielopolina. One additional species of Danielopolina inhabits deep waters 

in the mid-Atlantic Ocean. 

Decapoda: Shrimps, crabs, and lobsters. Anchialine 
shrimps include 18 genera from four superfamilies: 
Alpheoidea, Oplophoroidea, Palaemonoidea, and 
Procaridoidea. The galatheid crab Munidopsis 
polymorpha inhabits an anchialine lava tube in the 
Canary Islands. Two species of brachyuran crabs occur 
in anchialine habitats in the Galapagos Islands, while 
a third brachyuran inhabits anchialine limestone caves 
in the Solomon Islands. 

Troglomorphic Adaptations 

Most obligate anchialine taxa exhibit typical troglomorphic 
adaptations to a lightless, food-poor cave environment. 
These include varying degrees of eye loss and depigmenta- 
tion, increased tactile and chemical sensitivity, and larger but 
fewer eggs. Similar adaptations are also found in marine 
organisms inhabiting other lightless environments such as 
the interstitial and deep sea. Total depigmentation and loss 
of eyes are found in anchialine polynoid polychaetes, 
remipedes, halocyprid ostracods, epacteriscid and misoph- 

rioid copepods, cirolanid isopods, leptostracans, thermos- 
baenaceans, chaeotognaths, fishes, some shrimps, and 
numerous amphipods. Such pronounced troglomorphic 
adaptations usually occur in more primitive taxa that appa- 
rently have been restricted to caves for long time periods. 

As with deep-sea shrimps, many anchialine shrimps are 
scarlet to blood-red in color. Ten species of insular, Indo- 
Pacific anchialine shrimps, representing nine genera in five 
families, all possess red integumentary chromatophores. 
However, color display is variable between individuals within 
a population. In contrast, continental cave shrimps of the 
Indo-Pacific are colorless with the exception of Macro- 
brachium cavernicola, a red palaemonid from inland caves in 
India. The hippolytid shrimp Barbouria cubensis displays an 
intense red coloration when found in anchialine pools open 
to the surface but is colorless in totally dark caves. If these 
shrimps are moved from darkness to light, a color trans- 
formation occurs within a matter of minutes. 

Eyes are also present in many anchialine shrimps. Of 14 
species within the atyid genus Zyphlatya, pigmented eyes are 
found in 5 (mostly brackish or marine) species, suggesting a 
more recent adaptation to caves, while the remaining five 
(mostly freshwater) species lack eye pigment. Reduced but 

28 Anchialine Caves, Biodiversity in 

FIGURE 5 World map showing the distribution of the epacteriscid copepod genus Enantiosis. 

FIGURE 6 The peracarid crustacean Mictocaris halope inhabits anchialine 

caves in Bermuda. The order Mictacea was erected to include both cave and 
deep sea species. 

pigmented eyes are present even in some of the most primi- 
tive anchialine shrimps, such as Procaris hawaiana, Procaris 
chacei, Agostocaris williamsi, Yagerocaris cozumel, and 
Ventrocaris chaceorum. 

Primitive Characteristics 

Living fossils are organisms that occur in low numbers, have 
few close relatives, inhabit small geographic areas, and evolve 

FIGURE 7 A thermosbaenacean from Norman's Pond Cave in the Exuma 
Cays, Bahamas. 

at a very slow rate so that they have survived long periods of 
time with few changes. Anchialine stygobites such as the 
remipedes and the copepods Antrisocopia and Erebonectes fit 
these criteria and are regarded as living fossils. A number of 
anchialine crustaceans possess extremely primitive character- 
istics. Anchialine caves contain many of the most primitive 
known copepods in the orders Calanoida, Misophrioida, and 
Platycopioida. The platycopioid Antrisocopia prehesilis from 
Bermuda caves resembles in many ways with the theoretical 

ancestral copepod. Evebonectes nesioticus, also from Bermuda, 
is considered one of the most primitive calanoids due to its 
first antennae consisting of 27 segments (the highest number 
ever recorded), essentially unmodified mouthparts, and 
primitive legs. Misophrioid copepods are of considerable 
phylogenetic interest because they represent the first branch 
to diverge from the main podoplean lineage within the 
Copepoda. Aside from the order Platycopioida, copepods are 
divided into two superorders: Gymnoplea including the 
Calanoida and Podoplea with the remaining eight orders 
including the Misophrioida. The presence of numerous 
primitive and apparently ancient taxa in anchialine caves 
attests to the great age and long-term stability of this habitat. 
Thus, anchialine caves act as preserving centers for relict 
animals known nowhere else on Earth. 

Anomalous Biogeographical Distribution 

Despite being limited to caves, many anchialine taxa have 
a broad and highly disjunct distribution, thus raising 
questions as to their origins. A variety of genera contain 
species inhabiting caves on opposite sides of the Atlantic 
Ocean. These include the remipede Speleonectes, the anthurid 
isopod Curassanthura, the mysid Heteromysoides, the ostracod 

Anchialine Caves, Biodiversity in 29 

Danielopolina, the thermosbaenacean Halosbaena, the poly- 
chaete Namanereis, the shrimp Typhlatya, and the amphipods 
Gevegiella, Pseudoniphargus, and Spelaeonicippe. Such an 
amphiatlantic distribution suggests a Tethyan origin, with 
dispersal by plate-tectonic rafting. 

Other anchialine genera, previously known from caves 
in the Caribbean and Atlantic, are found in the Galapagos 
Islands, located 1000 km off the South American coast in 
the eastern equatorial Pacific. Included in this group are the 
shrimp Typhlatya, the ostracod Danielopolina, and the 
copepods Enantiosis and Expansophria. The 
derived anchialine fauna of the Galapagos must have entered 


the Pacific prior to closure of the Panama land bridge about 
3 to 5 million years ago. 

Even more complex zoogeographic relationships exist 
for other anchialine taxa. The primitive shrimp Procaris is 
known only from caves and anchialine pools on Hawaii in 
the mid-Pacific, Ascension Island in the South Atlantic, 
Bermuda in the North Atlantic, and Cozumel in the 
Caribbean (Fig. 8). Equally perplexing is the case of the 
misophrioid copepod genus Expansophria. It is represented 
by cave species from Palau and the Galapagos Islands on 

opposite sides of the Pacific and from the Canary Islands 
in the Atlantic. 

FIGURE 8 World map showing the distribution of the anchialine shrimp genus Procaris. 

30 Anchialine Caves 

FIGURE 9 The polychaete Pelagomacellicephala iliffei belongs to the family 
Polynoidae, which is known primarily from the deep sea, including bathyal 

and abyssal depths. 

Anchialine taxa, originally thought to be limited to or 
known primarily from the Atlantic and Caribbean, have 
recently been discovered in caves in the Cape Range of 
Western Australia. These include the remipede Lasionectes, 
the isopod Haptolana, the thermosbaenacean Halosbaena, the 
ostracod Danielopolina, and the amphipod Liagoceradocus. 
The wide range and overlapping distributions of these 
taxa suggest considerable age coupled with similar dispersal 

mechanisms and simultaneous colonization of the present 

Deep-Sea Affinities 

A considerable number of anchialine species have close 
relatives inhabiting the deep sea. Anchialine genera with 
affinities to present-day bathyal taxa include the galatheid 
crab Munidopsis, the polynoid polychaetes Giesiella and 
Pelagomacellicephala (Fig. 9), the mictacean Mictocaris, the 
ostracod Danielopolina, the amphipod Spelaeonicippe, and 
the misophrioid copepods Bosxshallia, Dimisophria, 
Expansophria, Misophria, Palpophria, and Speleophria. 


Boxshall, G.A. and D. Jaume (2000) Discoveries of cave misophrioids 
(Crustacea: Copepoda) shed new light on the origin of anchialine faunas. 
Zoologischer Anzeiger 239: 1-19. 

Fosshagen, A., G.A. Boxshall, and T.M. Iliffe (2001) The Epacteriscidae: a 
cave-living family of calanoid copepods. Sarsia 86: 245-318. 

Holthuis, L.B. (1973) Caridean shrimp found in land locked saltwater pools 
at four Indo-West Pacific localities (Sinai Peninsula, Funafuti Atoll, Maui 
and Hawaii Islands), with the description of one new genus and four new 
species. Zoologische Verhandelingen 128: 1-48. 

Humphreys, W.F. (2000) The hypogean fauna of the Cape Range Peninsula 
and Barrow Island, Northwestern Australia. In Subterranean Ecosystems 
(H. Wilkins, D.C. Culver, and W.F. Humphreys, Eds.). Elsevier Press, 
Amsterdam, pp. 581-601. 

Iliffe, T.M. (2000) Anchialine cave ecology. In Subterranean Ecosystems (H. 
Wilkins, D.C. Culver, and W.E Humphreys, Eds.). Elsevier Press, 
Amsterdam, pp. 59-76. 

Iliffe, T.M. (2004) Anchialine caves and cave fauna of the world 

lliffe, T.M., J. Parzefall, and H. Wilkens (2000) Ecology and species 

distribution of the Monte Carona lava tunnel on Lanzarote (Canary 

Islands). In Subterranean Ecosystems (H. Wilkins, D.C. Culver, and W.F. 
Humphreys, Eds.). Elsevier Press, Amsterdam, pp. 633-644. 

Pesce, G.L. (2004) Anchialine waters of Italy ( 

Riedl, R. (1966) Biologie der Meereshihlen, Paul Parey, Hamburg, 636 pp. 

Anchialine Caves 

Boris Sket 

Univerza v Ljubljani, Slovenia 

nehihaline (or anchialine) habitats are water bodies in 

hollows along the sea coasts where the influence of the sea 
may be felt and which are inhabited by some subterranean 
species. This usually indicates an underground connection of 
the cave or a pool with the sea. Such a habitat may contain 
seawater, but it primarily has layers of different brackish 
salinities. In exceptional cases, it has freshwater but is 
inhabited by some animals of the recent marine provenience 
(Sket, 1996). We must, however, exclude genuine sea caves— 
“spaces ... containing a high number of speleophilous forms 
of the rocky littoral with exclusion of any speleobiotic 
species” (Riedl, 1966). 

In principle, anchihaline habitats can occur everywhere 
along the coast, but hardly any have been noted and 
investigated outside the tropical or warm moderate climatic 
zones. Geologically, two very different types of coasts are 
often of interest. Some anchihaline caves are in karstified 
limestones, while others are in lava fields. Because the sea level 
along most coasts during the Ice Ages was approximately 100 
meters lower than now, the coasts may be hollowed out to 
such a depth. They often contain dripstone formations (such 
as stalactites and stalagmites), indicating their own previous 
continental and terrestrial nature. The limestones may be of 
different ages; some being old and primarily compact, with 
others being quite recently formed and still primarily porous 
coral limestones or little transformed coral reefs. Similarly, 
anchihaline habitats are also formed in lava flows, either in 
systems of cracks or in the tunnel-shaped lava tubes. 


Although well known since ancient times, the so-called blue 
caves in the Mediterranean (¢.g., Capri, Italy; BiSevo, Croatia) 
were never investigated speleobiologically, as their morpho- 
logy does not promise extensive anchihaline habitats in them. 
A number of localities with anchihaline caves have been 
discovered and investigated along tropical seas and in the 
Mediterranean. Since the middle of the 19th century, 
brackish habitats of some fish and shrimp species in Cuba 
have been known. Another historical site of this kind is 

Jameo del Agua, a segment of a lava tube on the island of 
Lanzarote (Islas Canarias). The 20th century began with the 
discovery of interesting fauna in some south Italian caves. 
Important anchihaline caves were later sampled in Lybia, and 
in islands of the Caribbean. 

Ried! (1966), who systematically studied sea caves of the 
Mediterranean, conceived the concept of the marginal cave 
(“Randhoehle” in German), which was similar to the 
present-day concept of the anchi(h)aline cave. Ried| defined 
the marginal cave by its position at the sea coast, the presence 
of brackish water, and the presence of stygobiotic species in 
it, but he also supposed that the origin of such caves was due 
to wave action and their gradual loss of connection to the sea. 
Thus, they should be a vector of the stygobiotic fauna 
originating in the sea and making it limnic. 

An important study by Holthuis (1973) followed Riedl’s 
studies and dealt with a number of shrimp species, including 
the now well-known red shrimps, as well as with a new 
concept of anchialine pools that was free of suppositions 
about the past and the future of the habitat and therefore 
easier to apply. However, Holthuis obtained his animals 
from open pools in Hawaii and on the Red Sea coast which 
contained some troglomorphic animals along with a very 
scarce selection of normal marine animals. In the period that 
followed, Stock and his numerous colleagues investigated a 
number of anchihaline caves in the Caribbean, mainly in 
regard to their faunistic and taxonomic features but also bio- 
geographically. This region appeared to be particularly rich in 
such caves and fauna (e.g., Stock, 1994). Stock suggested and 
substantiated the spelling change of anchialine to anchihaline. 

Sket studied the caves along the Adriatic coast and gave the 
first detailed quantitative data about ecological conditions in 
them (Sket, 1986). Later, he also specified the concept of 
the anchihaline cave habitat. Iliffe (2001) is noted for his 
sampling of a great number of caves in the Caribbean region 
as well as in the tropical islands of the Indopacific. His studies 
revealed a high number of new animal species representing 
an array of groups. In the meantime, a number of other field 
researchers and taxonomists were either sampling caves more 
locally or describing and studying collected animals. 

Yager (1981) discovered the new crustacean group 
Remipedia, probably the most astonishing discovery in 
anchihaline waters. About the time of this discovery, research 
intensified into the blue holes and inland anchihaline caves 
in the Bahamas. Por, Fosshagen and Boxshall studied 
Copepoda and Tantulocarida; Kornicker and Danielopol 
studied Ostracoda; Stock and Holsinger studied Amphipoda; 
Holthuis and others studied Caridea; and Stock and 
Botosaneanu studied Isopoda. Jaume noted euhaline 
anchihaline caves in the Mediterranean (e.g., Jaume and 
Boxshall, 1996), and Por described hyperhaline pools in 
the Arabian peninsula. A number of authors discussed the 
models and hypotheses regarding the colonization of 
subterranean habitats by animals in which the anchihaline 
caves played a not negligible role. 

Anchialine Caves 31 

(See Por, 1985; Sket, 1986, 1996; Iliffe, 2001.) The morpho- 

logy of the anchihaline caves and other cavities containing 
anchihaline habitats is of low importance for their fauna; 
however, it may be highly important for its accessibility 
(particularly for the time before SCUBA techniques were 
introduced) and to gain an understanding of its composition 
and distribution, as well as relations between species. 
Comparatively scarce are horizontal caves that either are 
filled with water along the entire corridor or exhibit series of 
pools with different connections between themselves and 
with the sea. Such caves can be either karst caves or lava 
tubes. Horizontal coastal caves may, however, be very exten- 
sive inundated cave systems, such as those being investigated 
along the Mexican coasts. 

More common are shallow, vertical caves. They may be of 
small dimensions, as are the natural wells on Adriatic islands 
(Croatia, within the former Yugoslavia), which are sometimes 
artificially elaborated as watering places for sheep. Similar are 
grietas in some Caribbean islands or cracks in lava fields of 
the Galapagos; however, accessible vertical parts are always 
connected with horizontal corridors or fissures that may 
connect groups of caves. Of large dimensions are cenotes, 
known mainly in the Yucatan peninsula (Mexico), which 
again are often extended by systems of horizontal corridors. 
An important characteristic of such habitats is that a signi- 
ficant part of the aqueous body is illuminated by daylight. 

Even more open are anchihaline pools in lava or in coral 
reefs. Such pools must be connected to the sea and may be 
interconnected by systems of more or less dark corridors or 
fissures, although the open space may prevail. Such pools are 
particularly well developed in Hawaii. 

Due to the effect of chained vessels, in caves close to coasts 
under the influence of tides, the tides may also be felt 
underground, although with a certain delay and decline. This 
means that in the mild cases, such as in the Adriatic coastal 
caves, the entire stratified column of water moves twice a day 
up and down, carrying the floating organisms with itself and 
remarkably changing the life conditions for substratum- 
bound animals, particularly in the reach of the halocline. The 
tides in the open Adriatic are only approximately 0.5 meters, 
but more extreme conditions may occur elsewhere. In the 
Bahamian blue holes, the effects of very important tidal 
currents (dangerous for a diver) have not yet been studied for 
their influence on living conditions. In the horizontal St. 
Paul Underground River corridor (Palawan, Philippines), the 
seawater moves with the tides for some kilometers in and out 
along the bottom. It is difficult, however, to imagine that an 
anchihaline fauna exists in such a cave. 

The aqueous body is seldom purely marine (euhaline) but 
such are present in some deeply set submarine caves that have 
to be very extensive to contain any stygobionts. However, 
extensive portions of the cave systems in the underground 
of the Bahamas (Porter), as well as some caves in the 

32 Anchialine Caves 

Mediterranean islands (Jaume) are functionally euhaline. 
When influenced by freshwater, the situation may be very 
different for a cave on a small island than one that is on 
the mainland coast. In any case, the relation between the 
fresh and marine part of the column is governed by 
the Ghyben—Herzberg principle: In a porous aquifer, if the 
freshwater overlies seawater, the interface will be below the 
sea level approximately 40 times the height of the water 
above sea level. On an island, the freshwater body will take 
on the form of a lens (Ghyben—Herzberg lens), the depth of 
which depends on the precipitation at the site and the speed 
of the water outflow toward the sea. The shape of the 
freshwater lens and the height of the underlying layers of 
seawater will change slowly during the year and from year 
to year. Less predictable and more complex is the situation 
on the coast of the continent (or of a larger island), where 
canalized (thus concentrated) water inflow from inland may 
be the result of net-like systems of corridors. This may 
complicate the salinity distribution and will sometimes 
change the salinity relations very rapidly. 

The transition from euhaline to limnic salinities in 
the column is never absolutely gradual. The depth of very 
gradually stratified or even homogeneous waters is regularly 
divided by a halocline that can be a few decimeters to a 
couple of meters thick. The halocline is a layer where the 
salinity abruptly changes (sometimes from polyhaline to 
nearly limnic). In genuine caves—such as in caves in 
general—the water temperature may be constant and equal 
throughout the column, approximately equal to the local 
yearly average. If the water body is exposed to the surface 
temperature influences, this may also cause a temperature 
stratification, changing its direction throughout the year. 
During the winter in moderate climates, the surface cooling 
in freshwater lakes causes the loss of stratification and 
penetration of cooler (z.¢., heavier) and well-aerated surface 
water toward the bottom. In anchihaline waters, the density 
changes caused by varying temperatures are much lower than 
the density differences caused by a small salinity difference. 
With cooling from 20 to 4°C, the density of pure water 
changes from 998 to 1000 g L™ (i.e., 2 g L), while with the 
increase of salinity from 10 to 20 ppt the density changes 
from 1009 to 1019gL™ (ie, 10gL"). This makes the 
water column in mixohaline anchihaline caves very stable 
and prevents mixing of layers, with consequences as 
explained below. 

Although the source of organic input can appear rather 
obvious, some sources have been also traced by sophisticated 
isotope methods (Pohlman et a/., 2000). In caves of Mexico, 
the soil particles may originate from a tropical rain forest, 
while in the Mediterranean these particles may be from the 
poor remains of soil in deep karren. The use of the holes as 
watering places, however, may cause the introduction of very 
high quantities of debris which is deposited on the bottom. 
Another source of organic matters is production in illumi- 
nated parts of caves. Rich planctonic algal communities may 

be present there, while rocky walls can also be overgrown by 
microscopic algae. The virtual contribution of autotrophic 
bacteria has only been inferred (supposed). Rich sources of 
oxydizable substances can be used in such a way. Some 
nitrate production (which may be by nitrite oxydation) has 
been noted within the halocline in a Mexican cave. On the 
other hand, remarkable hydrogen sulfide (HS) concentra- 
tions in many anchihaline caves are a strong basis for chemo- 
autotrophy; clouds of bacteria within the halocline and mats 
covering the bottom have been observed in some caves, and 
the sulfide-oxidizing mixotrophic bacteria of the genera 
Beggiatoa and Thiothrix have been identified. 

Wherever the input of organic substances is remarkable, 
they are deposited on the bottom and putrefy, which often 
causes total consumption of oxygen and an accumulation 
of hydrogen sulfide. If communication with the open sea is 
feeble, a thick lower layer of the water may be infested by 
this gas. A 250-mmol L™ concentration was established in a 
Mexican cave, while such remarkable concentrations have 
only been organoleptically established in Adriatic caves. Such 
an accumulation can be identified up to the water surface in 
extreme cases. 

The oxygen that is consumed in cave waters must be 
renewed from air on the surface, by an influx of precipitation 
water and water jets from inland, at the surface, or by mixing 
or diffusion from the open sea in deeper layers. Therefore, 
the surface layer is usually comparatively well aerated while 
the oxygen concentration diminishes towards the halocline; 
below the halocline it may rise again or decrease further. It is 
the rule that dissolved oxygen does not reach its saturation 
in anchihaline waters. Its depletion may be remarkable, 
probably total, although no analysis until now has been 
accurate enough to establish this. Such are the conditions 
close to the debris-covered bottom under a deep layer of 
hydrogen sulfide-rich water. In some other caves, the oxygen 
saturation is the lowest or close to zero within the halocline, 
the most stable water layer. Characteristically, the pH has also 
always been lowest in the halocline when measured. 


Sponges are represented very richly in true marine caves, 
sometimes with some deep-sea elements, such as the calcified 
Calcarea Pharetronida. This is particularly evident in some 
sack-like Mediterranean caves that capture and hold cold 
water throughout the year and also harbor small glass 
sponges (Hexactinellida) and the exceptional carnivorous 
Cladorhizidae (Asbestopluma hypogea Vacelet et Boury- 
Esnault) (Vacelet et al., 1994). Some species are claimed to be 
true stygobionts of anchihaline caves, such as the haplosclerid 
Pellina penicilliformis Van Soest and Sass in marine caves in 
Bahamas and the axinellid Higginsia ciccaresei Pansini and 
Pesce in Southern Italy; the latter is a typical anchihaline 
animal living in brackish water. 

Very little is known about gastropods in anchihaline caves. 
The presence of the tiny and mostly interstitial marine 
Caecum spp. is most probably in connection with some sand 
deposits within the cave. Only some of the apparently hydro- 
bioid snails in Adriatic caves have yet been identified. 

Among the few filter feeders are some tube worms occur- 
ring quite regularly at high salinities, such as Filogranula 
annulata (O.G., Costa) in Adriatic caves. Some stygobiotic errant 
polychaetes have been described, such as, for example, Gesiella 
jameensis (Hartmann-Schroeder) from the Canaries. Probably 
more sediment dependent are the tiny members of the family 
Nerillidae. It is remarkable that the Dinaric cave tube worm 
(Marifugia cavatica Absolon and Hrabe) has never been found 
in a cave within the influence of the sea. Some oligochaetes occur- 
ring in such caves do not seem to be closely related to them. 

Some species of sessile chaetognaths (family Spadellidae) 
occur in euhaline anchihaline waters, and at least the eyeless 
and pigmentless Paraspadella anops Bowman et Bieri from 
the Bahamas is supposed to be stygobiotic; some other 
species from caves are oculated. 

Fishes are among the longest known anchihaline animals; 
approximately 10 anchihaline species are known today. The 
family Ophidiidae is exapted for such habitats by their shy 
behavior, nocturnal habits, and/or deep-sea prevalence. The 
stygobiotic Lucifuga subterraneus Poey and Stygicola dentatus 
Poey have been recognized in Cuban grietas and similar 
localities since the middle of the 19th century. New, related 
species have recently been found in the region. The 
stygobiotic and euryhaline Ogi/bia galapagosensis (Poll and 
Leleup) from underground waters of the Galapagos Islands 
has variably reduced eyes and a closely eyed relative in the 
species O. deroyi (Van Mol), which dwells in littoral crevasses. 
In Yucatan caves, the ophidiid Ogilbia pearsei (Hubbs) is a 
neighbor of Ophisternon infernale (Hubbs), representing 
another family of often cryptically living fishes, Synbran- 
chidae. The synbranchid Ophisternon candidum (Mees), at up 
to 370 mm in length, is probably the longest cave animal. 
It inhabits brackish waters in northwestern Australia, 
accompanied by the gobiid Mylieringa veritas Whitley. It is 
remarkable that related (congeneric) species occur in 
anchihaline caves in mutually very remote areas. 


Generally, Crustacea are by far the richest and most diversely 
represented group in subterranean waters, and the same holds 
for anchihaline cave waters. The predominance of crustaceans 
may also be felt within a particular cave where the specimen 
number or the biomass relation between crustaceans and 
non-crustaceans may be 10:1 or higher. Among Crustacea, 
by far the richest in species are Amphipoda and Copepoda, 
although Decapoda (due to their size) are probably the most 
obvious. For Remipedia, the anchihaline caves are the only 
known habitat, and it seems to be the central one also for 

Anchialine Caves 33 

Remipedia are crustaceans with by far the most plesio- 
morphic trunk and appendages but with a highly specialized 
(diversified) anterior end of its body, a very nice case of 
mosaic evolution. They have up to now only been found in 
anchihaline caves in euhaline salinities, or close to them, 
usually below the halocline in poorly aerated water (typically 
less than 1 mg L™ oxygen). Remipedia are hermaphrodites. 
Up to 45-mm-long, centipede-shaped swimmers have been 
found in the tropical belt connecting Mexico, the Bahamas, 
and the Canaries and along the northwestern Australian 
coast; they seem to be absent in the Mediterranean. In 
some Bahamian caves, a number of species may occur 
together; Speleonectes lucayensis Yager was the first found and 

Copepods are the richest group of “Entomostraca”, which 
include approximately 25% of crustacean species either 
in the whole or in subterranean habitats, but only 14% of 
those in anchihaline waters. In subterranean freshwaters, 
Harpacticoida are several times richer than Cyclopoida, while 
Calanoida are negligible (approximately 440:150:10 in 
1986); in anchihaline waters, the harpacticoids are not so 
important, while calanoids are comparatively well re- 
presented. Each of these groups has some ecologically 
endemic families, or nearly so, such as Superornatiremidae 
in Harpacticoida, Epacteriscidae in Calanoida, and 
Speleoithonidae in Cyclopoida. According to existing data, a 
family may be also geographically extremely limited, such as 
the calanoid Boholinidae in the Philippines. Of particular 
interest is the small and relatively plesiomorphic group 
Misophrioida, an essentially (but not exclusively) deep-sea 
group discovered recently in Atlantic and Indo-Pacific 
tropical anchihaline caves. 

Ostracods are easily recognizable as tiny crustaceans with 
extremely shortened body hidden between two valves. The 
order Halocyprida is well represented with species of very 
diversely shaped and sculptured valves. Among them, the 
genus Danielopolina is interesting for its presence in 
anchihaline waters and in deep sea. 

Leptostracans are small malacostracans with a bivalved 
carapax on the anterior part of the body and stalked eyes. 
Speonebalia cannoni Bowman, Yager et Iliffe from a cenote- 
like cave in the Caicos Islands is eyeless. 

Decapods are the best known crustaceans represented in 
anchihaline caves, primarily by different kinds of shrimps. 
Members of the small family Procarididae have only been 
found in such habitats in the tropical Atlantic and in Hawaii; 
Agostocarididae seem to be restricted to the Caribbean 
region. Atyidae are a huge family of freshwater shrimps, 
represented at the sea nearly exclusively by some genera in 
anchihaline waters. The most remarkable is the genus 
Typhlatya, which is found with some closely related genera 
exhibiting a range of polyhaline to limnic species distributed 
circumtropically; they also exhibit different degrees of troglo- 
morphism. The atyid Antecaridina lauensis (Edmondson) 
alone is very widely distributed—Madagascar, Red Sea, 

34 Anchialine Caves 

Hawaii—which is testament to the dispersion possibilities 
of some shrimps outside the anchihaline habitat. This has 
been also proven for Palaemonidae, represented by the genus 
Macrobrachium, for example, which repeats some distrib- 
utional characteristics of the family Atyidae. The palaemonid 
Typhlocaris lethaea Parisi from Lybia is probably the largest 
cave invertebrate at all. A series of anchihaline genera belong 
to the essentially marine family Alpheidae. The alpheid 
Barbouria cubensis (Von Martens) in the Caribbean region 
has been known since the beginning of 20th century; 
the Indo-Pacific Parhippolyte uveae Borradaile, again one 
of the famous red shrimps, has been known from the end of 
the 19th century. Some Grapsidae crabs, such as Orcovita 
spp., have been found recently, although they are mostly not 
very troglomorphic. The only anomuran, the historical 
Munidopsis polymorpha Koelbel (family Galatheidae), from 
the inundated lava tube in Canary Islands, belongs to an 
essentially deep sea genus. 

Thermosbaenaceans are small (up to 5mm long) and 
delicate crustaceans with a short carapax under which the 
female broods her eggs; all are stygobiotic. Although the first 
species was found in hot (approximately 45°C) freshwater 
springs, and some species inhabit fresh continental waters, 
most are nevertheless anchihaline. The group is distributed in 
the Tethyan fashion, which will be described later. As an 
example, Halosbaena acanthura Stock is widely spread in 
the Caribbean, inhabiting interstitial and cave waters of 
brackish to hyperhaline salinities. Related species are known 
from the Canaries and northwestern Australia in almost 
fresh water. Monodella halophila S. Karaman, found in caves 
along the northeastern Adriatic coast, often most densely 
populates the least aerated water layers of a polyhaline 

Mysids are small and delicate shrimp-like crustaceans 
with a ventral brood chamber. It is essentially a marine 
group that inhabits anchihaline caves circumtropically, with 
some exceptional species in southern Italy. Such is also the 
distribution of the genus Spelaeomysis, which resembles the 
atyid shrimp Typhlatya in its geographical and ecological 
distribution and the span of morphological adaptations. 
The widely distributed and non-troglomorphic species S. 
cardisomae Bowman seems to penetrate the entire variably 
brackish part of the groundwater body in the Caribbean 
island of San Andres, occurring even in the putrefied and 
probably deoxygenated waters in the land crab burrows. 

Mictaceans slightly resemble the thermosbaenaceans. 
This is a small group of some deep-sea and two marine 
cavernicolous species in the Bermudas (Mictocaris halope 
Bowman and Iliffe and Bahamas. 

Isopods are one of the major groups of Malacostraca, 
characterized mainly by their foliaceous abdominal appen- 
dages (pleopods); they are very diverse in shape and size. 
Since early in the 20th century, some Typhlocirolana spp. 
from the Mediterranean basin have been known; some are 
anchihaline while some are from freshwaters. Similarly old is 

the history of Annina lacustris Budde-Lund from east Africa 
which is a very euryoecious, non-troglomorphic animal that 
also occurs in mixohaline anchihaline caves. All of these 
animals belong to the family Cirolanidae, which is essentially 
marine, and the great majority of freshwater species is 
stygobiotic. A number of genera and species have been found 
in anchihaline waters, particularly in the Caribbean, (such as 
Bahalana mayana Bowman in Mexico and the Yucatan), but 
also in the Mediterranean and in Indo-Pacific areas and 
elsewhere, such as Haptolana pholeta Bruce et Humphreys 
in western Australia. Similar is the frame habitat selection of 
the genus Cyathura (Anthuridae), although these animals are 
strongly sediment bound and burrow, while cirolanids are 
free walkers and even occasional swimmers. The stygobiont 
anthurids are only present in the tropics. 

The presence of tiny species of the families Microcerberidae 
and Microparasellidae in anchihaline waters such as artificial 
wells is mostly due to sand deposits as they are essentially 
interstitial animals. Genuine cave animals are evidently some 
members of the related (z¢., asellote) families of Janiridae 
(Trogloianiropsis lloberai Jaume) in the Balearic Islands; 
Stenetriidae (Neostenetroides schotteae Ortiz, Lalana et Perez) 
in Cuba; and Gnathostenetroididae (Stenobermuda mergens 
Botosaneanu and Iliffe) in the Bahamas, which were found 
in Caribbean to Bermudian anchihalines, as well as in 
Mediterranean caves. The family Atlantasellidae seems to be 
endemic to anchihaline caves of the tropical Atlantic 
(Bermuda and the Caribbean). 

Amphipoda seem to be by far the richest anchihaline group 
(more than twice the species number of any other crustacean 
group or three times all non-crustacean species), although 
the effect of numerous researchers cannot be excluded. 
The family level taxonomy of this group (counted to 6000 or 
even 8600 world species) is not yet settled, but it has to be 
used for the sake of surveying. The most characteristic group 
is comprised of Hadziidae and Melitidae with particularly 
high diversities of genera in the Caribbean region (e.g, 
Metaniphargus, Bahadzia) and Mediterranean (e.g., Hadzia) 
and with a number of scattered taxa elsewhere (e.¢., 
Liagoceradocus spp.); Western Australia seems to represent the 
third such diversity center (with the genus Nedsia, for 
example). The group as a whole, as well as some genera, are 
represented by species of different salinity preferences, from 
euhaline to limnic. On a smaller scale and limited to the 
Mediterranean and Atlantic, is the genus Pseudoniphargus; 
the majority of its species are freshwater but never very far 
from the sea. Taxonomicaly remote is the family Ingol- 
fiellidae; although including some marine species, it also 
includes ancient freshwater species in the center of Africa, for 
example. This family also has some anchihaline species. 

Some taxonomically and geographically isolated species 
seem to be more or less direct descendants of marine 
relatives, such as Antronicippe serrata Stock et Iliffe 
of Pardaliscidae, from anchihaline caves in Galapagos 

At least most species or groups mentioned above are 
primarily marine, conquering continental waters with more 
or less success. Different are some amphipods of the Dinaric 
(z.e., Adriatic) anchihaline caves. Some species of Niphargus 
are present there; NV. hebereri Schellenberg is moderately 
widely distributed, although not obligate, while the narrowly 
endemic NV. pectencoronatae Sket seems to be an obligate 
anchihaline species. Niphargids are a large group of essen- 
tially freshwater animals and only secondarily did they adapt 
to the brackish water; NV. hebereri is very euryhaline and also 
in other respects a very tolerant species. 


Different patterns of geographical distribution can be seen 
in anchihaline cave faunas. One of the most remarkable 
patterns is the so-called Tethyan distribution (Humphreys, 
2001), if historical grounds are emphasized, or circum- 
tropical distribution, if we suppose ecological connections. 
The Tethys Sea encircled the globe in the Mesozoic period, 
when the supercontinent Pangaea split into a northern 
Laurasia and southern Gondwana. It was only in the Tertiary 
when southern and northern continents came so close to 
each other that they disconnected this circumtropical seaway; 
note that even the Mediterranean was completely dry for 
some period in Miocene. A number of animal groups, 
even genera and some species, have distribution ranges that 
reflect the possible circumglobal distribution within the 
Tethys. We may presume that particularly in the tropical belt 
a number of Tethyan inhabitants have survived and many 
of them succeeded in colonizing marginal waters, avoiding 
in this way competition with modern sea fauna. Such are 
Remipedia, Thermosbaenacea, hadziid amphipods, the 
Typhlatya-related group of shrimp genera, and even the 
mysid genus Spelaeomysis, the thermosbaenacean genus 
Halosbaena, and some others. However, it is difficult to 
believe that some circumtropically spread species, such as the 
shrimp Antecaridina lauensis, could conserve its morpho- 
logical unity without a recent or nearly recent gene flow 
among populations; its genetic unity, however, has not been 
proven. An open question is also what occurred during the 
Messinian salinity crisis. Did the Tethyan elements tempo- 
rarily change for freshwaters? Maybe this is connected with 
the existence of some freshwater populations of some elements 
such as Hadzia fragilis S. Karaman in the Adriatic region. 

A certain number of those Tethyan elements character- 
istically skip the Mediterranean in their west—east distribu- 
tion, such as Remipedia and the genus Halosbaena; therefore, 
it is also possible that some Tethyan elements disappeared 
from the dried-out Messinian Mediterranean, which was 
later colonized again by only some of them. The third 
possibility is their disappearance during the colder climates 
in place after the Terciary. 

Other distribution patterns are more locally restricted. 
An instructive one is the paralittoral distribution area along 

Anchialine Caves 35 

the northeastern (Dinaric) coast of the Adriatic. Typical 
anchihaline taxa (such as Hadzia fragilis and Monodella 
halophila) are present along the coast, including the islands, 
in brackish waters. They are, however, absent in such waters 
within the gulf of Kvarner (Quarnero) which was not yet 
marine prior to the Pleistocene. These species are replaced 
in the Kvarner anchihaline caves by generally continental 
species, such as Niphargus arbiter G. Karaman, while Hadzia 
is represented there by freshwater populations. This is a clear 
evidence of a historical—not purely ecological—basis for the 
species distribution in anchihaline habitats. 


Unfortunately, for most anchihaline species the ecological 
inclinations are virtually unknown. Some groups are evidently 
represented only by euhaline or at least polyhaline species 
which are all unquestionable immigrants from the deep sea 
or close relatives of deep-sea animals: the glass sponges and 
carnivorous sponges in the Mediterranean sea caves, the 
anomuran Munidopsis polymorpha in the Canary Islands, and 
the mictacean Mictocaris halope in the Bermudas. There 
are also unique representatives of some marine groups such 
as chaetognaths or leptostracans (Paraspadella anops and 
Speonebalia cannoni); however, the completely anchihaline 
group Remipedia also seems to be polyhaline. 

The most anchihaline animals are to some degree 
euryhaline, with different optima more or less expressed. The 
distribution of shrimp species in a series of pools in Hawaii 
seems to depend mostly on different salinities. They are 
able to adapt their position in the water body to other local 
conditions, particularly the presence of competitors or 
predators. The physically delicate thermosbaenacean Monodella 
halophila, along the Adriatic coast, is limited to compar- 
atively high salinities, although its very close relative M. 
argentarii Stella inhabits freshwaters in the Italian Monte 
Argentario. The Adriatic caves are inhabited by some (most 
of them by two) amphipod species which may be predatory 
on Monodella. This might also explain the (supposed) 
absence of Monodella in some Caribbean islands where caves 
were probably not investigated meters deep and where 
hadziid amphipods were regularly found. Some essentially 
limnic species, such as the copepod Diacyclops antrincola 
Kiefer or the shrimp Troglocaris sp. may be present at higher 
salinities than the supposed marine derivatives Hadzia fragilis 
and Salentinella angelieri Ruffo et Delamare—Deboutteville, 
while the limnic Miphargus hebereri Schellenberg reaches 30 
ppt salinity, which is close to marine values. 

It was observed in some Adriatic anchihaline waters that 
at least some absolutely troglomorphic animals do not avoid 
sunlit parts of the water body if they are not forced to do so 
by competing surface animals. The population of Niphargus 
hebereri was the most dense at the water surface in a shaft-like 
cave. Such a behavior (as well as similar situations in conti- 

36 Anchialine Caves 

nental cave waters) teaches us that light and dark relations are 
not of direct importance for cave species. This also excuses 
us when we treat the open anchihaline pools jointly with 
anchihaline caves. The so-called red shrimps are very charac- 
teristic inhabitants of those pools, and they show quite 
different relations toward sunshine. The atyid Halocaridina 
rubra Holthuis also appears in open anchihaline pools during 
sunshine, while Antecaridina lauensis (Edmondson is more 
photophobic but is able to change its color from red to white 
translucent and back in response to light and disturbance. 
The hippolytid Parhippolyte uveae Borradaile hides in cracks 
after sunset and during cloudy days, while it reappears 
in ponds in masses when the sun strikes the water; it is 
also able to rapidly change its coloration. One has to add, 
however, that these observations were made in single 
localities, and we do not know if such was the behavior of the 
species in any conditions. 

Again, very characteristic is the sometimes apparently 
positive (in fact, probably neutral) relation of anchihaline 
animals toward low oxygen concentrations in water. These 
discussion has unfortunately some very serious limitations. 
First, due to some analysis problems at very low oxygen 
concentrations, we do not know if cave animals have been 
found in totally anoxic or just deeply suboxic water; however, 
the presence of H,S in a thick layer speaks in favor of a total 
lack of oxygen. Second, particularly high densities of some 
animal colonies within the anoxic (or suboxic) water layers 
are most probably a result of competition (or predation) 
avoidance; however, this speculative conclusion has never 
been subjected to an experimental verification. Anyway, at 
extremely low oxygen concentrations and in the presence of 
H,S (not quantitatively evaluated) euhaline Remipedia were 
found commonly or regularly; the mesohaline Monodella 
halophila, often; and the originally limnic Niphargus hebereri, 
quite so. Resistance to extremely low oxygen concentrations 
has been experimentally studied and verified in some 
freshwater cave crustaceans; it may be explained by their 
characteristically low metabolism. 

Thus, based on the evidence available, anchihaline animals 
are in general very euryecious (generalist) toward a number 
of ecological parameters. One of the most durable among 
stygobionts is Niphargus hebereri, which has been found in 
limnic conditions (kilometers far inland) and in mixohaline 
waters of up to 30 ppt salinity; in very food poor habitats 
and in deposits of decaying organic debris; in well- 
oxygenated waters as well as in apparently anoxic layers; in 
waters smelling strongly of HS; and in a sink polluted with 
mineral oil. 


Supposedly, the richest source of food is input from the earth 
surface, while some food may be produced in the cave itself, 
as mentioned above. It is safe to say that the currents reaching 
into a euhaline marine cave are bringing food particles as well 

as pelagic larvae. Therefore, such caves are usually populated 
with a depauperate marine fauna, and the stygobionts are to 
some degree excluded by competition. As the most modest 
users of organic matter and as ecological generalists (as shown 
above), they are able to populate the rest of the coastal 
underground. Some interesting studies of the organic matter 
transport within genuinely marine and anchihaline caves 
have been made but generalizations are still not possible. The 
sessile suspension feeders (filtrators) are nearly absent; some 
sponges and tube worms are on the border between the 
marine and anchihaline realm. Very few observations have 
been done about the behavior of those crustaceans that 
possess putative filtering structures, such as atyid shrimps 
and thermosbaenaceans. They seem to be predominantly on 
the bottom and collect organic detritus particles or bacteria 
from the sediment. Probably more pelagic is the rich assort- 
ment of anchihaline calanoid copepods. On the other hand, 
some inhabitants are proven predators, such as the procaridid 
shrimp Procaris ascensionis Chace, that feed on other shrimps. 
Little is known about feeding by Remipedia, but their 
maxillipeds are explicitly predatory. Most probably the 
majority of anchihaline amphipods are generalistic, feeding 
on detritus and carcasses as well as on live animal prey. 


The position of anchihaline habitats on the doorstep between 
the epigean and hypogean realms, as well as between marine 
and freshwater environments, makes them very intriguing for 
theoretical exploitation. Several theories about the coloniza- 
tion of the underground include anchihaline habitats, either 
implicitly or explicitly. However, the first question that has to 
be answered regards the origin of the anchihaline fauna itself. 
The marine provenance of the majority is out of the question 
if we look at the taxonomic composition. Only one local 
fauna is known for which a higher percentage of members are 
freshwater by origin (anchihaline fauna along the Adriatic 
coast). The fact that the band inhabited by it was out of the 
sea during the last glaciation is not peculiar to that area. 
Probably, it was the particularly rich continental cave fauna 
that succeeded to rule anchihaline waters and thus prevented 
intrusion of other biota. 

A much-discussed question has been the relation of 
anchihaline fauna to the deep sea. Their phylogenetic prox- 
imities have been discussed previously. At one time it was 
thought that anchihaline species were directly derived from 
deep-sea species. Later, data about deoxygenation of deeper 
sea layers during the geological past cast doubts on this 
possibility, and some researchers came to regard the deep-sea 
species and anchihaline species as two parallel lineages ori- 
ginating from littoral ancestors. However, cladistic analysis 
of some ostracod and some copepod groups have revealed 
that at least some anchihaline species in some regions may 
originate from deep-sea faunas (Danielopol, 1990). 

The anchihaline habitat may also be regarded as a door 
into continental waters. The regression model and the two- 
step model suggest that a marine benthic animal colonizes 
marine cave, adapts to the cave environment and to the less 
saline water, and becomes isolated from the sea after the sea 
regresses, which results in a limnic stygobiont. These and 
other models differ in some details which become theore- 
tically negligible (making all models complementary) if we 
consider the ecologically and genetically sound supposition 
of an active immigration. Active immigration into a new 
habitat would be furthered by a diversified gene pool and 
wide ecological potential of the species and would be forced 
to occur only by its own normally excessive reproduction and 
population expansion. Sea regression (stranding) in this 
context is not an act of repression; it implies only inter- 
ruption of the gene flow (with the marine or surface part of 
the population) and therefore faster specialization. Of course, 
this is equally important for understanding the origin of the 
surface freshwater fauna, which may have occurred the same 
way on the surface, and the origin of continental cave fauna, 
which occurred by immigration from surface freshwater 


It has always been a challenge to humans to enter the dark 
underworld; sometimes such a venture reveals the beauty of 
speleothems or offers the promise of a cool atmosphere in the 
heat of summer. So, many karst and lava caves have been 
exploited for tourism, and those that are close to coastal 
tourist resorts are particularly subject to such a possibility. 
Some of them contain anchihaline pools. 

The Sipun cave in Cavtat (near Dubrovnik, Croatia) was 
adapted for such visits decades ago, with the anchihaline pool 
at its end being shown as a special gem. Its comparatively 
very rich fauna of tiny crustaceans is not particularly interest- 
ing for most non-professionals but has nonetheless been 
greatly endangered by pollution produced by the visitors. 
Anchihaline caves play an important role in tourism and the 
economy of Bermuda; nevertheless, a number of them have 
been damaged or destroyed—along with their diverse 
fauna—by pollution or by quarrying activities. On the other 
hand, the lava cave Jameo del Agua in Lanzarote (Islas 
Canarias, Spain) has also been exploited for tourism and its 
jameito or cangrejo ciego (Munidopsis polymorpha) is an official 
symbol of the island that is used in various promotional 
campaigns. The gorgeous hypogean estuary St. Paul Under- 
ground River (Palawan, Philippines) is the central attraction 
of the St. Paul Subterranean National Park; unfortunately, 
there are no data on an anchihaline fauna in it. Very 
imposing is the 7-cm long anchihaline shrimp Typhlocaris 
lethaea from the cave Giok-Kebir, close to Benghazi (Libya); 
this might be the world’s largest troglobiotic invertebrate and 

Anchialine Caves 37 

could be an interesting subject for interested tourists. 
Although smaller, the beautifully colored anchihaline red 
shrimps can be remarkable enough to be given local verna- 
cular names such as opaeula (Halocaridina rubra) in Hawaii. 
They may be a subject of local taboos, such as pulang pasayan 
(Parhippolyte uveae) in the Philippines, or even of a kind of 
worship, such as ura buta (the same species) on the Fiji Island 
of Vatulele. No doubt, red shrimps might become also an 
attraction for ecotourists. 

Apart from species protection to serve tourist needs or 
religious habits, modern humans have became aware of the 
biological importance of such habitats. A system of anchi- 
haline pools on the Hawaiian island of Maui has been pro- 
tected as a natural reserve particularly for the rich anchihaline 
fauna. A flooded crack in the Sinai Peninsula has (or used to 
have) a similar status. 

See Also the Following Articles 
Anchialine Caves, Biodiversity in 


Danielopol, D.L. (1990) The origin of the anchialine fauna: the “deep sea” 
versus the “shallow water” hypothesis tested against the empirical 
evidence of the Thaumatocyprididae (Ostracoda). Bijdragen tot de 
Dierkunde 60(3-4): 137-143. 

Holthuis, L.B. (1973) Caridean shrimps found in land-locked saltwater 
pools at four Indo-west Pacific localities (Sinai Peninsula, Funafuti Atoll, 
Maui and Hawaii Islands), with the description of the new genus and 
four new species. Zoologische Verhandelingen 128: 1-48. 

Humphreys, W.F. (2001) Relict faunas and their derivation. In Subterranean 
Ecosystems (H. Wilkens, D.C. Culver, and W.E Humphreys, Eds.). 
Elsevier Press, Amsterdam, pp. 417-428. 

Iliffe, T.M. (2001) Anchialine cave ecology. In Subterranean Ecosystems (H. 
Wilkens, D.C. Culver, and W.E. Humphreys, Eds.). Elsevier Press, 
Amsterdam, pp. 59-73. 

Jaume, D. and G.A. Boxshall (1996) The persistence of an ancient marine 
fauna in Mediterranean waters: new evidence from misophrioid 
copepods living in anchialine caves. J. Natural History 30, 1583-1595. 

Pohlman, J.W., L.A. Cifuentes, and T.M. Iliffe (2001) Food web dynamics 
and biogeochemistry of anchialine caves: a stable isotope approach. In 
Subterranean Ecosystems (H. Wilkens, D.C. Culver, and W.E. Humphreys, 
Eds.). Elsevier Press, Amsterdam, pp. 345-356. 

Por, ED. (1985) Anchialine pools: comparative hydrobiology. In Hypersaline 
Ecosystems: The Gavish Sabkha (G.M. Friedman and W.E. Krumbein, 
Eds.). Ecological Studies: Analysis and Synthesis, Vol. 53, Springer, New 
York, pp. 136-144. 

Riedl, R. (1966) Biologie der Meereshoehlen [Biology of Sea Caves. Parey, 
Hamburg/Berlin, 636 pp. 

Sket, B. (1986) Ecology of the mixohaline hypogean fauna along the 
Yugoslav coast. Stygologia 2(4): 317-338. 

Sket, B. (1996) The ecology of the anchihaline caves. Trends Ecol. Evol. 
11(5): 221-225. 

Stock, J.H. (1994) Biogeographic synthesis of the insular groundwater 
faunas of the (sub)tropical Atlantic. Hydrobiologia 287, 105-117. 

Vacelet, J., N. Boury-Esnault, and J.-G. Harmelin (1994) Hexactinellid 
cave, a unique deep-sea habitat in the scuba zone, Deep-Sea Res. I 41(7), 

Yager, J. (1981) Remipedia, a new class of Crustacea from a marine cave in 

the Bahamas. / Crustacean Biol 1(3), 328-333. 


Susan W. Murray and Thomas H. Kunz 

Boston University 


Bats (order Chiroptera) are an ecologically diverse and geo- 
graphically widespread mammalian group. With over 1100 
species (Simmons, 2003), members of two suborders (Micro- 
chiroptera and Megachiroptera) constitute approximately 
one-fifth of all extant mammals. They range in size from the 
tiny 2-gram Kitti’s hog-nosed bat (Craseonycteris thonglogyai) 
to the large, Malayan flying fox (Preropus vampyrus) weighing 
about 1200 grams, with a wingspan of nearly 1.5 meters 
(Kunz and Pierson, 1994). Bats exhibit several life history 
traits that make them unique among mammals. Compared 
to small terrestrial mammals, for example, bats of similar size 
have very few young and long periods of pregnancy and 
lactation, and they may live up to 37 years. Differences in 
life-history traits between bats and other mammals are often 
attributed to the evolution of flight and echolocation 
(Crichton and Krutzsch, 2000; Kunz and Fenton, 2003). 

Bats are the only mammals that have evolved powered 
flight; thus, along with echolocation, flight has made it 
possible for bats to seek shelter in many different types of 
structures (¢.g., foliage, tree cavities, caves, rock crevices) that 
are not generally used by terrestrial mammals and to exploit 
a wide variety of food sources. The diets of bats include fruit, 
nectar, pollen, leaves, invertebrates (¢.g., insects, spiders, 
crustaceans, and scorpions), small vertebrates (e.g., frogs, 
birds, fish, and other mammals), and blood (Kunz and 
Pierson, 1994). 

Although echolocation has evolved independently in birds 
and mammals, the most sophisticated and diversified form of 
echolocation can be found in bats. Echolocation is used for 

prey detection and capture, for navigation, and in some 
instances for communication. Several species rely on a 
combination of vision, olfaction, and prey-generated sounds 
to locate food in addition to or instead of echolocation 
(Altringham, 1996). 

The roosting habits of bats are often highly specialized, 
with different species occupying tree cavities; spaces beneath 
exfoliating bark; unmodified foliage; leaves modified into so 
called “tents;” abandoned ant, termite, and bird nests; large 
and small caves; rock crevices; and a wide range of manmade 
structures, including mines, buildings, stone ruins, and 
bridges (Kunz, 1982; Kunz and Fenton, 2003). Caves 
alone provide a variety of structural substrates for roosting, 
including crevices, cavities, textured walls and ceilings, 
expansive ceilings, rock outcrops, and rock rubble on floors. 
In addition, the microclimates of caves that are occupied by 
bats can vary enormously, depending on latitude, altitude, 
depth, and volume, as well as the number, size, and position 
of openings to the outside. These variables can influence the 
amount of airflow, the presence of flowing and standing 
water, and daily and seasonal variations in atmospheric 
pressure, temperature, and humidity. Thus, the environ- 
mental conditions within caves may be hot, cold, dry, humid, 
still, or windy. 

This chapter highlights the biology of bats that typically 
roost in caves and cave-like structures. Specifically, we discuss 
why bats live in caves, where they are found, their roosting 
requirements, as well as conservation and management issues 
important to protecting cave-dwelling species. 


Cave bats are defined as trogloxenes, species that do not 
complete their entire life cycle within caves. Their ability to 
fly and echolocate has allowed microbats to exploit caves and 
similar subterranean habitats for roosts and to forage for food 
away from these structures. The vast majority of bats that 


40 Bats 

belong to the suborder Microchiroptera are able to 
echolocate and navigate in dark underground habitats and to 
feed in open fields or in dense forested areas. Rousette fruit 
bats (suborder Megachiroptera) may also roost in caves, but 
they rely on tongue clicks to produce audible sounds to help 
them navigate in the dark. The handful of other mega- 
chiropterans found roosting in caves are restricted to areas 
with enough light to make it possible for them to find their 
way in and out, as they largely rely on vision and not echolo- 
cation to navigate while in flight (Kunz and Pierson, 1994). 

Bats are virtually ubiquitous. They are known from all 
continents, except Antarctica, and from many oceanic 
islands. Most bats occur in tropical regions, where they are 
often the most diverse and abundant mammals present. The 
diversity of bats generally increases as one travels from the 
poles toward the equator, a pattern that is largely attributable 
to an increase in habitat complexity as latitude decreases 
(Findley, 1993; Kunz and Fenton, 2003). 

The distribution of cave bats not only depends on the 
presence of caves, but is also a consequence of specific 
roosting requirements (Kunz, 1982). For example, although 
the ghost bat (Macroderma gigas), the orange leaf-nosed bat 
(Rhinonycteris aurantius), and the large bent-wing bat 
(Miniopterus schreibersii) are all found in Australian caves, 
their roosting requirements and hence geographic distrib- 
utions are quite different (Baudinette et al, 2000). The 
orange leaf-nosed bat selects caves that are extremely hot and 
humid (28 to 30°C and greater than 94% relative humidity) 
and currently are known from only ten caves in Australia. In 
contrast, the large bent-wing bat can be found roosting at a 
broader range of temperatures and humidities and has one 
of the widest reported distributions of cave-roosting bats, 
encompassing southern Europe, Africa, southeast Asia, 
Japan, and Australia. The ghost bat is restricted in its distrib- 
ution to the Northern Territories of Australia, but it occupies 
far more caves and mines than the orange leaf-nosed bat. As 
one might expect, the ghost bat has roosting requirements for 
which humidity is not such a vital condition as compared to 
the orange leaf-nosed bat. 

Bats are found almost everywhere subterranean habitats 
exist. The distributions of cave-dwelling bats are determined 
largely by species-specific roosting requirements that vary 
depending on their ecology and evolutionary history. Local 
and global distributions and densities of bats that rely on 
caves for at least part of their life cycle are in turn determined 
largely by the distribution, quantity, and characteristics of 
available caves. For example, the Townsend’s long-eared bat 
(Corynorhinus townsendii), known only from North America 
(Fig. 1), is primarily found roosting in caves and cave-like 
structures (Wilson and Ruff, 1999). 


Caves are used by bats for a variety of reasons, including 
courtship and mating, raising young, and hibernating. 

FIGURE 1 Townsend’s long-eared bat (Corynorhinus townsendii) is an 

insectivorous, cave-dwelling bat endemic to North America. (Photograph by 

J.S. Altenbach.) 

Bats seek shelter during the day and disperse from these 
sites to forage for food at night. During the day, bats typically 
rest, groom, and often interact with their roost-mates. For 
example, a typical day for lactating female lesser long-nosed 
bats (Leptonycteris curasoae) involves resting quietly for up 
to 16 hours, interspersed with periodic grooming and 
nursing behavior. Although females usually roost together in 
caves during the day, they seldom interact with one another 
(Fleming e¢ a/., 1998). In contrast, the common vampire bat 
(Desmodus rotundus) forms long-term social bonds, and 
individuals groom one another as they interact socially while 
occupying cave and tree roosts (Greenhall and Schmidt, 
1988). In addition, many insectivorous species retreat to 
caves between feeding bouts, where they may cull the wings 
and heads of insects that were captured while foraging. 
Frugivorous species sometimes transport large fruits to caves 
where they cull soft pulp and where they can reduce the risks 
of predation. Some species that roost in foliage or tree cavities 
in the warm months hibernate in caves during the winter 
(Kunz, 1982; Kunz and Fenton, 2003). 

Courtship and Mating 

Several types of mating systems have been described for cave- 
roosting bats. Mating systems of bats and other mammals are 
often classified into three general categories: promiscuity, 
polygyny, and monogamy. However, bat mating systems 
cannot always be easily categorized into one of these groups, 
as they often depict a continuous spectrum of mating 
behaviors (Crichton and Krutzsch, 2000). 

Promiscuity is a type of mating system in which both 
males and females have multiple partners. Such a system is 
almost always highly structured, with some males siring more 
young than others. Promiscuity is common among temperate 
cave-dwelling species, possibly because of the limited time 
available for mating in autumn before individuals enter 
hibernation (Altringham, 1996). Males and females of many 
temperate species generally do not roost together during 

warm months but instead roost alone or in small groups. 
Assemblages of bats that gather at caves and mines in the 
autumn (referred to as swarming) may aid individuals in 
finding a mate. During the swarming season, bats are active in 
caves and mines at night, where males can often be observed 
displaying and chasing females. In the United Kingdom, 
male and female greater horseshoe bats (Rhinolophus ferru- 
mequinum) have a mating system in which males establish 
territorial sites inside caves and mines in early autumn. 
Females gather at these sites and selectively visit a series of 
different males on their territories (Crichton and Krutzsch, 

Polygyny, a mating system thought to be the most 
common in bats, is characterized by one male mating with 
several females (Crichton and Krutzsch, 2000). An example 
of this type of mating system can be observed in the greater 
spear-nosed bat (Phyllostomus hastatus). In this species, 
females roost in caves in small stable groups, often remaining 
together for 10 years or more. Because the females often form 
discrete roosting groups in solution cavities or “pot holes” on 
cave ceilings, it is easy for a dominant male to defend a group 
of females from intrusions by other males. By defending 
the females, or the roost cavity, a so-called harem male is 
often able to mate with several females. Sometimes these 
harem males are accompanied by a subordinate male who has 
positioned himself in the harem to assume a dominant role 
if the harem male should become injured or die. The risks 
and costs associated with mate-guarding behavior can be 
substantial. For example, a harem male greater spear-nosed 
bat may incur some injuries while defending the females or 
roost cavity and in the final analysis may sire only 60 to 90% 
of the young born to those females (Crichton and Krutzsch, 
2000). A similar pattern of mate guarding and courtship has 
been observed in the Jamaican fruit bat (Artibeus jamaicensis), 
which commonly roosts in caves on many of the islands in 
the West Indies and throughout Central and South America. 

Monogamy occurs when males and females form long- 
term pair bonds. This type of mating system has been 
described for only a few species of bats. Two examples are 
the African false vampire bat (Cardioderma cor) and the 
American false vampire bat (Vampyrum spectrum), both of 
whom are carnivorous, sit-and-wait predators. An extended 
period of parental involvement in which males provision 
both females and young may have contributed to the 
evolution of monogamy in these and other species (Crichton 
and Krutzsch, 2000). 

Rearing Young 

During pregnancy and lactation, females form maternity 
colonies, which are often located in separate places from 
roosts used by males. In most species of bats, the respons- 
ibility of raising young lies solely with females. Pregnancy 
and lactation are both energetically expensive events, thus 
females and their young can benefit from the heat generated 

Bats 41 

when they form clusters in the partially enclosed spaces often 
found in caves and cave-like structures. Roosting together in 
large clusters may reduce the energy expenditure of some 
individuals by up to 50%. When lactating females disperse 
from roosts in the evening to feed, they often leave their pups 
in a warm, incubator-like environment. Females incur high 
energy costs when they forage and return to the roosts one to 
three times each night to find and suckle their dependent 
young. Thus, assembling in warm places can help reduce the 
energy needed by small bats to remain homeothermic 
(Crichton and Krutzsch, 2000). 

Each spring, Brazilian free-tailed bats (Zadarida brasiliensis) 
migrate from Mexico to the southwestern United States 
to form large maternity colonies in caves and sometimes 
other structures. This species is thought to form the largest 
ageregations of mammals known to mankind, where nightly 
emergences sometimes exceed several million individuals 
from a single cave (Fig. 2). Each time a female Brazilian free- 
tailed bat returns from a feeding bout to suckle her young, 
she faces the daunting task of finding her own pup among 
the millions of babies that are left on the ceilings and walls of 
the cave. A mother bat begins this adventure by returning to 
the area in the cave where she left her pups before emerging 
to feed. Next, she uses vocal and olfactory cues to identify her 
own pups among the thousands or more that are present 
(Crichton and Krutzsch, 2000). Hungry pups will sometimes 
attempt to nurse from almost any female, although lactating 
females usually guard against milk stealing from unrelated 
individuals (McCracken, 1984). The investment that a 
mother bat makes in her pups is substantial, requiring 
quantities of food intake equal to about two-thirds of her 
body mass each night at peak lactation (Kunz et a/., 1995). 

Young Brazilian free-tailed bats grow rapidly from a diet of 
energy-rich milk. Mothers nourish their young with milk for 
several weeks, because young bats cannot fly and feed on 

FIGURE 2 Emerging column of Brazilian free-tailed bats (Zadarida 
brasiliensis) dispersing nightly from caves used as maternity roosts in 

northern Mexico and the southwestern United States during warm months. 
(Photograph by T-H. Kunz). 

42 Bats 

their own until their wings have almost reached adult 
dimensions. Within 6 weeks of birth, young free-tailed bats 
are able to fly and forage on their own. In contrast to most 
other mammals that typically wean their young at about 
40% of adult size, most insectivorous bat species suckle their 
young until they are about 90% of adult size (Crichton and 
Krutzsch, 2000). 


Bats have evolved behavioral and physiological mechanisms 
to avoid long periods of adverse weather and low food or 
water availability. Some species migrate to more suitable 
areas, but others use daily torpor, a controlled lowering of 
body temperature to conserve energy. Only temperate species 
in the families Vespertilionidae, Rhinolophidae, and 
Molossidae are known to hibernate in caves and mines 
(Fig. 3) (Kunz and Fenton, 2003). 

Hibernating bats rely on stored fat as their primary energy 
source during hibernation and are sustained on these reserves 
for upward of 6 to 8 months. Hibernation is an energy- 
saving strategy that is strongly influenced by the ambient 
conditions in a cave. When a bat is hibernating, low ambient 
temperatures lead to a decrease in metabolism. When the 
ambient temperature is too cold or too warm, bats typically 
arouse and move to another part of the cave. It is important 
for hibernating bats to occupy caves and mines that provide 
a variety of temperatures, because individuals often change 
roosting positions as the season progresses (Kunz and Fenton, 

During hibernation, bats lower their body temperature 
to within a few degrees of the ambient temperature, but 
individuals arouse periodically by producing heat employing 
non-shivering thermogenesis. Bouts of hibernation can last 
anywhere from a few days to several months. In areas with 
moderate winters, bats such as the greater horseshoe bat 
(Rhinolophus ferrumequinum) in the United Kingdom may 
feed on insects on warm winter nights. Arousals from deep 
hibernation are energetically costly, with a single arousal 
expending the energy equivalent of a bat spending 68 days in 

FIGURE 3 Small hibernating cluster of the cave myotis (Myotis velifer), a 
cave-dwelling species endemic to North America. (Photograph by T:H. 

deep torpor. Thus, if hibernating bats arouse too often, either 
because the microclimate is not optimal or from human 
disturbance, they may not have enough fat reserves to survive 
the winter (Kunz and Fenton, 2003). 


The decision about where to roost is critical to the survival 
and reproductive success of bats. The type of roost that a bat 
selects is influenced by its morphology, ecology, and physio- 
logical requirements and often reflects a compromise 
between the costs and benefits associated with a particular 
type of roost (Kunz, 1982). For cave-roosting species, the 
benefits of living in a cave usually outweigh any costs that 
they may otherwise incur. In the following section, we 
discuss the major costs and benefits considered critical for the 
selection of roosts by cave-dwelling bats. It is important to 
note that roosting requirements and relevant costs and 
benefits are not uniform for all species and may vary intra- 
specifically, depending on geographic location, reproductive 
condition, and/or season (Kunz and Fenton, 2003). 


Caves offer a wide range of benefits including a structurally 
and climatically stable environment, and protection from 
predators and adverse weather. Microclimate, specifically 
temperature and relative humidity, is arguably the most 
important factor in roost selection by cave-dwelling bats 
(Baudinette et al, 2000). Different bat species roost in a 
variety of microclimates within caves and mines, and this 
variation is often correlated with a bat’s body size, diet, 
phylogeny, and their ability to enter torpor (Kunz and 
Fenton, 2003). 

Compared to non-volant mammals, bats have high rates of 
evaporative water and heat loss, due in part to their relatively 
high surface-to-volume ratio, enhanced by the large surface 
of their naked wing membranes. At low relative humidities 
(<20%), bats may lose up to 30% of their body mass per day 
from evaporative water loss alone. This rapid dehydration 
can be lethal. Many bats select caves that have high relative 
humidity to help conserve water during the day (Kunz and 
Fenton, 2003). 

Bats are endothermic, meaning that they rely on the 
internal production of heat to maintain their body tempe- 
ratures within their thermal neutral zone. Maintaining 
homeothermic body temperatures requires a substantial 
amount of energy. At ambient temperatures above and below 
the thermal neutral zone, bats must expend energy to cool or 
warm themselves, respectively. Bats use at least four different 
strategies for conserving energy while in their roosts (Kunz 
and Fenton, 2003). Some species select roosts that have 
an ambient temperature within their thermal neutral zone. 
The California leaf-nosed bat, for example, often exploits 
geothermally heated mines to conserve heat during the 

winter. Other species form large colonies in parts of caves 
that have little airflow, leading to an increase in roost tempe- 
rature as the metabolic heat generated by the bats becomes 
trapped. The lesser long-nosed bat (Leptonycteris curasoae) 
in South America, the large bent-wing bat (Miniopterus 
schreibersii) in Australia, and the Brazilian free-tailed bat 
(Tadarida brasiliensis) in the southwestern United States 
are examples of cave-dwelling species that form colonies 
large enough to substantially increase the temperature of 
their roost environment. Still other species select colder roost 
environments that allow them to reduce their body tempera- 
ture and thus become torpid. Daily torpor not only reduces 
the amount of energy a bat expends in a day but also helps 
reduce water loss. Finally, some species form dense clusters 
that buffer individuals from changes in ambient temperature, 
a behavior that also reduces their energy expenditure. 

In addition to the energy savings that bats may experience, 
they can also benefit from social interactions promoted by 
cave living. For example, the environmental stability of caves 
can facilitate social interactions such as finding, attracting, 
and guarding mates; information transfer; and interactions 
that evolve through kin selection and/or reciprocal altruism 
(Crichton and Krutzsch, 2000). Females that roost together 
sometimes share information about feeding resources, such as 
the location of flowering and fruiting trees. Female greater 
spear-nosed bats (Phyllostomus hastatus) typically roost in 
caves where they form stable groups of unrelated individuals. 
Information transfer, presumably facilitated by vocal contact, 
may help females coordinate efforts to defend food patches 
from other bats (Wilkinson and Boughman, 1998). 

The common vampire bat (Desmodus rotundus), another 
highly social species, has evolved a system of sharing blood 
with both relatives and unrelated roost mates. Vampire bats 
must obtain a blood meal at least once every three days or 
they will invariably die. Females often share blood with 
roost-mates that are at risk of starving, but this sharing 
occurs only among individuals with whom they are closely 
associated. This is referred to as reciprocity (or reciprocal 
altruism) and occurs when the cost to the individual 
performing the altruistic act is less than the benefit to the 
recipient when such an act is later reciprocated (Greenhall 

and Schmidt, 1988). 


There are several potential costs associated with living in 
caves, most of which are related to living in large groups. 
Large numbers of bats that live in close physical contact with 
one another may be more prone to transmit certain diseases 
or increase the risk of parasitic infestations. High mite 
infestations on a bat, for example, may cause an increase in 
the amount of time an individual spends grooming, and thus 
increase is daily energy budget (Kunz, 1982). 

That bats often emerge synchronously from a cave may 
increase an individual’s risk of predation. Researchers have 

Bats 43 

documented birds of prey, such as owls, hawks, and falcons, 
swooping down into columns of bats that emerge nightly 
from caves and mines. As many as 14 bird species are known 
to feed on bats in Britain alone, with the most important 
predators being owls. Some birds even specialize on bats, 
such as the bat hawk (Machaerhamphus alcinus) in Africa 
and the bat falcon (Falco rufigularis) in Central and South 
America. Most predatory birds are territorial, so their 
numbers at any one cave are probably quite small, thus the 
impact on local populations is probably minimal. Other 
animals that sometimes prey on cave-roosting bats include 
snakes, raccoons, skunks, opossums, and other bats; even a 
frog has been observed preying on bats (Altringham, 1996). 
Few studies, however, have evaluated the impact that 
predators have on bat populations. 

The distribution of caves in most terrestrial landscapes is 
highly variable, and some may not be located near abundant 
food resources upon which bats depend. Bats that roost alone 
or in small groups in tree cavities and in foliage can often take 
advantage of food resources located near their roosts, but 
cave bats, especially those that form large colonies, more 
often must commute considerable distances to foraging sites. 
Because flight is energetically expensive, bats must make 
compromises between colony size and the amount of energy 
spent commuting to feeding sites and the energy that is 
conserved by selecting a roost that has microclimate 
conducive to energy and water conservation (Kunz, 1982). 

Local food resources may not be sufficient to support the 
energy and nutrient budgets of all individuals that form 
large cave colonies, thus some individuals must disperse 
considerable distances in order to secure their daily energy 
and nutrient requirements. Some maternity colonies of 
Brazilian free-tailed bats (Tadarida brasiliensis) may number 
in the millions, requiring some individuals to fly upwards of 
50 km each night to obtain their food. This may also be the 
case for other species that form large colonies in caves, such 
as the large bent-wing bat (Miniopterus schreibersii) and the 
lesser long-nosed bat (Leptonycteris curasoae) (Kunz and 
Fenton, 2003). 


In recent years, reductions in the numbers of cave bat popu- 
lations have increasingly concerned conservation biologists 
(Kunz and Racey, 1998). One of the major problems that 
places bat populations at risk is that they have relatively low 
reproductive rates and are unable to recover quickly from 
population declines. Cave bats face a variety of human 
threats that may vary in different regions of the world. 
Some threats reflect differences due to socioeconomic 
conditions, habitat types, and cultural attitudes toward 
bats. Notwithstanding, several successful approaches, such as 
habitat restoration and cave protection, have been employed 
to protect bats, their roosts, and food resources. Increasingly, 
most if not all geopolitical units (cities, states, countries) are 

44 Bats 

faced with issues related to increased mining and quarrying 
operations, spelunking, ecotourism, vandalism, sealing of 
caves and mines for safety reasons, and deliberate killing 
of bats. Other local issues include guano mining and over- 
collection of bats for scientific research which can have 
adverse affects on cave environments. The negative image of 
bats often portrayed by the media can best be overcome 
through better educational efforts. Lack of basic information 
about the natural history of most cave-roosting bats is a 
problem in many regions of the world. Thorough knowledge 
about the ecology and behavior of different species is 
essential if natural resource managers and politicians are to 
make informed decisions that affect the well-being of bats 
and the food and roost resources on which they depend 
(Kunz and Racey, 2003). 


Many cave-dwelling bats provide essential ecosystem services 
by helping to maintain forest diversity by dispersing seeds 
and pollinating flowers. Changes in bat diversity or 
abundance due to forest fragmentation or roost destruction 
can lead to the dysfunction of forest ecosystems. In addition 
to plant-visiting bats that disperse seeds and pollinate 
flowers, many insectivorous bats consume vast quantities of 
insects. Some insectivorous species feed on insects that cause 
significant damage to agricultural crops. The Brazilian free- 
tailed bat (Zadarida brasiliensis), for example, is known to 
feed on insects that cause millions of dollars in damage to 
corn and cotton crops in the United States each year. In 
addition, nearly everywhere that large quantities of guano 
have accumulated in caves, local communities have 
discovered its value as a fertilizer. In some parts of the world, 
guano is mined locally and sold commercially for fertilizer. 
Although guano mining is no longer common in most 
industrialized countries, it is still practiced in some under- 
developed countries (Kunz and Pierson, 1994; Kunz and 
Fenton, 2003). 

The organic input from bat guano (feces and urine) is 
essential for sustaining the health of cave ecosystems. Many 
cave-dependent organisms (e.g., fungi, arthropods, fish, 
salamanders) depend on bats to produce guano and thus 
provide critical food resources in an environment where 
other sources of organic nutrients are relatively scarce. Not 
only do bats defecate and urinate in caves, but some also 
discard culled wings of insects or seeds and bits of fruit and 
leaves as they feed, and thus supply energy and nutrients for 
a variety of cave organisms such as fungi, arthropods, fish, 
and salamanders (Kunz and Racey, 1998). 

Human Disturbance 

Humans enter caves for various reasons, including scientific 
research—exploration, shelter, tourism, mining, and even 
sometimes for collecting bats to eat. Whatever the intentions 

might be, these activities can have adverse consequences for 
bats (Kunz and Racey, 1998; Kunz and Fenton, 2003). 
Disturbing bats during the maternity period, whether they 
are handled or not, can cause pregnant females to abort their 
young or cause young to fall to the floor, leading to injury or 
certain death. Hibernating bats are particularly vulnerable to 
disturbances from human activities. When humans disturb 
hibernating bats, they often respond by arousing, which is 
energetically expensive. Non-tactile stimuli, such as light and 
noise, can also increase the activity of bats that hibernate in 
caves and thus lead to the depletion of valuable energy 
reserves (Kunz and Fenton, 2003). 

Habitat Destruction and Alteration 

Mining and quarrying activities can have adverse impacts 
on cave-roosting bats, because such activity often modifies 
the physical structure and microclimate of their subterranean 
habitats. Because many bats have very specific roosting 
requirements, such changes may cause bats to abandon these 
sites. Just as important is the fact that mining operations 
often use chemicals that are highly toxic. In some regions, 
modern gold-mining techniques, for example, use cyanide 
to extract gold from ore, and such practices have killed 
enormous numbers of animals, including cave bats, by 
contaminating water sources from which bats derive their 
food and drinking water. 

In some situations, bats have taken advantage of human 
technology by readily using manmade structures such as 
bridges and mines. Increases in the number of abandoned 
mines and bat-friendly bridges have also increased the 
abundance and distribution of some cave-dwelling species. 
However, reclamation of mines and the closing of others have 
also led to an increase in bat mortality when these structures 
are closed without first verifying their presence (Tuttle and 
Taylor, 1994). 

The habitat surrounding caves can be just as important as 
the environment within the cave itself. Many hibernating 
species, such as the endangered Indiana bat (Myotis sodalis), 
typically roost beneath exfoliating bark during the 
warm months but hibernate in caves during the winter. To 
survive a prolonged period of hibernation, bats must be 
able replenish their fat reserves following migration, thus 
productive foraging habitats located near hibernacula are 
essential to their success. The vegetation around a cave not 
only supports source populations of insect prey but also may 
buffer the interior of the cave from severe changes in wind 
flow and temperature. 

The Paradox of Vampire Bats 

A source of myths and legends, vampire bats offer a valuable 
lesson about the need to learn more about these and other 
species before it is too late to protect them from extinction. 
Three species of vampire bats range from northern Mexico 

through South America, but only the common vampire bat 
(Desmodus rotundus) is abundant enough to be considered a 
nuisance to humans and their livestock. All three species 
depend on a diet of blood, but they feed on a variety of 
different animals. Most of our knowledge of vampire bats 
comes from the common vampire bat, which specializes 
on mammalian blood as a source of food. Populations of 
vampire bats increased sharply in areas of Latin America 
following the introduction of livestock by European settlers 
over 500 years ago. Because the common vampire bat 
feeds on cattle and occasionally on humans, this species has 
become a pest in most of Central and South America. The 
economic loss due to cattle dying from bat-transmitted rabies 
alone is a major concern in many regions of Latin America 
(Greenhall and Schmidt, 1988). 

Lack of education and misguided attempts to control 
vampire bat populations have led to the mass destruction of 
these and other non-targeted species. Nonselective killing 
techniques, such as fire and gas (fumigating), have been used 
either because local landowners are often unaware of the 
differences between vampire bats and other species or because 
they are uninformed about the value of bats in general. 
Poisons, such as strychnine or anticoagulants, are often 
applied to the wounds on livestock because vampire bats 
return to wounds that they made the previous night 
(Greenhall and Schmidt, 1988). Selective approaches that 
concentrate on controlling vampire bats should be used 
whenever possible. Recent discoveries by researchers indicate 
that chemicals present in the saliva of common vampire bats 
have important medical benefits (e.g., reducing the risks of 
stroke and heart attacks in humans). Thus, a bat species that 
is considered a nuisance or public health threat by some 
segments of society may also offer enormous benefits to 

Many local, national, and international organizations have 
become engaged in efforts to support research on bats and 
have also helped to educate the public about the benefits of 
bats to humankind. Many cave organizations have joined this 
effort to protect cave bats. Television programs, newspaper 
articles, and other media must be used to promote the ecolo- 
gical value of bats and the importance of caves for sustaining 
many bat populations on a worldwide scale. Organizations 
such as Bat Conservation International (, 
Bat Conservation Trust (, the Lubee Bat 
Conservancy (, and the Organization for 
Bat Conservation ( are among a growing 
number of non-government organizations that are contrib- 
uting to these efforts. Notwithstanding, additional efforts are 
needed to help promote and protect the nearly 1100 species 
of bats known worldwide. 


We wish to acknowledge support from the American Society 
of Mammalogists and Bat Conservation International 

Beetles 45 

(S.W.M.), as well as the National Geographic Society, 
National Science Foundation, and the Lubee Bat Conser- 
vancy (T.H.K) for funding our research. We also thank Tigga 
Kingston and Wendy Hood for making suggestions on an 
earlier draft of this manuscript. 


Altringham, J.D. (1996) Bats: Biology and Behaviour, Oxford University 
Press, London, 262 pp. 

Baudinette, R.V., S.K. Churchill, K.A. Christian, J.E. Nelson, and PJ. 
Hudson (2000) Energy, water balance and the roost microenvironment 
in three Australian cave-dwelling bats (Microchiroptera). /. Comp. 
Physiol. B 170: 439-446. 

Crichton, E.G. and PH. Krutzsch, Eds. (2000) Biology of Bat Reproduction, 
Academic Press, San Diego, 510 pp. 

Findley, J. S. (1993) Bats: A Community Perspective, Cambridge University 
Press, Cambridge, MA, 167 pp. 

Fleming, T.H., A.A. Nelson, and V.M. Dalton (1998) Roosting behavior 
of the lesser long-nosed bat, Leptonycteris curasoae. J. Mammal. 
79: 147-155. 

Greenhall, A.M. and U. Schmidt, Eds. (1988) Natural History of Vampire 
Bats. CRC Press, Boca Raton, FL, 246 pp. 

Kunz, T.H., Ed. (1982) Ecology of Bats. Plenum Press, New York, 425 pp. 

Kunz, T.H. and E.D. Pierson (1994) Bats of the World: An Introduction. In 
Walker’ Bats of the World (R.W. Nowak, Ed.). Johns Hopkins University 
Press, Baltimore, MD, pp. 1-46. 

Kunz, T.H. and M.B. Fenton, Eds. (2003) Bat Ecology, University of 
Chicago Press, Chicago, IL. 

Kunz, T.H. and PA. Racey, Eds. (1998) Bat Biology and Conservation, 
Smithsonian Institution Press, Washington, D.C., 365 pp. 

Kunz, T:H., J.O. Whitaker, Jr, and M.D. Wadanoli (1995) Dietary 
energetics of the Mexican free-tailed bat (Zadarida brasiliensis) during 
pregnancy and lactation. Oecologia 101: 107-115. 

McCracken, G.E. (1984) Communal nursing in Mexican free-tailed bat 
maternity colonies, Science 223: 1090-1091. 

Simmons, N.B. (2003) Bats. In Mammal Species of the World: A Taxonomic 
and Geographic Reference, 3rd ed. (D.E. Wilson and D.M. Reeder, Eds.). 
Smithsonian Institution Press, Washington, D.C. 

Tuttle, M.D. and D. Taylor (1994) Bats and Mines. Resource Publication 
No. 3, Bat Conservation International, Austin, TX, 42 pp. 

Wilkinson, G.S. and J.W. Boughman (1998) Social calls coordinate foraging 
in greater spear-nosed bats. Animal Behav. 55: 337-350. 

Wilson, D.E. and S. Ruff, Eds. (1999) The Smithsonian Book of North 
American Mammals. Smithsonian Institution Press, Washington, D.C., 

745 pp. 


Oana Teodora Moldovan 
Emil Racovitza Speleological Institute, Cluj, Romania 


As part of the Insects class, representatives of the order of 
Coleoptera usually have a sclerotized body with sclerotized 
forewings that are leathery or horny and modified to act as 
rigid covers (elytra) over the membranous, reduced, or even 
absent hindwings. The mouthparts are adapted for cutting, 

46 Beetles 

nibbling, and chewing, and the antennae have usually 8 to 11 
articles. The male genitalia are retractable, and the females do 
not posses an ovipositor. Beetles are the most numerous of all 
insects, with more than 300,000 described species—almost 
one-third of all known animals (Brusca and Brusca, 1990). 
Big or small, they are everywhere, occurring in all environ- 
ments, including caves, lava tubes, cracks and fissures in a 
massif, mesovoid shallow substratum (MSS) in limestone, or 
in different other rocks as schist, gneiss, granodiorites, 
basalts, quartzits, grits, etc. 

Of the 40 families of the order, 15 have species in the 
underground world: the aquatic Dryopidae, Dytiscidae 
(predaceous diving beetles), Elmidae, Hydrophilidae (water 
scavenger beetles), and Noteridae, as well as the terrestrial 
Carabidae (ground beetles), Cholevidae, Curculionidae 
(weevils), Histeridae, Pselaphidae, Staphylinidae (rove beetles), 
Merophysidae, Ptiliidae, Scydmaenidae, and Tenebrionidae 
(darkling beetles) (Fig. 1). Even underground they are the 
best represented of all animals, with around 2000 species. 

It is not surprisingly that one of the first discovered cave 
animals was a beetle, in Postojnska Jama (Slovenia). In 1831, 
Ceé observed specimens that looked very much like ants on 
the beautiful stalagmites. Count Franz von Hohenwart sent 
the material to the Austrian naturalist Schmidt, who described 
Leptodirus hohenwarthi, an amazing terrestrial species 
displaying high degree of adaptation to cave life (Fig. 2). 


Several adaptations of cave beetles to darkness, low and 
heterogeneous in time and space food input (at least in 
temperate regions, where the food is brought inside caves 
through the cracks network, only in some seasons and mostly 
by water), and a relatively constant climate are characteristic 
of cave life, but the degree of morphological, anatomical, 
behavioral, and physiological adaptation is not similar in all 
species. One of the first morphological changes to occur 
among beetles during colonization of caves is a loss of 

Pselaphidae 1% Aquatic beetles 
4% 2% 
Histeridae Others 
1% 0% 
Cholevidae Carabidae 
31% 60% 

FIGURE 1 Proportion of cave beetle families in the group of Coleoptera. 
(Adapted from Juberthie and Decu, 1998.) 

a at * ee oe wo 

FIGURE 2 Leptodirus hochenwarthi (Schmidt, 1832). (Photograph by Valika 

Ku8tor, Slovenia.) 

pigmentation. The cuticle becomes thinner, and the color of 
the cave beetles given by the benzoquinones is a red-brown. 
The most evident morphological change is the reduction or 
complete lack of eyes. The ocelles and rhabdomeres can 
disappear completely, and reduction can also affect the 
optical center, much more pronounced than for stygobitic 
fishes and decapod crustaceans. 

Elongation of the body and antennae, which become 
slender and longer, is linked to their preferred habitat (the 
network of cracks) and serves as compensation for the lack 
of eyes. Longer antennae mean also longer mechanical, gusta- 
tory, olfactory, and hygrosensitive sensilla, as well as enhance- 
ment of the receptory surface. Cave beetles have larger 
reception surface compared with their epigean relatives. 

Life in caves makes the use of the wings impossible. Wings 
are completely lost in some species, and the elytrae are fused 
together. Moreover, in highly specialized species, under the 
elytrae is located a compartment containing air for humidity 
regulation and that causing a false physiogastry, similar to 
the bulging abdomen of ants and termites filled with lipid 

The anatomical internal modifications are especially due 
to the scarcity of food and its uneven distribution through- 
out the year; therefore, adapted species have developed a fat 
body containing huge vesicles filled with fat, proteins, and 
glycogen that allow survival during several months of fast. 
There is also a modification of the exocrine glands, observed 
on some Cholevidae. The soil species have a big sternal 
exocrine gland (secreting pheromones) that disappears com- 
pletely in cave relatives, being replaced by smaller unicellular 
glands. The secreted substances switch from a mixture of 

very volatile and less volatile ones for soil species to less 
volatile ones, perceived only at small distances (the special 
group of cuticular hydrocarbons), in subterranean species. In 
caves, the presence of food can be an attractant at long 
distances, while the small-distance pheromones act in place, 
releasing the mating behavior and even the laying of eggs. 
This behavior saves the energy needed for the production of 
offspring. Even the larvae save energy by not feeding during 
the development, as they are protected by small “houses” 
of clay. 

Several breeding experiments were done in the Pyrenean 
Cave laboratory of Moulis (France) (Deleurance-Glacon, 
1963) on beetles with different degrees of adaptation. The 
most adapted have low fecundity, with a reduction in the 
number of ovarioles. The females lay few eggs or only one, 
bigger, with more vitellinic reserves; also, the time required 
for egg hatching is longer. The larval stages and larval life are 
reduced, and the time spent as a pupa increases (Fig. 3). The 
life cycle varies; French species spend 4 to 5 years as adults, 
and the American Ptomaphagus species spend only 2 years 
(Peck, 1986). 

The activity of epigean species depends very much on the 
day/night and seasonal alternations. For cave beetles in the 
absence of light and other 24-hour environmental cues, 
periods of activity and rest to do not have a daily rhythm. 

The tropical caves with important food input are 
inhabited by less adapted beetles, especially those caves with 
large deposits of guano. The process of adaptation of 
guanophilic and guanobitic species is slowed down. In these 
regions, at low altitude the species are troglomorphic, and 


FIGURE 3 Development cycles of underground beetles with different levels 
of adaptation: 1, endogean; 2, less adapted; 3, very adapted. 

Beetles 47 

true troglobitic beetles typically inhabit high mountain caves 

(Sbordoni et al., 1977). 


The beetles that have colonized the underground world were 
preadapted or exapted; they were terrestrial and nocturnal, 
preferring moist habitats, frequently the fissures of soil, bark, 
or entrances to caves. The moment of cave occupation is 
still debated, but it is often related with the Quaternary 
glacial—interglacial periods or with the natural processes of 
empty environment colonization. Peck (1981) proposed the 
following scenario for colonization of caves in the Grand 
Canyon (between 1160- and 1580-m altitude) by 
Ptomaphagus hirtus. The first species aroused during the 
interglaciary period after the Illinois glaciation (235,000 to 
185,000 BP) and dispersed underground during the next 
glaciation. Then, again, new cave species appeared in the next 
interglaciary period (150,000 to 90,000 BP). Colonization 
with troglophilic or troglobitic species was followed by 
isolation during the interglaciary periods which separated the 
populations, and new cave species appeared. This hypothesis 
has been tested and validated with alloenzyme paleodating 
and paleotemperature measurements. 

The subterranean beetles spread on all continents and on 
some islands, not only in karstic areas (Fig. 4). The richest 
regions are the temperate ones. The glaciated areas usually do 
not have cave beetles, but some troglobitic species have 
recolonized the underground habitats. There are parts of 
the world where the beetle fauna is very diversified, such 
as the Mediterranean karst. The reasons for the separation of 
different populations with wide ranges of distribution can be 
paleogeographical, paleoclimatical, or ecological. European 
and North American species are the best known, given the 
number of specialists and long biospeleological tradition, but 

FIGURE 4 World distribution of cave beetles (in gray). (Adapted from 
Juberthie and Decu, 1998.) 

48 Beetles 

also the favorable climatic conditions. The extreme north 
and south areas were repeatedly covered with ice, and the 
limits of the glaciers are very well reflected today in the limits 
of the troglobitic beetle distribution. 


The family Carabidae includes the Trechinae, representing 
1047 species in 105 genera. These have colonized all the 
subterranean habitats, from soil to caves. The cavernicolous 
species have been classified into two morphological types 
(anophthalm and aphenopsian), corresponding to how 
advanced they are in the adaptive evolutionary processes 
(Fig. 5). The anophthalm type characterizes the endogean 
and some cave species: depigmented, reduced or no eyes, 
anterior body and appendages slightly elongated. The other 
type, aphenopsian, is a very evolved species that is eyeless 
with a very long body, antennae, and legs and very thin 
cuticle. Typical for this last type are Aphaenops in Europe 
and Neaphaenops and Mexaphaenops in North America. But, 
one of the most amazing adaptations concerns the sensory 
equipment, with development of olfaction through increased 
numbers of antennae receptors and lengthening of the 
mechanical trichobotries on the elytrae. Some Trechinae are 
polyphagous predators, such as the French Aphaenops, which 
has a diet consisting of adults and larvae of other beetles, 
springtails, flies, sometimes millipedes, crickets, diplurans, 
etc. and which hunts in caves and MSS (Juberthie and 
Decu, 1998). Others are very specialized on a single prey, 
such as Rhadine subterranea from Great Onyx Cave (United 
States), which feeds exclusively on the eggs and young of the 
cricket Hadenoecus subterranea. The anatomy and behavior of 
this beetle are surprisingly adapted for locating and stealing 
the eggs deposited one at the time beneath the surface of silt. 
Chemical substances left by the cricket release the search 
behavior of the beetle, and when the prey is found a hole is 

FIGURE 5 Different Trechinae adaptations to caves: (a) anophthalm and (b) 
aphenopsian. (Adapted from Ginet and Decou, 1977.) 

dug and the contents of the eggs are devoured (Mohr and 
Poulson, 1966). Some species can be rare in caves, only 
sometimes being found on the sandy banks of subterranean 
rivers or during flooding seasons. The Trechinae are very 
mobile and active in the search of the food, probably cover- 
ing large areas in fissured massifs. In this group, the steps of 
evolution and the way in which adaptation has occurred can 
be observed very well thanks to the presence of many species 
outside, in moist or dark habitats (such as under rocks or 
moss) and deep in the soil. The troglomorphic features of 
the cave Carabidae are very diversified; some of them are 
explained by adaptation to different underground compart- 
ments, but others represent the original contribution of some 
phyletical line or the years spent as cave inhabitants (Fig. 6). 

Another rich group is the Cholevidae, especially the 
subfamily of Leptodirinae, with 138 genera and 562 species. 
Leptodirinae occur only in European caves, being replaced in 
North America by the smaller but very interesting subfamily 
of Ptomaphaginae. As for the previous group, the Cholevidae 
have colonized all the subterranean habitats, but they con- 
sume organic or decomposed matter and are not predators. 
Cave clay or moonmilk that contains bacteria, fungi, and 
algae is important to their diet, and these beetles can 
sometimes be found in huge numbers on these substrates. 
Species of this group have different morphologies according 
to their degree of adaptation to cave life; four morphological 
types are accepted: (1) bathyscioid, for humicolous, endo- 
genous, and some less specialized cave species with more or 
less globular forms of body and short appendages; (2) 
pholeuonid, characterizing specialized forms with longer, 
slender bodies and appendages and false physiogastry; (3) 
leptodiroid, for highly specialized species, such as Leptodirus, 
with extremely elongated legs and antennae and a very small 
anterior part of the body; (4) scaphoid (from scaphe, Greek for 
“boat”), which are also highly specialized and very similar 
to the previous group but with a different form of the body, 
like a boat (Fig. 7). Not enough data are available regarding 
predators of these beetles. It is generally accepted that they do 

FIGURE 6 Head, antennae, and first pair of legs of (a) Clivina subterranea 
and (b) Italodytes stammeri showing the differences, respectively, between 
new colonizers and old colonizers of caves. 

FIGURE7 Different types of Leptodirinae as adaptation to caves: (a) 
batyscioid; (b) pholeuonid; (c) leptodiroid; (d) scaphoid. (Adapted from 
Ginet and Decou, 1977.) 

not have as many as their epigean relatives, but some cases 
of predation from harvestmen and pseudoscorpions have 
been reported, and is probably more pronounced on the eggs 
and larvae than on the adults. Other Coleoptera families 
have few troglobitic representatives but do have interesting 
adaptations to subterranean life. 

The aquatic beetles (Fig. 8a) were not as successful in 
colonizing hypogean habitats as the terrestrial ones. The first 
stygobiontic beetle was mentioned in France at the beginning 
of the 20th century. Today, 21 genera with 31 species are 
known to inhabit cave streams, springs, or wells, usually in 
warmer climates. Besides the typical adaptations of cave 
beetles, aquatic beetles have other adaptations, such as 
switching between swimmers and crawlers, different methods 
for obtaining air through cuticular respiration or tracheated 
elytral respiration, pupation at the bottom of subterranean 
waters, and smaller size (from 1.1 to 4.5 mm). 

The Curculionidae is the best represented family of beetles 
on the surface, but there are few cave inhabitants, with no 
typical cave adaptation. These species display different 
degrees of specialization to a deep soil environment, as the 
eyeless Troglorhynchus monteleonii from a cave in Central Italy 
(Osella, 1982). It is not a question of adaptation for cave life 
but rather one of a different degree of specialization to a deep 
soil environment (Juberthie and Decu, 1998). The cave in 
this case is more of a trap. 

Troglophilic and guanobitic specialization (many species 
are myrmecophylous and termitophylous) of the Histeridae 
can explain the low number of cave species; the first 
cave specimen was discovered in Turkey (Fig. 8b). Most 
subterranean species are troglophilic and guanobitic. Thirty- 
five genera of Pselaphidae have troglobitic representatives 
(Fig. 8c). The origin of the temperate troglobitic pselaphids 
can be traced to the Tertiary, under similar conditions as 
in the humid and relatively cold forests of the African 
mountains, where forms with small eyes and no wings are 
largely spread in the humus. 

Most Staphylinidae are from the Mediterranean region 
(Morocco, Algeria, Spain, Italy) or nearby (Romania and 
the Canary and Madeira Islands), but generally they are 
troglophilic (Fig. 8d). 

Beetles 49 

FIGURE 8 Cave beetles: (a) aquatic Morimotoa gigantea (Uéno, 1957) (from 
Japan); (b) histerid Spelaeabraeus agazzii (Moro, 1957) (from Italy); (c) 
pselafid Decumarellus sarbui (Poggi, 1994) (from Romania); (d) staphylinid 
Domene vulcanica (Orom{ and Hernandez, 1986) (from Canary Islands). 

Other families of cave beetles are associated with guano 
deposits and therefore do not share the same morphological 
adaptations as the cave species. 


Cave beetles generally live in relatively stable climates with 
constant temperature, an atmosphere saturated with water 
vapor, and no air currents. There is a link between the 
presence of beetles in caves and food input in spring and 
autumn (for temperate regions). It has been observed that 
populations migrate between the network of cracks and the 
caves, depending on the presence of food and on climatic 
parameters. The fissures and cracks offer a more stable 
habitat than the big passages and rooms that have large 
volumes of air and are more or less in direct contact with the 
natural entrances. Very sensitive to any change in the 
conditions of their environment, cave beetles quickly initiate 
a behavior, even if it is only to run. 

50 Beetles 

) vey 

Number of individuals 

- - : - - 
Jan Mar May Jul Sep Nov "Jan Mar May Jul Sep NovJan Mar Moy Jul Sep Nov 

1968 1969 1970 

FIGURE 9 Numerical variation of Pholeuon moczaryi in Vadu Crisului Cave 
(Romania), in the period 1968 to 1970 as a consequence of flooding periods 
(indicated by arrows). (Adapted from Racovita, 1971.) 

Interesting examples are provided by studies from 
Transylvania (Romania), France, and the United States. 
Variations of the number of individuals at different time 
during a year were determined to be influenced by air tempe- 
rature or the level of the subterranean stream. The example 
in Fig. 9 relates the increase of water level to the withdrawal 
of the local population of Pholeuon (Racovita, 1971). 

Determining the size of cave beetle populations has been a 
concern of coleopterologists. Individuals captured at a given 
time have been marked and then recaptured after a period of 
time. Estimates can then be made by comparing the number 
of recaptured marked individuals with the total number of 
all marked ones. For example, the Neaphaenops tellkampfi 
in Mammoth Cave has been estimated to be 750,000 
individuals (Barr and Kuehne, 1971). In France, the popu- 
lation of catopid Speonomus in a MSS station was estimated 
at 1,000,000 individuals for two species, and the population 
of trechine Aphaenops was estimated to be 100,000 
individuals (Juberthie and Decu, 1998). 

The presence of two or more beetle species in the same 
cave usually means that one of them is dominant, especially 
if they compete for the same food resources. The species can 
choose different places in the underground system and the 
predators different niches. In the Transylvanian caves, two 
genera of Leptodirinae coexist: Pholeuon and Drimeotus. 
With very few exceptions, Drimeotus is limited primarily to 
near the entrances and, in small numbers, in the network 
of cracks, while the Pholeuon are numerous throughout 
the entire cave. In Mammoth Cave, beside the numerous 
Neaphaenops tellkampfi, there are other two dominant 
trechine with different distributions: Pseudoanophthalmus 
menetriesii, which is found in the upper level, and P striatus, 
which is found along the subterranean river. Mammoth Cave 
has one of the largest and most complex subterranean 
biological communities of all known caves, with no less than 
six trechine species, one pselaphid, and one catopid. In 
White Cave, the animal’s entire life is dependent on crickets 
and to a lesser extent on a few pieces of rotten wood (Barr 
and Kuehne, 1971); a generalized food web for this com- 


Hadenoecus subterraneus ey 

ROTTING V Neaphaenops 
woob 5 al acs Selo 

snails, isopods, millipeds, collembolans, diplurans, catopid beetles 

spiders, opilionids, pseudoscorpions, trechine beetles 

FIGURE 10 Food web in White Cave. (Adapted from Barr and Kuehne, 

munity is shown in Fig. 10. On the partially decomposed 
cricket guano live saprophytic snails, isopods, millipedes, 
springtails, diplurans, and catopid beetles. These are eaten by 
spiders, harvestmen, pseudoscorpions, and trechine beetles 
(it has not been determined whether or not the predators are 

highly specialized). 


The diversity of species makes this group of cave inver- 
tebrates one of the best for testing many of the hypotheses 
concerning adaptation strategies during colonization of 
empty places on Earth. Cave beetles are also very precious 
to the natural biodiversity of the world. One species can 
populate one cave or one massif, so the degree of endemism 
is very high. On the other hand, in many regions the 
biospeleological explorations are only just beginning, and the 
task of the coleopterologist is to find and describe newly 
discovered species. The importance of their studies lies 
primarily in the development of knowledge regarding 
conservation management measures to be taken for more 
vulnerable or rare species. Experiments carried out in a 
Romanian show cave indicate that the presence of tourists 
has eliminated the troglobitic beetles from visited parts of 
the cave and has also influenced the yearly dynamics of the 
leptodirin population (Fig. 11) in the non-visited part. 

Rare and having a strange morphology, cavernicolous 
species have attracted the attention of collectors, especially in 
Europe; however, laws to prevent overcollection of specimens 
from caves seem to have had little effect on this trade. 


Barr, T:C. and R.A. Kuehne (1971) Ecological studies in the Mammoth 
Cave system of Kentucky. I. The Ecosystem. Annis. Spéol. 26(1): 47-96. 

Brusca, R.C. and G.J. Brusca (1990) Invertebrates. Sinauer, Sunderland, 
MA, 922 pp. 

Deleurance-Glagon, S. (1963) Recherches sur les coléoptéres troglobies de la 
sous-famille des Bathysciinae. Annls. Sc. Nat. 5 (1): 1-172. 

5 z 
2 3 
- tom 
6 8 
iy ° 
8 g 
€ i=] 
s = 
2 3 
& & 
Yr oO ve & Mg Number of tourists 
—@— Number of Drimeotus 
1997 - 1998 

FIGURE 11 Variation in the number of Drimeotus in Ursilor Cave 
(Romania) as a function of tourist periods. 

Ginet, R. and V. Decou (1977) Initiation a la Biologie et a lEcologie 
Souterraines, Delarge, Paris. 

Juberthie, C. and V. Decu, Eds. (1998) Encyclopaedia Biospeologica, Vol. I, 
Société de Biospéologies, Paris. 

Mohr, C.E. and T.L. Poulson (1966) The Life of the Cave, McGraw-Hill, 
New York. 

Osella, G. (1982) I Curculionidi cavernicoli italiani (Riassunto). Lavori Soc. 
Ital. Biogeogr. 7(1978); 337-338. 

Peck, S.B. (1981) Evolution of cave Cholevinae in North America 
(Coleoptera: Leiodidae). In Proc. 8th Int. Congr. Speleology, Bowling 
Green, KY, Vol. 2, pp. 503-505. 

Peck, S.B. (1986) Evolution of adult morphology and life-history characters 
in cavernicolous Ptomaphagus beetles. Evolution 40(5): 1021-1030. 

Racovita, G. (1971) La variation numérique de la population de Pholeuon 
(Parapholeuon) moczaryi Cs. de la grotte de Vadu-Crisului. Trav. Inst. 
Spéol. “E. Racovitza” 10: 273-278. 

Sbordoni, V., R. Argano, V. Vomer, and A. Zullini. (1977) Ricerche sulla 
fauna cavernicola del Chiapas (Messico) e delle regioni limitrophe: grotte 
explorate nel 1973 e nel 1975. Criteri per una classifizatione 
biospeologica delle grotte. In Subterranean fauna of Mexico, part III, 
Accad. Naz. Lincei ed. 

Uéno, S.-I. (1957) Blind aquatic beetles of Japan with some accounts of the 
fauna of Japanese subterranean waters. Arch. Hydrobiol., 53: 250-296. 

Behavioral Adaptations 

Jakob Parzefall 

Zoologisches Institut und Zoologisches Museum der Universitét Hamburg 

nimals living in darkness have to compete for food, 

mates, and space for undisturbed reproduction just as 
their epigean conspecifics do in the epigean habitats, but 
there is one striking difference: In light, animals can use 
visual signals. Thus, important aspects of behavior driven by 
visual signals cannot apply in darkness. The question arises, 
then, of how cave dwellers compensate for this disadvantage 
in complete darkness. This article uses several examples to 
compare various behavior patterns among cave dwelling 
populations with epigean ancestors. 

Behavioral Adaptations 51 


Potential cave dwellers must have the sense organs and 
behavior necessary to find food and to reproduce in caves. 
Such animals may be said to be preadapted to cave life, and 
in fact some of these animals can survive in the darkness 
without behavioral adaptations and can reduce behavioral 
characters not necessary in the caves. In contrast, some cave 
dwellers have improved sense organs and have acquired 
behavior adapted to their extreme habitats. 

Food and Feeding Behavior 

Suitable food sources and quantity vary from cave to cave. In 
general, cave animals depend upon food brought in from 
outside and are omnivorous. With some exceptions, most 
caves do not have an abundance of food compared to above- 
ground habitats. Food sources can be widely distributed or 
concentrated in patches, and their occurrence is mostly 
unpredictable; therefore, food-finding abilities have to be 
improved, and food must be stored to ensure survival during 
long starvation periods. 

When the blind fish population of Astyanax fasciatus from 
Pachon Cave was studied in competition experiments 
conducted in darkness, it was found that they retrieved 80% 
of small pieces of meat distributed on the bottom of an 
aquarium, whereas the epigean fish got only 20% (Hiippop, 
1987). Among amblyopsid fish, which comprise six species 
in four genera, the ability to detect invertebrates at low prey 
densities in the dark is much better for the cave-living species 
Amblyopsis spelea than the troglophile Chologaster agassizi. 
When one Daphnia was introduced into a 100-L tank, the 
A, spelea found the prey hours before the C. agassizi did. In 
addition, the maximal prey detection distance is greater in 
cave species. Daphnia was detected by Typhlichtys subterra- 
neous within 30 to 40 mm and by C. agassizi within 10 mm. 
(Poulson, 1963). 

The reaction to prey by salamanders has been studied in 
the facultative cave-living Pyrenean salamander Euproctus 
asper, which has fully developed eyes, and the blind Proteus 
anguinus. Both species react to living and dead chironomids. 
Even in light, where £. asper can use their visual sense, P 
anguinus required less time to initiate the first snapping 
response to dead prey. When the time between the start of 
an experiment and the first snap at prey was divided into 
pre-approach and approach phases, it turned out that the 
difference found could be attributed to the pre-approach 
phase (Fig. 1a). Living prey was detected more quickly than 
dead prey in both species, but E. asper needed more time in 
the darkness than did P anguinus (Fig. 1b,c). These data 
show that P anguinus is well adapted to search prey on the 
basis of chemical and mechanical information. In contrast, 
E. asper demonstrated a more directed, visually dominated 

52 Behavioral Adaptations 

Dead prey, 

“7 b) Live prey, 
20 light 


Live prey, 
40 darkness 

Time to first snapping /min. 

Euproctus Proteus 

FIGURE 1 Snapping response in Euproctus asper and Proteus anguinus. The 
mean time interval between the start of the experiments and the first snap at 
prey was divided into a pre-approach phase (hatched bars) and an approach 
phase (open bars) in three different experimental treatments. Standard 
errors of the mean are shown on top of the bars. (From Uiblein, E, Durand, 
J. B, Juberthie, C., and Parzefall, J., Behav. Proc., 28, 33-40, 1992. With 


approach behavior with live prey in light and can switch to a 
more active, widely foraging mode with live prey in darkness 
and dead prey in light. This young cave colonizer seems less 
adapted to the dark but is capable of foraging successfully in 
both epigean and hypogean habitats. 

Compensating for the unpredictability of food quality and 
quantity also results in physiological adaptations. Cave 
animals are able to survive for long periods without food— 
nearly one year for invertebrates and up to several years for 
caves fishes and salamanders (Vandel, 1964). In A. fasciatus, 

the cave fish are able to build up enormous fat reserves. A 
1-year-old cave fish fed ad libitum had a mean fat content of 
37 % fresh body mass compared to 9% in epigean fish under 
the same conditions (Hiippop, 1987). 

Reproductive Behavior 

Having found enough food to reach sexual maturity, the next 
problem to be solved by cave dwellers is finding a sexual 
partner in the darkness. Subsequently, they need behavior 
patterns that provide effective fertilization in the absence of 
any visual orientation. In species with high population 
densities, it is easy for the male to find a female. The male of 
the galatheid crab Munidopsis polymorpha of the marine cave 
Jameos del Aqua on Lanzarote in the Canary Islands receives 
a chemical signal sent by the molting female ready for 
reproduction (Parzefall and Wilkens, 1975). 

In terrestrial invertebrates, a comparable situation has 
been observed: The females of the cave crickets Hadenoecus 
subterraneus and H. cumberlandicus that are ready to mate 
release an olfactory attractant. Normally, several males reach 
the attractive female at the same time. They can transmit 
information about their high reproductive fitness by sending 
tactile signals through the air by their elongated antennae as 
it is done by both of the cricket species. In M. polymorpha the 
male emits rhythmic water waves with chelipeds (Fig. 2). 
They must repeat these signals several times. The female 
eventually decides which male is good, accepts the sperm 
transfer (or not), and escapes. 

A comparable situation has been found in the live-bearing 
poeciliid fish Poecilia mexicana, which lives in a high- 
population-density cave habitat. The males check conspecific 
females by nipping at the genital region. The females ready 
for reproduction produce a species-specific chemical signal 
and attractant for about 3 days during a cycle of about 28 
days. A female accepting a male stops swimming and allows 
the copulation. Normally, bigger males are preferred on the 
basis of visual signals. Only the cave fish female is able to 
perform this behavior in darkness (Fig. 3); she does so by 
switching from the visual system to a lateral line system (the 
fish have one lateral line system only) (Parzefall, 2001). 

FIGURE 2 The male Munidopsis polymorpha (right) displays with cheliped 
shaking in front of a female. (From Parzefall, J. and Wilkens, H., Ann. 
Spéléol., 30, 325-335, 1975. With permission.) 

a ) Poecilia mexicana 

Behavioral Adaptations 53 

10004 Female choice in light 
— 800 
® n=11,p<0.05 : 
= 600 n=l ;p<0.02 n=11;p<0.05 
> 200 
large d small largest small” large o* small 0” 
Epigean Pop. Rio Oxolatan Epigean Pop. Milky river Cave Pop. PS XI 
100074 Poecilia mexicana 
Female choice in darkness 
_ 800 
® n=12.N.s. n=20;n.s. n=20,p<0.02 
2 600 : 
2 100 ee eae CE TEE 
& 200 
larged small @ large small o* large o small o” 
Epigean Pop. Rio Oxolatan EpigeanPop. Milky river Cave Pop. PS Xl 

FIGURE 3 Female choice behavior in different populations of Poecilia mexicana in (a) light and (b) darkness. The female had the choice to swim to the 

compartment of a big or a small mature male, and the time spent in a male compartment was measured. A clear partition prevented direct contact. The middle 
line in the box plot represents the median; the upper end of the box, the 75% value; the lower end of the box, the 25% value. The whiskers represent the 90% 
and 10% values, respectively. (From Parzefall, J., Environ. Biol. Fishes, 62, 263-275, 2001. With permission.) 

In the above-mentioned crab and cricket species, data 
about sexual behavior in their epigean relatives are lacking, 
so we cannot determine whether the reproductive behavior 
has changed in adaptation to the dark habitat. For Poecilia 
mexicana, comparative data on epigean conspecifics reveal 
that in the epigean habitat visually orientated sexual displays 
are lacking, in contrast to other species of the genus such 
as P velifera, P latipinna, and P reticulata. So, PR mexicana 
seems to be preadapted to cave life and has improved their 
reproductive fitness in the dark by means of a special female 
choice behavior based on a lateral line system (the fish have 
one lateral line system only). 

Species with lower population densities in cave habitats, 
such as the characid fish Astyanax fasciatus or the salamander 
Proteus anguinus, attract conspecifics from chemical signals 
transmitted in the water. Comparative studies with the 
epigean proteid Necturus maculosus have demonstrated that 

this information is species specific (Parzefall et al, 1980). 
The animals also constantly deposit a substance while in 
contact with the substrate and at communal resting 
places. This substance is individual specific but does not 
provide any detailed information about sex or reproductive 
state; it merely brings members of the species together. For 
recognition of sex and reproductive state, Proteus requires 
direct contact. When sexually motivated, a male establishes 
a territory that a female may enter only after direct 
body contact. The male sends a chemical signal by fanning 
his tail against the female and from time to time begins 
to walk away. The female follows and nips at the genital 
region of the male. After a short walk, they stop and the 
male deposits a spermatophore, which the female retrieves 
and places in her cloacal region. Unfortunately, comparative 
data in the epigean salamander Necturus maculosus are 

54 Behavioral Adaptations 

Aggressive Behavior 

Aggressive behavior consists of different patterns of 
threatening postures and attacks followed by fights. This 
behavior has various functional aspects and is absolutely 
necessary in darkness; therefore, it has to be adapted to cave 

COMPETITION FOR FOOD Food competition results 
in food territories for groups, pairs, or solitary animals. 
Within a group, a limited food supply can lead to aggressive 
encounters. In general, defending food resources is only 
adaptive when the costs are not higher than the potential 
incoming energy of the food. The majority of data available 
regarding aggression among cave and epigean populations 
is for the characid fish Astyanax fasciatus. The epigean 
form is widely distributed in Mexico. When undisturbed, the 
epigean fish defends small territories of 10 to 20 cm, de- 
pending on body size, by fin spreading, snake swimming, and 
Ramming (Fig. 4). In the laboratory, epigean fish of both 
sexes display the entire aggressive pattern. The subdominant 
fish demonstrates submission by a head-up position and 
trying to hide or escape. The fights can be very strong; in 

smaller aquaria that offer no place to hide the death of the 
subdominant fish can result. 

Aggressive behavior depends on optical releasers. Using 
dummies of different types, it has been found that natural 
shape and locomotion are important visual signals. Tests with 
infrared video have shown that the epigean fish is not able to 
perform aggressive patterns in complete darkness (Hausberg, 
1995). They do not establish territories at all. From such data 
we can conclude that epigean fish, when colonizing caves, 
were no longer able to perform aggressive encounters. In the 
blind populations of the Pachon, Piedras, and Yerbaniz caves, 
Hausberg (1995) noted a high percentage of fish with 
injuries on fins and scales. In an experiment with the Pachon 
population, the number of injured fish increased in the 
absence of food. The aggressive behavior observed within this 
fish includes defending small territories of a few centimeters 
by biting, circling, and tail beating (Fig. 5). Also in these 
experiments, a striking difference in swimming behavior was 
observed; the fish that were regularly fed glided slowly 
through the water of the entire aquarium without initiating 
aggressive encounters against conspecifics. When food was 
lacking, the locomotor activity decreased. The fish mostly 
hovered at the bottom and rhythmically flicked their fins; 

FIGURE 4 Aggressive patterns in the epigean Astyanax fasciatus. (a) Aggressive fine ercection; the head-down position of the fish on the right expresses a higher 
aggressive motivation. (b) Snake swimming is shown by the fish on the right and aggressive fin erection by the fish on the left. (c) The fish on the left is 
ramming against the one on the right. (d) Both fish show circling and tail beating. (From Parzefall, J. and Hausberg, C., Mém. Spéleol., 28, 153-157, 2001. 

With permission.) 

FIGURE 5 Ramming and circling in the Pachon cave population of 
Astyanax fasciatus. (From Parzefall, J. and Hausberg, C., Mém. Spéleol., 28, 
153-157, 2001. With permission.) 

fish entering the small area of a few centimeters were attacked. 
The territory size was correlated with the aggressiveness of 
the fish, and the aggressive patterns differed from those 
shown in epigean fish. The cave fish has developed an 
aggressive behavior with signals that are only effective in close 
body contact. 

Among cave-living invertebrates, the galatheid crab 
Munidopsis polymorpha of the marine cave Jameos del Aqua 
on Lanzarote in the Canary Islands feeds mainly on diatoms 
on lava rocks. The animals keep a minimal distance from one 
another according to the length of their second antenna 
(Fig. 6). Any closer than this distance, and Munidopsis attacks 
with its extended chelipeds and by snapping its pincers. This 
behavior does not depend on optical releasers but on water 
movements (Parzefall and Wilkens, 1975). The aggressive 
patterns of Munidopsis are very similar to the one described 
for the deep-water, bottom-living, epigean galatheid Munida 
sarsi. The author believes that, despite their naturally dim 
environment, vision is still the primary sense involved in the 
aggressive behavior of this species, so it appears that the 
aggressive behavior of galatheids is effective in light and in 
darkness, with no striking differences. 

anguinus, studies in the laboratory have revealed that males 

In the blind cave salamander Proteus 

show aggressive behavior and territoriality for only a very 
short reproductive period. Normally, the animals rest under 

Behavioral Adaptations 55 

FIGURE 6 Aggressive interactions among Munidopsis polymorpha. (From 
Parzefall, J. and Wilkens, H., Ann. Spéléol., 30, 325-335, 1975. With 


FIGURE7 Aggressive behavior of a Proteus anguinus male against an 
intruder in his territory: (a) body contact (chemical identification); (b) tail 
beating; (c) biting. (From Parzefall, J. Z. Tierpsychol., 42, 29-49, 1976. With 


stones in groups of both sex without any aggressive reaction. 
When a male became sexually active, it begins to control its 
conspecifics by contacts with his snout and allows only 
females in the reproductive state to remain in the hiding 
place. Intruders will be attacked by tail beating, ramming, 
and biting in close body contact (Fig. 7). After being 
attacked in a particular territory, Proteus avoids that territory 
for several days on the basis of chemical cues on the substrate. 
In the Poeciliid fish Poecilia mexicana, males use aggressive 
behavior to establish a size-dependent rank order within a 

56 Breakdown 

mixed school. The females have a reproductive cycle of about 
28 days and are attractive to males within the first 3 days 
of the cycle. The dominant male controls the females by 
nipping in response to an attractive female. In the field, 
the pair separates from the shoal and become more or less 
sedentary. The male nips and tries to copulate while also 
defending the female. During aggressive encounters with 
more or less equal-sized males, small males use a female-like 
body coloration to try to sneak copulations (Parzefall, 1979). 
The population of P mexicana that colonized a limestone 
cave in Tabasco (Mexico) does not school, and the males 
do not fight. In laboratory studies with epigean fish and cave 
fish having functional eyes, a quantitative, genetically based 
reduction of aggressive patterns and schooling has been 
demonstrated (Parzefall, 1979). The reaction are highly 
variable within the population. Some of the cave fishes tested 
seemed unable to understand the attacks and answered by 
nipping and copulation attempts. It seems that aggression in 
these cave fishes is a disadvantage, because fighting males risk 
losing the opportunity for contact with an attractive female 
in darkness. 


Studies of behavior in cave dwellers have revealed complex 
systems of responses to visual, chemical, and tactile stimuli. 
Many animals can survive in complete darkness with no 
visual signals. The use of weak electric signals among cave 
dwellers has not been detected. In some cases, an existing 
behavior (such as the aggression exhibited by Astyanax 
fasciatus) has changed to a more effective behavioral system. 
These changes are always based on existing above-ground 
behavior, and no completely new behavioral character has 
been found in cave animals. 

See Also the Following Articles 
Morphological Adaptations 


Hausberg, C. (1995) Das Aggressionsverhalten von Astyanax fasciatus 
(Characidae, Teleostei): Zur Ontogenie, Genetik und Evolution der 
epigdischen und hypogidischen Form [unpublished disseration]. 
University of Hamburg, Germany, p. 139. 

Hiippop, K. (1987) Food finding ability in cave fish (Astyanax fasciatus). Int. 
J. Speleol. 16: 59-66. 

Parzefall, J (1979) Zur Genetik und biologischen Bedeutung des 
Aggressionsverhaltens von Poecilia sphenops (Pisces, Poeciliidae): 
Untersuchungen an Bastarden ober- und unterirdisch lebender 
Populationen. Z. Tierpsychol. 50: 399-422. 

Parzefall, J (1989) Sexual and aggressive behaviour in species hybrids of 
Poecilia mexicana and Poecilia velifera (Pisces, Poeciliidae). Ethology 
82: 101-115. 

Parzefall, J. (2001) A review of morphological and behavioural changes in 
the cave molly, Poecilia mexicana, from Tabasco, Mexico. Environ. Biol. 
Fishes 62: 263-275. 

Parzefall, J. and C. Hausberg (2001) Ontogeny of the aggressive behaviour 
in epigean and hypogean populations of Astyanax fasciatus (Characidae, 
Teleostei) and their hybrids. Mém. Spéleol. 28: 153-157. 

Parzefall, J.and H. Wilkens (1975) Zur Ethologie augenreduzierter 
Tiere: Untersuchungen an Munidopsis polymorpha Koelbel (Anomura, 
Galatheidae). Ann. Spéléol. 30: 325-335. 

Parzefall, J., J.P. Durand, and B. Richard (1980) Chemical communication 
in Necturus maculosus and his cave-living relative Proteus anguinus 
(Proteidae, Urodela). Z. Tierpsychol. 53: 133-138. 

Poulson, T.L. (1963) Cave adaptation in amblyopsid fishes. Am. Midl. Nat. 
70: 257-290. 

Vandel, A. (1964) Biospéologie 1964: La Biologie des Animaux Cavernicoles. 
Gauthier-Villars, Paris. 


Elizabeth L. White 

The Pennsylvania State University 

ave roofs and walls are rarely stable. Rockfall pieces range 

from small isolated blocks to complete ceiling collapse. 
The term breakdown refers to the masses of rock fragments 
found mostly on cave floors. Breakdown is frequently 
mentioned in the literature and is ubiquitous in most caves. 
This article considers breakdown mechanics and outlines a 
number of geological processes that could set the stage for 
breakdown. Many of the field observations were made in the 
Flint Ridge Section of Mammoth Cave and in the historic 
sections of Mammoth Cave located in south-central 
Kentucky. Most of the observations for folded-limestone 
caves were made in Appalachian Mountain caves of 
Pennsylvania and West Virginia. The observations for the 
younger Tertiary and Pleistocene limestone caves were made 
on Mona Island (off the coast of Puerto Rico) and on San 
Salvador Island in the Bahamas. 


Perhaps the most common breakdown feature is the 
breakdown-littered cave floor as shown in Fig. 1. Piles of 
breakdown are unsorted and highly permeable. Layering is 
undistinguishable or nonexistent. From this reference point 
one can distinguish small-scale breakdown features that 
are the various types of breakdown blocks themselves and 
large-scale features that are cavern features consisting of (or 
generated by) breakdown processes. 

Small-Scale Features 

Breakdown can be classified by the relationship of individual 
blocks to the bedding of the parent bedrock: 
Block breakdown consists of masses of rocks with more 
than one bed remaining as a coherent unit. 
Slab breakdown consists of fragments of single beds. 
Chip breakdown consists of small rock chips and shards 
derived from the fragmentation of individual beds. 

FIGURE 1 Breakdown-littered cave floor. 

This classification has the advantage that breakdown 
observed in the field can be properly classified without 
speculation as to its origin; however, it has the disadvantage 
of also being a function of the limestone lithology. 
Thus, limestone fragments of a given intermediate size might 
be blocks if derived from a thin-bedded limestone or slabs 
if derived from a massively bedded limestone. In general, 
however, this classification has been found useful for the areas 

Block breakdown can be massive; blocks measure up to 
tens of meters on a side and are usually bounded by bedding 
planes along the bedding and by joint planes across the 
bedding. Slab breakdown has a plate shape, with slab thick- 
ness being controlled by the thickness of the beds; the width 
of individual slabs varies from tens of centimeters to many 
meters. Chip breakdown ranges in size from centimeters to 
tens of centimeters; the shape of chip breakdown is variable 
and dependent on the process that created the breakdown. 
Crystal wedging , frost pry, and closely spaced joints produce 
very angular chunks, whereas pressure-induced spalling and 
mineral replacement produce flatter, more irregular shards. 

In the Flint Ridge Section of Mammoth Cave, slab 
breakdown is the most common and is distributed through 
all levels of the cave. Block breakdown occurs where major 
roof collapse has taken place and where dividing walls have 
fallen between coalescing vertical shafts. An extensive break- 
down has occurred in the upper gallery of the Great Salts 
Cave Section. This passage is floored with block and slab 
breakdown to a depth of 12 or more meters for a distance of 
more than a kilometer. The largest breakdown block so far 
observed in this passage is a single block 19 m long, 4.5 m 
wide, and 1 m thick. 

Large-Scale Features 

Although breakdown blocks form a variety of features in 
caves ranging from a few scattered blocks to major collapsed 
passages, it is useful to describe two types of features: termi- 

Breakdown 57 

nal breakdown and breakout domes. Terminal breakdown 
occurs at the end of collapsed major cave passages. Eroding 
valleys on the surface have the effect of causing collapse 
(breakdown) in the caves below. Massive breakdown that 
completely occludes a cave passage is referred to as terminal 
breakdown. In a number of caves, artificial entrances have 
been created where the cave passage would have intersected 
the surface topographic valley. In the Central Kentucky 
Karst, a terminal breakdown is the most common terminator 
of passages. Often the terminal breakdown contains 
sandstone as well as limestone fragments where the collapse 
has extended upward to the overlying caprock. Major trunk 
passages beneath the sandstone-capped plateau were once 
continuous feeder conduits carrying groundwater from the 
Sinkhole Plain to the south and east of the plateau to Green 
River in the north. These formerly continuous passages 
have been truncated by ceiling collapse. Some are actual 
intersections of the passages with the surface; others have 
collapsed at depth. The present-day configuration of the cave 
system is due in large part to these random features of 
collapse. Similar terminal breakdown occurs in many other 
caves with and without caprock. 

Breakout domes, among the most remarkable of cavern 
features, are the huge rooms that form as a result of major 
ceiling collapse. Some of these, such as Chief City in 
Mammoth Cave, have floor dimensions of more than 100 m 
and ceiling heights of 30m. The Rumble Room in the 
Rumbling Falls cave system under Fall Creek Falls State Park 
in Tennessee has a ceiling about 20 stories tall. This breakout 
dome is the largest in the eastern United States and the 
second-largest breakout dome known in the United States. 
Other breakout domes include Rothrock Cathedral in 
Wyandotte Cave (Indiana), the entrance room in Hellhole 
Cave (West Virginia), the entrance room in Marvel Cave 
(Missouri), Devil’s Sinkhole (Texas), and Salle de la Verna in 
Pierre Saint-Martin (France). The details of the enlargement 
mechanism are less clear in Devil’s Sinkhole than in the 
others mentioned, although it also has the beehive-shaped 
room and the gigantic debris cone typical of all breakout 

Careful examination of many breakdown areas reveals a 
continuum of sizes, from very large breakdown rooms to 
small, roughly circular or elliptical breakdown areas in cave 
ceilings. The features at the small end of the scale are 
sometimes only 3 m in diameter and involve only one or two 
beds. The morphological term breakout dome describes all 
such features, regardless of their size. 

Debris piles vary in size from dome to dome, but in those 
domes that are accessible the volume of debris is much 
smaller than the enclosing volume of the dome. Because the 
bulk density of the debris cone is considerably less than that 
of the original bedrock, it is apparent that large quantities of 
material must have been removed. Large breakout domes 
must therefore have formed at a time when water was actively 
circulating near their base. The dome could then enlarge by 

58 Breakdown 

a mechanism of solution action on fallen blocks with con- 
current stoping of the sides. The dome itself is usually 
circular or elliptical in contour. The top is often capped by a 
single massive bed. 


Breakdown is generally assumed to be formed by simple 
mechanical bed failure due to gravitational load. Proposed 
failure mechanisms include brittle fracture of incompetent 
beams (White and White, 1969, 2000) and failure by 
inelastic creep (Tharp and Holdrege, 1994; Tharp, 1995). 
The brittle fracturing provides the simplest model for break- 
down occurrence—the concept dates back to the work of 
Davies (1951), who based his model on the mining litera- 
ture. The model assumes a rectangular passage formed in 
well-bedded limestone. A small amount of elastic sag of the 
unsupported roof beds causes these beds to separate slightly. 
Figure 2 shows the parameters of the fixed-beam model and 
the dome shape of the stress distribution. The beds act as 
fixed beams across the cave passage. For the ceiling to be 
stable, the bending strength of the beams must be greater 
than the gravitational force acting on the weight of the 
unsupported span. Thicker beds are stronger than thin beds. 
There will be a critical thickness (¢cpry) for any given passage 
width at which the strength of the bed is just sufficient to 
support its weight. When the mechanics are worked out, 
only the length of the beam (Z) and the critical bed thickness 
(tcrrr) remain. The width of the beam cancels out. The beam 
length (Z) is set equal to the passage width, while the extent 
of the ceiling bed along the axis of the cave passage does 
not enter the calculation. Figure 3 shows the roof stability 
according to the fixed-beam model. For a fixed beam, the 
critical thickness is: 

t = _ pl? 
coal 2S(cos 8) 


where p is the density of the bedrock (in kg/m®), ® is the 
bedding dip (in degrees), and S is the flexural strength (in 
MPa). If the ceiling beds are not supported at both sides of 
the cave passage, they are treated as cantilever beams for 
which the critical thickness is: 

FIGURE 2 Sketch showing the parameters of the fixed-beam model. 

foIT = 2S(cos 8) 2) 
The fixed-beam model implies a completely elastic response 
of the ceiling beds. It does not allow for plastic deformation 
and long-term creep that could lead to bed failure in the 
absence of any geologic triggers. It also has no time depend- 
ence; a stable ceiling would remain stable until some geologic 
process destabilized it. Numerous occurrences of recent 
breakdown have been reported. Some have occurred near 
entrances where freeze-thaw cycles may be responsible. Some 
have occurred deep in the cave but in areas of active vadose 
water. However, a fraction of recent breakdown occurrences 

are in dry passages with no obvious triggering mechanism. 

There have been four documented roof failures in 
Mammoth Cave in the past century. Three were massive 
rockfalls, but the fourth involved a plastically deformed 
ceiling slab that had been mapped in detail. The slab 
displayed extensive plastic deformation in the 1960s. 
Sometime in the early 1970s it fell. 

A more comprehensive model allows for inelastic creep 
(Tharp and Holdrege, 1994; Tharp, 1995). Materials break 
through a mechanism of crack propagation. Tharp’s model is 
based on the propagation of micro-cracks, which allow defor- 
mation and creep. The crack propagation velocity is given by: 

vee(Z) (3) 

where X; is the stress intensity (MPa m°”), K;, is the fracture 
toughness of the bedrock, and c is a constant related to the 

activation energy for crack movement. The parameter c is 
given by: 

140 : 


00, = Ceiling 


80 7 stable 
60 - 
| jen Ceiling 
20; : unstable 7 

AA ; ¥] a { 
wr acl | 
0 Leese | t L L - 
5 10 15 20 25 30 35 

FIGURE 3 Cave roof stability according to the fixed-beam model for the 
special case of horizontal beds. Typical values of shear stress for Paleozoic 
limestones range from 12 MPa (1700 psi) to 18 MPa (2600 psi). 


c= Voer? (4) 

where His the enthalpy of activation = 67-147 kJ/mole; R is 
the gas constant = 8.3145 J/K; Vo and mare fitting constants. 

Propagation of micro-cracks allows inelastic deformation 
and also a time-to-failure. Using the Tharp model, the 
time-to-failure is the time scale of crack propagation through 
beds of nominal thickness. The time frame can range from 
thousands of years to as much as 1 million years. Thus, the 
time to failure can be in the same range as the age of the cave 
passage, implying that breakdown can occur at any time, 
even in the absence of geologic triggering processes. 

The Tharp model introduces fracture toughness as another 
parameter in addition to the flexural strength for determin- 
ing whether particular beds will collapse. The Paleozoic 
limestones of the eastern United States (where many of the 
breakdown investigations have been made) are dense, fine- 
grained rocks. Coarse fracturing occurs along joints and 
bedding plane partings. But, within the rock mass there is 
little to inhibit crack propagation, and these rocks break 
mainly by brittle fracture. It is for this reason that the fixed- 
beam model has worked so well. 

Porous and vuggy rocks, such as the Tertiary limestone 
beds of the Caribbean, may have a lower flexural strength but 
they have higher fracture toughness, because pores and vugs 
inhibit crack propagation. The caves on Mona Island in 
Puerto Rico have large, relatively flat chambers with little 
breakdown because of the toughness of the porous, young 
limestone beds. Because of the inability of the limestone beds 
to propagate cracks, roof spans of 30 m or more are found 
throughout these caves. 


Crystal Wedging or Mineral-Activation To Initiate 

Many breakdown areas in caves with extensive sulfate 
minerals (primarily gypsum) suggest that crystal wedging and 
replacement of limestone by gypsum are important factors 
in this type of cavern collapse. Features that are characteristic 
of mineral-activated breakdown are: (1) walls and ceilings 
fractured in irregular patterns, often with visible veins of 
gypsum following the fractures; (2) breakdown consisting of 
thin, irregular splinters and shards of bedrock; and (3) curved 
plates of bedrock ranging in size from a few centimeters to 
more than a meter hanging from the ceiling at steep angles 
cemented only by a thin layer of gypsum. Microscope 
examination of thin sections of the bent beds shows that the 
sagging and bending are due to the direct replacement of 
limestone by gypsum. Figure 4 shows the curved breakdown 
slabs in Turner Avenue in Mammoth Cave. Another charac- 
teristic feature is the collapses that take the form of 
symmetrical mounds with coarse, irregular blocks at the base 
grading upward into a rock flour at the top. 

Breakdown 59 

FIGURE 4 Thin plates and fragments, some held to the ceiling by gypsum, 
in the Upper Turner Avenue in Mammoth Cave (Kentucky). Total width of 

image is approximately 1 m. 

Crystal wedging produces a subset of chip breakdown. 
Chip breakdown consists of rock fragments that are smaller 
than individual bedding plane slabs and can result from 
many processes, including purely mechanical ones. Crystal 
wedging breakdown appears to be of two types. Type I 
consists of angular rock fragments broken on sharp planes 
that cut the bedding planes. Type I breakdown, with fractures 
filled with gypsum, results from mechanical wedging due 
to crystallization of the gypsum. Similar rock fragments are 
found near cave entrances, where they result from frost 
action. Type II breakdown is more complex. The fragments 
and plates are angular and sharp and are fractured across the 
usual zones of weakness—bedding planes and joints. Many 
of the fragments, only a few centimeters on a side and less 
than a millimeter thick, crush like broken glass when walked 
upon. These irregular plates are the signature of the crystal 
wedging process. The limestone bedrock is shattered and 
intermixed with gypsum so that the passage walls become 

piles of rubble. 

Geologic Processes That Initiate Breakdown 

Within the context of either the fixed-beam or Tharp model, 
any geologic processes that lengthens the beams or converts 
fixed beams into cantilever beams can move the ceiling beds 
from stable to unstable configurations. Beam thickness, 
flexural strength, and fracture toughness are properties of the 
bedrock and do not change during evolution of the cave 
passage. A list of triggering processes, not necessarily 
complete, is as follows: 

Passage enlargement below the water table. Phreatic 
passages continue to enlarge as water flows through 
them. If the hydrologic conditions are such that the 
passage is not drained, it may continue to enlarge 
until it becomes mechanically unstable 

Removal of buoyant support. By Archimedes’ principle, the 
ceiling beds of a water-filled cave passage are buoyed 

60 Breakdown 

upward by a force proportional to the ratio of the 
density of water to the density of the bedrock. For 
limestone with a typical density of 2.65 g/cm*, 35% 
of the buoyant support of the ceiling is lost when the 
cave is drained. 

Effects of base-level back flooding. During the time when 
the emergent cave passage is in the flood-water zone, 
rises and falls in base level alternatively fill and drain 
the cave passage. Additional dissolution at this time, 
particularly dissolution along ceiling joints, can turn 
fixed beams into cantilever beams and destabilize the 

Action of vadose water. Formation of vertical shafts, 
solution chimneys, and solutionally enlarged fractures 
by the action of undersaturated vadose water often has 
the effect of cutting ceiling beds, thus changing fixed 
beams into cantilever beams. 

Ice wedging. Caves that draw in cold winter air can have 
freezing conditions some distance inside. When water 
moving through joints and bedding plane partings 
freezes, then the expansion creates enough force to 
fracture the bedrock. 

Crystal wedging. Replacement of calcite in the bedrock by 
other minerals can exert a wedging effect. Because 
gypsum has a greater volume than the calcite it replaces, 
enough force is generated to fracture the bedrock. 

Role of Breakdown in Speleological Processes 

Both geologic triggering and slow creep of beds under load 
assure that breakdown can occur at any time during the 
evolutionary history of a cave passage; however, breakdown 
processes are most active during the enlargement phase of 
cave development and during the decay phase of cave 
development. The role of breakdown in the enlargement 
phase includes the following: 

Breakdown during the enlargement phase exposes more 
limestone surfaces and thus increases the rate of 

Upward stoping by breakdown processes can create large 
chambers, if actively circulating water removes the 
breakdown blocks at floor level. 

Upward stoping along fracture zones with removal of 
fallen blocks results in the formation of stoping shafts; 
Sétano de las Golondrinas in Mexico is an 
outstanding example. 

Breakdown in master conduits, particularly during the 
floodwater stage, can provide a support structure for 
groundwater dams. Silt and clay that deposit behind 
the blockage seal the dam, raise hydraulic heads 
upstream, and thereby generate a hydraulic gradient 
for the formation of new tap-off passages. 

Breakdown continues to play a role during the stagnation 
and decay phases of cave development as follows: 

Breakdown processes can stope upward to interconnect 
previously isolated cave levels into an integrated 
system of passages. 

Truncation of cave passages by the formation of terminal 
breakdown is a dominant process in the breakup of 
continuous conduits into the fragments characteristic 
of the decay stage of cave development. 

The final phase in the decay of caves is the passage 
collapse that takes place when the eroding land 
surface intersects the stress dome in the rocks of the 
cave ceiling. 

The final residue of a cave is a rubble zone consisting largely 

of breakdown. 


Davies, W.E. (1951) Mechanics of cavern breakdown, Natl. Speleol. Soc. 
Bull. 13: 36-43. 

Tharp, T.M. and T. J. Holdrege (1994) Fracture mechanics analysis of 
limestone cantilevers subject to very long term tensile stress in natural 
caves. In First North American Rock Mechanics Symposium Proceedings 
(PP. Nelson and S.E. Laubbach, Eds.). A.A. Balkema, Rotterdam, pp. 

Tharp, T.-M. (1995) Design against collapse of karst caverns. In Karst 
Geohazards (B.F. Beck, Ed.). A.A. Balkema, Rotterdam, pp. 397-406. 
White, E.L. and White, W.B. (1969) Processes of cavern breakdown, Natl. 

Speleol. Soc. Bull. 31: 83-96. 

White, E.L. and White, W.B. (2000) Breakdown morphology. In 
Speleogenesis: Evolution of Karst Aquifers (A.B. Klimchouk, D.C. Ford, 
A.N. Palmer, and W., Dreybrodt, Eds.). National Speleological Society, 
Huntsville, AL, pp. 427-429. 

Burnsville Cove, Virginia 

Gregg S. Clemmer 

Butler Cave Conservation Society, Inc. 


For more than half a century, cavers have pushed, explored, 
and mapped the caves of Virginia’s Burnsville Cove. Located 
near the Bath/Highland County border approximately 80 
km west of Staunton, VA, this sparsely populated area is 
known for its rural, scenic character. The systematic study 
and mapping of Breathing Cave by Nittany Grotto of 
The Pennsylvania State University, beginning in 1954, 
marked the first organized effort to chart what was then 
the largest and best known cave in Burnsville Cove. Initial 
work produced an overland topographic survey, passage 
cross sections and longitudinal profiles, and an analysis of 
the cave’s formation relative to stratigraphic folds and faults 
(see Fig. 1). 

Bevin Hewitt’s dramatic aqualung dive into the Mill Run 
spring in 1956 and his discovery of Aqua Cave beyond fueled 
interest in finding more caves in the Cove. Ike Nicholson’s 


‘ 7 ne 

-7 “WooozeLt 

Burnsville Cove, Virginia 61 






* ~ 

STONE cave \\ 




FIGURE 1 (A) Map of Burnsville Cove showing the drainage system, location of caves, and large surface depressions. 

discovery of Butler Cave 2 years later (Clemmer, 2001) and 
the rapid discovery of more than 16 km of large cave passages 
there, including the Sinking Creek trunk, confirmed that 
Burnsville Cove possessed vast underground secrets. 

The 1982 Burnsville Cove Symposium summarized the 
geology and cave descriptions of the Burnsville Cove 
reported to that date. Nicholson and Wefer (1982) described 
five caves—Boundless, Breathing, Butler, Better-Forgotten, 
and Aqua—as being part of an underground integrated 
karst drainage system situated between Jack Mountain to 
the west and Chestnut Ridge to the east. These caves, 
although not connected by human transit, comprised the 
Butler Cave—Sinking Creek Cave System, described at 
that time as containing approximately 35 km of mapped 

Over the past two decades, additional discoveries and 
mappings have greatly expanded the extent and understand- 
ing of this cave system. With more than 88 km of mapped 
cave passages in Burnsville Cove, this isolated corner of the 

Old Dominion ranks as one of the primary karst regions in 
the United States. 


Burnsville Cove sits astride the Highland/Bath County line, 
20 km southwest of the village of McDowell, VA. Bordered 
by Jack Mountain on the west and Tower Hill Mountain to 
the east, the Cove is a broad, dual-synclinal valley, split 
longitudinally by a folded anticlinal ridge known as Chestnut 
Ridge. The region measures approximately 8 km long and 
5 km wide and plunges geologically to the northeast. 

The Helderberg limestones of Silurian—Devonian age 
make up the exposed karst of the Cove. The stratigraphy of 
the Helderberg in not a uniform calcareous sequence, a fact 
that significantly impacts the speleogenesis of the area (Hess 
and White, 1982). The Helderberg, which sits atop the 
shaley Tonoloway limestone member of the Cayuga Forma- 
tion, is about 150 to 160 m thick. From bottom to top, the 

62 Burnsville Cove, Virginia 


© 1997 BCCS, Inc. 

; . 7 


oy + 


vA ee 


1000 1500 

FIGURE 1 (continued) (B) outline map showing principal caves in Burnsville Cove in relation to each other and to the geologic structure. (Maps courtesy of 

the Butler Cave Conservation Society.) 

sequence consists of the Keyser limestone (100 m thick), the 
Coeymans (12 m), the New Scotland (5 m), and the Becraft 
Cherty limestone (40 m). Of these members, the Keyser 
contains three thin layers of insoluble sandstone (lower and 
upper Clifton Forge sandstones and Healing Springs 
sandstone) that significantly impact cave development in 
Burnsville Cove. With water rarely able to breach these 
layers, cave passages are often floored and/or roofed by these 
sandstones for long distances, lending a characteristic flat 
ceiling to many of the galleries. 

The geologic structure of Burnsville Cove is not straight- 
forward. The two synclinal valleys so separated by the folded 
anticline of Chestnut Ridge extend northeast from the little 
village of Burnsville. Sinkholes dot the pastures and woods, 
particularly the upper (southwest) portion of Burnsville Cove, 
where Burnsville Sink harbors the entrances to Butler Cave. 
At the Water Sinks 5.5 km to the northeast, a dramatic stream 
insurgence flows into small Water Sinks Cave. Water drainage 
in almost all other circumstances in Burnsville Cove is sub- 

surface (Davis and Hess, 1982), discharging at several large 
springs on the Bullpasture River 8 km northeast of Burnsville. 

Aqua Cave is the primary resurgence for Sinking Creek 
Valley, a western synclinal valley containing the Boundless, 
Butler, Breathing, Better Forgotten, Barberry, Buckwheat, 
Blind Faith, Battered Bar, and Helictite caves, as well as the 
Burnsville Turnpike/Black Canyon portion of the Chestnut 
Ridge cave system. Emory Spring, a road-covered karst spring 
1.5 km upriver from the Aqua Cave resurgence contains no 
known cave but may be the resurgence for waters in recently 
discovered Helictite Cave. Cathedral Spring, located on the 
Bullpasture River 800 m downstream of the Aqua Cave 
resurgence, is the primary outflow for the caves of the eastern 
synclinal valley: Burns, Robins Rift, and the Bobcat—Blarney 
Stone portion of the Chestnut Ridge Cave System. Of key 
note here is the drainage divide in the Chestnut Ridge Cave 
System made possible by the North/South Trunk cutting 
transversely through the Chestnut Ridge anticline. Waters in 
the Burnsville Turnpike (possibly the largest contiguous 

underground chamber known in Virginia, averaging 20 m 
wide by 15 m high and extending more than 1200 m, with 
the most massive segment measuring 45 m wide and 425 m 
long) continue into the narrow confines of Black Canyon 
and sump at the 622 Sump only to reappear in Aqua Cave 
and flow into the Bullpasture River. Waters in the Cyanide 
Canyon section and all of the Blarney Stone section of the 
system emerge at Cathedral Spring. 


Well-known Breathing Cave was mined for saltpeter during 
the Civil War. It gained prominence after World War II as a 
popular sport cave with members of the National Speleolo- 
gical Society (NSS). Interest spread among NSS members 
and other cavers to investigate Mill Run Spring on the 
Bullpasture River. Bevin Hewitt’s discovery of Aqua Cave 
in 1956 amply demonstrated that more cave waited to be 
found. Ike Nicholson’s discovery of Butler in 1958 attracted 
a considerable number of experienced cavers to Burnsville 
Cove. The August 1958 Sand Canyon camp expedition 
(Nicholson and Wefer, 1982) “recon-mapped” 4500 m of 
new cave. Dozens of leads abounded. Nittany Grotto joined 
the survey in November 1958 and quickly mapped more than 
8 km of intricate passages off the upstream and downstream 
trunk. Upstream discoveries in 1959 took cavers beyond 
Penn State Lake to the distant, joint-controlled passages of 
what later became known as Mbagintao Land. Far down- 
stream, two parallel sumps—Last Hope and Rats’ Doom— 
blocked Butler’s advance to the northeast, but four years later 
an obscure side lead at Kutz Pit Junction revealed the muddy, 
remote galleries of Marlboro Country. 

Overland surveys from Butler to Breathing plus the added 
data from both cave maps pointed to a possible connection. 
By 1967, a host of cavers were pushing from both caves to 
connect (Nicholson and Wefer, 1982). Despite new disco- 
veries in Breathing and repeated efforts at digging in both 
caves, no connection has been found and the caves remain 
more than 150 m apart. 

On Chestnut Ridge, a small pit discovered in 1959 was 
pushed in hopes of extending Butler downstream. It quickly 
degenerated into a vertical crawl of daunting proportions. 
Ten years later, cavers returned to this Better Forgotten Cave 
and hammered through the vertical crawl to find more than 
a 300 m of large trunk passage ending in a terminal sump. 
Upstream remained choked in breakdown. 

As the 1960s ended, a number of cavers, hoping to protect 
the pristine nature of Butler Cave, formed the Butler Cave 
Conservation Society (BCCS), ushering in the first private 
cave conservation organization in the United States. In 1975, 
the BCCS purchased the 65-acre tract of land containing the 
Nicholson entrance to Butler Cave. 

Exploration in Burnsville Cove slowed in the 1970s. The 
Robins Rift dig was a dynamic project in the relatively 
caveless eastern synclinal valley and eventually led to 

Burnsville Cove, Virginia 63 

approximately 600 m of discovery, but surface digs in the 
Cove yielded little significant cave. Remapping in Butler 
took priority and despite a kilometer of virgin cave 
discovered in Mbagintao Land, some began to feel that 
Burnsville Cove had yielded most of its secrets. 

In 1979, members of the Shenandoah Valley Grotto 
visited an obscure blowing cave on Chestnut Ridge first 
located by David Nicholson in 1957. In a series of gritty, 
exhausting trips commencing in early 1982, they dropped 
deep into the heart of the ridge, following good air and a 
small, contorted stream passage in this cave they called 
Bobcat. Their discovery of the North/South Trunk in 1983 
led to new, exciting discoveries of extensive, large caves. 
Finding day trips too short, highly fatiguing, and vastly 
inefficient for mapping in such remote passages, they began 
camping in the cave. From 1984 to 1990, 27 separate under- 
ground camps explored and surveyed more than 14.5 km of 
virgin cave, and ultimately Bobcat was given the distinction 
of being the deepest cave in Virginia. 

In 1989, Ron Simmons conducted a series of cave dives 
into the constricted fissures of Cathedral Spring. After 
widening the cherty, underwater conduit, Simmons mapped 
a larger, descending passage to a depth of 46 m. With 290 m 
surveyed and the cave continuing its plunge, Simmons dis- 
continued the exploration in the face of serious hypothermia, 
extended decompression times, and special gas requirements. 

In March of 1991, ridge cavers dug into a small sink on 
the eastern flank of Chestnut Ridge and found Blarney Stone 
Cave. A twisting, wet passage similar to that encountered 
in Bobcat led to an extensive, multilevel, decorated gallery of 
cave passages exceeding 6 km in length. Blarney Stone and 
Bobcat were connected in a dual team effort in August, 1994, 
thus forming the 22.5-km Chestnut Ridge Cave System. 

Digging for new caves continued, and in 1993 Barberry 
Cave (Schwartz, 1999) was discovered in an obscure pasture 
sinkhole 1 km east of Burnsville. A second entrance dug the 
following summer provided more comfortable, safer entry to 
a cave that had now been explored to a length of nearly 2 km. 
In November 1994, Ben Schwartz and Mike Ficco pushed an 
extremely low water crawl more than 30 m to find a large 
trunk passage headed to the northeast. Survey here added 
3 km of cave to Barberry, but the Barberric Crawl soon had 
everyone considering a second series of cave camps. The first 
campers into Barberry, however, became temporarily trapped 
for several days when high water completely sealed the 
Crawl. Despite a media frenzy, the group engineered an 
intrepid self-rescue only to have the entrance later placed off- 
limits by the landowner. 

Undeterred, cavers resumed digging once again, this time 
on the land of caver Nevin W. Davis. After drilling a 25-cm 
hole 21 m down into the ceiling of the large Barberry trunk 
passage beneath the Davis farm, cavers began excavating and 
shoring to install a large metal tank endwise. Further drilling 
and blasting, accompanied with a full complement of specta- 
cular setbacks, eventually opened the Big Bucks Pit entrance 

64 Burnsville Cove, Virginia 

to Barberry Cave in 1996. Meanwhile, an ongoing dig on 
the slopes above the Water Sinks broke into virgin cave in 
March, 1996. Named Helictite Cave for its pretty display of 
such formations in the entrance area, the cave led to a rabbit 
warren of joint-controlled passages that are still being 

Strong air also lured cavers into Burns Chestnut Ridge 
Cave. Extremely arduous trips to “bottom” this tight cave 
had exhausted and frustrated a succession of caving teams for 
more than three decades, but a concerted effort led by Gregg 
Clemmer, Nevin Davis, and Tom Shifflett finally revealed 
a sizeable stream passage 200 m below the entrance. Subse- 
quent trips pushed the depth to 240 m, surpassing the 
Chestnut Ridge Cave System as Virginia’s deepest. 

Given this success, digging for caves in Burnsville Cove 
accelerated. Buckwheat Cave was opened after an hour's 
effort in March 1998. Pushing low crawls and following air 
brought mappers to within 4 m of connecting with the far 
downstream end of Barberry. Cave diggers found Blind Faith 
Cave the following March after digging down 4m into a 
blind sinkhole. The cave has been mapped to more than 
1 km of passage and drains toward Woodzell Sink. 

A dig in April of 2000 yielded an improbable entrance to 
the surprisingly extensive Battered Bar Cave. Atop a narrow 
karst saddle on the west flank of Chestnut Ridge, the cave 
currently is mapped to 2072 m of passage. Two months after 
this discovery, a utility lineman climbed up a roadside bank 
1.5 km east of Burnsville to check a telephone repeater and 
discovered a blowing hole. This By-the-Road Cave is now 
gated and managed by the BCCS in a unique agreement with 
the Virginia Department of Transportation, which owns the 

Breathing Cave 

Breathing Cave is one of the best known and most visited 
noncommercial caves in Virginia. Until the discovery of 
Butler Cave, it was the largest known cave in the state 
(Douglas, 1964). Partially mapped several times after 1945, 
the cave was more completely surveyed by Nittany Grotto of 
The Pennsylvania State University in the late 1950s who 
charted it to a length of 7.3 km (Holsinger, 1975). Devel- 
oped in the Keyser Limestone, Breathing was mined for 
saltpeter during the Civil War. A challenging, heavily joint- 
controlled, parallel maze cave, it divides 30 m inside the large 
sinkhole entrance into the Historic Section or saltpeter mine 
on the left and the Main Section on the right. Here, to the 
right, several kilometers of parallel, interconnected passages 
trend southeast, ending in a series of very low, narrow, wet 
crawls one surveyor termed pseudopsyphons. At the extreme 
terminus, the cave approaches the Good News Passage in 
Butler Cave (Nicholson and Wefer, 1982). Breathing Cave is 
currently being resurveyed by the Gangsta Cavers of the 

mid-Atlantic area. The cave is a registered National Natural 

Aqua Cave 

Aqua Cave was discovered by Bevin Hewitt in July 1956 by 
diving 8 m horizontally into Mill Run Spring. A low airway 
was subsequently blasted from the left ceiling of the spring, 
giving cavers without dive gear a sporting access to 2 km of 
bracing, fairly large river cave. The recent discovery of an 
extensive upper level, the Big Brother section, puts the cave 
at over 2700 m in length. 

Butler Cave 
With over 25 km of mapped passage, Butler Cave ranks as 

the longest cave in Burnsville Cove. Formed in the Keyser 
and Tonoloway limestones, extensive portions of the cave are 
developed between the upper and lower Clifton Forge 
sandstones of the Keyser, giving some passages a distinctive 
flat ceiling over great expanses (Hess and White, 1982). The 
cave underlies the western synclinal valley of the Cove, a fact 
that in large part determines its passage layout. Entering the 
spacious, central Trunk Channel at Sand Canyon, visitors 
follow the axis of the syncline as they walk up- or down- 
stream. Infeeder branch passages intersect primarily from the 
west. A number of wet weather streams course through these 
side passages to the Main Trunk. All waters in Butler Cave 
have been dye-traced to their resurgence at Aqua Cave (Davis 
and Hess, 1982). Although not as spectacularly decorated as 
other caves in the Cove, Butler awes with its sheer volume. It 
remains a cave of great challenges, be they a novice’s first 
glimpse of the Moon Room or a veteran's long trek to the 
remoteness of such exotic destinations as Djibouti or the 
Doom Room. Like Breathing Cave, Butler Cave is a registered 
National Natural Landmark. The cave is owned and managed 
by the BCCS. A second entrance to the system, accessed by 
a culvert pipe, was dug in 1998. (See Figs. 2 to 5.) 

Better Forgotten 

Better Forgotten is an aptly named, tight, muddy, multidrop 
cave on the west flank of Chestnut Ridge near the Bath/ 
Highland County line. Its 12-m pit entrance leads to a series 
of narrow vertical drops as the cave develops down dip. This 
eventually intersects a 580-m long section of stream trunk 
passage that ends downstream in a sump (Holsinger, 1975). 
The cave is 1200 m long and reaches a depth of 130 m. The 
stream has been dye-traced to its resurgence at Aqua Cave 
on the Bullpasture River. Better Forgotten Cave is owned by 
the BCCS. 


Boundless Cave opens in a small sink 800 m southwest of the 
entrance to Butler Cave. Trending northeast, the passage is 

FIGURE 2 The Glop Slot is a very narrow squeeze at the bottom of the 

entrance pit at the Nicholson Entrance to Butler Cave. Until the opening of 

the SOFA entrance, all who entered the cave had to pass through it. 

[GURE 3 This camp at Sand Canyon in Butler cave was occupied by a 

seven-man crew in August 1958, and most of the easily accessible passages 
in the cave were mapped and photographed. Notice the use of cotton clothes 
and provisions still in their original cardboard packaging. These practices 
would be frowned upon today. 

characteristically very low and filled with sand and cobbles. A 
small stream has been traced to its resurgence in Aqua Cave 
more than 6 km to the northeast (Holsinger, 1975). 

Opened by digging a large, air-blowing sink located at the 
western base of Tower Hill Mountain, Robins Rift Cave 
quickly developed a notorious reputation for entrance 
instability. At least four separate cave digs over the last 30 
years have attempted to keep this cave open to visitation. 
With more than 600 m of passage surveyed, it stood as the 

FIGURE 4 Sometimes it is just difficult to get up in an environment of 
perpetual darkness. This photograph from the 1958 camp in Butler Cave 
shows equipment of the time. Today, the inflatable air mattress would be 
replaced by a Therm-a-Rest pad or equivalent and the down sleeping bag 
would be a synthetic fiber-filled bag which would remain warm even if it gets 
a bit damp. 

SIGURE 5. Ike Nicholson, the discoverer of Butler Cave, admires the crystal- 
filled dry pools at Crystal Craters. This photograph was taken during the 
1958 camp in the cave. 

largest cave in the eastern synclinal valley of the Burnsville 
Cove until the discovery of Blarney Stone Cave in 1991. 
Water in Robins Rift has been traced to the Cathedral Spring 
resurgence (Davis and Hess, 1982). 

Bobcat Cave was previously known as Chestnut Ridge 
Blowing Cave (Douglas, 1964). The two entrances to Bobcat 
are situated about 10 m apart near the top of Chestnut Ridge 
just south of the Bath/Highland County line. The 550-m- 
long entrance series is a muddy, contorted, plunging slot 
following a small stream that has been dye-traced to 
Cathedral Spring. This stream is finally intersected by 
Tombstone Alley, a dry, paleo-overflow segment that leads to 

66 Burnsville Cov 

the North/South Trunk, a passage ranging from 6 to 21 m 
wide and up to 12 m high (Rosenfeld and Shifflett, 1995). 
One of three large trunk passages in the cave, the North/ 
South Trunk winds through large breakdown and exceptional 
displays of aragonite trees. To the north, pits interrupt the 
trunk passage, which appears to terminate at voluminous 
SVG Hall. A 30-m lead climb here, however, leads to the 
blowing Porpoise Passage, which crosses over the Chestnut 
Ridge anticline. Beyond, the 6-m Mud Piton Climb leads to 
24-m Damart Drop and a second pitch of 11 m, Polypro Pit. 
This drops into the second trunk passage consisting of Sixth 
of July Room to the south and the highly decorated Jewel 
Cave/Big Bend area to the north. Maret’s Lead out of Sixth 
of July rambles through big rooms floored in slippery mud 
and challenging down climbs for 600 m, ending in a sump 
190 m beneath the entrance. This has been dye-traced to the 
Aqua Cave resurgence. A small, blowing infeeder here leads 
to Black Canyon, 800 m of washed, scrambling stream cave 
in very dark limestone. A sporting up climb through a 
cascade intersects the beginning of the Burnsville Turnpike, 
the third trunk and by far the largest. Extending more than 
1200 m with widths approaching 44 m and heights up to 
30 m, the Turnpike is one of the biggest and most remote 
underground passages in Virginia. The Turnpike ends in 
uptrending breakdown, with a stream entering from a lower 
level of breakdown. The North/South Trunk south of its 
intersection with Tombstone Alley is a pleasant walk of 
300 m toward the Camp Room. Off the southeast side of the 
Camp Room, a steeply plunging down-dip lead drops into 
the Shamrock Dome area of the cave. A series of muddy slots 
and down climbs leads to Satisfaction Junction and the 722 
Sump, the deepest point in the system. Southwest from the 
Camp Room, but on the same relative level, the North/South 
Trunk continues as a series of rambling, up and down climbs. 
A tight slot with air can be followed through collapsed 
breakdown to the South Lead Terminus, a comfortable room 
ending in more breakdown. A small, dug hole under one 
ledge drops into the Blarney Stone continuation of the 
system. The cave is considered one of the most demanding in 
the Old Dominion. (See Figs. 6 to 10.) 

Blarney Stone Cave was discovered by digging open an 
obscure sink on the eastern slope of Chestnut Ridge in 
March of 1991. Its muddy fissure entrance series—four short 
rope drops and a series of challenging down climbs—is a 
shorter version of the Bobcat entrance series. After 300 m, a 
small overflow tube leads to larger cave. On an upper level, a 
5-m aid climb across a 20-m deep shaft accesses an impressive 
paleotrunk named Ghost Hall. Highly decorated with 
stalagmites, stalactites, totem poles, and columns, Ghost Hall 
leads south to expansive Upper Ghost Hall. Black Diamond 
Crawl, a small stream crawl with black gravel, also exits the 
south end of Ghost Hall. Black Diamond intersects a pit that 

FIGURE7 An anthodite formation called The Elk Horn in Bobcat Cave; 
the white vertical piece at the bottom is about a foot long. 

descends to the Pearly Gates, named for a beautiful and 
prolific display of cave pearls. Moon River extends upstream 
and downstream below this point for nearly 600 m. 
Numerous side passages abound here, the most notable being 

An anthodite formation in Bobcat Cave. 

the Stairway to Heaven, an extensive series of challenging 
upclimbs that rise more than 150 m. North of Ghost Hall, 
the cave winds through the totem poles of Leprechaun 
Forest. Extraordinary crystalline white chandeliers decorate a 
delicate section of the wall. A large walking passage rambles 
to the north, eventually finding the obscure lead beyond the 
Earthworks to the cave’s connection with Bobcat Cave. 

Barberry Cave has three excavated entrances, all but one 
closed. The entrance to Big Bucks is an excavated 21-m shaft 
dropping into an impressive 23-m-high trunk passage. At 
the bottom of the pit, this decorated, spacious stream trunk 
extends 670 m to the north and 550 m to the south and is 
aligned with the south end of the Burnsville Turnpike. The 
stream ends in a deep sump, but the trunk passage continues 
another 150 m to massive breakdown. The Woway, a sizeable 
side lead entering the main trunk passage from the west, 

a - 

RE 9 Crystal formations found in a section of Bobcat Cave called the 

North/South Trunk. 

0 The main camp in Bobcat Cave; the crew is gathered around 

the cooking area. 

extends via watery passages to within 120m of Butler 
Cave. An air-blowing lead at the end of this very tortuous 
passage still holds promise of a connection with Burnsville 
Cove’s biggest cave. Barberry Cave is 5.31 km long. (See Figs. 
11 to 13.) 


Morphine Waterfalls. The main stream in Barberry Cave, Bath 

County Virginia, in flowing down the bedding before cascading over a 6 foot 
waterfall. The main passage in this area is 40 feet wide and 70 feet high and 
has massive flowstone decorating the walls. 

Buckwheat Cave, another excavated entrance cave, plunges as 
a walking stream passage into a series of low water crawls, 
blocked by massive breakdown. Coming to within 4 m of 
portions of Barberry Cave, Buckwheat drains a small part 
of the western flank of Chestnut Ridge. To date, 670 m of 
cave have been mapped in Buckwheat to a depth of 42 m. 
Digging is ongoing for a connection to Barberry Cave. 

Blind Faith was discovered the year after Buckwheat Cave by 
digging a 4-meter shaft in a small sink in the next wooded 
valley 600 m north of Buckwheat. A series of crawls and 
challenging down climbs eventually drop into a going stream 
passage. This degenerates downstream in an extremely low, 
downtrending passage. Upstream, the cave winds for several 
hundred meters along the western flank of Chestnut Ridge 
but stops well short of connecting to nearby Buckwheat. 
More than 1000 m of cave have been charted in Blind Faith 
to a depth of 48 m. 

Great White Wow in Barberry Cave, Bath County, Virginia. 
The large spacious passage in the Barberry Trunk Trends southwest until it 

reaches an intersection where this grand formation is suddenly encountered 
by the visitor. The first word spoken by its discoverers was used as part of the 


Battered Bar Cave is located about 450 m north of the Blind 
Faith entrance on the edge of a deep sink corresponding to 
the terminal end of the Burnsville Turnpike in the Chestnut 

la = So si) > ees i Pres 
2s > = = mA a 
FE ai CRS de as PE os “ 

FIGURE 13 “Sidewalk superintendents” gather around Big Bucks Pit. The 
third entrance to Barberry Cave began as a drill hole followed by a 6-foot 
diameter shaft hand “dug” down to the top of the 80-foot high ceiling in the 
cave. In this photo, cavers are gathered around the top of the shaft, peering 
down to observe the “digging” activity at the bottom of the shaft. 

Ridge cave system. A narrow, 18-m pit leads to an even 
tighter slot that slopes down 8 m to the top of a slippery 
12-m-deep shaft. A steep, muddy up climb leads to a third 
drop of 4m. Beyond, through massive breakdown blocks, 
the cave opens up dramatically. In the first big room, a fissure 
leads down to the Ramp, a steeply inclined 30-m-long 
chuteway floored with breakdown lingering at the angle of 
repose. Two walking passages extend to the south from the 
bottom of the Ramp, the left-hand passage dividing after 
some 125 m into left and right branches. The left branch 
approaches the downstream end of Blind Faith Cave, ending 
near an unusual folded limestone feature called the Stone 
Rainbow. The right-hand branch continues south 200 m, 
plunging dramatically to a sump in a passage covered with 
pure white sand. The right-hand trunk at the bottom of the 
Ramp climbs through breakdown, then extends for 300 m 
into a maze of small passages ending in breakdown. 
Underneath the Ramp, a stream can be followed for 150 m 
to massive breakdown, much of which has yet to be pushed. 
Any connection with the Burnsville Turnpike is approx- 
imately 250 m beyond. The cave has been surveyed to more 
than 2040 m of passage at a depth of 125 m. 


Helictite Cave is formed near the top of the Helderberg 
limestone sequence, dissolved mainly out of the New 
Scotland Limestone. A vast maze cave of tubes and canyons 
with one major paleostream passage, Helictite possesses dra- 
matic examples of dogtooth spar, cave pearls, helictites, and 
slickensides. With more than 11 km of cave mapped since its 

Burnsville Cove, Virginia 69 

discovery in 1996, Helictite is not a typical Cove cave. Its 
drainage has not been dye-traced. (See Figs. 14 to 19.) 

Burns Chestnut Ridge 

Burns is the deepest cave in Burnsville Cove at 240 m. The 
entrance series of low, sinuous, muddy crawls; tight, body- 
sized cracks; and plunging slot canyons to the bottom of the 
cave is one of the most arduous 500 m of cave in the United 
States. An impressive stream canyon below the 198-m level 
soon sumps upstream but flows north to a series of cascades 
at the 213-m level. Here, a high lead some 12 m above the 
stream leads to nearly 2 km of walking passage, eventually 
giving access to the rarely visited Cathedral River. This 
passage was mapped downstream for 335 m to a point where 

FIGURE 14 A cluster of helictites, the signature of Helictite Cave, Highland 
County, Virginia. When this cave was first entered some of the first 

formations seen were helictites. As exploration and surveys continued it was 
realized that this cave was far richer in this type of formation than other 
Virginia caves. This photo is part of a large cluster of helictites located just a 
hundred feet inside the entrance. (Photograph taken by Arthur N. Palmer. 
With permission.) 



clutch of cave pearls in a rimstone pool in Helictite Cave. The largest pearl 
is about the size of a golf ball and is with 9/1000 of an inch of being perfectly 
symmetrical. (Photograph taken by Arthur N. Palmer. With permission.) 

a es 

Traversing Dagger Pit, Helictite Cave, Highland County, 
Virginia. One of the connections to the 7-mile maze cave was found across 
the top of a pit named Dagger Pit for the blade-like pendants in the bottom 
of the pit. A traverse line enables a safe passage across the top and 40 feet 

above the “daggers” below. 

the water came to within 10 cm of the ceiling. The down- 
stream cave approaches Robins Rift to within 300 m. A high, 
dribbling infeeder in the dry, walkable upper level holds 
potential for a connection to the southernmost end of 
Blarney Stone Cave. 

The step across, Helictite Cave, Highland County, Virginia. 
Maneuvering through breakdown and canyon passage in Helictite Cave 

requires a daring step or two. The Step Across is a seemingly simple 
maneuver that often causes some consternation. 

Found in June of 2000, the entrance is a recent sink collapse 
located in the eastern synclinal valley 800 m southwest of 
Robins Rift. Strategically placed to offer access to the sumped 
upstream portion of Burns Chestnut Ridge Cave, By-the- 
Road is currently mapped to an ongoing in-cave dig. 
Considerable air blows from the cave in hot weather. By-the- 
Road is gated and managed by the BCCS at the request of 
the Virginia Department of Transportation. 

A preliminary report on cave fauna of Burnsville Cove was 
published in the Burnsville Cove Symposium (Holsinger, 
1982) and itemized 11 invertebrates and 8 vertebrates. Two 
species, the amphipod Stygobromus conradi and a beetle 

FIGURE 18 Pool spar, Helictite Cave, Highland County, Virginia. One of 
the crystals of calcite is spar sometimes called Dogtoothed Spar. In Helictite 
Cave, Highland County, the Dogtoothed Spar in this photo is located in a 
pool and is about 20 feet long and 6 feet wide. (Photograph taken by Arthur 
N. Palmer. With permission.) 

FIGURE 19 The Slickenslides Room, Highland County, Virginia. Helictite 
Cave has areas where faults are intersected by passage development. Here the 
Slickenslides of a fault has dropped into cave passage below. This makes for 

a dramatic passage where the scrapings or slickenslides are easily recognized 
on the ceiling and floor of this passage. (Photograph taken by Arthur N. 
Palmer. With permission.) 

(Pseudanophtalmus), are endemic to the Burnsville Cove. A 
more recent report (Hershler et a/, 1990) noted the dis- 
covery of a new species of aquatic snail, Fontigens morrisoni, 
citing Butler Cave as one of only two locations. An updated 
report describes the discovery of a new species of springtail 
(Arrhopalites) from Butler Cave, since found in another 
Virginia cave outside the Cove. The appearance of unknown 
animal tracks deep in the North/South Trunk of Bobcat Cave 
and the finding of a “raccoon-like” skeleton beneath the 
Camp Room proved a startling discovery. Photographs and 
castings were subsequently identified as belonging to Martes 
pennanti, which is a fisher unreported in Virginia for almost 
two centuries. 

A preliminary report on cave mineralogy of the Burnsville 
Cove was also published in the Burnsville Cove Symposium 

Burnsville Cove, Virginia 71 

(White, 1982) and characterized the secondary mineral 
deposits as sparse though widely dispersed. Recent dis- 
coveries in selected areas of the Chestnut Ridge System, 
Barberry, and Helictite caves reveal an astounding array of 
helictites, aragonite trees, anthodites, cave pearls, and moon 
milk. Sediment studies in Butler Cave have found evidence 
of magnetic reversal as well as iron-fixing filamentous 
bacteria in a brownish-yellow layer of goethite. 

After a half century of systematic investigation, it is 
now understood that the caves of the Burnsville Cove are 
hydrologically connected. Despite the physical barrier of 
the Chestnut Ridge anticline, the improbable presence 
of the North/South Trunk in the Chestnut Ridge System 
provides a key underground connection between the Butler 
Cave-Sinking Creek drainage of the western syncline and 
the caves of the less understood eastern syncline of the 
Burns—Blarney Stone—Cathedral Spring drainage. Despite 
the connection of Bobcat and Blarney Stone caves in 1994, 
connections between other caves of Burnsville Cove remain 
elusive. When realized, such achievements will only enhance 
a cave/karst region already recognized for its international 


Clemmer, G.S. (2001) That cave just had to be there. In Virginia Cavalcade 
(J.A. Campbell, Ed.). The Library of Virginia, Richmond, VA. 

Davis, N.W. and J.W. Hess (1982) Hydrogeology of the drainage system, 
Burnsville Cove, Virginia. In Burnsville Cove Symposium, (W.B. White 
and J.W. Hess, Eds.). Adobe Press, Albuquerque, NM. 

Douglas, H.H. (1964) Caves of Virginia. Virginia Cave Survey, Falls Church, 

Hershler, R., J.R. Holsinger, and L. Hubricht (1990) A revision of the North 
American freshwater snail genus Fontigens (Prosobranchia: Hydrobiidae). 
Smithsonian Contr. Zool. 509. 

Hess, J.W. and W.B. White (1982) Geomorphology of Burnsville Cove 
and the geology of the Butler Cave-Sinking Creek system. In Burnsville 
Cove Symposium (W.B. White and J.W. Hess, Eds.). Adobe Press, 
Albuquerque, NM. 

Holsinger, J.R. (1975) Description of Virginia Caves. Bull. No. 85, Division 
of Mineral Resources, Charlottesville, VA. 

Holsinger, J.R. (1982) A preliminary report on the cave fauna of Burnsville 
Cove, Virginia. In Burnsville Cove Symposium, (W.B. White and J.W. 
Hess, Eds.). Adobe Press, Albuquerque, NM. 

Nicholson, I.K. and FW. Wefer, (1982) Exploration and mapping of the 
Sinking Creek system. In Burnsville Cove Symposium. (W.B. White and 
J.W. Hess, Eds.). Adobe Press, Albuquerque, NM. 

Rosenfeld, J.R. and T. E. Shifflett (1995) The Caves of Burnsville Cove, 
Virginia. In Underground in the Appalachians: A Guidebook for the 1995 
NSS Convention (C. Zokaites, Ed.). National Speleological Society, 
Huntsville, AL. 

Schwartz, B. (1999). Exploring Barberry Cave, NSS News September, 

White, W.B. (1982). Mineralogy of the Butler Cave-Sinking Creek system. 
In Burnsville Cove Symposium. (W.B. White and J.W. Hess, Eds.). Adobe 
Press, Albuquerque, NM. 


ig S. Clemmer 

Butler Cave Conservation Society, Inc. 

he thorough exploration and survey of an extensive cave 

system demands that all participants “push the limits” of 
the cave to its “bitter end.” Such idealistic, oft-used phrases 
employed by cavers reflect a deeply held philosophy—an 
ethic some would say—that goal-oriented, expedition-style 
caving requires careful preparation, long-term dedication, 
and extensive stamina. “Push trips” to the bottom or to the 
far reaches of vast, complex cave systems challenge all three 
requirements. Planning entails cooperation, competency, 
specific goals, a mountain of gear, and the occasional 
kilometer of rope. Participation impacts everything from 
bank accounts and vacation time to jobs and marriages. 

It is a given that everyone enters the cave in superb 
physical and mental condition. But, what happens when the 
mountain of gear is consumed, when the kilometer of rope is 
rappelled, when the strongest caver is exhausted, and the cave 
still goes down and down, on and on? In the years since the 
founding of the National Speleological Society, Inc., in 1941, 
U.S. cavers have continued to push the limits. In the first half 
century of the Society’s existence, for a variety of reasons— 
personal comfort, novelty and intrigue, or simply because the 
cave went on and on—cavers in a few instances resorted to 
underground camps to pursue their respective goals. Cavers 
in Europe confronted the same challenges. Exploration in 
Switzerland’s Holloch expanded to an underground camp in 
1949. After four cavers were trapped by high water for 9 days 
in 1952, all exploration camps in Holloch became winter- 
time endeavors (Courbon et a/., 1989). The widely acclaimed 
1952 descent into Pierre Saint-Martin employed a 5-day 
underground camp but ended tragically with the death of 

Marcel Loubens (Tazieff, 1953). Gouffre Berger became 
the first cave to break 1000 m in depth, a feat realized in 
1956 that owed much to the staged underground camps 
of 1954-55 at 494 m and 860 m (Cadoux, 1957). A 1955 
expedition into the Cigalere also employed staged cave 
camps, but, instead of pushing the cave ever deeper, explorers 
confronted a daunting series of waterfalls as they ascended 
into the mountain (Casteret, 1962). 

The evolving European model for underground expedi- 
tion camping utilized advance supply teams to rig pits, 
lay phone lines, transport mountains of gear, and establish 
camps. Single-rope techniques being unknown, pits were 
negotiated by cable and rope ladders backed up by belay. The 
early exploration of Utah’s Neff Canyon Cave in October of 
1953 followed this European model and utilized a support 
crew to aid four cavers on a 33-hour trip to the bottom of the 
cave. Engaged in a rivalry with a local climbing club, the 
group carried in a large amount of gear—sleeping bags, cable 
ladders, 150 m of rope, field phones, and coils of wire—to 
support their effort. But, after enduring a “fitful sleep in cold, 
cramped quarters” in their unsuccessful attempt to find 
the cave’s deepest point, the explorers emerged “completely 
exhausted.” One member of the support crew spent a week 
in the hospital suffering “utter fatigue.” They chalked up 
their failure to “bulky packs and unmanageable gear in the 
narrow, jagged passageways” (Green and Halliday, 1958). 

A few months later, Floyd Collins’ Crystal Cave Expedi- 
tion (C-3 Expedition) electrified the caving community with 
a sensational attempt in Flint Ridge to push the far reaches 
of Kentucky’s most extensive cave. With movie cameras 
rolling and backed by a full complement of sponsors, 
fawning reporters, and radio broadcasters, the cavers entered 
the cave with ambitious goals for a week underground. Metal 
“Gurnee cans” protected gear through tight, rocky crawlways. 
Field phones connected remote sections of the cave with 
the surface support crew. Experts in cave biology, geology, 
hydrology, medicine, and meteorology accompanied the 


74 Camps 

hard-charging explorers, all eager to measure the cave, the 
cavers, and the phenomena therein. The expedition ended 
with numerous official reports detailing everything from 
sleeplessness and mild shocks from ring voltages in the 
phone system to the morale-boosting effects of candy and 
tobacco being delivered from the surface. The festive, self 
congratulatory tone at the expedition’s end ignored the large 
amounts of trash buried or burned in the cave. Several kilo- 
meters of abandoned telephone wire would litter the passages 
for decades (Brucker and Lawrence, 1955). Thankfully, the 
“success” of the C-3 Expedition was never repeated, but in a 
subsequent report 2 years later one participant recommended 
“simplicity in all phases of trip organization” as a future goal, 
warning that too often “success is judged by size instead of 
actual accomplishment” (Smith, 1956). 

In contrast, the Butler Cave Camp of August 1958 
commenced as a closely held secret. Not eager to get 
“scooped” and exploring mostly in blue jeans and wool shirts, 
the seven-man crew eagerly pushed deep into this Virginia 
discovery without field phones, surface support crew, or 
scientific agenda. At the end of an exciting, tiring week, they 
exited with 5 km “recon-mapped” and a collection of superb 
color slides, but they buried their trash and spent carbide in 
the cave. 

Youthful exuberance and naiveté characterized the August 
1962 cave camp in Indiana’ Sullivan cave. With lofty 
goals to map the cave, sample the soil for microbes, conduct 
psychological surveys on participants, and clean up extensive 
vandalism, the teenagers elected to spend 2 weeks camping in 
the cave despite its relatively close proximity to the entrance. 
Field phones connected them to the surface. Equipped with 
sleeping bags on cots and supplied with a double-burner 
Coleman stove, canned goods, rye bread, and even fresh 
vegetables (celery, carrots, and lettuce), the explorers endured 
a miserable, cold existence in wet, muddy clothes despite five 
complete changes of underground wardrobe. Although they 
mapped 2 km of passage, the young “Sullivaneers” discovered 
only half of it to be virgin cave. One participant characterized 
part of their mapping as “a comedy of errors.” 

Vastly more significant discoveries rewarded a two-person, 
week-long camp in Ellison’s Cave, Georgia, in 1969. Despite 
the dramatic failure of their only stove at base camp more 
than 250 m below the surface, this man and woman team 
stomached cold food and dank conditions to survey almost 
4 km of cave without field phones and surface crew, all with 
minimal impact to the cave (Smith, 1977). 

A bizarre example of cave camping occurred in 1972 in 
Midnight Cave, TX, when one man spent 6 months under- 
ground. Dismissed as nothing more than a publicity stunt by 
some, the venture did garner enormous attention, including 
a feature article in National Geographic. The subject entered 
the cave ostensibly to investigate the long-term psychological 
and physiological effects of solitary confinement and sensory 
deprivation. Amenities included a canopied sleeping area on 
a wooden platform, extensive incandescent lighting, field 

phone, books, and record player. At the end of his time 
underground, the relieved cave dweller declared his trial a 
success as the “longest beyond time experiment in history” 
(Siffre, 1975). 

Ongoing exploration of Wind Cave in South Dakota 
employed an underground camp in 1972. Situated near 
the Master Room, the relatively comfortable camp was 
supported by a surface crew, stocked by supply teams, and 
connected to the surface by field phones. To thwart hypo- 
thermia, participants toyed with the novel notion of running 
heat lamps in the camp on 480 V piped down the telephone 
cable. Despite the discovery of major extensions to the cave, 
no one favored a second underground camp the following 
year. “The logistics of running a base camp, although 
successful, were very difficult and time consuming,” wrote 
one organizer. Henceforth, the survey reverted to “long, 
single-day trips from the surface” (Scheltens, 1988). 

All of this experience gained was lost on organizers of 
Project SIMMER when 118 cavers descended on Simmons- 
Mingo Cave in West Virginia in October 1973. Ambitiously 
planned much like a military operation with a chain of 
command, mess tent, and administration tent, the expedi- 
tion ultimately consumed 10 hours of preparation to every 
hour actually spent underground. Planners managed to lay 
more than 15 km of wire for field phones, then touted their 
work as the “world’s largest in-cave communication 
network.” The Gurnee can of the C-3 Expedition morphed 
into a “Carts can,” a stovepipe and plywood contraption used 
to haul gear into the cave. Plagued by poor sleeping bags, 
wet clothes, ringing phones, and a miserable camp spot, 
the crews mapped less than a kilometer of cave. Project 
SIMMER had profited nothing from the C-3 experience of 
20 years earlier and never issued a final expedition report. 
Participants even abandoned the phone wire in the cave. Yet, 
beyond these disappointments, the overall underground 
manager opined at the end of the experience that for deep or 
remote cave exploration, “the small camp-in party [would be] 
more efficient than the larger, more formally organized 

American deep-caving expeditions to Mexico also began 
camping underground. European participants with extensive 
expedition experience contributed a wealth of knowledge 
toward maintaining a comfortable, efficient subterranean 
camp. Prolonged underground stays were begun in the 
mid-1960s and by the late 1970s had pushed the reaches of 
vast, deep, technically difficult caves (such as Sotano de San 
Agustin, La Grieta, Sumidero Yochib, Sotano del Rio Iglesia, 
and Sistema Purificacion) far beyond the reaches of conven- 
tional day-trip caving (Stone, 1978). (See Figs. 1 to 4.) 

Outside the warmer climes of Mexico, occasional cave 
camping in the temperate latitudes of the United States 
generated little appeal for second attempts. Despite some 
glowing declarations of expedition success, the cave campers 
of Neff Canyon, C-3, Butler Cave, Sullivan Cave, Ellison 
Cave, Wind Cave, Project SIMMER, and even a successful, 

URE 1 The chamber in which camp 3 is located in Sotano de San 
Agustin; this is part of the cave system Sistema Huautla in Oaxaca in 
southern Mexico. 

£2 A meal is being prepared at camp 3 in Sotano de San Agustin. 

comfortable camp deep in Fern Cave in Alabama brokered 
little enthusiasm to repeat their adventures. Pushing caves to 
their limits in the chillier continental 48 states went back to 
being brutal day-trip endeavors. Remote underground camps 
were best left for the warmer caves south of the border. The 
1983 discovery of a large cave system under Chestnut Ridge, 
near Burnsville, VA, provided the impetus for yet another 
group to try an extended underground camp in the United 
States. More than 20 km of challenging, decorated, virgin 
galleries rewarded those who endured the cold, sloppy, 
tortuous entrance series of Bobcat Cave (see Figs. 5 to 7). Yet, 
exhaustion and the real threat of hypothermia limited all 
efforts to safely extend exploration via increasingly longer day 
trips. With no other choices, cavers with decades of grueling 
experience grudgingly confronted the possibility that 
camping underground was the only feasible way to continue 
the survey. Given the Spartan experiences a generation earlier 
at nearby Butler Cave, few relished the idea. No cave in 
the United States had ever been continuously pushed and 
mapped in such a manner. Nevertheless, over the next 10 


Cooking area of the White Lead Room in La Nita, which is part 

of Sistema Huautla in Oaxaca in southern Mexico. 

URE 4 A caver is entering survey data into a programmable calculator, 

which will indicate where he is in relation to an already mapped passage in 
a connection attempt. This is at camp 2 in Nita Nanta, which is part of 
Sistema Huautla in Oaxaca in southern Mexico. 

years, more than 15 km of passageways were explored and 
surveyed via 27 underground camps in Bobcat Cave, culmi- 
nating in the 1994 connection with 7-km-long Blarney 
Stone Cave. 

76 Camps 

FIGURE 5 Bobcat Cave; another view of the cooking area after several days 
of use. 

FIGURE 6 Bobcat Cave; one caver’s area in the main camp chamber. He is 

using a hammock instead of a sleeping pad on the ground. 

To be fair, few caves offer the isolation and daunting 
physical challenges that justify camping underground, but 
Bobcat Cave did, and once the decision to camp had been 
made the question became one of how to thwart hypo- 
thermia, obtain adequate nourishment, maintain endurance, 
and still safely and efficiently get the cave competently 
explored and mapped. Custom nylon coveralls were the first 
big difference from previous camps. Until the early 1980s, 
experienced American cavers—with rare exceptions—went 
underground clothed primarily in cotton and wool. “Farmer 
John” coveralls ruled the day. Wet suits were tight and 
uncomfortable but battled hypothermia and provided 
protection far better than blue jeans and corduroy jackets or 
the wool sweaters and flannel shirts of earlier times. Comfort 
counted, and the eternal cold of soggy cotton and smelly 
wool when still kilometers from the entrance begged for 
garments promising warmth, agility, and the ability to stay 
dry despite the wearer’s body heat. 

Nylon did just that. Participants in the first Bobcat camp 
purchased yards of the fabric, adapted a borrowed coverall 
pattern, and sewed their own. Cave packs evolved the same 
way. The bulky, battered metal towing cans of the C-3 
Expedition and Project SIMMER never even came under 

FIGURE 7 Bobcat Cave; another caver’s camp area where a hammock is also 


consideration. Instead, long, cylindrical, flexible nylon duffel 
bags, also self-sewn, performed admirably. With a tether 
on one end for upright attachment to seat harnesses when 
ascending or descending drops, a handle in the middle for 
grasping in crawls and crevices, and back straps for carrying 
over long distances, the “camp duff” proved invaluable for 
getting gear and food into camp. Double or triple thicknesses 
of trash bags protected food, clothes, and sleeping bags from 
devastating leaks. Sucking out the air from such a packed bag 
before tying it off provided additional space. Two decades 
later, nylon packs and coveralls enjoy almost universal use in 
caving and even exude their own fashion statements, having 
spawned a cottage industry in custom cave gear, vertical rigs, 
and personalized repair using a variety of incredibly durable 
fabrics. Changing into dry, warm camp clothes upon 
reaching camp boosted morale, especially if one’s body heat 
aided the process; thus, wickers and polypropylene replaced 
cotton and wool undergarments. 

Sleeping underground, though, had always been a 
prolonged struggle against chill and dampness. Cotton or 
down sleeping bags were dismal failures, but lightweight 
fiberfill or synthetic bags worked nicely when laid on a foam 
pad (or, for example, a Therm-A-Rest pad) atop a reflective 
ground cloth. A stocking cap kept head and ears warm 
all night. Some campers even wore gloves. A dry change 
of socks, bound up in small plastic bags, assured dry feet 
even when moving about camp in wet, muddy cave boots. 
An extra polypropylene top and bottom, properly bagged, 
provided the luxury of a pillow. Hammocks, although 
favored by some on Mexican expeditions, were quickly 
abandoned after a fitful night tossing in the damp, 48° F chill 
of Bobcat Cave. The camp site itself needed to be relatively 
level, spacious enough for sleeping quarters and a commu- 
nity kitchen and eating area, and fairly close to reliable water. 
A drop of iodine per gallon of water accomplished water 
purification. The latrine was located in respectable proximity 
to camp, dug into a clay bank. 

Eating revolved around breakfast and dinner, supple- 
mented during the day by personal preferences (energy bars, 
Gorp™, cheese, candy, premade sandwiches, even a baked 

potato). Freeze-dried food covered most menus, being far 
tastier than the wretched examples of the past and 
significantly lighter than canned goods. (Note: In desert caves 
or where water is scarce, canned goods could be a significant 
supply for both food and water.) Tea, coffee, sugar, salt, 
oatmeal, dried fruit, pepper and other spices, and even luxury 
condiments were easily stuffed in zip-lock bags and buried in 
the depths of packed sleeping bags. 

Aside from ropes, climbing gear, bolt kits, and survey gear, 
community camp gear on the initial trip included a small 
white gas stove with repair kit; several full, secured fuel 
bottles; a cooking pot for hot water; three or four collapsible 
plastic gallon jugs; first-aid kit; and trowels and toilet paper, 
all divided among the participants. Outside of replenishment 
items, these were secured in the cave from camp to camp. 
Luxury items ranged from washcloths and personal journals 
to cards and a harmonica. Carbide provided 90% of the 
lighting, with candles around camp adding an intimate touch 
and saving acetylene. 

Cavers know that adaptability and incentive remain a 
vital part of pushing the limits. Future camps may very well 
embrace caving LED lamps, for which rumored 50-hour 
burn times on one set of four D cells would surely lighten 
camp duffs of pounds of bulky calcium carbide on the way in 
and spent carbide on the way out. Every Bobcat expedition 
entered the cave as a small, self-contained team. Never did 
more than nine cavers (three teams of three) participate; six 
proved the average. Surface crews lounging in administration 
tents fielding phones attached to kilometers of wire strung 
through near-virgin cave were never an option and would 
probably violate the conservation ethic of today’s cavers. 
Instead, with a safety contact just a few kilometers from the 
entrance, the expedition entered the cave with competent 
associates on the surface aware and available. In the years 
since, cave camping has remained a seldom-used tool of 
American cavers. 

The recent multi-sump isolation of camping on a tarp 
suspended above water deep in Mexico’s Sistema Huautla is 
surely the extreme (Stone et a/., 2002), but the continued 
success of camps in making new discoveries such as in 
Kentucky’s Fisher Ridge, New Mexico’s Lechuguilla, and 
Virginia's Omega System is a tribute to cavers’ adaptability to 
the challenging extremes of the caves they continue to push. 

See Also the Following Articles 
Recreational Caving ¢ Exploration and Light Sources 


Brucker, R.W. and J. Lawrence, Jr. (1955) The Caves Beyond: The Story of 
Floyd Collins’ Crystal Cave Exploration, Funk & Wagnall’s, New York. 

Cadoux, J. (1957) One Thousand Meters Down, George Allen & Unwin, 

Casteret, N. (1962) More Years Under the Earth, Neville Spearman, London. 

Courbon, P., C. Chabert, P. Bosted, and K. Lindsley (1989) Aélas of Great 
Caves of the World, Cave Books, St. Louis, MO. 

Green, D.J. and W.R. Halliday (1958) America’s deepest cave. NSS Bull. 20, 

Castleguard Cave, Canada 77 

Scheltens, J. (1988) Windy City Grotto at Wind Cave. NSS News January, 

Siffre, M. (1975) Six months alone in a cave. Natl. Geogr. March. 

Smith, M.O. (1977) The Exploration and Survey of Ellison’s Cave, Georgia, 
Smith Print and Copy Center, Birmingham, AL. 

Smith, P. (1956) Seven principles of effective expedition organization. VSS 
Bull. 18, 46-49. 

Stone, B. (1978) Underground camps for deep caves. AMCS Activities 
Newsl. 8, 37-45. 

Stone, B. and B. Ende, with P. Monte (2002) Beyond the Deep: The Deadly 
Descent into the World’ Most Treacherous Cave. Warner Books, New York. 

Tazieff, H. (1953) Caves of Adventure. Harper & Brothers, New York, NY. 

Castleguard Cave, Canada 

Derek Ford 
McMaster University 

astleguard Cave is the longest cave system currently 

known in Canada (20 km) and the foremost example 
anywhere of a cavern extending underneath a modern glacier 
(Fig. 1). It displays many striking features of interactions 
between glaciers and karst aquifers, a complex modern 
climate, rich mineralization, and a troglobitic fauna that has 
possibly survived one or more ice ages beneath deep ice cover 
in the heart of the Rocky Mountains. 


The cave is located in the northwest corner of Banff National 
Park, Alberta, very close to the Continental Divide. The 
region is one of rugged alpine mountains with many horn 
peaks, cirques, and U-shaped valleys typical of intensive 
glacial erosion, plus a few small but high plateaus. The range 
of elevation is from 1500 m asl in the floors of trunk valleys 
to summits at 3500 m. Mean annual temperatures are 0 to 
—14°C across this height range. Natural boreal forests extend 
up to ~2100 m, passing into grass and low shrub tundra and 
then alpine desert generally above 2400 m. The Columbia 
Icefield is a plateau ice cap 320 km? in area and 200 to 300 m 
thick, the largest remaining ice mass in the Rocky 
Mountains. Valley glaciers radiate up to 10 km out from it 
today. Ice thickness and extent were much greater during 
the major glaciations, when the glaciers extended 100 km 
or more from the icecap, with only the mountain peaks 
protruding as nunataks. 

The karst rocks are resistant carbonates of Cambrian 
age. The Cathedral Formation (>560 m thick) is massively 
bedded, very resistant crystalline limestone that contains the 
cave. Above it, the Stephen Formation (80 m) is a limestone 
shale that can block much descending groundwater but leaks 
readily through some major fractures (7.¢., it is an aquitard). 
It is overlain by further thick-bedded limestones and dolo- 
stones, the Eldon and Pika Formations. The summit strata 

78 Castleguard Cave, Canada 


3000 m 


2500 m 




FIGURE 1 Schematic section through Castleguard Mountain, Alberta, Canada, showing the geological formations, the location of Castleguard Cave, 

Castleguard II, and the Meadows. 

are mechanically weaker shales, sandstones, and dolostones. 
Beneath the Icefield, around the cave and north of it, these 
rocks dip regularly south-southeast at 4 to 6°. South of the 
cave and parallel to it there is a sharp downfold in the 
Cathedral rocks that caused some slippage of bedding planes 
(thrusting) to the north. A valley is excavated along the 
downfold, with a glacier from the Icefield at its head and the 
Castleguard River starting at the glacier snout. Castleguard 
karst groundwater drainage reaches the River via 60 or more 

On the surface, the Cathedral limestones host a suite 
of small but typical alpine karst landforms such as karren, 
solution and suffosion dolines, and vertical shafts. They are 
particularly well seen in the Meadows, a broad, shallow valley 
north of the cave mouth (Fig. 1). Many of these features were 
overridden and lightly eroded by glaciers during a minor 
readvance—the “Little Ice Age,” which occurred during the 
past 500 years. The glaciers are now receding. Meltwater 
streams sink underground around their edges or in the 
Meadows. At places, streams can be heard cascading down 
shafts still concealed beneath the flowing ice. 


Castleguard Cave is a textbook example of a meteoric water 
dissolutional cave in limestone. Cavers can enter it only at 
its downstream end at 2010 m asl in the north wall of 
Castleguard River valley, more than 300 m above the valley 
floor. From there, the cave ascends 386 m to terminations 
underneath the Icefield, where explorers are farther from 
their only entrance (and exit) than in any other known cave. 
There are three distinct morphologic sections. The Headward 

Complex is comprised of inlet passages beneath the modern 
Icefield that were created by repeated glacial blockage and 
rerouting of sinking waters in the past. The passages are 
plugged by glacier ice or debris today. Younger vadose shafts 
pass down through them and become blocked by constric- 
tions or debris below. The Downstream or Entrance Complex 
includes low tunnels in two major bedding planes and 
created by flooding and obstructions by glacier ice in 
Castleguard Valley in the past. Finally, the Central Cave is a 
sequence of remarkably long, straight conduits created where 
one master bedding plane is intersected by a pair of vertical 
joints that are linked by a sedimentary dike (Grottoes Dike) 
crossing them (Fig. 2A); in the bedding plane there is some 
evidence of crushing and shearing, indicating that differential 
slip opened it up a little, permitting groundwater to 
penetrate at its juncture with the joints and dike. 

The cave possibly originated as a single phreatic loop 
beneath the Stephen impermeable cover rocks that descended 
more than 370m below a paleowatertable and then re- 
ascended to ancient springs just below the Meadows. More 
certainly, as Castleguard Valley was entrenched below the 
Stephen Formation, the cave became enlarged to nearly its 
modern dimensions, as shown in Fig. 2B. It is then com- 
prised of two shallow, principal loops with vadose canyon 
entrenchments up to 20m deep at their upstream ends, 
grading downstream into phreatic tubes 4 to 5 m in diameter 
and of beautiful circularity. The downstream loop discharged 
into Helictite Passage in the Entrance Complex by a vertical 
lift (phreatic shaft) of 24 m. Following further entrenchment 
of the Valley, the main cave headwaters were diverted into 
a lower cave (Castleguard II) and residual waters drained 
through constricted undercapture passages in the bottoms 

> drained phreatic passage 
"xy vadose entrenchment 

Castleguard Cave, Canada 79 

FIGURE 2 (A) The initial phreatic passages in the Central Cave and Downstream Complex showing the master bedding plane and intersecting vertical 

fractures that guided them. They discharged into the Entrance Complex bedding planes via P24, a vertical shaft up a small fault. (B) The cave at the close of 
the principal enlargement phase. Drawdown vadose canyons supplied water to a succession of shallow phreatic loops. (C) The modern cave; small invasion 

vadose streams have cut shafts and underfit trenches in the drained galleries and are lost into impenetrably small, undercapture passages continuing on down 

into Castleguard II. 

of the loops (Fig. 2C). The undercaptures channel local 
invasion waters passing through the leaky Stephen rocks 
today, but the Central Cave and Headward Complex are 
essentially abandoned hydrologic relicts. The Downstream 
Complex, however, can be flooded with terrifying rapidity 
when waters pour out of another, quite independent lifting 
shaft within it (Boon’s Blunder) which fills the first 
kilometers of low passages entirely and discharges the waters 
through the explorers’ entrance. 

Modern Hydrology 

Modern waters drain underground from the glacier soles, 
alpine karst, and Meadows to a set of springs extending 3 km 
downstream from the Big Springs, which are a trio of drama- 
tic overflows 15 to 40 m above the valley floor (Fig. 1). The 
waters flow through the putative series of inaccessible caves 
(Castleguard II), as illustrated in Fig. 3. Artesian Spring, the 
lowest in elevation, is perennial. As the annual melt season 
progresses, upstream springs such as Gravel and Watchman 

80 Castleguard Cave, Canada 

Pot Karst Underflow 

—-» Waterflow 
---» Airflow 

FIGURE 3 The proportional flow model of the inaccessible Castleguard II cave system, its feeder invasion vadose shafts, and discharge springs that Smart 
(1983) derived from quantitative hydrological studies. Cross-sectional areas of individual passages are proportional to their share of the total groundwater flow 

measured in the system, and their complicated pattern of interconnections is deduced from dye tracing and flood overflow behavior. 

begin to flow. The Big Springs, 100 m higher than Artesian 
Spring, have a maximum discharge >7 m’s! and handle 
average summer floods. Their capacity is exceeded when 
there is very strong melting at the head of the Meadows and/ 
or on the icefield. Groundwater then backs up in the aquifer 
until it floods Boon’s Blunder and the Downstream Complex 
of the cave, 270 m above Big Springs. Cave discharges can be 
>5 m°s1. Tracer dye placed in glacier edge sinkholes can 
reach the Big Springs 4+ km distant and 750 m lower in as 
little as 3.5 hours. This is again a textbook example—here, of 
a dynamic alpine karst aquifer. The large number and great 
height range of its springs are attributed to repeated 
disruptions such as debris pluggings during the glaciations. 

Cave Sediments, Speleothems, and Dating 

Subglacial boulders, gravel, and sand were swept or bulldozed 
into the head of the cave. Many have been partially cemented 
by calcite and later eroded, indicating a long history of filling 
and removal. Throughout the Central Cave are remnants of 
three partial fillings with varved silts and clays, separated by 
phases of erosion or calcite deposition. They are deposits of 
glacier flour, settled out of suspension on occasions when the 
cave became backflooded with subglacial waters because the 
Castleguard Valley was filled with flowing ice. 

Despite its location beneath glaciers or alpine desert, the cave 
beyond the modern entrance flood zone is well decorated 
with speleothems, chiefly very pure, white calcite. There has 
been much speculation about mechanisms for its deposition 
in the absence of any sources of soil CO, overhead. There is 
one double layer of cave pearls that are all edge-rounded 
cubes 5 to 7 mm in diameter, a unique deposit. In the warmer 
central sector (temperature of +2 to +3°C, relative humidity of 
295%), there are small evaporative aureoles of aragonite, 
huntite, hydromagnesite, gypsum, mirabilite, and epsomite; 
evaporation in these extreme conditions is due to strong 
drafts blowing through the cave when it is not in flood. 

Although most speleothems appear to postdate the latest 
phase of varve deposition, there are many remains of older 
stalactites and stalagmites that suffered erosion during one or 
more varve floods. Uranium series dating and paleomagnetic 
studies of some of the oldest examples show that the cave 
became relict (z.¢., Castleguard II was already well developed) 
more than 780,000 years ago. The varved clays are younger. 
This antiquity is typical of multilevel alpine caves. 


Castleguard Cave is a chilly place; however, because it passes 
through a big mountain (Fig. 1), the geothermal heat flux is 

able to warm the central, most insulated parts around the 
Grottoes to approximately 3°C. In winter, this is much warmer 
than outside temperatures. The cave then functions as a 
chimney, with the warmer central cave air flowing upstream 
into the sole of the Icefield with dense, cold air pouring in 
through the explorers’ entrance to replace it. The cold, dry air 
freezes residual pools of summer flood water in the Entrance 
Complex, giving cavers a 300-m belly crawl over dusty ice. 
The first liquid water is encountered only 1 km inside the cave. 
There is a subsidiary daily effect because the draft is strongest 
and coldest immediately before dawn, not a good time to be 
coming out through the ice crawls. In summer, the dynamic 
situation is reversed as cold, moist air from the interior flows 
down and out through the explorers’ entrance into the warm 
exterior. Thick hoarfrost is deposited onto the chilled entrance 
walls until the first floods of spring arrive to remove it. 


Harsh as conditions are, the cave has a small population of 
animals that prefer it or are entirely dependent on its 
protection. The first, troglophiles, are packrats nesting above 
the floodwater limits in the Entrance zone. They have their 
own private entrances from the Meadows overhead that 
are too small for human cavers. Two species of blind and 
unpigmented (troglobitic or fully adapted) crustaceans live in 
pools in the Central Cave, where the water apparently never 
freezes. An isopod, Salmasellus steganothrix, is known else- 
where in the Canadian Rockies. An amphipod, Stygobromus 
canadensis, is known only in Castleguard. It is possible that 
the cave served as a subglacial refuge for these species during 
the last glaciation or longer. A packrat nest and other organic 
material were found in the Headward Complex in 1983 and 
carbon-dated to ~9000 years BP, indicating that in early 
postglacial times the Icefield had receded from at least part of 
it, thus providing some food for the troglobites downstream. 


Castleguard Cave has proved to be a fine laboratory for 
the study of dissolutional cavern genesis and of groundwater 
flow under glacier cover. It appears that most passages in the 
Cave that are large enough for humans have now been found 
and mapped, except perhaps in the Headward area. The 
Boon’s Blunder flooded shaft has been dived to —10 m, where 
it enters a large, waterfilled tube ascending gently up the 
stratal dip. The underlying, very dynamic complex of 
Castleguard I that drains much of the Columbia Icefield 
today continues to defy all attempts at physical exploration. 
Under the Meadows and around the summit massif of the 
mountain more than 100 mapped sinkholes plus unmea- 
sured subglacial streams feed water into Castleguard III, 
another almost entirely unknown system that may contain 
both extensive fossil and active galleries. There is great 
potential here for future generations of speleologists. 

Cave, Definition of 81 


Ford, D.C., Ed. (1983) Castleguard Cave and Karst, Columbia Icefields 
Area, Rocky Mountains of Canada: A Symposium. Arctic Alpine Res. 
15(4): 425-554. 

Ford, D.C, Lauritzen, S.-E., and Worthington, S. (2000) Speleogenesis of 
Castleguard Cave, Rocky Mountains, Alberta, Canada. In Speleogenesis: 
Evolution of Karst Aquifers (A. Klimchouk, D.C. Ford, A.N. Palmer, and 
W. Dreybrodt, Eds.). National Speleological Society of America, 
Huntsville, AL, pp. 332-337. 

Muir, R.D. and D.C. Ford (1985) Castleguard. Parks Canada Centennial 
Publication, Dept. of Supply and Services, Gatineau, Quebec, 242 pp. 

Smart C.C. (1983) The Hydrology of the Castleguard Karst, Columbia 
Icefields, Alberta, Canada. Arctic Alpine Res. 15(4): 471-486. 

Cave, Definition of 

William B. White 

The Pennsylvania State University 

David C. Culver 

American University 


Human kind has been involved with caves for millennia. 
Caves were places of shelter for early humans in many parts 
of the world. They have served as places of worship for many 
societies in many times and places. Caves have been used as 
storehouses, as munitions factories, and as resting places for 
the dead. Caves play a prominent role in the myths and 
legends of many cultures throughout recorded history. Caves 
are secret places. Small children make “caves” by draping 
blankets over furniture. In contemporary society, caves 
frequently appear in movies and in cartoons. Show caves 
continue to draw thousands of visitors each year. Caves are 
also of great interest to scientists and explorers. Because caves 
are voids in rock, they are considered geological features, and 
indeed many textbooks (¢.g, Ford and Williams, 1989) 
firmly defend this point of view. However, caves are more 
than their geology because of their interaction with people 
and with organisms. One textbook (White, 1988) recognizes 
their human appeal by defining caves as “a natural opening in 
the Earth, large enough to admit a human being, and which 
some human beings choose to call a cave.” 


Caves are enticing, awaking an interest in many to see “where 
it goes.” Although the first cave explorers are lost in the mists 
of history, cave exploration as a specialized human activity 
dates from the middle of the 19th century (Shaw, 1992). 
Earlier authors include discussions of caves in general 

82 Cave, Definition of 

travelogues and descriptions of regions. The most famous of 
these in Valvasor’s 1656 book, The Glories of the Duchy 
of Carniola (an area that is part of present-day Slovenia). In 
North America, Horace C. Hovey, Luella Owen, and a few 
others wrote popular accounts of their expeditions. The 
first modern speleologist is said to be Adolf Schmidl, who 
explored many Slovenian caves in the mid-19th century. 
Somewhat later came Edouard A. Martel in France who is 
considered the father of cave exploration. Unlike his 
American counterparts, Martel organized serious caving 
expeditions into the large caves and deep pits of the Pyrenees, 
the Alps, and many other places in Europe, an activity that 
also required inventing the technology of cave exploration as 
he went along. 

Cave exploration by organized caving societies was well 
underway in Europe in the early 20th century. Organized 
exploration came later in the United States when several 
caving groups formed the National Speleological Society in 
1941. During the past 50 years, cave exploration has blossomed 
into a recreational activity for thousands of individuals 
throughout the world and is organized into hundreds of 
caving clubs, national societies, and specialized scientific 
organizations. Most cavers take their explorations seriously 
and spend substantial time in preparing maps and writing 
detailed reports. Exploration and surveying of caves are 
among the few remaining activities where useful contribu- 
tions to knowledge can be made by nonprofessionals. 

Much of the allure of caves resides in their remoteness 
and wilderness character even beneath urban sprawl. This is, 
in part, because the underground landscape with its total 
darkness and unusual shapes of rock and mineral deposits is 
so alien compared with the familiar surface landscape. Caves 
are remote in the sense of the time and effort required to 
explore them. The farthest reaches of a large cave system may 
be only a few kilometers from the entrance as the crow flies 
and may be no more than 10 or 15 km as the caver crawls. 
And, of course, the outside world is only a few tens or 
hundreds of meters away, vertically, through solid rock. 
However, reaching the farthest corners of a large cave system, 
doing a bit of exploring and surveying, and returning to the 
entrance may require 24 to 36 hours. Or, it may require 
several days and an underground camp. In the same time one 
could have traveled comfortably across the continent on a jet 
plane, attended a conference on the opposite coast, returned 
home, and been clean and well-rested to boot. The remote- 
ness of the cave arises not from distance but from the time 
needed to traverse it, from the obstacles that must be 
overcome, and from the sense of the strange and the bizarre 
in the cave landscape, a landscape duplicated nowhere on 
the Earth’s surface. Caving, from this point of view, is truly 
an esthetic or wilderness experience. It requires solitude, a 
leisurely pace, and a sense of absorption into the environ- 
ment. The emergence of the caver into the misty air under 
the bright stars of a summer's night is indeed the return from 
another world. 


The cave environment may be described as dark, wet, neutral 
to mildly alkaline, and oxidizing. The variations of these 
parameters are much less than the variations of similar 
parameters on the Earth’s surface. Chemical reactions under 
these very precisely controlled conditions permit the growth 
of unusual minerals and the growth of crystals of exceptional 
size. Mineral deposits take on the form of stalactites, 
stalagmites, flowstone, and other forms known collectively as 
speleothems. Because these deposits are nourished by water 
seeping down from the surface, changes in the climate and 
vegetation on the surface leave their signatures in the growth 
bands of the speleothems. The deposits of caves have become 
an important source of paleoclimatic information. 

Because caves are void spaces, they tend to fill up with 
various materials collapsed or washed in from the surface. 
Debris piles near cave entrances preserve archeological and 
paleontological deposits. Stream-borne deposits of clays, silts, 
sands, and gravels record past flood conditions. Because the 
filling of caves takes place slowly over very long periods of 
time and because still older deposits are preserved in the high 
abandoned levels of cave systems, cave deposits are a history 
book for the ice ages. 

Caves, of course, do not form in isolation. Every cave is 
related to a drainage system that now or at some time in the 
past carried water on its way from an inlet point to an outlet 
at a big spring. The underground pathways in limestones 
through which water moves from sinkholes and sinking 
streams to the outlets at springs are known as conduits. 
Conduit systems are very long but much of them is invisible. 
Conduits may be water filled or may have no humanly 
accessible entrances. Another definition of caves is that they 
are those conduit fragments that are accessible to human 
exploration. Careful inspection of the size, shape, and 
patterns of caves as well as details of solutional sculpturing on 
the cave walls provides much information on the present or 
past behavior of the groundwater flow system. 

Much of the value of caves to the geological sciences is that 
they preserve records over long periods of time. Streams erode 
and deepen surface valleys and in so doing destroy the stream 
channel, flood plain, and valley shape that was there before. 
In contrast, caves deepen by forming new passages at lower 
levels, leaving the old levels high, dry, and well preserved. 


We can also look for a definition of caves by considering what 
animals consider caves to be dwelling places. As habitats, 
caves have several distinct environmental properties. In 
temperate zones in the summer, they tend to be cooler; 
conversely, they tend to be warmer in the winter. This 
characteristic of temperature buffering is shared not only by 

what we would call true caves by any definition, but also by 
rock overhangs and shelters as well. Some of the fauna of true 
cave entrances such as phoebes and swallows are found in 
rock overhangs as well. If we define caves as an area of 
temperature buffering, then not only will rock overhangs be 
included but so will many manmade structures such as cellars 
and the underside of bridges. 

A more appropriate environmental parameter to consider 
is darkness. Species isolated in the darkness of caves evolve 
a characteristic morphology, including loss of eyes and 
pigment. The presence of eyeless, depigmented species in a 
habitat would be one way to define a cave from an organism's 
point of view. What would be the characteristics of such a 
definition? First, the cavity would have to be large enough to 
have a zone of darkness. This would eliminate many short 
cavities. That is, the length of the habitat must be great 
enough relative to the diameter of the opening so that 
sunlight does not penetrate to the far reaches of the habitat. 
In practical terms, this length ranges from a few meters to 
hundreds of meters. This definition would exclude many 
open-air pits, which are formed the same way caves are—by 
dissolution of carbonate through the action of carbonic and 
sulfuric acid. 

The second characteristic of such a habitat would be that 
it had been around long enough for animals either to evolve 
in situ or for animals to migrate into the habitat. It seems 
likely that most caves and karst areas are old enough to have 
an eyeless, depigmented fauna. Some of the youngest caves 
known are on carbonate islands such as San Salvador Island 
in the Bahamas. These caves, less than 125,000 years old, 
have an eyeless, depigmented cave fauna. In glaciated regions, 
the caves may be quite old but have only been ice free for 
perhaps 10,000 years. In this case, there are few if any 
terrestrial species that are either depigmented or eyeless. 
There are typically eyeless, depigmented aquatic species. These 
may have survived underneath the ice in unfrozen freshwater 
or have colonized from unglaciated areas. In any case, 10,000 
years is almost certainly not long enough for species to evolve 
eyelessness in situ, a process that probably takes between 
100,000 and 1,000,000 years (Culver et al., 1995). Caves in 
evaporates (gypsum) and lava form much more quickly (and 
disappear more quickly). There are few reports of eyeless, 
depigmented species in gypsum caves, but lava tubes often 
have a rich fauna. The reasons for this are complex, but one 
of the factors is that the lava fields themselves are consid- 
erably older than a lava tube. Species found in lava tubes only 
a few thousand years old almost certainly migrated there 
from other lava tubes and cavities in the lava. 

A third characteristic of such a habitat is that, if darkness 
combined with the presence of eyeless species is how it is 
delineated, then it will include a wide variety of aquatic 
habitats in darkness that have no connection with our 
intuitive idea of a cave. These include the underflow of rivers, 
marine and freshwater beaches, and any subsurface water. 
Many of these habitats, collectively termed interstitial or 

Cave, Definition of 83 

alternatively permeable and small cavities (as opposed to 
permeable large cavities of caves), have a rich subterranean 
fauna of eyeless, depigmented species. At least in principle, 
we can recognize the differences between species in inter- 
stitial habitats and species in caves. While they share a lack of 
pigment and eyes, interstitial species tend to have shorter 
appendages, smaller body size, and a more worm-like appear- 
ance than do cave species (Coineau, 1999). In practice, it is 
difficult to distinguish species from many taxonomic groups. 
For example, snails occur in both habitats but there is little 
to distinguish the two groups morphologically. In areas with 
carbonate rocks, we can also distinguish the two by the size 
of the cavity. The simple, nonscientific definition of caves as 
natural cavities in a rock that can be entered by people does 
not help in this case. Cave animals can obviously thrive in 
cavities that are too small to be entered by people. Cavity 
diameters exist in a continuum, with a breakpoint between 
laminar and turbulent flow cavities of approximately 1 cm. 
The transition between laminar and turbulent flow is also an 
important biological transition, and one that in a general 
sense separates interstitial and cave habitats. 

Thus, we can define a cave from a biological point of view 
as a cavity, at least part of which is in constant darkness, with 
turbulent water flow and with eyeless, depigmented species 


When we define caves in terms of human access rather than 
in terms of geologic setting, caves can be found in many 
different geologic settings and have been formed by many 
kinds of processes (Fig. 1). Caves can be formed by purely 
mechanical processes. Bulk masses of rock can be fractured 
and shifted by tectonic processes such as faults. Likewise, 
rock masses can break along fractures and then pull apart by 
the rock masses sliding under the influence of gravity. In 
either case, void spaces formed between the rock masses are 
called tectonic caves. Tectonic caves tend to form in hard, 
massive, brittle rocks such as sandstones. They tend to be 
small, typically a few tens of meters, although some reach 
lengths of hundreds of meters. 

Boulder piles, if the boulders are sufficiently large to allow 
humans to explore the pore spaces between the boulders, 
can form caves. These openings are called talus caves. Because 
talus caves are simply the interstices between a pile of boulders 
that may or may not be completely roofed over, talus caves 
really do not have a definable length. Some talus caves in 
the Adirondack Mountains have been claimed to be several 
kilometers in length, although this requires a fairly generous 
definition of a cave. Rocks may be differentially eroded to 
produce cave-sized openings. The categories of erosional 
caves are defined by the erosion process. 

Sea caves are formed by wave action on sea cliffs. Fractures 
in the rock produce zones of weakness that focus the attack 
of the waves. Sea caves are formed in many kinds of resistant 

84 Cave, Definition of 



FIGURE 1 Types of caves. 

rocks in many parts of the world. Fingal’s Cave on the Isle 
of Staffa in the Scottish Hebrides (made famous by 
Mendelssohn in his Hebrides Overture) is in columnar basalt. 
Sea caves form in granite on the coast of Maine. Many sea 
caves are found on Santa Cruz and Anacapa Islands off the 
coast of southern California (Bunnell, 1988, 1993). Sea caves 
extend for distances from a few meters to a few hundred 
meters into the sea cliff. Frequently, the caves consist of an 
inner chamber that is much larger than the connection 
passage to the open sea. Access to a sea cave may be possible 
at low tide, while the cave may be flooded at high tide. Some 
sea caves have blow holes in the back that spurt water when 
tides or storm surges rush into the cave. 

Aeolian caves are formed by wind action. These are common 
in the arid regions of the American Southwest. Wind-blown 
sand scours sandstone cliffs and sculpts out chambers a few 
meters to a few tens of meters in diameter. A typical aeolian 
cave is a bowl-shaped chamber carved in solid rock. Ceiling 
heights vary from 1 to 2 meters. Often the entrance is a 
smoothly carved hole in the cliff much smaller than the 
chamber inside. 

Rock shelters are borderline caves. In places where a 
resistant bed of rock, typically sandstone, overlies weaker 

beds of rock, typically shales, the weaker rock can be eroded 
away, leaving the resistant rock beds to form a roof. Rock 
shelters are usually small, a few meters to a few tens of meters 
in depth although they can be tens to hundreds of meters 
wide. As caves, rock shelters usually do not extend to total dark- 
ness; however, such shelters were habitats for early humans. 
They are often rich repositories of archeological material and 
sometimes are referred to as caves by archeologists. 

Fine-grained, poorly consolidated sediment can be swept 
away by stormwaters. In such locations as Badlands National 
Monument in western South Dakota, sediments flushed by 
stormwaters produce small caves in the loosely consolidated 
silts and clays. These are known as suffosional caves. 

Of great interest to cave explorers are lava tubes and other 
volcanic caves. Lava tubes form on the sides of volcanoes 
where streams of lava freeze over, drain, and leave behind, 
buried in the lava flow, open tubes with diameters of meters 
to tens of meters and lengths up to many kilometers. The 
processes that form lava tubes are entirely different from the 
processes that form solution caves in limestone and other 
rock but they are caves in the same sense that other natural 
openings of human size are caves. Lava tubes are also habitat 
for organisms. Volcanic areas that have discharged fluid 

basaltic lavas usually have systems of lava tubes—;for 
example, Hawaii, southern Oregon, and northern California. 
Lava Beds National Monument in northern California 
provides easily accessed examples (Waters et al., 1990). 

When glaciers melt, the meltwater drains down into the 
glacier through fissures. This water moves along the base of 
the glacier and, because it is slightly warmer than the freezing 
point of water, it gradually carves out long tunnels that open 
at the front of the glacier forming the rivers that drain the 
glaciers. Glacier caves are ice tunnels with floors of rock and 
walls and ceilings of ice. When the surface of the glacier is 
below freezing, the tunnels drain and become open to explo- 
ration. When the glacier is melting, the tunnels are often 
filled with water. 

Most important are the solution caves. These form by 
chemical dissolution of the bedrock by circulating ground- 
water. They come in a great variety depending on the type 
of rock and the source and chemistry of the water that does 
the dissolving. Most solution caves are formed in limestone, 
a rock consisting largely of calcium carbonate, or dolomite, a 
rock consisting largely of calcium magnesium carbonate. 
These would be called limestone caves or dolomite caves. In 
arid regions, the more soluble rock, gypsum, calcium sulfate, 
is exposed at the land surface, and one finds gypsum caves, 
for example, in west Texas, western Oklahoma, and New 
Mexico, where gypsum rock occurs at the land surface. Caves 
easily form in salt and a few other highly soluble materials, 
but such caves are rare because salt does not survive at the 
land surface except in a few extremely arid regions. 

Figure 1 shows the variety of sources for the water respons- 
ible for the development of caves. Most of the caves are 
dissolved by the movement of ground water in contemporary 
drainage basins. In coastal regions, the mixing of fresh 
groundwater with saltwater produces an aggressive solution 
that can dissolve out caves. Some caves (for example, the 
large caves of the Black Hills of South Dakota) are formed 
from hot water rising up from deep within the rock. Carlsbad 
Caverns in New Mexico and other caves of the Guadalupe 
Mountains have been formed by sulfuric acid derived from 
the oxidation of hydrogen sulfide migrating upward from the 
oil fields to the east. 

In summary, caves form in a great variety of rocks by a 
great variety of geological and chemical processes. Each has 
its importance to geology. However, the common theme that 
binds this diverse collection of cavities together is their 
interest to human explorers and their use as habitat by cave- 
adapted organisms. 


Bunnell, D.E. (1988) Sea Caves of Santa Cruz Island. McNally and Loftin, 
Santa Barbara, CA, 123 pp. 

Bunnell, D.E. (1993) Sea Caves of Anacapa Island. McNally and Loftin, 
Santa Barbara, CA, 207 pp. 

Coineau, N. (2000) Adaptations to interstitial groundwater life. In 
Subterranean Ecosystems (H. Wilkins, D.C. Culver, and W.F. Humphreys, 
Eds.). Elsevier Press, Amsterdam, pp. 189-210. 

Cave Dwellers in the Middle East 85 

Culver, D.C., T.C. Kane, and D.W. Fong (1995) Adaptation and Natural 
Selection in Caves. Harvard University Press, Cambridge, MA, 223 pp. 

Ford, D.C. and PW. Williams (1989) Karst Geomorphology and Hydrology. 
Unwin Hyman, London, 601 pp. 

Shaw, T.R. (1992) History of Cave Science. Sydney Speleological Society, 
Sydney, Australia, 338 pp. 

Waters, A.C., J.M. Donnelly-Nolan, and B.W. Rogers (1990) Selected caves 
and lava-tube systems in and near Lava Beds National Monument, 
California. U.S. Geol. Survey Bull. 1673, 102 pp. 

White, W.B. (1988) Geomorphology and Hydrology of Karst Terrains. Oxford 
University Press, New York, 464 pp. 

Cave Dwellers in the Middle 

Paul Goldberg 

Boston University 

Ofer Bar-Yosef 
Harvard University 

his article offers information about prehistoric caves in 

the Middle East. Caves are a source of fascination for 
the public as well as many scientists. In this short entry 
we will attempt to briefly describe and summarize some of 
the discoveries in Middle Eastern caves and what those 
discoveries reveal about human behavior. 


Caves figure strongly in the archeological record, through 
most archeological time periods and across most continents. 
Caves not only served as loci of habitation where daily 
activities such as food processing and sleeping took place but 
also functioned as gathering places for spiritual and religious 

Prehistoric caves—particularly those in the Middle East— 
are special sedimentary environments (Fig. 1). In essence, 
because of the overall low energies of the depositional 
processes, caves serve as excellent sedimentary traps: What- 
ever is brought into the cave tends to stay there. As a result, 
caves can preserve faithful records of past environments, as 
well as past human activities. Environmental information 
is conveyed by the presence of macro- and microfaunal 
remains—particularly the latter. Plant remains, an additional 
environmental indicator, are scarce, although microbotanical 
remains such as pollen and phytoliths can be found. Phyto- 
liths tend to be better preserved and abundant and provide 
insights not only about past local and regional conditions but 
also about human activities, such as the gathering of plants 

for fuel, bedding, and food. 

86 Cave Dwellers in the Middle East 


_ aden 
@ Aleppo 



Jerf Ajla 
Narh Ibrahim 

Ksar Akil of: ®Yabrud 


Mediterranean = 


eTor Sabiha 
®. \. 
Tor Faraj, 

Tabun Skhul 

® Sites 
® Cities/towns 

I 1KM 
0 100 

O.B-Y & I. K. 

FIGURE 1 Map of Middle East and the location of major prehistoric cave sites in the area. Only a few caves are discussed here. 

In addition, anthropogenic deposits also reveal the nature 
of past human activities at a cave by incorporating animal 
bones and stone artifacts. These objects inform us about 
hunting and butchering practices, the manufacture and use 
of curated stone tools, and the use and function of space (¢.¢., 
working areas, fireplaces, dumping, and sleeping zones). 

The above statements are pertinent to many cave sites 
throughout the world. Middle Eastern caves, however, 
benefit from their location at the crossroads to human past 
and present migrations and thus offer the opportunity to 
monitor and document important phases in human 
evolution, particularly with regard to the origin of modern 
humans and the demise of Neanderthals. 


Most of the caves in the Middle East are situated either in 
the Mediterranean climatic zone (Fig. 1), particularly in a 
belt close to the present Mediterranean coastline, or in the 
drier, steppic and desertic region. The Mediterranean 
climate, which includes distinct climatic gradients from west 
to east and from north to south is characterized by warm, dry 
summers and cool, wet winters in which 500 to more than 
1000 mm of precipitation can fall. Similarly, the transition 
from steppe to desert is marked by reductions in rainfall from 
500 to less than 100 mm. 

Most of the caves outlined in this section developed in 
limestone and formed during the late Tertiary/Quaternary 
period under phreatic conditions and commonly as enlarge- 
ments along joints. Vadose expansion is more recent, 
resulting in a vaulted or domed morphology with vertical 
chimneys that extend to the surface above several of the 
caves. Many of these same caves have depressions or 
sinkholes that are situated beneath the chimneys. In any case, 
modern-day karstic activity is generally negligible, and the 
remains of formerly more extensive and well-developed 
dripstone features signify karstic conditions that were much 
different from what occurs today. 


Depositional and post-depositional processes operating in 
karstic caves are relatively well known but less so in pre- 
historic caves; however, recent research, particularly in 
Middle Eastern prehistoric caves, has begun to reveal a high 
degree of marked complexity. Material of geological origin 
may accumulate in the interior in a number of ways, 
including: (1) gravity-derived rockfall from the walls and 
roof of the cave; (2) aqueous processes associated with fluvial 
and phreatic deposition, or runoff; (3) colluviation of soils 
derived from the surfaces above and outside the cave; and (4) 
aeolian deposition of sand and silt. 

Middle Eastern caves contain distinct biogenic contribu- 
tions, such as bird and bat guano. In addition, coprolites and 

Cave Dwellers in the Middle East 87 

organic matter remains produced by carnivores—particularly 
hyenas—are quite common. In many caves, the presence of 
millimeter-sized rounded stones indicates distinct gastrolith 
input by pigeons. Plant remains, such as grass and wood, can 
be washed and blown into the cave, or alternatively brought 
into the cave by past human occupants for fuel, bedding, 
or shelter. 

Anthropogenic contributions to cave sedimentation in 
this region are noteworthy and tend to be unnoticed in many 
prehistoric cave settings throughout the world. These 
accumulations consist of primarily bone and shell remains, as 
well as the buildup of ashes, organic matter, and charcoal that 
are associated with fireplaces and burning activities. 
Additional traces of fine-grained soil and sediment that were 
tracked into the cave by humans and other animals are subtle 
but important. 

Once within the cave itself, deposited material is 
commonly modified by a number of processes that are often 
penecontemporaneous with deposition. Deposits are some- 
times modified by wind, runoff, and dripping water or 
subjected to burrowing by animals or to trampling by 
humans and other occupants. Mineral and organic residues 
are sometimes moved by human or animal activity from their 
original location within the cave. Bone and lithic discard 
evidence of activities and indications of removal, dumping, 
and trampling of material associated with the cleaning or 
maintenance of hearths are not uncommon. 

Prehistoric caves in this region are damp and act as 
sinks for water, organic matter, and guano and accordingly 
tend to act as chemical engines whereby numerous mineral 
transformations take place. These secondary alterations 
commonly include the precipitation (and dissolution) of 
carbonates that form speleothems, layered travertines, or the 
so-called cave breccia, which represent calcite-cemented 
clastic sediments typically consisting of inwashed soil 
material. Another characteristic type of mineral alteration 
involves the formation of several types of phosphate 
minerals, along with the formation of opal and the 
breakdown of clays. 

Palaeoenvironmental and palaeoclimatic changes expressed 
in the faunal records, as well as by depositional and post- 
depositional accumulations and removals, are registered in a 
rather coarse chronology. Speleothems in Nahal Soreq Cave 
and a cave near Jerusalem provided detailed records of 
palaeoclimatic fluctuations of the last 170 thousand years. 
These karstic caves were rarely penetrated by humans, and 
these oscillations follow the pattern known from other 
localities of the northern hemisphere. 


Prehistoric caves in the region produced the basic 
archaeological sequence of at least the past 400 thousand 

88 Cave Dwellers in the Middle East 

years (Fig. 1); several examples are illustrated in the following 
text, with comments related to the themes discussed above. 
The oldest known cave deposits were exposed in Umm 
Qatafa, Tabun, (Fig. 2a,b), and the Yabrud IV rockshelter, 
where the lowermost layers—more than 400 thousand years 
old—contained a core-and-flake industry, a phenomenon 
that intersects the Acheulian sequence. The Late (or Upper) 
Acheulian is known from several caves. This industry is 
characterized by handaxes or bifaces, but these tools are not 
quantitatively the most dominant tool type. The Late (or 
Upper) Acheulian is followed by the Acheulo—Yabrudian, 
where the combination of bifaces and scrapers, often shaped 

FIGURE 2a ‘Tabun Cave, Mount Carmel. Shown in this photograph (taken 
during the 1969 season) are bedded sandy and silty sediments that were 
blown into the cave from the adjacent coastal plain. These contain Lower 
and Middle Palaeolithic industries, including the Upper Acheulian (UA), 
Acheulo-Yabrudian (AY), and Mousterian (M). Note the strong dip of the 
lowermost sediments that plunge into a subsurface swallow hole. This 

subsidence took place during the earlier phases of deposition, as the bulk of 
the sediments from the middle part of the photograph upwards are roughly 
horizontal. The uppermost part of the photograph (see Fig. 2b) is composed 
of interbedded clay and ashy deposits, punctuated with numerous blocks of 
roof fall. The dating of the deposits here is a subject of debate, but estimates 
using thermoluminescence (TL) dating on burned flints suggest that the 
deposits down to the top of the large hole on the left are close to 350 
thousand years old (Mercier et al., 1995). 


FIGURE 2b Tabun Cave, Mount Carmel. This photograph is from the 
upper part of the cave, above the uppermost ledge in Fig. 2a. In contrast to 
the lower deposits, these are largely anthropogenic in nature and consist of 
lighter and darker bands of ash and red clay, respectively. These anthro- 
pogenic layers are overlain by massive reddish clay that contains decimeter- 
sized blocks of roof fall. The latter attests to an enlargement and eventual 
opening of the chimney that leads up to the surface of Mount Carmel. It is 
through this chimney that the clayey serra rossa soils were washed into the 

on thick flakes, appear in Tabun, Hayonim, Qesem, 
Zuttiyeh, and the Yabrud I rockshelter, as well as in open air- 
sites in the El-Kowm basin (Northeast Syria). This entity 
occupies the Levant between the Acheulian sequence and the 
Mousterian or Middle Palaeolithic. Its geographic distribu- 
tion indicates an origin in the northern Levant. 

Cave sites with Middle Palaeolithic (250-270 thousand to 
48-46 thousand calendar years BP) remains have produced 
a wealth of evidence, as well as a large number of human 
burials and isolated human bones. The earlier deposits 
contain skeletal remains identified as archaic Modern 
humans (also known as the Skhul-Qfazeh group; Fig. 3). 
Occurring stratigraphically above these human remains are 
those of southwestern Asia Neanderthals that were found in 
Kebara (Fig. 4), Amud, Dederiyeh, and Shanidar. Based on 
the evidence, both human populations demonstrated good 
hunting skills, the use of fire, and the ability to procure raw 
materials for making stone tools from a radius of 5 to 20 km 
around the sites. Mobility between lowland and highland 
areas has been easier to trace in Lebanon, South Jordan, and 
the Zagros because of the greater topographic relief in these 

Among Upper Palaeolithic and Epi-Palaeolithic (48-46 to 
11.5 thousand calendar years BP) cave occupations, the best 
known are (1) Ksar ’Akil, with an unusual 18-m sequence of 
Upper Palaeolithic layers; (2) the few Levantine Aurignacian 
sites (e.g., Yabrud II, El-Wad, Kebara, and Ucagizili); and (3) 
those attributed to the Natufian culture that immediately 
preceded the earliest farming villages. The Natufians camped 
in caves, built rooms inside the main chambers, and buried 
their dead. Their use of caves was probably more intense than 
that of their predecessors. During the prehistoric periods, 

FIGURE 3 Qafzeh Cave, Lower Galilee. This cave is famous for its series of 
Middle Palaeolithic human burials in layers XVII to XXI. The morpho- 
logical analyses identified the remains as archaic Modern Humans and are 

therefore considered as coming out of Africa. These layers, dated by 
thermoluminescence (TL) from about 95 to 115 thousand calendar years BP 
(Valladas et al., 1988), consist of fine-grained angular rock fall that has been 
reworked by slopewash and colluvial processes. The darker band in layer XXI 
results from manganese enrichment associated with a subsurface spring that 

was operational during that time. 

2 a ‘ 
\ a rs + eu. ‘ 

FIGURE 4 Kebara Cave, Mount Carmel. Situated about 12 km south of 
Tabun, Kebara is considerably younger, dating to about 60 to 70 thousand 
calendar years BP. This view of the southeast corner of the cave shows mostly 
Upper Palaeolithic (UP) deposits in the walls on the right and left and 
Middle Palaeolithic (MP) sediments on the floor of the excavation. The 
Upper Palaeolithic deposits are comprised of finely bedded and laminated 
silt and sand-sized material (S) that has washed back into the cave via runoff. 
It commonly includes sand-sized aggregates that have been reworked from 
older sediments near the entrance. A large piece of limestone roof fall (Ls) in 
the wall to the left shows that some gravity deposition also takes place. 
Interestingly, the other portions of this block have been subjected to 
diagenesis in which the original dolomite block has been transformed into a 
number of phosphate minerals, including francolite, crandallite, 
montgomeryite, and leucophosphite (Weiner et al., 1993). Punctuating 
these geogenic deposits are isolated hearths (H) representing some 
anthropogenic material within the mostly geogenic sediments. The Middle 
Palaeolithic deposits, on the other hand, are mostly anthropogenic and 
consist of hearths and hearth products (charcoal, organic matter, and ash). 
Some of these hearths are revealed by a white area (see the base of the 


Cave Dwellers in the Middle East 89 

caves were often used for camping by entire groups and rarely 
used as stations for performing specific tasks. One of the best 
examples is the use of caves at higher altitudes for short-term 
camps by hunters. 

Since the Neolithic period (11,500 to 7500 calendar years 
BP) the use of caves changed dramatically. Several caves 
served as camps for special artisans, but in more than one 
case the caves became sacred and were utilized, as in Nahal 
Hemar and Nahal Qana caves, for storage of paraphernalia or 
other sacred objects. Larger caves could have served as loca- 
tions for ceremonies, as indicated by a few caves in Turkey. 

During the Chalcolithic and Bronze ages (7500 to 3200 
calendar years BP), caves were employed for various 
purposes. Certain karstic caves, such as Peqi’in, served as 
burial grounds, while others functioned as storage facilities, 
animal pens, and even refugia. During the Bronze Age, caves 
continued to be used in similar ways, and a unique example 
is the warrior burial in a cave near the Jordan Valley. Finally, 
a common use of most caves during the last millennium BC 
and the first two millennia AD was by shepherds who often 
spent the late fall and winter months in these protected 


Ayalon, A., M. Bar-Matthews, and A. Kaufman (2002) Climatic conditions 
during marine oxygen isotope stage in the eastern Mediterranean region 
from the isotopic composition of speleothems of Soreq Cave, Israel. 
Geology 30: 303-306. 

Bar-Matthews, M., A. Ayalon, and A. Kaufman (1997) Late Quaternary 
paleoclimate in the eastern Mediterranean region from stable isotope 
analysis of speleothems at Soreq Cave, Israel. Quater. Res. 47: 155-168. 

Bar-Matthews, M., A. Ayalon, A. Kaufman, and G.J. Wasserburg (1999) 
The eastern Mediterranean paleoclimate as a reflection of regional events: 
Soreq Cave, Israel. Earth Planetary Sci. Lett. 166: 85-95. 

Frumkin, A., D.C. Ford, and H.P. Schwarcz (1999) Continental oxygen 
isotopic record of the last 170,000 years in Jerusalem. Quater. Res. (New 
York) 51: 317-327. 

Karkanas, P., O. Bar-Yosef, P. Goldberg, and S. Weiner (2000) Diagenesis in 
prehistoric caves: the use of minerals that form in situ to assess the 
completeness of the archaeological record. /. Archaeol. Sci. 27: 915-929. 

Mercier, N., H. Valladas, G. Valladas, and J.-L. Reyss (1995) TL dates of 
burnt flints from Jelinek’s excavations at Tabun and their implications. /. 
Archaeol. Sci. 22: 495-509. 

Schiegl, S., S. Lev-Yadun, O. Bar-Yosef, A.E. Goresy, and S. Weiner (1994) 
Siliceous aggregates from prehistoric wood ash: a major component of 
sediments in Kebara and Hayonim caves (Israel). Israel J. Earth Sci. 
43: 267-278. 

Schiegl, S., P. Goldberg, O. Bar-Yosef, and S. Weiner (1996) Ash deposits 
in Hayonim and Kebara Caves, Israel: macroscopic, microscopic and 
mineralogical observations, and their archaeological implications. /. 
Archaeol. Sci. 23: 763-781. 

Valladas, H., J.L. Reyss, J.L. Joron, G. Valladas, O. Bar-Yosef, and B. 
Vandermeersch (1988) Thermoluminescence dating of Mousterian 
‘Proto-Cro-Magnon’ remains from Israel and the origin of modern man. 
Nature 331: 614-616. 

Weiner, S., P. Goldberg, and O. Bar-Yosef (1993) Bone preservation in 
Kebara Cave, Israel, using on-site Fourier transform infrared 
spectrometry. J. Archaeol. Sci. 20: 613-627. 

Weiner, S., S. Schiegl, P. Goldberg, and O. Bar-Yosef (1995) Mineral 
assemblages in Kebara and Hayonim caves, Israel: excavation strategies, 
bone preservation, and wood ash remnants. Israel J. Chem. 35: 143-154. 

90 Chemoautotrophy 


Annette Summers Engel 
The University of Texas at Austin 

N& all life on Earth depends on sunlight and photo- 
synthesis for organic carbon and cellular energy; however, 
the absence of light energy does not preclude life, as chemo- 
synthesis provides another source of sustainable energy. 
Reactive rock surfaces and mineral-rich groundwater provide 
an assortment of potential energy sources for specialized 
microorganisms that gain cellular energy from the chemical 
oxidation of inorganic compounds and convert inorganic 
carbon sources into organic carbon. While chemosynthetic 
microorganisms are found in nearly every environment on 
Earth, they are most abundant in habitats where darkness 
prevails and competition with photosynthetic organisms is 
eliminated. Significant chemosynthetic populations have 
been reported at deep-sea hydrothermal vents (Deming and 
Baross, 1993), within the deep terrestrial subsurface (Stevens 
and McKinley, 1995), and from caves (Sarbu et al, 1996). 

Caves form distinctive habitats with complete darkness, 
relatively constant air and water temperatures, and a depleted 
supply of easily degradable organic matter. Because darkness 
precludes photosynthetic activity, most cave ecosystems are 
dependent on allochthonous organic material for energy. 
Indeed, the limited organic matter found in caves is usually 
derived from dead photosynthetic material, having been carried 
into the subsurface system via air currents, speleothem 
dripwaters originating from the surface, or stream drainage 
or occurring as guano from cave-dwelling organisms. Early 
(bio)speleologists demonstrated that microorganisms often 
colonize caves but assumed their presence was only as 
secondary degraders and food sources for higher organisms 
(Caumartin, 1963). They also suggested that chemosynthetic 
activity was not evident or was limited in most cave systems; 
however, the recent discoveries of complex cave ecosystems 
and chemosynthetic microorganisms encourage modifica- 
tions to these original hypotheses. 

Current research has shown that chemosynthetically based 
cave ecosystems form as a consequence of energy-rich com- 
pounds (such as hydrogen, reduced iron, or hydrogen sulfide) 
being present in the groundwater or cave sediments. The 
microorganisms utilize the available chemical energy in the 
cave that would otherwise be lost to the system. In this 
manner, chemosynthesis is a rich alternative energy source 
for cave organisms, resulting in higher ecosystem biodiversity 
compared to most non-chemosynthetically based cave eco- 
systems that rely on inconsistent and limited inputs of 
organic carbon and as such have relatively low species 
diversity and population densities. This article focuses on 
chemosynthesis and chemosynthetically based ecosystems 

found in caves and summarizes current methodology and 
major microbial groups with respect to their evolutionary 
affiliations and metabolic pathways. The distribution of 
chemosynthetic microbial groups in several cave ecosystems 
is emphasized, and the article concludes with a discussion of 
subsurface chemosynthetically based ecosystems and suggests 
relevant aspects of future research. 


The physiological mechanisms for capturing chemical energy 
during chemosynthesis are diverse, and there are several 
descriptive qualifiers that define an organism based on its 
carbon and energy sources. Carbon for cellular growth 
originates from either (1) converting inorganic carbon (CO,, 
HCO, ) to organic carbon as an autotroph (“self-feeding”), or 
(2) from assimilation of organic carbon initially produced by 
autotrophs. The distinction between a chemosynthetic and a 
photosynthetic organism is based on whether the initial 
source of energy is from inorganic chemicals (tho) or light 
(photo). During chemosynthesis, microorganisms gain energy 
by transferring electrons from one chemical (electron donor) 
to another (electron acceptor). Typically, they use compounds 
present in groundwater or they colonize rock and mineral 
surfaces to mine essential nutrients. Microbes that gain 
energy through chemosynthesis and fix inorganic carbon are 
chemolithoautotrophs (literally, a self-feeding rock-eater). 
Chemical electron donors include, but are not limited to, 
molecular hydrogen or reduced sulfur compounds; organic 
molecules (organo), such as acetate or formate, can also be 
used. However, an organism is no longer classified as a 
chemolithoautotroph if organic compounds are used to gain 
energy or as a carbon source. Organisms that gain cellular 
energy from chemical transformations but use organic carbon 
compounds for their carbon source are chemoorganotrophs. 
Heterotrophs use organic carbon for cellular energy and 
carbon sources. Several studies have shown that chemolitho- 
autotrophs can grow if organic carbon is present as mixo- 
trophs, in which both chemolithotrophy and heterotrophy 
are expressed simultaneously. For the purposes of this article, 
strictly chemolithoautotrophic microbial processes will be 
considered from caves. 

As microorganisms are classified based on the electron 
donors they prefer for their specific metabolic activities, they 
are also classified on the basis of oxygen requirements and 
whether they respire aerobically, anaerobically, or ferment, 
all of which relates to electron acceptor utilization (Table I). 
The majority of well-characterized microorganisms require 
oxygen (aerobes), whereby oxygen serves as the terminal 
electron acceptor for metabolic processes yielding energy. 
Common aerobic reactions include sulfur oxidation, iron 
oxidation, and ammonia oxidation. In more reducing 
environments, traditionally where organic carbon is 
consumed rapidly, microbes that do not require oxygen 

Chemoautotrophy 91 

TABLEI Examples of Chemolithoautotrophic Energy Reactions and Carbon Fixation Pathways 

Energy Reaction Metabolic Process Major Genera Electron Electron Carbon 
Donor Acceptor Pathway* 


4H, + CO, — CH, + 2H,0 Methanogenesis Methanoseata, Methanococcus Hy CO, A-CoA 

2CO, + 4H, — CH;COOH + 2H,O Acetogenesis Acetobacterium H, CO, A-CoA 

4H, + SOF + 2H* > H,S + 4H,O Sulfate reduction Archeoglobus H, sop (S,0; ) A-CoA 

4H, + SOF + 2H* > H,S + 4H,O0 Sulfate reduction Desulfovibrio, Desulfobacter H, sO; A-CoA, RCA 

FeOOH + 2H* + '/2H, > Fe** + 2H,O Iron reduction Geobacter, Shewanella H, Fe**as FEOOH ~~ Calvin 

2NO, + 5H, > N, + 4H,O + 20H” Denitrification Pyrolobus, T: denitrificans Hy), H,S, $° NO, Calvin, A-CoA 


H, + 1/20, > H,0 Hydrogen oxidation Alcaligenes, Aquifex H, O, Calvin, RCA 

H,S + 20, > SOF + 2H* Sulfur oxidation Sulfolobus, Thiobacillus H,S, S° O,, NO, RCA, Calvin 

2FeS, + 70, + 2H,O — 2FeSO, + 2H,SO, Sulfur oxidation T. ferrooxidans S*, FeS, O; Calvin 

4Fe** + O) + 10H,O — 4Fe(OH); + 8H* Iron oxidation T. ferrooxidans Fe** O,, NO, Calvin 

2Mn** + O, + 2H,O > 2MnO, + 4H* Manganese oxidation — Shewenella Mn** O, Calvin 

NH; + 20, > NO, + 2H* + H,O Ammonia oxidation Nitrobacter, Nitrosomonas NH,,NO, O, Calvin 

CH, + 20, > CO, + 2H,O Methane oxidation Methanomonas CH, O, RMP 

* A-CoA, acetyl—coenzyme A pathway; RCA, reverse citric acid cycle; Calvin, Calvin—Benson cycle; RMP, Ribulose monophosphate cycle. 

(anaerobes) use a variety of alternative electron acceptors for 
respiration in a sequence of energetic, reduction reactions 
that occur along thermodynamic (and redox) gradients 
(Fig. 1). Terminal electron acceptors are reduced in the 
following order: 

O, > NO; > Mn* — Fe* > SOF > CO, 

Most anaerobic microbes are obligate anaerobes, but 
microaerophilic organisms require low concentrations of 
oxygen, while facultative organisms can grow in the presence 
or absence of oxygen, using different electron acceptors if 
necessary and if they are available. Some of the more impor- 
tant anaerobic redox reactions include nitrate or dissimi- 
latory nitrogen reduction (NO, — N,), dissimilatory sulfate 
(or sulfur) reduction (SOZ — H,S), ferric iron reduction 
(Fe** — Fe**), and methanogenesis (CO, — CHy) (Table I). 
Microorganisms involved in these anaerobic processes grow 
as chemolithoautotrophs or chemoorganotrophs. 


Today chemolithoautotrophs play important roles in global 
chemical and ecosystem processes. They serve as catalysts for 
reactions that would not otherwise occur or would proceed 
slowly over geological time. In addition to being primary 
producers in an ecosystem, chemolithoautotrophic micro- 
organisms couple the carbon cycle to other element cycles for 
their metabolism and growth. As such, they are intimately 
involved in the carbon, nitrogen, and sulfur cycles, and some 

Aerobic Pathway Anaerobic Pathway 


CO, and biomass 

FIGURE 1 Schematic representation of chemolithoautotrophic, energy- 

yielding metabolic reactions. Aerobic microorganisms, such as iron- 
oxidizing bacteria, use dissolved oxygen (O2) to generate energy. In contrast, 
anaerobes use alternative terminal electron acceptors, with carbon dioxide 
(CO)) utilization during methanogenesis (CH, production) indicating the 
most reducing conditions. CO, is required for all chemolithoautotrophic 
metabolic reactions, and H, is required by most anaerobes. Aerobic and 
anaerobic metabolism can occur simultaneously, although there is usually 
spatial and temporal separation of the reactions within a microbial habitat. 
For instance, aerobic microorganisms colonize the uppermost portions at the 
air—water interface, while anaerobes occupy inner regions of a mat or water- 
sediment interface. 

92 Chemoautotrophy 

are also associated with redox cycling of trace elements, such 
as iron, manganese, arsenic, and lead. 

Their involvement in current biogeochemical cycling 
suggests that chemolithoautotrophs have been crucial in the 
transformation of chemical energy from one reservoir to 
another during Earth’s history. The habitats of early Earth 
were virtually devoid of organic compounds, and the early 
atmosphere was extremely reducing. Just as today, however, 
early Earth would have had an assortment of electron donors, 
and various metabolic processes, albeit simple, could have 
existed. Life that first evolved would have been anaerobic and 
dependent on inorganic substrates for energy and growth, 
perfect for the evolution of chemolithoautotrophic 

Chemolithoautotrophs use a variety of carbon fixation 
pathways: the Calvin—Benson cycle (also known as the 
reductive pentose phosphate cycle), the acetyl-coenzyme A 
cycle, or the reverse citric acid cycle. Based on evolutionary 
relationships of the 16S (small-subunit) ribosomal RNA 
(rRNA) gene and the subsequent “tree of life” (Woese, 1987) 
(Fig. 2), autotrophic carbon fixation pathways are distinctly 
related with the phylogenetic position of the respective 
organisms. Only members from the domains of Bacteria and 
Archaea are capable of chemolithoautotrophy (plants are 
photoautotrophs). Although the last common ancestor of 
Bacteria and Archaea has not been elucidated, none of the 
ancestral, deeply branching lineages are phototrophic, 
supporting the suggestion that chemosynthesis probably 
preceded photosynthesis on Earth (Schidlowski, 2001). 
Carbon fixation pathways among these ancient groups, 
however, differ and may have been one of the evolutionary 
factors in developing the lineages of Bacteria and Archaea 



(Fig. 2). The archaeal thermophiles, such as Thermoproteus 
and Pyrodictium, fix organic carbon via the reverse citric acid 
cycle. Anaerobic chemolithoautotrophic microbes, such as 
some sulfate reducers and acetogenic bacteria, as well as 
methanogens (Archaea) fix carbon by the reductive acetyl— 
coenzyme A (CoA) pathway (Ljungdahl-Wood pathway). 
Following the evolution of photosynthesis and the buildup 
of oxygen in the atmosphere estimated at ~2.5 billion years 
ago, aerobic microorganisms evolved. Most Proteobacteria 
(purple bacteria) are aerobes and fix carbon via the 
Calvin—Benson cycle as chemolithoautotrophs, although 
some Proteobacteria (e.g., sulfate reducers) fix carbon 
using the acetyl-CoA cycle or reverse citric acid cycle (Fig. 2, 
Table I). Other microbes that use the Calvin—Benson cycle 
include all photoautotrophs, some anoxygenic chemolitho- 
autotrophic microorganisms of the domain Bacteria, and 
Ribulose bisphosphate 
carboxylase/oxygenase (RuBisCO) is the signature enzyme in 
the Calvin—Benson cycle, and today RuBisCO is probably 
the most abundant protein in the world. As a consequence, 

photoautotrophic eukaryotes. 

the Calvin—Benson Cycle is the principal mechanism for 
biologically mediated reduction of CO, to organic carbon 
operating on Earth today. 


Electron donors are essential for chemolithoautotrophic 
growth (Table I). For microbes other than chemolitho- 
autotrophs (¢.g., heterotrophs, chemoorganotrophs), organic 
carbon molecules serve as electron donors used to gain 
cellular energy. One of the most energetic chemolitho- 


i Animals 
Entamoebae Slime 
Euryarchaeota Molds : 
Crenarchaeota Fungi 
Gram Methano- Methano- Extreme 
: positive ; sarcina Halophil Plants 
Proteobacteria acteria bacterium Sey 
Thermoproteus 2 
Thermoplasma Ciliates 
Cyanobacteria Sulfolobus 
Flavobacteria . 

Pyrodictium ens Flagellates 
bien Trichomonads 
nonsulfur Korarchaeota 

@ Calvin-Benson Cycle Diplomonads 

A Reverse Citric Acid Cycle 
@ Acetyl CoenzymeA Cycle 

FIGURE 2 Schematic of the tree of life showing the evolutionary relationships of the domains Bacteria, Archaea, and Eukarya, as well as the carbon fixation 

pathways for specific lineages. 

autotrophic electron donors is molecular hydrogen. Microbes 
that use hydrogen gas use CO) as an electron acceptor, either 
to make methane (methanogenesis) or acetic acid (aceto- 
genesis). Sulfate-reducing microbes also compete for 
hydrogen, but some can use small organic molecules as elec- 
tron donors. Hydrogen oxidizing bacteria utilize hydrogen 
aerobically. Additional electron donors include reduced 
sulfur compounds (e.g., hydrogen sulfide, elemental sulfur), 
ammonia (NHj), and reduced iron (Fe**) and manganese 
(Mn). The following sections describe chemolithoauto- 
trophic microbial groups based on required electron donors. 


Hydrogen gas accumulates from the anaerobic breakdown 
of organic molecules by fermentative (heterotrophic) 
bacteria or at some locations from volcanic gases, and serves 
as an important energy source for aerobic and anaerobic 
archaeal and bacterial chemolithoautotrophs. The anaerobic 
microbial groups that require hydrogen as an electron donor 
follow predictable redox chemistry and reflect alternative 
terminal electron acceptor utilization (Fig. 1). 

In most cases, hydrogen is rapidly consumed under 
anaerobic conditions by methanogens and sulfate reducers; 
however, if anaerobic growth is slower than production of 
hydrogen by heterotrophs or chemoorganotrophs, then 
hydrogen will diffuse into the aerobic environment, where it 
can be used by aerobic hydrogen-utilizing bacteria. Chemo- 
lithoautotrophic, aerobic, hydrogen-oxidizing bacteria grow 
with hydrogen as the electron donor and oxygen as the 
electron acceptor and fix CO, via the Calvin-Benson cycle or 
reverse citric acid cycle. A wide variety of Gram-negative and 
Gram-positive bacteria are hydrogen oxidizers, including 
Alcaligenes, Hydrogenobacter, Pseudomonas, and Aquifex. 

Iron or manganese reduction during anaerobic respiration 
typically occurs by chemoorganotrophs, but it is also known 
for some chemolithoautotrophs. Ferric iron is used as an 
exclusive electron acceptor by dissimilatory iron-reducing 
bacteria that use hydrogen as their electron donor; some 
iron reducers require organic carbon compounds such as 
acetate or formate. One of the most studied iron reducers, 
Shewanella putrefaciens, is a chemoorganotroph that can 
also use MnO, as a sole electron acceptor. Iron reduction in 
aquatic sediments is an important anaerobic ecological 
process, and the formation of iron sulfides in marine and 
freshwater has been attributed to biological iron reduction. 

Dissimilatory sulfate-reducing bacteria are obligate 
anaerobes that use hydrogen for the reduction of sulfate 
to hydrogen sulfide. Sulfate reducers are divided into two 
broad physiological subgroups. Group I dissimilatory sulfate 
reducers, including the genera Desulfovibrio, Desulfomonas, 
Desulfotomaculum, and Desulfobulbus, use lactate, pyruvate, 
ethanol, or fatty acids as carbon sources. Group II genera 
oxidize acetate and include Desulfococcus, Desulfosarcina, 

Chemoautotrophy 93 

Desulfobacter, and Desulfonema. There are also hyper- 
thermophilic sulfate-reducing Archaea belonging to the genera 
Archeoglobus. Certain species from the bacterial groups are 
capable of growing chemolithoautotrophically with hydrogen 
as the electron donor, sulfate as the electron acceptor, and 
CO) as the sole carbon source. Chemolithoautotrophs use 
the acetyl-CoA pathway, while the reductive citric acid cycle 
is used by chemoorganotrophs. If sulfate concentrations 
are high, sulfate-reducing bacteria completely oxidize 
fermentation by-products to CO}. In low-sulfate anaerobic 
environments, however, sulfate-reducing bacteria compete 
with methanogens for hydrogen and organic compounds. 
The Archaea Thermoproteus, Acidianus, and Desulfurolobus, 
common in acidic hot springs, are chemolithoautotrophic 
sulfur reducers. 

Acetogenesis results in less overall cellular energy than 
methanogenesis, but both metabolic groups are typically 
found in similar habitats. Acteogenic bacteria, such as 
Clostridium and Acetobacterium, are obligate anaerobes that 
form acetate from the oxidation of hydrogen using the 
acetyl-CoA pathway for CO) fixation. Hydrogen is the 
common electron donor, but donors can also be from sugars, 
organic acids, and amino acids. Many acetogens also reduce 
nitrate and thiosulfate and are tolerant of low pH, making 
them more metabolically versatile than other anaerobes. 

Methanogens are exclusively Archaea and are one of the 
most common anaerobic microbial groups. Methanogens 
oxidize hydrogen as the electron donor while reducing CO; 
to methane during anaerobic respiration. Methanogenesis 
from CO, reduction is done with the acetyl-CoA pathway, 
but formate, acetate, alcohols, carbon monoxide, and even 
elemental iron can serve as alternative electron donors. There 
are seven major groups of methanogens based on their 
physiology and 16S rRNA gene sequence comparisons, with 
the genera Methanobacterium, Methanococcus, Methanomicro- 
bium, Methanosarcina, and Methanolobus having the most 
identified species. Chemolithoautotrophic methanogenesis 
occurs with hydrogen and CO, only, although growth is 
less efficient than those utilizing acetate or other organic 
compounds. Methanogens are ubiquitous in most anaerobic 
environments and are usually found in close association with 
decomposing organic material. 

Reduced Inorganic Sulfur Compounds 

High concentrations of reduced sulfur compounds are toxic 
to most organisms. Hydrogen sulfide gas, in particular, 
reacts with operational biomolecules to form nonfunctional 
complexes that inhibit respiration; however, hydrogen 
sulfide, elemental sulfur, thiosulfate, polythionates, metal 
sulfide, and sulfite serve as electron donors for chemolitho- 
autotrophic sulfur-oxidizing bacteria and Archaea. (Note: For 
the purposes of this discussion, any microbe capable of 
oxidizing any reduced sulfur compound will be referred to 

94 Chemoautotrophy 

as a sulfur oxidizer). One difficulty for sulfur-oxidizing 
microbes is competing with chemical oxidation of reduced 
sulfur compounds, specifically hydrogen sulfide. As a result, 
most sulfur-oxidizing bacteria occupy aerobic/anaerobic 
interfaces where sulfide and oxygen meet and are therefore 
chemotactic. Interface or gradient growth is important in 
some ecosystems where sulfur-oxidizing bacteria form 
symbiotic relationships with animals, such as at deep-sea 
hydrothermal vent sites. 

The earliest microbiological research regarding chemo- 
lithoautotrophic metabolism was done with sulfur-oxidizing 
bacteria (Winogradsky, 1887). Some sulfur-oxidizers thrive 
in low pH environments (acidophiles), while others require 
neutral pH conditions. Sulfur-oxidizing bacteria can form 
large microbial mats or thick biofilms in a range of habitats, 
including acid mine drainage and mine tailings, sulfidic 
thermal springs, marine sediments, and sewage sludge. There 
are several different types of sulfur oxidizers, including those 
belonging to the genus T/iobacillus and the morphologically 
conspicuous sulfur bacteria such as Beggiatoa, Thiothrix, 
Thioploca, Thiomicrospira, and Thiovulum, among others. 
Virtually all of these colorless sulfur bacteria (as compared to 
the green and purple sulfur bacterial groups) are Gram- 
negative and some deposit intracellular elemental sulfur 
when grown on hydrogen sulfide. While sulfur-oxidizing 
bacteria are by nature chemolithoautotrophic, using the 
Calvin—Benson cycle for carbon fixation, some can be 
chemoorganotrophic, heterotrophic, or mixotrophic with 
respect to their carbon source. Typically, molecular oxygen is 
the terminal electron acceptor, although some are capable of 
microaerophilic growth within sulfide-oxygen gradients 
using nitrate. 

The use of reduced sulfur compounds as electron donors 
and nitrate as an electron acceptor is a form of denitrification 
and results in the production of NO, N,O, or N>. Denitrifi- 
cation can also occur if nitrate is used as an alternative 
electron acceptor during anaerobic respiration, but with 
organic carbon as the carbon source. Thiobacillus denitri- 
ficans is a common denitrifying bacterium. 

Ammonia and Nitrite 

Nitrification includes the oxidation of both ammonia and 
nitrite and results in the formation of nitrite and nitrate, 
respectively. The formation of nitrate requires a two-step 
process in which ammonia is first oxidized to nitrite by 
ammonia oxidizers, and then nitrite is oxidized to nitrate by 
nitrite oxidizers. The two groups of nitrifying bacteria, 
belonging to the family Nétrobacteraceae, typically work 
synchronously in a habitat as chemolithoautotrophs that use 
the ammonia or nitrite as electron donors and CO, as the 
sole source of carbon. In some rare instances, small organic 
compounds can also be used. Genera of ammonia oxidizers 
have the prefix Nitroso-, including Nitrosomonas and. Nitro- 

sospira, whereas genera of nitrite oxidizers have the prefix 
Nitro-, including Nitrobacter and Nitrospira. Nitrification 
occurs at the aerobic/anaerobic interface, but nitrifying 
bacteria have a high affinity for oxygen. Microbiological 
ammonia and nitrite oxidation can occur over a broad range 
of conditions, including in acid soils, and at low- and high- 

Iron and Manganese 

As with sulfur oxidation, bacteria can also gain energy from 
ferrous iron (Fe**) or manganous manganese (Mn”*) oxida- 
tion. In the past, iron and manganese oxidation was attrib- 
uted to many different bacteria based on the accumulation of 
iron or manganese minerals associated with cellular material; 
however, recent investigations have found that most of these 
bacteria do not gain energy from mineral buildup. Iron 
rapidly oxidizes to ferric (Fe**) iron at neutral pH, so 
successful iron oxidizers live in low pH environments, such 
as acid mine drainage, acid springs, mine tailings, or acid 
soils containing sulfide minerals such as pyrite. Chemolitho- 
autotrophic iron-oxidizing bacteria include the acidophilic 
aerobe Thiobacillus ferrooxidans. Despite research attempting 
to isolate other chemolithoautotrophs, the most common 
iron oxidizers, Leptospirillum and Gallionella, and some 
manganese oxidizers, such as Arthrobacter and Hyphomicro- 
bium (except H. manganoxidans), are chemoorganotrophs. 
Ecologically, iron and manganese oxidizers are important 
for their detoxification of the environment by lowering the 
concentration of dissolved toxic metals. 

One-Carbon Compounds 

There is some controversy over whether microbial processes 
requiring growth on reduced one-carbon (C-1) compounds 
as the sole carbon and energy source represent true chemo- 
lithoautotrophy; C-1 compounds used as electron donors 
are biogenic, thereby not fulfilling the “tho component of 
chemolithoautotrophy. However, the theory of cycling C-1 
compounds by potential chemolithoautotrophs is gaining 
appeal because understanding CO, and CH, dynamics 
is important for the study of carbon budgets of aquatic 
ecosystems in surface and subsurface environments. 

Methanotrophs and methylotrophs utilize methane or 
methanol for cellular energy and growth. There are two 
physiological groups of methanotrophic bacteria, both 
obligate aerobes. Type I methanotrophs include the genera 
Methylomonas, Methylosphaera, and Methylomicrobium and 
fix carbon by assimilating formaldehyde in the ribulose 
monophosphate pathway. Type II methanotrophs fix 
formaldehyde in the serine pathway, and genera include 
Methylocystis and Methylosinus. 

Growth on carbon monoxide by carboxidobacteria 
(carbon-monoxide-oxidizing bacteria) occurs with a range 

of electron acceptors, including oxygen, elemental sulfur, 
sulfate, nitrate, or ferric iron. Members of the genera Pseudo- 
monas, Bacillus, Alcaligenes, and Clostridium are known for 
carbon monoxide oxidation. 


Before discussing the microbial groups found in caves, a 
description of how chemolithoautotrophy and chemolitho- 
autotrophic microorganisms can be characterized is warranted. 
Brief discussions of various methods are presented in the 
following sections. A word of caution: Virtually any method 
of microbial identification or characterization has some 
type of methodological error associated with it that should 
be considered when interpreting results. The techniques 
described are just a few of the possible microbiological, 
ecological, and geochemical methodologies that could be 
used for more thorough characterization of the diversity and 
distribution of chemolithoautotrophic microbial populations 
in caves. Without question, a combination of different tech- 
niques, while not always feasible, does provide the best 
possible results. 

Traditional Methods, Including Culturing and 
Molecular Methodology 

One of the most common microbiology techniques involves 
growing chemolithoautotrophic microorganisms in culture 
by providing cells with all the essential nutrients, including 
required electron donors and acceptors, required for growth. 
Indeed, many chemolithoautotrophic microbes have been 
discovered by this technique, and for many years this was the 
only method to obtain chemolithoautotrophic microbial 
strains. Even when the metabolic requirements are known, 
however, laboratory enrichment and pure culture isolation 
have proven difficult, and most microbes still remain un- 
culturable. In fact, most microorganisms in nature have not 
been grown in culture, and it has been estimated that less 
than 1% of known microbes can be cultured with current 
techniques. Standard culturing often introduces a selective 
bias toward microorganisms that are able to grow quickly or 
to utilize substrates provided in the medium more efficiently 
(r-strategists). Culturing, although highly selective and diffi- 
cult for unknown or poorly characterized microorganisms, is 
still a very reliable way to decipher what biogeochemical 
activities may be microbially mediated. 

Recent advances in molecular techniques can circumvent 
these problems and provide culture-independent methods 
from which microorganisms can be identified based on the 
evolutionary relationships of 16S rRNA gene sequences. 
However, the polymerase chain reaction (PCR) used to 
amplify environmental 16S rRNA sequences may result in 
biased species diversity due to uneven extraction of DNA 
or PCR-amplification bias due to genome size. Although 

Chemoautotrophy 95 

phylogenetic affiliations are assigned for organisms in natural 
samples, metabolic activity must be implied, especially for 
novel or uncharacterized groups; therefore, additional 
ecological and molecular methods are required to determine 
if known or novel organisms in an ecosystem are capable of 
chemolithoautotrophic growth. 

Most microbiological research conducted with cavewater, 
sediment, or mineral samples has focused on the census of 
microorganisms colonizing a particular habitat. Many pure 
and mixed culture isolates have been obtained, including 
iron- and manganese-oxidizing bacteria, ammonia-oxidizing 
bacteria, sulfur-oxidizing bacteria, and countless heterotrophs 
and chemoorganotrophs (Table II). PCR-based techniques 
used to describe cave microbial populations demonstrate that 
chemolithoautotrophic systems have high biodiversity. 
Recent molecular research has been done in cave systems 
dominated by sulfur-oxidizing bacteria (Table I), even 
though culture-independent methodology allows for 
wide-ranging surveys of many different cave habitats. These 
molecular studies indicate that closely related microbial 
groups are found in geographically separated cave systems 
and may suggest that colonization of the subsurface, 
particularly sulfur-based environments, is not an isolated 

Metabolic (Biochemical) Assays 

If pure or mixed cultures are obtained, microorganisms can 
be characterized or identified based on a variety of growth- 
dependent assays. Biochemical assays can also be done to 
identify metabolic activities in natural samples without 
cultures. Assays include, but are not limited to, measuring 
the presence or absence of enzymes involved in degradation 
of substrates in the growth medium, the production of 
gases (e.g., hydrogen, carbon dioxide, hydrogen sulfide), and 
utilization of specific carbon substrates to test for mixo- 
trophic growth, as well as determining the ability to grow 
under different physicochemical conditions, such as growth 
over a range of pH values or temperatures or even under 
varying oxygen concentrations. For many organisms, a few 
key assays are all that is required for identification or charac- 
terization. Some more sophisticated assays include measuring 
enzymatic activity or membrane lipids such as phospholipid- 
derived fatty acids (Boschker and Middelburg, 2002). 
Additionally, dehydrogenase, adenosine triphosphate (ATP) 
or total adenylate nucleotides, and RuBisCO presence and 
activity are important for characterizing chemolithoauto- 
trophic microbial populations. Most metabolic assays 
conducted with cave microorganisms have been done on 
heterotrophs or chemoorganotrophs. To date, only microbes 
from Movile Cave have had RuBisCO activity measured, and 
various populations of sulfur-oxidizing bacteria from Florida 
springs and Lower Kane Cave have been investigated using 
an assortment of biochemical assays. Future studies of cave 

96 Chemoautotrophy 

TABLE II Chronological Summary of Studies Describing Chemolithoautotrophic Microorganisms Found in Various Caves 

Cave Study Date(s)* Chemolithoautotrophic Metabolic Processes? Approach* 

Movile Cave, Romania 1986-2003 Sulfur oxidation, methanotrophy, methanogenesis, M, C, SI, DNA, BA, RI 
ammonia oxidation 

Grotta Azzurra, Italy 1986-1999 Sulfur oxidation C, M, SI 

Frasassi Caves‘, Italy 1995, 2000-2003 Sulfur oxidation M, C, SI, DNA, RI 

Tito Bustillo Cave, Maltravieso 1997, 1999-2003 Ammonia oxidation, sulfur oxidation, iron and C, DNA 

Cave and others, Spain manganese oxidation 

Zoloushka Cave, Ukraine 1994, 2001 Sulfate reduction, denitrification, sulfur oxidation, C, SI 
iron oxidation 

Kugitangtou caves, Turkmenistan 1994, 1997, 2001 Sulfur oxidation, sulfate reduction, iron oxidation M, C, DNA 

Anchialine caves, Mexico 1997, 2002 Ammonia oxidation, sulfur oxidation, methanotrophy M, SI 

Cueva de Villa Luz, Mexico 1994-2002 Sulfur oxidation, sulfate reduction M, C, DNA 
Bungonia Caves, Australia 1973-1994, 2001 Iron reduction, sulfate reduction, sulfur oxidation M, C, SI 
Bundera Sinkhole, Australia 1994-1999 Sulfur oxidation, ammonium oxidation SI 

Nullarbor caves, Australia 1994, 2001 Nitrite oxidation, sulfur oxidation M, DNA, 

Caves in France, Romania, USA 1963, 1987, 2000-2002 Iron and manganese oxidation M, C, DNA 
Florida Aquifer caves 1965, 1992-2003 Sulfur oxidation M, C, DNA, BA 
Cesspool Cave, Virginia 1986, 2001 Sulfur oxidation M, C, DNA, RI 
Mammoth Cave, Kentucky 1977, 2000-2003 Ammonia and nitrite oxidation C, DNA 

Parker Cave, Kentucky 1988, 1998 Sulfur oxidation M, DNA 
Lechuguilla Cave, New Mexico 1994, 2000, 2003 Iron and manganese oxidation, nitrite oxidation C, BA, DNA 
Lower Kane Cave, Wyoming 1999-2003 Sulfur oxidation, sulfate reduction, iron reduction, M, C, SI, DNA, BA, RI, FISH 

methanogenesis, iron oxidation 

“Some representative years for which published results can be found in the literature. 

> Listed in order of dominance, if known. 

“Approach used to identify and to verify chemolithoautotrophic processes. M = microscopy; C = culture techniques; SI = stable isotopes; DNA = phylo- 
genetic relationships based on 16S rRNA analyses; BA = biomarker assays; RI = radioisotope experiments; FISH = fluorescence in situ hybridization 

4 Also known as Cueva del Azuffre or Cueva de las Sardinas. 
© Includes the Grotte di Frasassi and Grotta Sulfurea system 

chemolithoautotrophic populations should consider the 
inclusion of more biochemical assay work. 

Stable Carbon Isotope Ratio Analysis 

Stable isotope ratio analysis (SIRA) methods can be used to 
understand ecosystem energetics and carbon fixation path- 
ways. The creation and alteration of organic compounds, 
reflecting carbon fluxes in an ecosystem, control the isotopic 
composition of organic matter (Boschker and Middelburg, 
2002). The two carbon isotopes of importance in studying 
carbon cycling are '*C and '°C. Incorporation of carbon into 
living tissues causes significant kinetic isotope fractionation, 
such that the lighter isotope ('*C) is preferentially assimi- 
lated, leaving behind the heavier isotope ('°C). This 
discrimination for '"C is largely due to kinetic effects 
caused by the irreversible enzymatic CO,-fixing reaction that 
assimilates CO, into the carboxyl group of an organic acid. 
In the Calvin-Benson Cycle, for example, RuBisCO is mostly 
responsible for isotope discrimination. Differences in the 
isotopic composition are expressed in terms of the delta 

(5)-notation of a ratio of '°C/'*C in a sample relative to a 
standard, measured in per mil (%o): 

8'X (%o) = [soto : 1 x 1000 
C/”C standard 

The overall isotopic fractionation is determined by 
subtracting the isotopic composition of the carbon of an 
organism from the isotopic composition of the carbon 
reservoir, with most chemolithoautotrophic fractionation 
values ranging between —20 and —40%o. Excretion, respira- 
tion, and heterotrophic carbon cycling are (for the most part) 
considered negligible processes for carbon fractionation, and 
the isotopic composition of heterotrophic organic matter will 
be the same or slightly higher than the source organic carbon. 
Due to the diversity in carbon-fixation pathways, there are 
a wide variety of carbon isotope fractionation values that 
exist for microbial populations within an ecosystem. But 
chemolithoautotrophically fixed carbon in general has lower 
5'°C values than the more well-studied photoautotrophs. 

Carbon stable isotope ratio analysis has been used in 
Movile Cave in Romania, where microbial mat 5'°C values 

suggest that the dominant microbes, sulfur-oxidizing bac- 
teria, are chemolithoautotrophs (Sarbu e¢ a/., 1996). These 
chemolithoautotrophic bacteria serve as the main energy and 
food base for the cave ecosystem, evident by the ~26 to 30%o 
fractionation between the mat and the bicarbonate, and the 
similar 5'°C compositions of grazers (-39 to —46%o) and 
carnivores (—40 to —45%o) to the mats (-41 to —-45%o). The 
similar isotopic ratios suggest that grazers consume the mats, 
and are in turn consumed by carnivores. The “you are what 
you eat” principle has been applied to ecological studies in 
the Frasassi Caves and Grotta Azzurra in Italy, Lower Kane 
Cave in Wyoming, and an anchialine cave system in Mexico. 
These analyses not only show microbial chemolithoauto- 
trophic activity, but also evaluate ecosystem level dynamics 
and food quality for higher trophic level animals in caves. 
Moreover, the carbon isotope values for the Movile Cave 
system are distinct from surface derived organic carbon, that 
have 5'°C values of —20 to —30%o, thereby providing a useful 
tool to distinguish among organic carbon sources. 

Marker Genes 

While comparison of 16S rRNA gene sequences of various 
organisms in a community provides a phylogeny, additional 
molecular techniques can exploit the fact that specific 
operational genes are required for certain metabolic activities. 
Marker gene information can identify a specific metabolic 
activity, and, because of the potential for greater sequence 
variation than for 16S rRNA, marker genes provide increased 
phylogenetic resolution among closely related organisms. For 
example, a gene codes for the enzyme ammonia mono- 
oxygenase (amoA) used for chemolithoautotrophic ammonia 
oxidation; this amoA gene can be used as a marker gene to 
test if ammonia oxidation is possible in natural samples. To 
date, only a few studies have used marker genes to research 
Movile Cave microbial communities, specifically methano- 
trophs. Microarrays, comprised of microchips with attached 
nucleotide probes, can also be used to study metabolic 
activity and microbial diversity in complex systems (Guschin 
et al., 1997). Probes can be designed for marker genes, as well 
as for species-, genera-, group-, or domain-specific 16S rRNA 
targets. The number of probes used in an array is limited only 
by the size of the microchip. Microarray methods circumvent 
polymerase chain reaction (PCR)-based problems involving 
amplification bias and molecular interactions that form 
chimeric sequences. The utility of microarrays is limitless, but 
at present no cave studies have incorporated this methodology. 

Radioisotope Experiments 

Microbial activity in natural or culture samples can be 
measured in situ using radioisotope-labeled (radiolabeled) 
substrates, and both carbon fixation and energy pathways for 
chemolithoautotrophy can be quantified. These experiments 
provide activity rate estimates without knowing how many 

Chemoautotrophy 97 

organisms are in the sample and allow a sensitive, quantita- 
tive assessment of biogeochemical transformations in nature. 
The rate of certain metabolic processes can be determined by 
measuring the uptake of one or more compounds over time 
and the subsequent buildup of other compounds. Terminal 
electron acceptors are the most common radiolabeled 
substrates. For instance, sulfate reduction can be monitored 
using radiolabeled **SO{ compounds, whereby loss of 
§O7 would indicate consumption and a gain of *S? as 
the byproduct of sulfate reduction would provide sufficient 
evidence that sulfate reduction is occurring. 

These same principles can be used to determine carbon 
fixation rates. Carbon molecules, such as bicarbonate or 
acetate, are labeled with C, and the evolved “CO, and C 
biomass are measured to show rates of substrate uptake and 
utilization among different organisms within an ecosystem. 
Even when multiple radiolabeled substrates are combined 
in an experiment, however, there are several problems with 
carbon radioisotope studies, with the most significant being 
how to decipher the results. Carbon fixation or consumption 
can be overestimated or underestimated due to sampling 
perturbation or poor understanding of the community 
structure. Because natural samples can be extremely sensitive 
to disturbance, the experiment may skew results and demon- 
strate that one metabolic process is dominant, when in fact it 
is just less sensitive to disturbance than the other processes. 
Usually the sensitive nature of cave habitats prevents in situ 
radiolabeled experiments. Additionally, some microbes in a 
sample could be capable of facultative or mixotrophic 
growth, acting as chemolithoautotrophs in nature, but when 
provided with a rich organic carbon molecule they will use it 
as chemoorganotrophs or heterotrophs. Detailed culturing or 
phylogenetic analyses of the microbial communities should 
be coupled to radioisotope studies to aid interpretation. 
Recovering the DNA from cellular biomass that incorporated 
radiolabeled substrates can be done to investigate the active 
microorganisms of interest, as well. 

To date, radioisotope techniques have been used to 
describe microbial processes from only a few caves, including 
Movile Cave, the Frasassi Caves, Lower Kane Cave, and 
Cesspool Cave (Virginia). The measured rates of autotrophic 
productivity from these cave systems, based on incorporation 
of '4C-labeled bicarbonate, demonstrate that carbon fixation 
in Movile Cave is the highest, similar to values obtained from 
surface aquatic systems in open oceans, lakes, or streams. 
Moreover, radiolabeled acetate and leucine, organic molecules 
used by chemoorganotrophs and heterotrophs, suggest that 
Lower Kane Cave has high rates of heterotrophic producti- 
vity but an order of magnitude less than autotrophic primary 


Cell morphologies and the structural relationships between 
cells and inorganic substrates can be visualized directly by 

98 Chemoautotrophy 

light microscopy or transmission or scanning electron micro- 
scopy (TEM or SEM, respectively). But, while microscopy 
can provide a wealth of observational information, such as 
microbial mat structure and the nature of biomineralization, 
microscopy is not a reliable technique to identify a microbe, 
even with morphologically conspicuous cell types. Addi- 
tionally, general microscopy does not provide any 
information regarding metabolic activity. 

Combining microscopy with stains that highlight certain 
cells or metabolic activities can be a useful way to identify 
actively metabolizing cells, as well as to determine biomass. 
Acridine orange (AO) and 4’,6-diamidino-2-phenylindole 
(DAPI) are commonly used fluorescent stains, but they bind 
to all nucleic acids in both alive and dead (or inactive) cells. 
Staining for respiring cells using 2-(p-iodophenyl)-3-(p- 
nitrophenyl)-5-phenyltetrazolium chloride (INT) is much 
more useful to estimate the number of metabolically active 
cells in a sample. 

Much of the earlier microbiological work on cave 
sediments and microbial mats involved microscopy in one 
form or another. In particular, most studies report cell counts 
and morphology based on phase contrast or fluorescence 
microscopy using AO or DAPI. While this may prove 
useful to identify morphologically conspicuous bacteria, this 
approach is generally not adequate to characterize most 
microbial groups. Some interesting studies have been done 
using electron microscopy, showing that microbial cells can 
be found on nearly every surface of a cave. Unfortunately, 
little information about the microbial communities can 
be gained from these strictly observational investigations. 
With respect to chemolithoautotrophs in caves, the best 
use of traditional microscopy is to combine it with other 
geochemical, ecological, and microbiological methodologies. 

Fluorescence in situ hybridization (FISH) combines 
fluorescence microscopy with more specific probes that can 
characterize a microbiological sample. FISH probes are 
similar to those used in microarrays, with a majority of them 
being based on 16S rRNA targets. FISH combined with 
microautoradiography (FISH-MAR) can not only identify 
a microbe but also reveal what the microbe may be doing 
in nature by combining genetic probes with radiolabeled 
substrate incorporation experiments (Ouverney and 
1999). As FISH-MAR combines the best 

aspects of multiple techniques, it will be an extremely useful 


techniques for describing chemolithoautotrophy in the 


Cave ecosystems generally reflect energy and nutrient 
limitations. Most caves are fed by surface streams that have 
had photosynthetically derived organic matter and nutrient- 
rich sediments washed into them. Microorganisms are also 
brought into a cave with surface debris and deposited with 

sediments. Historically, studies have shown that most 
microorganisms in caves are not chemolithoautotrophic but 
instead are translocated soil heterotrophs, chemoorgano- 
trophs, or fecal coliform bacteria from contaminated surface 
water and detritus. The roles of microorganisms in these 
ecosystems that receive allochthonous input are to break 
down complex organic material and to serve as a food source 
for higher trophic levels. Other bacteria washed into caves 
from meteoric drip waters have been associated with bio- 
mineralization processes of calcium carbonate, iron, or 
manganese speleothem formation. 

Alternatively, caves with very little to no allochthonous 
input contain chemolithoautotrophic bacteria that colonize 
material, such as sediments, with high concentrations of 
inorganic chemical compounds. Moreover, groundwater rich 
in dissolved reduced compounds, such as CO, or hydrogen 
sulfide, can also discharge into caves and be used as energy 
sources by chemolithoautotrophs. Relatively few studies have 
documented the occurrence, distribution, and community 
composition of chemolithoautotrophic metabolic groups in 
these karst systems compared to higher tropic level animals, 
although the importance of chemolithoautotrophy in 
caves has recently been demonstrated. Since the mid-1990s, 
investigators have employed numerous microbiological, 
geochemical, and molecular techniques to understand 
specific chemolithoautotrophic groups in cave systems (Table 
II). To date, a few ecosystem-level studies have been initiated 
to describe multiple microbial groups and to attempt to 
integrate the activities of different microorganisms with 
respect to each other (Table II). It is hoped that these recent 
studies will serve as examples for future multidisciplinary 
projects. The following section discusses investigations 
relating to bacteria involved in sulfur, iron, and nitrogen 
cycling in caves. 

Sulfur Bacteria 

Ata glance, Table I demonstrates that sulfur oxidizers are the 
only group of cave chemolithoautotrophic bacteria studied 
rigorously. Prior to and just after the 1986 discovery of the 
chemolithoautotrophically based groundwater ecosystem in 
Movile Cave, few studies described sulfur-based microbial 
populations in caves. Much of the early work was simply 
observational or involved culturing, especially of non- 
chemolithoautotrophs. It was after the Movile Cave research 
was initiated that speleologists and microbiologists began 
focusing their studies on other sulfur-based cave microbial 
communities. Nearly all of the caves with sulfur-oxidizing 
bacterial populations have springs with hydrogen-sulfide-rich 
groundwater that discharge into them and are termed sulfidic 
caves. Because of hydrogen sulfide oxidation, sulfuric acid is 
generated and promotes local limestone bedrock dissolution. 
This cave formation process is known as sulfuric acid speleo- 
genesis, first described from Lower Kane Cave by Egemeier 

(1981) (Fig. 3). 

Chemoautotrophy 99 

FIGURE 3 Main trunk passage in Lower Kane Cave in Wyoming. White, 
filamentous microbial mats dominated by sulfur-oxidizing bacteria colonize 

shallow sulfidic water, beginning at the lower right corner (water flows from 
the lower right to upper left). This microbial mat extends for approximately 
20 m, with an average thickness of 5 cm. Piles of gypsum surround the 
stream (especially on the left), formed from the replacement of limestone 
during sulfuric acid speleogenesis. 

The first studies describing sulfur-oxidizing bacteria were 
driven by the need to understand the role of chemolithoauto- 
trophs in the establishment of complex cave ecosystems in 
sulfidic caves. While SIRA and radioisotope studies were 
used initially to characterize chemolithoautotrophy, less than 
half of the other known sulfidic cave populations have been 
described using these methods (Table II). 

Sulfur-oxidizing bacteria have been found in both aqueous 
and subaerial habitats in sulfidic caves. Filamentous bacteria 
form thick microbial mats or biofilms within cave spring 
orifices, pools, or streams (Fig. 4). One main exception is in 
Movile Cave, where a thin microbial mat floats at the surface 
of the groundwater (Fig. 5); an anaerobic microbial biofilm 
mixed with clay sediment also covers the submerged cave 
floor. Phylogenetic analyses of cave microbial mats from 
Lower Kane Cave, the Frasassi Caves, Cesspool Cave, Cueva 
de Villa Luz (Mexico), and Parker Cave (Kentucky) show 
sulfur-oxidizing bacteria closely related to Thiovulum, 
Thiothrix, Thiobacillus, Thiomonas, Thiomicrospira, and 
Achromatium. Based on culturing and microscopy, Thiothrix, 
Beggiatoa, and Thiomicrospira have also been reported from 
submarine caves in Italy. SIRA and culturing suggest that 
the majority of all the investigated microbial mats thus far 
contain chemolithoautotrophic sulfur oxidizers, although 
FISH-MAR experiments with radiolabeled sulfur com- 
pounds will provide more effective tests of chemolithoauto- 
trophy in the future. Isotopic values for aqueous microbial 
mats from Lower Kane Cave and the Frasassi Caves range 
from —22 to -39%o. The microbial mats from Movile Cave, 
however, have ratios that range from —41 to -46%o, indicat- 
ing that in addition to sulfur oxidation, methanogenesis and 
methanotrophy may also be active. 

Colonization of predominately sulfur-oxidizing bacterial 
mats by other metabolic groups, such as anaerobic sulfate 

FIGURE 4 White, subaqueous microbial mat from Lower Kane Cave in 
Wyoming. A thin, white biofilm connects filaments to each other in flowing 

water. Endemic snails (Physa spelunca) graze the mats. Streambed substrate is 

made of chert cobbles (upper left). 


FIGURE 5 Partially flooded passage in Movile Cave in Romania. A white 
filamentous microbial mat floats on the surface of the water. A grid was 
constructed to encourage additional mat growth. 

reducers or methanogens, has only recently been addressed, 
particularly in Movile Cave and Lower Kane Cave. Sulfate- 
reducing bacteria have been reported in anaerobic cave water 
and aqueous sediments from several sulfidic caves and springs 
(Table II). Sulfate reducers are important for recycling sulfur 
compounds by generating supplemental hydrogen sulfide 
that sulfur-oxidizing bacteria can use; however, more research 
must be conducted to differentiate between chemolithoauto- 
trophic sulfate-reducing bacteria and the more common 
chemoorganotrophic sulfate reducers. 

Microbial biofilms on subaerial cave-wall surfaces 
(snottites, or microbial draperies) have been described from 
the Frasassi Caves, Cueva de Villa Luz, Cesspool Cave, Lower 
Kane Cave, and Kugitangtou caves (Turkmenistan). While 
some of the largest snottites reportedly reach up to 10 cm in 
length or more, microbial biofilms typically occur as discon- 
tinuous patches of insoluble crusts with sub-centimeter-long 
mucus-like droplets suspended from the crusts (Fig. 6). Most 

100 Chemoautotrophy 

FIGURE 6 Discontinuous cave-wall biofilms and crusts (dark patches) from 
Lower Kane Cave in Wyoming. Crusts form on gypsum (light areas), and 

condensation droplets and mucus-like drops are suspended from the crusts. 
Elemental sulfur is typically associated with the crusts. 

of the snottites, droplets, and crusts have pH values between 
0 and 3. Biofilms from the Frasassi and Kane Caves have 
5'°C ratios of —36 to —39%o, suggesting that carbon has been 
produced from chemolithoautotrophy. Laboratory isolates 
and phylogenetic studies of subaerial surfaces in these caves 
demonstrate that strains are closely related to Thiobacillus, 
Sulfobacillus, and Acidimicrobium. Cultures of thiobacilli 
from cave-wall biofilms in Cesspool Cave have been used in 
dissolution experiments to determine rates of potential, 
microbially mediated cave modification, and enlargement. 
Sulfur bacteria are also associated with elemental sulfur, 
gypsum, and other mineral deposits in caves and may 
mediate precipitation of these minerals. 

Iron and Manganese Bacteria 

Microorganisms associated with iron and manganese cycling 
are the second most studied cave bacteria to date (Table II). 
According to Caumartin (1963), iron plays a significant 
role in the microbiology of a cave, and most cave sediment 
contains iron bacteria. Indeed, microorganisms associated 
with iron and manganese cycling have been found in cave 
sediments based on microscopy and culturing, with chemo- 
lithoautotrophic Gallionella and heterotrophic Leptothrix and 
Crenothrix being the most common genera described. Fossil, 
or encrusted, sheaths and stalks of these bacteria have been 
observed in stalactites, sediments, and corrosion crusts 
from several caves, including caves from Iowa, the Black Hills 
in South Dakota, Lechuguilla Cave in New Mexico, 
Kugitangtou caves in Turkmenistan, Grand Caymen island 
caves, and Nullarbor Plain caves in Australia. However, it is 
difficult to distinguish between chemical and biological iron/ 
manganese precipitates, and it is not evident from mineral- 
ization, fossil or active, whether an organism was chemo- 
lithoautotrophic. Therefore, the microbial ecology of these 
bacterial groups in caves remains speculative, and their role 

in cave ecosystems, as potentially minor primary producers 
or detoxifiers, also remains unclear. 

Nitrogen Bacteria 

Nitrification is an important chemolithoautotrophic 
pathway in nitrogen-limited freshwater streams and lakes, 
enhancing overall nitrogen availability. This process may also 
be important in aquatic cave ecosystems, although little work 
has been done. One study, for instance, based on SIRA of 
nitrogen compounds in anchialine caves in Mexico, indicates 
that nitrification may be the primary energy source for that 
ecosystem, in addition to sulfur and methane oxidation. 

The formation of nitrate minerals in caves, such as niter 
and nitrocalcite (saltpeter), has been attributed to nitrifica- 
tion, but the process is still highly debated. Nitrate minerals 
are typically found in dry cave sediments, such as in the 
southeastern United States. Nitrogen enters the cave in 
various forms, including atmospheric gas, nitrates from 
fertilizers, organic matter in soil, and from bat and rat guano. 
Ammonium is produced by nitrogen-fixing bacteria in soils 
overlying limestone and is brought into the caves via bedrock 
fissures or is washed in with stream sediments. Nitrosomonas 
in cave sediments oxidizes ammonia to nitrite, and then 
Nitrobacter oxidizes nitrite to nitrate, influencing the 
precipitation of nitrate minerals. Isolates of Nitrobacter have 
been obtained from Mammoth Cave in Kentucky in sedi- 
ments rich in saltpeter, suggesting geomicrobiological 
involvement, but additional investigations have been limited 
to date. 

Other Metabolic Groups 

Methanotrophy and methanogenesis, while expected to 
occur in some groundwater systems with high concentrations 
of dissolved methane, has only been studied to a large extent 
from Movile Cave. Enrichment and isolation techniques, 
'3CH,-labeling studies, and '*C-DNA analyses were used to 
detect methanotrophs from both type I and II groups, with 
type I dominating. Methanogens were also detected, based 
on 16S rRNA gene sequence analysis in Movile Cave 
anaerobic microbial mats. 


Although Winogradsky described chemolithoautotrophic 
metabolism from sulfidic springs more than 100 ago, it has 
been estimated that chemolithoautotrophs do not produce 
efficient energy and therefore could not be significant 
primary producers for an ecosystem, especially when com- 
pared to photosynthetically produced energy. However, in 
environments where light does not serve as the primary 
energy source for life (such as at the deep-sea hydrothermal 
vents, the deep terrestrial subsurface, and caves), photosyn- 

thesis is not possible. Chemolithoautotrophy can efficiently 
produce energy when competition is limited, and, as a result, 
chemolithoautotrophic organisms are prolific in the sub- 
surface. Moreover, the subsurface is a refuge for organisms to 
escape the harsh environmental surroundings of the surface, 
such as ultraviolet-light bombardment and temperature 
and moisture fluctuations. Chemolithoautotrophic micro- 
organisms colonizing the subsurface may serve as analogs for 
what life may have been like on early Earth or the type of life 
that may be found on other planetary bodies. In summary, 
estimation of the importance of chemolithoautotrophy to 
ecosystem level processes is being revised. 

Chemolithoautotrophically based ecosystems were first 
described from the deep-sea hydrothermal vent ecosystems 
(Deming and Baross, 1993). Now, chemolithoautotrophi- 
cally based ecosystems also include deep marine habitats 
associated with hydrocarbon cold seeps, estuarine sediments, 
and caves. Besides the free-living chemolithoautotrophic 
microorganisms from deep marine environments, primarily 
Beggiatoa and Epsilon Proteobacteria, animals are the most 
interesting constituents living at mid-ocean ridges. The 
vestimentiferan tubeworms (Riftia pachypila) and the clams 
Calyptogena magnifica have bacterial endosymbionts that 
oxidize hydrogen sulfide and fix CO). Deep-sea sediments 
exposed to hydrocarbon seeps have mussels with methano- 
trophic symbionts, and some mussels (Bathymodiolus 
thermophilus) also have sulfur- and methane-oxidizing 
bacterial symbionts. Clams, including Solemya and Thyasira, 
that live in brackish water and estuarine sediments have 
sulfur-oxidizing symbionts. The relationships between 
chemolithoautotrophic microorganisms and animals are 
beneficial to both organisms, as the animals provide pro- 
tective habitat and the bacteria supply nutrients and energy. 

While reduced sulfur compounds provide a rich energy 
source for chemolithoautotrophic growth in hydrothermal 
vents, cold seeps, and marine sediments, hydrogen-based 
chemolithoautotrophy has been described from deep 
aquifers. Microbes, including sulfate-reducing bacteria, 
acetogens, and methanogens, live at great depths into the 
earth, although most are considered chemoorganotrophs 
living off of scant organic debris that filters in from the 
surface. In isolated aquifers, however, geochemical produc- 
tion of hydrogen supports methanogenic communities 
(Stevens and McKinley, 1995). No higher trophic levels, 
including microscopic eukaryotes, have been reported from 
these methanogenic microbial ecosystems. 

One of the most extreme examples of a highly evolved, 
chemosynthetically based ecosystem is from the Movile 
Cave, a peculiar cave that serves as an access point to a large 
sulfidic aquifer in southeastern Romania (Sarbu et a/., 1996); 
33 new cave-adapted taxa have been identified from 30 
terrestrial invertebrate species (24 endemic), as well as 18 
species of aquatic animals (9 endemic). Similarly, the sulfur- 
based ecosystems from the Frasassi Caves, Cueva de Villa 
Luz, and sinkholes in Florida also have many different cave- 

Chemoautotrophy 101 

adapted animals dependent on energy produced by chemo- 
lithoautotrophic bacteria. Although Lower Kane Cave is 
also sulfur based, it has a relatively simple ecosystem of few 
terrestrial and aquatic species, with one of them being 
endemic (Fig. 4). The anchialine cave ecosystem of the 
Yucatan Peninsula is of additional interest because recent 
work shows that the ecosystem may be based on chemolitho- 
autotrophic ammonia oxidation. It is not known at present 
whether cave animals from any of these cave ecosystems 
contain sulfur-, ammonia-, or methane-oxidizing endo- 
symbionts, and this could be a future research direction. 


Caves are important and relatively accessible habitats to 
study chemolithoautotrophic metabolism. In the absence 
of organic carbon compounds, reactive rock surfaces and 
mineral-rich groundwater in the subsurface provide an 
assortment of potential energy sources for chemolithoauto- 
trophic microorganisms. In comparison to the research done 
with macroscopic organisms in caves, the distribution of 
chemolithoautotrophic microbial groups and associated 
biogeochemical processes occurring in caves have been 
addressed only infrequently, although the methodology 
exists. Of those investigations describing cave microbial 
communities, many have been simply observational, and the 
physiological capabilities of microorganisms were inferred 
based on the utilization of a specific substrate in cultures or 
from 16S rRNA gene sequence phylogenetic affiliations. To 
date, few stable isotope or radioisotope studies have been 
done to describe chemolithoautotrophy in cave ecosystems, 
and combined microscopy and molecular studies have only 
recently been initiated. Unfortunately, without future 
interdisciplinary investigations, understanding the true 
biogeochemical structure and function of chemolitho- 
autotrophs in subsurface habitats will remain limited and the 
significance of chemolithoautotrophs in cave ecosystems 
uncertain. However, with continuing technological advances 
and increased understanding of microbial physiology and 
ecology, exciting new discoveries are being made almost daily, 
and undoubtedly more chemolithoautotrophically based 
cave ecosystems will be identified. 

See Also the Following Articles 


Boschker, H.T.S. and J.J. Middelburg (2002) Stable isotopes and biomarkers 
in microbial ecology. FEMS Microbiol. Ecol. 1334: 1-12. 

Caumartin, V. (1963) Review of the microbiology of underground 
environments. Natl. Speleol. Soc. Bull. 25: 1-14. 

Deming, J. and J. Baross (1993) Deep-sea smokers: windows to a subsurface 
biosphere? Geochim. Cosmochim. Acta 57: 3219-3230. 

Egemeier, S.J. (1981) Cave development from thermal waters. Nael. Speleol. 
Soc. Bull. 43: 31-51. 

102 Clastic Sediments in Caves 

Guschin, D.Y., B.K. Mobarry, D. Proudnikov, D.A. Stahl, B.E. Rittmann, 
and A.D. Mirzabekov (1997) Oligonucleotide microchips as genosensors 
for determinative and environmental studies in microbiology. App/. 
Environ. Microbiol. 63: 2397-2402. 

Ouverney, C.C. and J.A. Fuhrman. (1999) Combined microautoradio- 
graphy—-16S rRNA probe technique for determination of radioisotope 
uptake by specific microbial cell types in situ. Appl. Environ. Microbiol. 
65: 1746-1752. 

Sarbu, S.M., T.C. Kane, and B.K. Kinkle (1996) A chemoautotrophically 
based cave ecosystem. Science 272: 1953-1955. 

Schidlowski, M. (2001) Carbon isotopes as biogeochemical recorders of life 
over 3.8 Ga of Earth history: evolution of a concept. Precambrian Res. 
106: 117-134. 

Stevens, T.O. and J.P. McKinley (1995) Lithotrophic microbial ecosystems 
in deep basalt aquifers. Science 270: 450-454. 

Winogradsky, S. (1887) Uber Schwefelbakterien. Botanishe Zeitung 
45: 489-610. 

Woese, C.R. (1987) Bacterial evolution. Microbiol. Rev. 51: 221-271. 

Clastic Sediments in Caves 

Gregory S. Springer 
Ohio University 

his article examines clastic sediments in caves, including 

such familiar deposits as cave mud and breakdown. 
Clastic sediments are composed of particles that have been 
eroded from preexisting rocks and subsequently deposited. 
The origins of clastic sediments are discussed together with 
the processes by which they are transported and accumulated. 
The latter phenomena are directly related to cave sediment 
stratigraphy, the layering of deposits, which is discussed in 
relation to the interpretation of sedimentary cave deposits. 
Representative cave deposits are presented as examples of 
how clastic sediments accumulate and are interpreted. 


Clastic sediments are volumetrically the most common 
deposits in caves. Clastic sediments are composed of rock 
fragments physically eroded from preexisting rocks and 
subsequently deposited. The majority of clastic particles in 
cave sediments are detrital grains eroded from land surfaces 
and carried into caves by streams, mass wasting, wind, wave 
action, and ice. The resulting cave deposits are similar to 
deposits created by equivalent processes on the surface. 
The term allochthonous deposit is used to identify sediments 
derived from outside of a cave, while autochthonous deposits 
are composed of particles eroded from cave walls and later 

deposited within the same cave. The mineral compositions of 
sediment grains can be used to infer whether deposits are 
autochthonous or allochthonous when the mineral content 
of surface rocks and a cave differ. The processes that trans- 
ported and deposited individual packages of clastic cave 
sediments can often be inferred from sedimentary structures, 
textures, stratigraphy, and facies analyses. In general, the 
processes that create deposits of clastic sediments in caves are 
strongly affected by large-scale trends in climatic, geologic, 
and geomorphic variables. As such, clastic sediments in caves 
may yield valuable information concerning cave genesis and 
long-term histories of earth system processes, including 
landscape erosion, flooding, and response of river networks 
to climate changes. Clastic sediments in caves are also 
valuable repositories of organic remains and artifacts and 
have attracted the concerted attention of archeologists and 
paleontologists since the 18th century. Clastic sediments in 
caves are particularly valuable to scientists because slow 
weathering rates in caves often allow for longer and better 
preservation of materials than overlying landscapes. 

Geomorphic Perspective 

Clastic sediments in caves record the processes that form 
cave passages and interactions of cave passages with other 
landforms. Stream caves are integrated components of 
surface drainage networks and contain sediments that record 
long-term transport of sediment eroded from hillsides and 
cave walls. As a result, cave sediments preserve a record of 
landscape and climate processes because stream caves are 
important conduits of water, solutes, and sediment that 
operate as substitutes for or in conjunction with surface 
streams. Changes in hillslope, surface stream, or climate 
processes cause localized sediment storage or erosion within 
stream passages. Changes are recorded by passage morpho- 
logy and sedimentary deposits, which are analogous to 
surface stream morphology and terraces. As with surface 
streams, changes in cave stream behavior are best recorded by 
clastic sediments because the principle expressions of changes 
in earth system processes are changes in the quantity and 
caliber of sediment carried by streams. In contrast to the 
dynamic interplay of cave streams and surface landforms, 
hypogene caves formed by deep-seated waters operate 
independently of surficial processes. Passage geometries are 
representative of autogenic processes and clastic sediments 
record cave development but offer little information about 
contemporaneous surface or climate processes. The origin of 
clastic cave sediments can be deciphered because passage 
morphology is often related to unique physical processes and 
allied sedimentary processes. Thus, cave passage morphology 
and clastic sediments can be used to interpret cave genesis, 
the relationship of the cave to other landforms, and changes 
in earth surface processes. 

Geomorphic, process-based studies that examine clastic 
sediments can be divided into two categories: (1) mechanistic 

studies that seek to understand the individual mechanisms 
by which clastic sediments are produced, transported, and 
deposited, and (2) interpretive studies that infer knowledge 
about large-scale processes from sedimentary records. 
Mechanistic studies serve as the foundation for interpretive 
studies, because they unravel how stratigraphic packages 
form and their relationships to large-scale earth systems. 
Modern geomorphology studies pair the two study types by 
identifying driving and resisting forces that cause clastic 
sediment deposition and erosion. For example, the size of 
sediments carried in a stream is proportional to the water 
velocity and flood size. Climate processes supply these 
driving forces in the form of rain and snowmelt. The resisting 
forces are friction caused by sediment grains and channel 
boundaries. A decrease in precipitation and temperature 
will typically decrease sediment transport while increasing 
sediment supply. The net result is increased sediment depo- 
sition on cave streambeds, called aggradation, during which 
time the cave stream deposits more sediment than it removes 
from a cave. By inference, such aggradational deposits can be 
used to recognize past climate changes. However, cave sedi- 
mentary deposits may be extremely complex because many 
cave streams occupy passages that are not much bigger than 
the stream; deposits may be extensively reworked and eroded 
because of the limited space available. Thus, interpreting the 
significance of clastic cave sediments is impossible without 
first understanding the mechanisms that produce, transport, 
and deposit sediment as well as their overall context. 


Sediment Production 

Clastic sediments are made from fragments of preexisting 
bedrock that have been broken apart by weathering (Fig. 1). 
The processes that fragment source rocks are diverse and 
rarely unique to caves. These processes include (1) chemical, 
(2) kinetic, and (3) physical phenomena. Chemical processes 
are those phenomena associated with corrosion or precipita- 
tion. Such sediments are created when grains of quartz, clay, 
dolomite, and chert are liberated from bedrock by wall 
dissolution. For instance, streambeds of many caves contain 
large quantities of gravel- and cobble-sized chert, which is 

Kinetic processes are those phenomena associated with 
motion, such as erosion in floodwaters. Erosion mechanisms 
include quarrying, whereby bedrock is eroded by floodwaters 
and. corrasion. Corrasion involves water accelerating grains 
of sand grains or cobbles and driving them into bedrock 
surfaces. Corrosion is especially effective at carving potholes 
in cave floors. Physical processes involve stresses that fracture 
rocks by tension or compression. The classic example is roof 
collapse caused by an increase in passage size. Shattering by 
ice crystals, known as frost wedging, is an important physical 

Clastic Sediments in Caves 103 

FIGURE 1 Laminated silts overlying sorted sands and basal gravels in 
Mammoth Cave in Kentucky. The 1.7-million-year-old sediments accu- 
mulated as the cave stream aggraded when the Green River was choked with 

sediment. The river sediments accumulated because continental glaciers had 
fed large quantities of sediment into the proto-Ohio River, which the Green 
River feeds. The accumulation of sediment lowered the local channel slope, 
a key driving force for sediment transport. (Photograph courtesy of Art 

process at middle and high latitudes, particularly in cave 

Breakdown is extremely common in caves and occurs 
where passage widths exceed the strength of ceiling beds and 
where walls are undercut by erosion (Fig. 2). The process 
is intimately related to tension generated above voids and 
separation of rock beds from one another along bedding 
planes and subhorizontal fractures. Breakdown piles, which 
are clastic deposits, can easily exceed the total volume of all 
other deposits in a cave. 

Sediment Transport and Deposition 

Allochthonous and most autochthonous sediments are 
repeatedly mobilized and deposited by flowing water in cave 
streams. Stream flow processes sort cave sediments by size 
because extremely small and extremely large particles are 
difficult to mobilize. As a result, clastic sediments deposited 
by streams are often composed of a narrow range of sizes. For 
instance, a bed of sand may lie atop a bed of gravel (Fig. 1). 
In contrast, sand, gravel, cobbles, and boulders are churned 

104 Clastic Sediments in Caves 

FIGURE 2 Breakdown clasts floating in gravel derived from an overlying 
sinkhole in Indiana. Collapses are commonly found where passages abut 
sinkholes or valley walls because ceiling strength is least where beds of rock 
are not continuous. (Photograph courtesy of Art Palmer.) 

together in the deposits of creep, rockslides, and debris 
flows because the deposits are capable of moving virtually 
everything available. The controls on stream and mass 
wasting processes are different as are interpretations of their 

Most cave streams only move significant quantities of 
clastic sediments during floods. The water velocity necessary 
to move an individual particle can be crudely estimated using 
the Hjulstrom diagram (Fig. 3). Examining the Hjulstrom 
diagram, it can be seen that fine sand is the easiest material 
to mobilize, while clay and boulders require substantially 
higher velocities. Once a particle has been mobilized it 
will continue to move at velocities below those required to 
initiate movement. Fine-grained sediments will not cease 

movement until water velocities are extremely slow or 
chemical processes cause flocculation. 

The sum of all sediment carried in a stream is called load. 
Individual grains may move as suspended load or bedload. 
Suspended load is particles borne within the water column 
by turbulence; suspended particles settle out of the water 
column as floods recede and typically consist of clay, silt, or 
sand. Bedload moves by rolling, toppling, and saltation 
(similar to bouncing). Most clastic sediments are carried as 
suspended load. 

Sediment will accumulate in stream passages if driving 
forces are incapable of moving all sediment supplied to a 
channel. Long-term deposition may be recorded as banks 
of sediments on channel margins or thick deposits beneath 
streambeds. These deposits may form when climate becomes 
drier or excess sediment is supplied to caves or nearby rivers 
by tributary streams, glaciers, or landsliding (Fig. 1). 


Deposits of clastic sediments usually contain multiple beds of 
unique appearances, origins, and ages. Large caves possess 
multiple generations of deposits scattered among passages for 
which the ages may differ by thousands or millions of years; 
therefore, long geomorphic records can be constructed by 
combining relatively short geomorphic records of many 
individual deposits. This requires placing deposits in a 
chronological sequence, which is typically accomplished 
through paleomagnetic, radiometric, cosmogenic isotope, 

Hjulstrom Diagram 

it Grain Size Classification 

i Erosion and Transport 

= Zone 

2 1.0 




5 S 

S 0.1 Transport Deposition 
= Zone 

0.001 0.01 0.1 1 10 100 1000 

Grain Diameter (mm) 

FIGURE 3 Hjulstrom’s diagram demonstrates the relationship between grain size and the velocity necessary to mobilize, transport, and deposit grains. Grains 
will remain in motion at velocities below those required to mobilize them; however, silt and clay are generally so small that they will remain suspended in water 
at extremely slow water velocities (left side). In contrast, coarse grains such as boulders can only be mobilized and transported under the highest velocities. The 
differences in mobilization and transport velocities cause sediments to be sorted by size. 

and related dating techniques. Sediment ages define a narra- 
tive sequence of cave development, but substantive scientific 
investigations do more than reconstruct cave histories. 
Changes in cave sedimentation patterns and materials reflect 
large-scale changes in earth system processes, which are of 
great interest to geologists, archeologists, climatologists, and 
biologists. These changes are brought about by changes in 
regional driving forces imparted by climate and tectonics, 
which generate local driving forces that influence sediment 
production, transport, and deposition (Fig. 1). 

Cave sediments generally adhere to the laws and precepts 
of traditional stratigraphic studies. The laws of superposition, 
original horizontality, and cross-cutting relationships are 
violated in unique ways, but not so frequently as to prevent 
the application of standard interpretation techniques. Key 
considerations are to be aware of complicating phenomena 
and to carefully scrutinize deposits for anomalies. Violations 
of the law of superposition are probably the most common 
problems encountered. The law of superposition states that 
vertically stacked sediments become younger with increasing 
height. The law is commonly violated in caves when cave 
streams carve openings beneath sediments and the opening is 
later filled by younger sediment. 

Sedimentary Facies 

The word facies encompasses an incredible diversity of con- 
cepts used to correlate, differentiate, and interpret sediments 
and rocks. Facies are widely used tools for interpreting 
depositional chronologies and reconstructing geomorphic 
histories. A sedimentary facies is a distinctly unique body of 
sediment that can be identified on the basis of appearance, 
composition, texture, or sediment sizes. Common types of 
facies that are useful for clastic cave sediments include 
depofacies, lithofacies, and textural facies. A depofacies 
possesses unique sedimentary structures and stratigraphic 
successions for which the origins are derived from deposi- 
tional processes. For instance, imbricated gravels are charac- 
teristic of bedload deposition within a streambed. Individual 
lithofacies possess unique mineralogy or rock fragments. For 
instance, granite cobbles carried from headwater streams 
form identifiable deposits in limestone caves. The granite- 
bearing deposits are different from deposits with no granite 
that are created by streams draining non-granitic terrain 
overlying the cave and can be used to identify stream 
provenance. Textural facies are comprised of beds with 
unique sedimentary structures. 

Identifying unique depofacies can be difficult where 
multiple processes create similar deposits. A widely identified 
depofacies is the abandonment suite, which is comprised of 
stream sediments that become finer toward the top of a 
section. Active stream tiers commonly carry ample bedload, 
which accumulates to varying depths on passage floors. A 
passage ceases to receive bedload when lower level tiers form. 
Floodwater velocities and depths in the older, higher tier 

Clastic Sediments in Caves 105 

gradually decrease as the lower passage incises downward or 
is itself abandoned. Lower velocities translate to finer 
sediments in suspension; hence, a typical abandonment 
succession from bottom to top is gravel, sand, silt, clayey silt. 
Abandonment depofacies have been recognized throughout 
the world, but they are similar to aggradational depofacies, 
which form by completely different processes. 

A cave stream will aggrade or build up the streambed if a 
downstream component is obstructed by collapse or 
sediment accumulates in downstream channels. The driving 
force of stream gradient decreases when the downstream bed 
rises and stream velocities and sediment transport rates 
decrease to cause sediment accumulation. The decrease in 
velocities is, in turn, responsible for a decrease in mean 
particle size as sediment accumulates; deposits become finer 
upward just as they do in abandonment suites. Extensive 
ageradational deposits in Mammoth Cave in Kentucky have 
been traced to episodes of cooler climate, whereas abandon- 
ment suites are the result of normal passage abandonment. 
Discerning the different origins of these fining-upward 
sequences requires the use of secondary information. For 
instance, bed aggradation may cause sediment on streambeds 
to touch cave ceilings. If sediment shields part of a ceiling 
from chemical attack, the protected area will project down- 
ward into accumulating sediments as a roof pendant. 
Chemical attack may eventually disconnect the pendant 
from the ceiling and leave it intact within the surrounding 
sediments as convincing evidence of aggradation as opposed 
to abandonment. Multiple lines of evidence should be used 
when defining facies for use in scientific investigations. 


Scientists examine clastic cave sediments to determine: (1) 
depositional environment, (2) sediment origins, (3) long-term 
deposition or erosion trends, (4) relationships of sediments 
to cave or landscape development, and (5) relationships of 
fossils or artifacts to cave processes. Each of these topics is 
interrelated with the others. Geomorphologists use knowl- 
edge of depositional histories to decipher how caves and 
landscapes have behaved as functions of tectonism and 
climate. But, archeologists and paleontologists may desire 
knowledge of how certain objects came to rest in a cave, 
whether objects have been disturbed, how objects relate to 
one another, and how objects have been altered since 
placement. The diversity of sediments and their uses means 
that research goals vary markedly from one study to another, 
but all cave studies share a commonality in that subterranean 
research techniques are modified versions of techniques 
developed for surface deposits and landforms. Analyses begin 
with sediment description. The investigative process can be 
illustrated by examining representative clastic cave sediments 
and their interpretation. 

106 Clastic Sediments in Caves 

Representative Clastic Sediments and Analyses 

Thick deposits of sediments accumulate in cave passages 
when a cave stream becomes incapable of transmitting 
available sediment. Cave streams may experience sympathetic 
aggradation if the stream it drains to rises or cave passages are 
blocked. Stream gradients decrease in both cases because 
raising the downstream channel elevation decreases stream 
gradient, which decreases stream velocity. However, 
sediments may also accumulate in abandoned, upper level 
stream passages by passive infilling during floods. The 
sedimentary package of clays, silts, and sands shown in Fig. 4 
was deposited following emplacement of a log jam 300 m 
inside a cave. Initially, the channel bed was composed of 
course sandstone cobbles carried from nearby mountain 
slopes. The obstruction significantly decreased upstream 
floodwater velocities and induced deposition of fine-grained 
sediments throughout the 300-m-long passage. Multiple 
point bars of fine sand and coarse silt record channel 
aggradation, as opposed to passive infilling. Laminae within 
individual beds have variable inclinations, and inclined scour 
surfaces record bank erosion between episodes of point bar 
deposition. Deposits formed by passive infilling are typically 
more homogenous and lack cross-cutting, inset channels 
because there is no through-flowing stream to cut and fill 

Base-level streams are large streams fed by tributary 
streams such as caves. Large cave systems are often associated 
with base-level streams, and the most prominent example 
is the intimate association of 550 km of Mammoth Cave 
passages with the Green River of Kentucky. Incision rates 
of base-level streams are particularly valuable for studies 
of stream evolution and neotectonism. Incision rates are 

FIGURE 4 A complex packages of silts, sand, and gravel deposited upstream 
of a log jam in a West Virginia cave. The stream aggraded (built up its bed) 
behind the log jam. Aggradation, as opposed to passive infilling, is recorded 

by the inset, cross-cutting sand and gravel deposits, which are typical of 
meandering streams. The white lines identify scour and deposition surfaces 
within the deposit. 

calculated by dividing the height of sediments above a stream 
by their age. However, sediments originally deposited in 
caves beneath a surface stream yield only a minimum incision 
rate; the total amount of downcutting is greater than the 
height of the cave above the present stream because the 
surface stream had to incise some unknown distance down to 
the cave and then the additional vertical distance seen today. 
Similarly, sediments deposited above the target stream yield 
a maximum bound. Sediments deposited above base 
level, such as those shown in Fig. 4, can be recognized by 
sedimentary structures typical of gravity flow and 
intermittent exposure. Useful sedimentary evidence includes 
multiple generations of mud cracks, desiccation chips, 
and evidence of meandering with cut-and-fill structures. 
Sediments deposited beneath base level or in phreatic caves 
will not display evidence of multiple wetting and drying 
episodes (e.g, lack mudcracks between individual beds). 
Phreatic sediments may be extremely coarse if the passage 
acted as stream conduit but may also be extremely fine 
grained with well-defined laminea (Fig. 5). 

Flood histories and short-term changes in river behavior 
are recorded in some caves close to large rivers. Caves in 
riverbanks may fill with sediment and sympathetically 
agerade if the nearby riverbed rises in response to climate 
changes. The small cave shown in Fig. 6 is almost entirely 
filled with laminated red clays deposited when the nearby 
river was choked with sediment produced during a glacial 
epoch. The clays accumulated in the cave, which was 
temporarily in the phreatic zone, but are now being eroded 
by occasional floods created by the same river that deposited 
them. The more recent floods have left thin deposits of 
silt and sand throughout the cave, which can be recognized 
by their brown colorations, mudcracks, cross-bedding, and 

. Ses a -"_ a) 

FIGURE 5 Laminated clays deposited in a large cave that formed well below 
the water table in New Mexico. Cave development was aided by sulfuric 
acid, which altered some of the phreatic sediments to white clay known as 
endellite (top). (Photograph courtesy of Art Palmer.) 

GS-99-29 L 
Ocm, ,—1985 or 1996 a GS-99-29 

40 | 
| A 
7 : snail shell 
140 | Ocm = 
160 a! ) 


Clastic Sediments in Caves 107 




15cm4 | 

breakdown Ee 

Symbol Key 

Pisolith-like balls of silt 
Rip-up clasts 

Flood deposits (sand) 

Flood deposits (silt, sand) 
Quietwater deposits (clay, silt) 

FIGURE 6 Plan view of a small, riverside cave filled with sediments deposited by a West Virginia river. Stratigraphic columns include a thick deposit of red 
clay (partially shown in GS-99-29 and GS-01-004), interpreted to have been deposited during aggradation of the river. Floodwaters are now occasionally 
refilling the cave and eroding the aggradational sediments while depositing thin veneers of sand and silt. The flood sediments can be used to infer the frequency 

of large floods along the river. 

organic deposits. Both depofacies (specifically, phreatic 
versus flood facies) provide useful information about the 
short-term history of the river. 

The types of deposits being creating in cave passages will 
change in response to many variables. Recognizing facies 
differences allows inferences to be drawn about geomorphic 
and climatic histories. A common facies succession is the 
transition from traction to gravity sediments. As shown in 
Fig. 7, massive deposits of breakdown and collapse debris 
may overlie well-sorted and bedded stream sediments. Such 
facies successions are especially common in cave entrances 
wherein stream sediments may represent warm periods 
between cooler climates. The breakdown and collapse debris 
are created during cooler climates by frost wedging. 
Archeologists often find that artifacts are preferentially found 
in association with particular facies in cave entrances. 

FIGURE 7 Stream sediments beneath breakdown and _ frost-wedging 
deposits. The stream sediments reflect an episode during which streams 
washed sediment into the cave via a nearby entrance. Subsequently, frost 
wedging and collapse have covered the fluvial sediments with a deposit of 
unsorted sediment (top of column). The deposit contains objects of 
archeological significance, hence the excavation and tags. (Photograph 
courtesy of Art Palmer.) 

108 Closed Depressions 


Campy, M. and J. Chaline (1993) Missing records and depositional breaks 
in French Late Pleistocene cave sediments. Quater. Res. 40: 318-331. 

Courty, M.A., P. Goldberg, and R. I., Macphail (1989) Soil Micromorpho- 
logy in Archaeology, Cambridge University Press, London, 344 pp. 

Ford, D.C. and PW. Williams (1989) Karst Geomorphology and Hydrology, 
Unwin Hyman, London, 601 pp. 

Granger, D.E., D. Fabel, and A.N. Palmer (2001) Pliocene—Pleistocene 
incision of the Green River, Kentucky, from radioactive decay of 
cosmogenic ?6Al and '°Be in Mammoth Cave sediments. Geol. Soc. Am. 
Bull. 113(7): 825-836. 

Sherwood, S.C. and P. Goldberg (2001) A geoarchaeological framework for 
the study of karstic cave sites in the Eastern Woodlands, Midcont. J. 
Archaeol. 26(2): 145-167. 

Springer, G.S., J.S. Kite, and V.A. Schmidt (1997) Cave sedimentation, 
genesis, and erosional history in the Cheat River Canyon, West Virginia. 
Geol. Soc. Am. Bull. 109(5): 524-532. 

White, E.L. and W.B. White (1968) Dynamics of sediment transport in 
caves, Bull. Natl. Speleol. Soc. 30: 115-129. 

Closed Depressions 

Ugo Sauro 
University of Padova 


Closed depressions are the dominant and most distinctive 
forms of the karst landscape. The following types of closed 
depressions can be observed in karst areas (listed according to 
increasing size): 

1. Rain pits or rain craters (1 cm to about 3 cm in size) 

2. Solutions pans or kamenitza (centimeter- to meter-scale 

3. Dolines; also called sinkholes (diameters ranging from 
some meters to over 1 km) 

4. Compound depressions (hundreds of meters to a few 

5. Poljes (reaching up to some tens of kilometers) 

Dolines, compound depressions, and poljes link the 
surface and the underground fissure and cave systems. 


The solution doline (also called a sinkhole) is considered 
to be the diagnostic form of the karst landscape and is also 
defined as the index form. Dolines are the most common 
closed depressions in the Carso di Trieste, which is known as 
the classical karst because it is regarded as the most typical 
karst region. Dolines commonly have circular to subcircular 
plan geometry, and a bowl- or funnel-shaped concave profile. 

Their depths range from a few decameters to a few hundred 
meters, and their inner slopes vary from subhorizontal to 
nearly vertical. 

The Slav term doline means “depression” and in a broad 
sense includes channels and hollows of different types, such 
as fluvial valleys, dry valleys, blind valleys, poljes, and karst 
dolines. The name dolina was first applied in 1848 by Morlot 
to a circular, closed, karst depression. Subsequently, Cviji¢, 
in 1895, extended use of the term to all circular, closed 
depressions in karst areas. Gams (2000) further suggested 
that the word doline could be replaced by the term kraska, to 
highlight its importance as a morphological marker of the 
karst landscapes. 

Scientific investigation of dolines began in the middle 
of the 19th century, during construction of the Southern 
Railway in the Austro-Hungarian Empire through the Karst 
of Trieste. At that time, the most common theory was that 
dolines resulted from the collapse of cave roofs. If that were 
true, the natural hazard for a railway crossing the karst was 
high. The railway engineers, on the other hand, came to a 
different conclusion that most dolines originated through 

Dolines link the surface and the underground drainage 
systems. The absence of still water, such as lakes or ponds, at 
the bottoms of most dolines reveals that water is lost to the 
underground karst system, most commonly to caves. 
Through the sole observation of surface forms a doline seems 
very simple: It is a bowl- or funnel-like depression with one 
or a few swallowing cavities at the bottom which are covered 
by soil. Cross sections outcropping in road cuts and quarries, 
however, reveal that dolines are actually rather complex. In 
order to understand how dolines and, in general, all large, 
closed karst depressions develop, it is necessary to examine: 
(1) their morphology and size, (2) their locations and 
relationships with the topographic and geomorphologic 
settings, (3) their structures, (4) their hydrological behavior 
and related solution processes, (5) other processes that play a 
role in their evolution, (6) examples of evolution, and (7) 
peculiar morphologies that occur under specific environ- 
mental conditions. In the following sections, all of these 
aspects will be discussed. 

Morphology and Size of Dolines 

Most dolines show a circular or slightly elliptical plan 
geometry (commonly, the main axis/minor axis ratio is 
<1.5 and the depth/width ratio is about 0.1). The three- 
dimensional form of dolines may be compared to a bowl 
(nearly hemispherical depression), a funnel (nearly conical 
depression), or a well (nearly cylindrical depression). In some 
doline populations, more complex forms, such as truncated 
conical depressions with flat bottoms and star shapes, are 
common. The bowl shape is the most typical doline shape 
and is four to ten times more common than funnel-shaped 
dolines (Cvijié, 1895). The well-shaped doline is uncommon. 

Soil-covered forms show an almost circular plan geometry. 
Depressions that developed on bare rock are more irregular 
and commonly show sharp contours that follow the main 
fracture systems. 

The dolines described by Cviji¢é as “holotypes” were 
subsequently carefully analyzed by Sti8tersi¢ (1994), who 
recognized that “...the dolines ... are not as regular as many 
simple morphometric methods presume ... some of these 
irregularities are probably due to the roughness of the 
neighbouring terrain ... the others are probably due to the 
greater dynamism of the whole doline which does not permit 
the ‘bowl’ to achieve a regular shape.” Sti8terSi¢é (1994) also 
identified the presence of three distinct concentric areas 
within each single doline: (1) a flat central area, which is 
generally covered by soil; (2) a ring of steep slopes; and (3) an 
outer belt with a more gentle slope. 

A population of dolines rarely shows the entire spectrum 
of possible sizes attained by these karst depressions, because 
smaller forms are not mapped. The morphometric 
parameters of the dolines measured in the Carso di Trieste 
indicate that there is not a continuous, asymptotic size 
distribution; therefore, it is possible to recognize different 
subpopulations. The smaller forms are relatively shallow, and 
it is difficult to identify them through field observation 
alone. Within a doline population, there are a few forms that 
can be distinguished from the typical doline of the 
population because of some particular characteristics, such as 
elongation, asymmetry, steepness of the slopes, depth, etc. 
Morphometrical methods and analyses of various doline 
populations are illustrated in a number of papers (Bondesan 
et al., 1992). 

Locations and Relationships with the Topographical 
and Geomorphological Settings 

Dolines commonly occur in populations with variable 
numbers of individuals and with different densities, which 
ranging from a few to over 200 per square kilometer. Within 
a population, the density depends on the surface slope. On 
subhorizontal or gentle slopes, the density of dolines is higher 
than on steep slopes. Dolines are commonly lacking on 
very steep slopes. Two major types of karst plateaus can be 
distinguished on the basis of the spatial distribution of 
dolines: one that is characterized by isolated dolines (the 
classical karst type) and another that is characterized by 
honeycomb systems of dolines. In many karst plateaus, 
intermediate situations can be observed. With respect to 
the distribution of dolines, it is possible to recognize the 
influence of major fracture systems. The center of the doline 
is commonly set at the intersection between two or more 
fractures. Rows of dolines commonly follow fault lines, 
especially when the dislocations are linked to extensional 
tectonics that bring different rock units into contact. In karst 
plateaus where a net of dry valleys can be distinguished, 
dolines are aligned along the valley bottoms. 

Closed Depressions 109 

The Structure of Dolines 

In the bare karst, such as at elevations above the timberline, 
it has been observed that the structure of closed depressions 
is controlled by the macro- and microstructural properties of 
the soluble rocks. For example, in Monte Canin (Julian Alps, 
between Italy and Slovenia) the properties of the rock yield 
a high density of steep-sided, rocky, nival depressions, or 
Rotlict. In the typical bowl- or funnel-shaped dolines, the 
rock is normally covered by soil and other surface deposits. The 
structure of the dolines in the soil-covered karst is visible only 
through artificial cuts. Some basic characteristics of dolines 
can be inferred from the analysis of many cross-sections. In 
particular, it can be clearly observed that the doline structure 
begins to develop before the onset of the surface depression. 
Many small depressions in the hard rock, which are 
completely filled by soil and surface deposits and go 
undetected if one looks only at the surface, are visible. These 
small forms have been termed embryonic cryptodolines or 
subsoil dolines. Some of these structures host paleosols, beside 
the present-day soils. The presence of paleosols in the filling 
sediments indicates that the development of cryptoforms 
took place over a long time and that cryptodolines have acted 
as sediment traps since the beginning of their formation. In 
the early phases, the trapping action was effective for soils, 
which show less erosion than in the surrounding area in 
correspondence to the cryptodepression. When the crypto- 
dolines evolved as surface forms, the trapping action also 
became efficient for other materials, such as rock debris from 
the slopes, soil sediments, aeolian dust, volcanic ashes, etc. In 
the Mediterranean karst, these fillings are commonly red (terra 
rossa, or “red soil”), and the color is the result of weathering 
processes. In the dolines of the classical and dinaric karst, the 
filling commonly consists of a slightly reddish brown silt, 
which is primarily composed of weathered, loess-like sedi- 
ment of aeolian origin. The filling may be up to 10 m thick. 
The profile of some dolines exhibits multilayered filling. 
The sketch of the structure of an ideal middle latitude 
doline (Fig. 1) shows that the solid rock/filling materials 
interface is not a simple boundary, but rather a complex, 
involute surface that consists of many rounded bulges in the 
soluble rock separated by crevices and fissures that become 
narrower inward. These covered solution forms, called rounded 
karren, develop through the slow flow of the water solution at 
the rock/filling interface. The structures of dolines in dif- 
ferent karst areas show large variability depending on the role 
that is played by different factors, such as lithology, structural 
conditions of the rock, topography, climate conditions, and 
features inherited from previous morphodynamical events. 

Hydrological Behavior and Related Solution 
Processes in Dolines 

A simple doline is a hydrological form that can be compared 
to a first-order valley segment. Its concavity is an expression 

110 Closed Depressions 










FIGURE 1 Schematic showing the structures of a doline and a cryptodoline. The epikarst in the host rock, the fillings, and the soil are outlined. The secondary 
porosity of the soluble rock is greater at the bottom of the dolines. The main morphodynamic and hydrologic processes are also provided. 

of the convergence of water toward the bottom which is a 
transitional point between surface and underground 
hydrologic networks. The physical structure of the doline is 
the framework of its hydrological functioning and, because 
this functioning is also related to the genesis and evolution 
of the form, the doline may be considered the surface 
expression of a peculiar three dimensional hydrostructure. 
The geomorphological setting and the structure of the 
dolines allow the distinction of three major hydrostructural 
types (Fig. 2): (1) point recharge doline, (2) drawdown doline, 
and (3) inception doline. 

A point-recharge doline begins to form when a fluvial 
net loses water to the cavities of a soluble rock. It is necessary 
for protocaves to begin to form inside the rocky mass, 
connecting surface recharge points to an underground 
network. Once developed, a protocave focuses both drainage 
and solute removal at the points of transition from the 
surface to the underground network, thus causing the 
development of a surface depression. The point-recharge 
doline is commonly located along a rather dry hydrographic 
net and shows transitional features between the depression 
of a blind valley and a common drawdown doline (Fig. 2A). 
It is often elongated and skewed in the direction of the 

flow, with the upstream side being gentler than the down- 
stream side. 

The drawdown doline is the most common and typical 
karst form. To understand the drawdown dolines it is 
necessary to consider the behavior of the water in the rock 
zone just beneath a nearly horizontal or gently sloping 
surface. Near the surface, the rock fractures tend to open by 
tensional relaxation; therefore, the soil water seeps through 
the fractures and enlarges them through dissolution. With 
time, secondary porosity develops inside the shallow rock 
layer near the surface which then becomes saturated in wet 
periods. If, in some points of this layer, the water is able to 
find and enlarge routes to an underground network, flow 
paths converging to the leakage zones are thus established. 
As a consequence, the water velocity increases centripetally 
toward each leakage point, and more water molecules are 
brought into contact with the rock surface near the major 
drain. There, the rock is dissolved more rapidly, secondary 
porosity increases faster than in the peripheral zone, and the 
hydraulic conductivity increases. Consequently, a depression 
tends to form at each leakage zone, the center of which 
deepens faster than in the surrounding areas. The drawdown 
doline is, therefore, the product of the resulting positive 

Closed Depressions 111 


7 — ied _E 
f? _ - 
oa tT 7 ie i yoo Sane : =I 5 FIST = 
iss en c YO Pill a ok x 
ae ee eee eo I: > 
a, —- q 
ast AN m 
tly i: 
fF ' A 7 } a | ay) te] Ts 134 on 2 
airs ie Ot ses Ses ea 
[er z *,) ‘ : 2 = + 
Thad 2 i a 
aan “9s? + 
\\ Z py Oy. . . \ a 
oc ce tea. ; ao : { é riod = 
ii RA 1 ay, es: ft ca alt ( = 
> m i ; x 
ba he, ( iN * f. - Al \ | < ) eet \ 
wane: “ Me SKA: te 12 
g \ aa eer Prous 
Oo Ca aa a = ee | ( a oS | Roa ef at | =F | <> ae fe eae I a | Be) 
z = 2 


FIGURE 2 The three main types of solution dolines. (A) In the point-recharge doline, the strongly asymmetric hollow is linked by one main sink, fed primarily 
by surface and soil water. (B) In the drawdown doline, the depression results from focusing of the dissolution inside the water infiltration zone of the rock 
through centripetal convergence of the mainly subsurface water held inside the epikarst (z.¢., the upper zone of the soluble rock presents a greater secondary 
porosity). (C) In the inception doline, the depression also originates from a centripetal convergence of water, but this occurs inside a preexisting 
hydrogeological structure and is triggered by a change of hydraulic conductivity of the rocky mass, influenced by lithological and structural factors. 

112 Closed Depressions 

FIGURE 3 A line of funnel shaped dolines in the Monti Lessini along the 
bottom of a dry valley (Venetian Prealps, Italy). These inception dolines 

develop just above a lithology change. 

feedback resulting from the interrelations between the 
hydrological and the solute processes acting inside the outer 
layers of a fractured soluble rock. 

Gams (2000) calls the dissolution process at the central 
part of a doline accelerated corrosion. The hydrostructure 
corresponding to this central part is comparable to the 
funnel-shaped depression created by a pumping well in the 
water table of a phreatic aquifer. The outer rock layer, which 
is characterized by high secondary porosity, is shallower in 
the peripheral areas and thicker in the central parts of the 
dolines. It hosts a hanging aquifer, which is nearly saturated 
during wet climate periods and nearly empty during dry 
periods. The existence of the epikarst with the characteristics 
of a water reservoir was recognized through the study of the 
hydrological regime of karst springs. 

The inception dolines develop from the interception by 
the epikarst of a hydrogeological structure formed inside 
the rock and previously triggered by a change in hydraulic 
conductivity of the rocky mass. A change of lithology or the 
presence of an impermeable layer, such as clay or chert, may 
cause the formation of a hanging aquifer. If connections 
such as fractured zones exist and allow water to overcome an 

obstacle, a hydro-structure similar to the drawdown dolines 
develops, yielding a focused dissolution both above and 
below the bottleneck. Above, the fractures of the rock are 
enlarged to fissures; below, small shafts develop. This type 
of doline differs from the drawdown doline because of the 
more marked lithological and structural control and because 
of an early evolution of the hydro-structure, which is 
not necessarily dependent on the epikarst. The models 
illustrated in Fig.2 do not consider the role of surface 
deposits or filling materials. These components are impor- 
tant, because the water is absorbed by soil, surface materials, 
and filling sediments before it reaches the fissures in soluble 
tock. The fillings may host small aquifers nested above the 
epikarst and act as pads that slowly release water downward, 
thus influencing with their physical character both the flow 
velocity at the cover—rock interface and the water regime of 
the epikarst and the main karst springs. 

Others Processes in Doline Evolution 

Even if the major process of doline formation is dissolution 
through a differential mass wasting of the rock, the following 
processes also play roles in the shaping of the depression: 
soil forming processes and other weathering processes, slope 
processes, overland flow processes, the capacity of the closed 
depression to trap different types of sediments, and the 
processes of evacuation of filling materials The soil acts as a 
filter and as an insulating layer with respect to the outer 
environment. Soil releases solutions rich in humic acids and 
air with a high content of carbon dioxide. It also yields 
fracture-filling material. 

Frost shattering of the rock is a very important weathering 
process. In the fillings of most mid-latitude dolines, a 
variable amount of angular rock debris is produced through 
this mechanism. During the cold phases of the Pleistocene, 
some dolines were completely filled with rock debris. Slope 
processes influenced by the gravitational force, such as creep 
and solifluction, are responsible for soil thinning, for the 
formation of a debris cover along the slopes, and for its 
thickening in the bottom areas. The overland flow processes 
are responsible for erosion, the washing of soils and loose 
material, and for the deposition of colluvial sediments in 
the bottom area of the depression. Once formed, a closed 
depression acts as a trap for different types of materials that 
are carried by wind, rain, etc. 

In the mid-latitude karst, the doline filling contains 
variable amounts of loess-like deposits that were transported 
by the wind during cold Pleistocene phases, volcanic ashes, 
sands, and silt deposited by rain, etc. The fine-grained 
material is commonly deposited on the entire karst surface 
and then accumulates by overland flow into the central part 
of the depression. Many authors consider the filling material 
to bea residue of limestone dissolution. Even if in some karst 
areas this may be true, most of the mid-latitude doline fillings 
consist of allochthonous materials, with the exception of 

angular rock debris due to im situ frost shattering. The filling 
material may be evacuated through the karst conduits 
through different processes, such as mass wasting of the rock 
debris by solution, subsurface flow, liquefaction of the loose 
material, or piping. 

Examples of Evolution 

Some examples will help to understand the evolution of 
dolines. Point-recharge dolines are present in the gypsum 
karst of the Santa Ninfa Plateau (Sicily, Italy). There, chains 
of dolines follow the pattern of a fluvial network that 
developed on the impermeable cover and was overprinted on 
the gypsum. Inside some chains, the bottom of the upstream 
doline is at a lower altitude than that of the progressively 
downstream doline. The last upstream doline marks the end 
of a blind valley. It is clear that each closed depression became 
inactive after the development upstream of a new swallow 
hole. So, the speed of bottom deepening is strictly linked to 
the activity of the swallow hole. 

In the Waitomo district of New Zealand, populations of 
point-recharge dolines occur in the framework of interstratal 
karstification. It is possible to reconstruct the transitions 
from a fluvial network in an impervious rock to a honeycomb 
system of dolines that developed in the underlying limestone 
units and the subsequent recovery to fluvial network in the 
impervious rock unit previously overlain by the limestones, 
following their chemical erosion. Gunn (1986) proposed 
a model based on five sample areas, each representing a 
different stage of the transition. The model outlines the 
important roles played by both allogenic recharge and the 
location of previous subterranean drainage structures. 

A small, nearly horizontal area of the classical karst, which 
was surveyed at very high resolution (scale 1:1000; contour 
interval 1 m; spot elevation, as the bottoms of some depres- 
sions, checked at the resolution of 0.1 m), is characterized by 
drawdown dolines, with some small and shallow forms that 
are difficult to see in the field (Fig. 4). Their main morpho- 
metric parameters reveal three main subpopulations: (1) 
small and shallow dolines less than 20 m in diameter and 0.4 
to 2m in depth, (2) medium sized dolines 12 to 50 m in 
diameter and 2 to 7 m in depth, and (3) large dolines, 50 to 
120 m in diameter and 8 to 15 m in depth. From the ratios 
between the numbers of individuals of each group it is 
possible to infer that most of the embryonic surface dolines 
will probably abort prior to becoming medium-sized forms. 
Their hydrological functioning is at the limit between 
triggering or not the positive feedback that would allow them 
to evolve into typical dolines. The coexistence of the three 
subpopulations also suggests different ages for the forms, 
which are probably related to the starting times of their 
evolution during favorable morpho-climatic phases. The 
smaller and shallower dolines could be the result of the karst 
morphogenesis that began at the end of the last cold period, 
which was triggered by the preexistence of cryptodolines. 

Closed Depressions 113 

Seven rock terraces were cut in the Montello (Venetian 
Prealps of Italy) neotectonic anticlinal morphostructure 
consisting of Upper Miocene conglomerate during the 
tectonic uplift. The terraces subsequently became the sites of 
doline morphogenesis (Fig. 4). Depending on the ages of the 
terraces, it is possible to analyze populations of drawdown 
dolines that developed in very similar geomorphological 
environments and now represent different evolutionary 
stages. A “standard” doline ” for each paleosurface was 
reconstructed from morphometrical analysis of the dolines of 
each terrace (with the exception of the lowest one, where no 
recognizable dolines exist) (Fig. 5). The series show that the 
standard forms are not the expression of a regular and linear 
growth, probably because of the different influence of some 
local factors on each terrace. Nevertheless, there is a general 
trend toward increasing sizes and depths of the dolines, from 
the lower and younger terraces to the upper and older ones. 
What is clear is that deepening process was faster than the 
widening process. What was increasing in a perfectly linear 
way was the sum of the volumes of the dolines for an area 
unit. The depressions tended to evolve from bowl-like to 
dish-like forms. The flat bottoms of the older and larger 
depressions result from thick fillings that formed inside the 
hollows. In the lower terraces, the dolines are isolated spots 
on a nearly flat surface, but on the upper terrace their 
boundaries are shared and a honeycomb karst morphology is 
almost reached (Fig. 4). 

In the Southern Monte Baldo (Venetian Prealps of Italy) 
are dolines that differ in size and character. Many host very 
thick fillings; some open dolines filled up to the rim look like 
amphitheaters with flat bottoms, which are open on one side. 
The fillings commonly consist of angular rock debris 
supported by a loess-like silt matrix. Some relicts of large egg- 
carton-like structures show that a population of big dolines, 
which probably developed in the late Tertiary, had been 
largely dismantled by the periglacial processes during the 
cold phases of the Pleistocene. During the Interglacials, the 
karst morphogenesis prevailed against other competitive 
processes, and so, many dolines survived up to the present 
day, even though their morphology has been strongly 

The Velebit Mountains of Croatia have different types of 
dolines. In the central plateau, the dolines are very large and 
deep and are funnel shaped, almost without covers and 
fillings and with open shafts on the slopes and at the bottom. 
In the southern plateau, the depressions are smaller and 
partially filled by rock debris and loess-like deposits. The 
main difference between the two groups is represented by 
the lithology of the host rocks: The larger rocky dolines 
developed in a massive limestone breccia characterized by 
low sensitivity to frost shattering, while the smaller and 
partially filled dolines developed in a limestone more 
susceptible to gelivation. The larger depressions remained, 
therefore, filling free, and during the winter they trapped 
large amounts of wind-transported snow. The large amount 

114 Closed Depressions 

500 m 

Classical Karst near Villa Opicina 

Montello: VII terrace 

FIGURE 4 Digital elevation model of two doline areas. (Top) Classic Karst near Opicina (Triest). (Bottom) seventh terrace in Montello Hill (Venetian Prealps). 
Morphometric parameters and distribution of dolines are more homogeneous in the Montello area than in the Opicina example. In the latter, the dolines show 
a larger variability of morphometric values. (Modified from Ferrarese and Sauro, 2000.) 

Closed Depressions 115 


FIGURE 5 Models of the standard doline for every Montello terrace, ordered from the youngest (T2) to the oldest (T7). The model is drawn using the average 
value of each morphometric parameter. The horizontal scale is the same for all drawings, while the vertical scale is slightly decreasing from T2 to T7. 

116 Closed Depressions 

of meltwaters accelerated the evolution of the forms, which 
reach huge sizes. 

In the Eyre Peninsula of South Australia, clusters and lines 
of dolines developed on a stack of fossil coastal fore-dunes 
of Middle and Late Pleistocene age, now consisting of 
calcarenites. Some dolines occur high in the local topography 
and are also aligned in groups. Their distribution is 
influenced by the diversion of groundwaters in fractures 
that cut into the pre-Pleistocene basement (granites) and the 
predominant action of solution processes in the limestone 
above such zones. This is probably related to an earlier mass 
wasting of the sand during its diagenesis in the open fissures 
of the underlying rock with a local increase of porosity of the 
newly formed rock above. The influence of a buried surface 
on the development of hydrostructures and related forms is 
referred to as underprinting by Twidale and Bourne (2000). 

The above examples demonstrate the important roles played 
by various factors and processes such as the morphostructural 
setting of the soluble rock, its lithology and density of discon- 
tinuities, its susceptibility to frost shattering, the qualities of 
the rock units above and below, the characters of the erosional 
surface on which the karst forms began to develop, the morpho— 
climatic system and its relative changes, and the occurrence 
of such events as falling volcanic ashes, among others. 

Populations of Dolines Linked to Specific 
Environmental Conditions. 

In humid tropical climates, some populations of closed 
depressions show peculiar characteristics that distinguish 
them from doline populations of the mid-latitude karst. In 
particular, the closed depressions are often larger than typical 
dolines and in the topographical maps show a star-shaped 
figure. The cartographic representation of a tropical karst 
landscape with closed depressions resembles a classic karst 
landscape with dolines, but with an inverted relief. In fact, 
most of the rounded, concentric figures drawn by the map 
contour lines are conical or tower-like hills encircling the 
concentric star-shaped contours of a closed depression 
(Fig. 6). The egg-carton shape of the basin is characterized 
by a great difference between minimum and maximum 
depths (differences in elevation between the bottom of the 
depression and the lowest and highest points of the 
perimeter). This type of tropical doline, called a cockpit from 
their Jamaican name, has very steep slopes, usually from 30 
to 40°, and is coated by a thin soil cover. Often, the rock 
outcropping on the steeper slopes shows evidence of 
biokarstic weathering, such as honeycomb alveolar cavities, 
and locally of carbonate deposition. At the bottom, the 
fillings are shallow, and, sometimes it is possible to observe a 
flat, sponge-like rocky bottom, blackened by biokarstic 
weathering, and a sharp change of gradient at the contact 
with the surrounding slopes. During the wet season, the 
shallow holes at the bottom of some depressions behave like 
springs, and temporary ponds may form. 

The morphological differences between the normal 
solution dolines and the tropical cockpits are mainly due to 
the presence of the valley-like depressions entrenched on the 
slopes of the cockpits, which determine the star-like form of 
these sinks. These incisions in the slopes, even if similar to 
some morphological types of the fluvial landscapes, are 
solution forms. Their presence in the tropical dolines, and 
not in the middle-latitude dolines, is due to such different 
factors as: (1) the larger dimensions of the slopes of the 
tropical dolines, (2) the poorer development and role of the 
epikarst, (3) the minor influence of limestone debris and soil 
cover, and (4) the more active overland flow and biokarstic 
weathering that act on the slopes. 

Similar to tropical dolines, some populations of mid- 
latitude dolines, such as those of the coastal belt of Dalmatia 
(Croatia), may be interpreted as inherited forms, which 
first developed in subtropical humid conditions during 
the late Tertiary. The survival of inherited forms was 
possible where the dismantling processes competing with the 
solution processes, such as frost shattering, have played a 
secondary morphogenetic role due to both peculiar environ- 
mental conditions and a low susceptibility of the rock to 
gelifraction. Also similar to the tropical dolines, some very 
large dolines of the Central Velebit plateau can be explained 
as old, inherited forms that were not considerably altered by 
the periglacial processes of the cold Pleistocene phases 
because they developed in a limestone breccia resistant to 
frost shattering, 

Summary of Solution Dolines 

The solution doline may be considered the most typical karst 
geo-ecosystem that links the surface and the underground 
networks. A doline is both a three-dimensional form, and a 
multicomponent geosystem consisting of several components 
of variable thickness, such as soil, surface deposits, fillings, 
and epikarst. If the development of a doline is triggered by 
a positive feedback in the synergical interactions of 
hydrological and solute processes, other cooperative and 
competitive processes may also play an important role in its 
evolution. In particular, the soil plays a cooperative function 
by enriching the water with carbon dioxide. The fillings, 
on the contrary, commonly act as a competitive factor by 
slowing water seepage into the epikarst and, if carbonates are 
present in their composition, by increasing the water pH, 
thus decreasing potential dissolution of the underlying solid 
rock. Without the presence of soil cover it is not possible 
for a doline to attain a typical bowl- or funnel-like shape, as 
documented by the irregular depressions of the bare high 
mountain karst. Soil and others surface deposits, therefore, 
play a fundamental role in the evolution of the forms. 
Dissolution is the dominant morphogenetic process in the 
cockpits of the tropical karst. If the soil cover is very thin on 
the steep slopes, it can be replaced by a rock layer that is 
strongly corroded by biological processes. 

Closed Depressions 117 

FIGURE 6 Sketch of a tropical cockpit. The star-shaped depression is encircled by conical hills separated by dry valleys and gaps. The planar shape of the 

watershed line is polygonal. 


In the karst areas it is possible to find other types of dolines 
that do not directly originate in the solution of rock at or 
near the surface. The main types are (1) collapse dolines, (2) 
subsidence dolines, (3) intersection dolines, and (4) cover 

dolines (Fig. 7). 

Collapse Dolines 

Collapse dolines result from the collapse of a cave. They show 
different shapes, with vertical or overhanging walls. Collapse 
dolines that allow entry into a subterranean system are also 
called karst windows, because they open into the under- 
ground environments (Fig. 8). The best known karst 
windows are the dolines of Skocjanske Jame in Slovenia 
(Fig. 9), where the Reka River enters into a large cave feeding 
the aquifer of the classical karst and the spring of Timavo 
40 km away. 

The range of sizes of the collapse dolines varies from a 
few meters to hundreds of meters. The Velika dolina (large 
doline) of Skocjanske is about 500 m in diameter and 164 m 
in depth. The Blue Lake and Red Lake dolines of Imotski 
(Croatia) are among the most spectacular collapse dolines in 
the world (Fig. 10). The main axis of the Blue Lake doline is 
about 1.5 km; the diameter of the Red Lake doline is about 
400 m at the lake surface, and its total depth, including the 
submerged part, is 520 m. These very large forms are not 
the result of a unique collapse episode, but of a number of 
collapses related to oscillations of the water table. Most of the 
collapse dolines do not provide access to a subterranean 
system because of collapse debris at the cavity bottom. 

A particular type of collapse dolines are the cenotes, open 
depressions in the coastal karst belts giving access to the 
water table and resulting from the collapse of the roof of 
submerged cavities. The best-known cenotes are in Yucatan 
and in the Caribbean Islands. The evolution of cenotes and 
of the related cave systems is often due to both efficient 

118 Closed Depressions 








FIGURE7 The four main types of dolines that do not directly originate by solution of rock at or near the surface: (A) collapse doline, resulting from the 

collapse of the roof of a cave; (B) subsidence doline, formed by the settling of an insoluble rock following solution of the underlying soluble rock; (C) 

intersection doline, originating from the emptying of fillings of an old fossil cave due to intersection with the topographical surface; (D) cover doline, which 
has developed in incoherent rocks burying a soluble rock or partially filling a karst depression. 

dissolution by the brackish waters of the coastal belts and 
oscillations of the water table. Some cenote populations 
could be interpreted to be intersection dolines. 

Subsidence Dolines 

Subsidence dolines are closed depressions caused by the sub- 
siding of an area. They may be due to upward propagation 
of an initial collapse of the roof of a deep cavity with 
development of a breccia-pipe structure; when this structure 
reaches the surface a closed depression with an unstable 
bottom is formed. Subsidence closed basins are also found on 
insoluble rocks overlying soluble rocks. If the underlying 
rocks are eroded by solution processes, the rocks above 
subside. These hollows, which may host lakes, are frequent 
on rocks overlying very soluble minerals, such as gypsum 
and salt. 

Intersection Dolines 

Intersection dolines are depressions formed when an old 
cave system, partially or totally filled with sediments, is cut 
by the topographic surface as a consequence of the lowering 
of such a surface by chemical erosion. The opening of these 
fossil caves reactivates these old hydrostructures, leading 
to evacuation of the fillings and development of closed 
depressions. These forms are relatively common in the 

classical karst, where chains of depressions or slender 
depressions also some hundred meters long exist. On the 
bottom of such a doline, it is easy to find relicts of cave 
fillings and stalagmites. The recent building of a highway has 
cut many of these forms and they have been labeled roofless 
caves by Mihevc (2001); here, the term intersection dolines is 
preferred because they are no longer underground forms. 

Cover Dolines 

Cover dolines are the closed depressions formed in 
incoherent materials such as alluvial deposits, glacial drift, 
superficial deposits, soil sediments, etc. These forms, also 
commonly called alluvial dolines, have different character- 
istics according to the type, continuity, and thickness of the 
deposits and the presence of a well-developed karst relief 
under the deposits or of a still incipient karst hydrology, 
among other factors. The flat bottoms of large solution 
dolines, consisting of filling deposits, and the alluvial plains 
of a large river may be considered as end members of the 
various development environments for these dolines. The 
influence of buried karst hydrostructures on the formation of 
cover dolines is a clear example of underprinting. Various 
processes may play a role in the evolution of these forms, 
such as water infiltration, washing and piping, suffosion (in 
the sense of erosion from below), suction of the sediments 
caused by oscillations in the water table, liquefaction of the 

Closed Depressions 119 

FIGURE 8 A karst window in the Rakov Skocijan cave system in Slovenia. 
In the foreground, the natural arch represents a remnant of the cave roof. 

sediments and the linked mud flows, and changes in volume 
caused by freeze-thaw cycles. 

The two main models for the development of cover 
dolines are (1) upward migration of a cavity, and (2) sub- 
sidence, as determined by the gradual rearrangement (a sort 
of inner creeping) of the cover material induced by its 
mass wasting through the karst net. If the cover material 
is sufficiently cohesive, the mass wasting by the water 
infiltrating underground produces a cavity with an arched 
roof; the arch gradually migrates upward, where it reaches the 
surface and causes the sudden opening of an ephemeral cover 
doline with nearly vertical walls. The subsidence cover 
dolines form in less cohesive sediments. 

In addition to the forms of dolines described above are 
other types originating by different processes, such as dolines 
along fault lines caused by seismotectonic movements and 
anthropogenic dolines. Also, craters similar to dolines were 
created by shelling during World War I in the chalky 
limestone of the Venetian Pre-Alps and are an interesting 
anthropogenic type. The crater-like depressions behave as 
drawdown dolines because of the strongly fractured rock 
fracturing caused by the explosions. 

FIGURE 9 The karst windows on the first part of the subterranean course of 
the Reka river in the Skocjanske Jame system (Slovenia). 


Both compound, and polygenetic closed depressions can be 
observed in the karst landscapes. The large closed depressions 
that do not show a doline morphology are also referred to as 
uvala. A compound hollow is a form that originated from the 
fusion of more simple forms. If more dolines coalesce 
together, an irregular depression develops, sometimes with a 
lobate perimeter. A polygenetic sink is a closed depression 
that clearly evolved through both the karst process and 
another morphogenetic process. The most frequent types are 
the (1) tecto-karstic hollows, (2) the fluvio-karstic hollows, and 
(3) the glacio-karstic hollows. 

A tecto-karstic hollow is a closed basin that developed 
inside a tectonic depression that also evolved through karst 
processes. A fluvio-karstic hollow may be considered the 
closed ends of blind valleys that evolved through both fluvial 
and karstic processes. Glacio-karstic basins are common 
forms in the alpine high mountain karst. These forms were 
formed by both karst processes and glacial abrasion and are 
of two main types: (1) the high plateau type, often elongated 
and similar to a bathtub, and (2) the glacial cirque type, 

120 Closed Depressions 

The depth from the rim to the lake is over 250 m, and the depth of the lake 
itself is greater than 250 m. 

which developed in the bottom of a glacial cirque. Larger 
glacio-karstic depressions can be over 1 km long. Their 
bottoms are often occupied by rock drumlins and till 


Poljes are the largest closed depressions observed in the karst 
terrains. In 1895, Cviji¢ defined a polje as a “large karst 
depression, with a wide, flat and nearly horizontal floor, 
completely enclosed between steep slopes.” This simple 
definition is insufficient to explain the genesis of this type of 
form and the role of the karst processes in its character- 
ization. To understand this form it is necessary, as for dolines, 
to consider both the structure of the form and its dynamic— 
in other words, to consider the entire geosystem expressed by 
the form. Poljes present a large variability, and the 
development of most of them cannot be explained through 

the karst process only, because they are polygenetic forms, 
resulting from the combination of a number of processes. 

Typical of this form are the poljes of the dinaric karst, 
which show large, flat, and nearly horizontal floors, from one 
to some tens of kilometers long. The sides, usually sloping at 
about 30°, connect with the bottoms at a sharp angle. Closed 
depressions of this type are also present in others karst 
regions of the world and are referred to by various names, 
such as campo or piano in Italy, plans in France, and hojos 
in Cuba. 

The main environmental peculiarity of this type of form is 
the common absence of a permanent lake inside the basin; 
an ephemeral lake without surface outlets may form and 
disappear during the seasonal cycle in relation to the precipi- 
tation regime. The lake, if present, represents the transition 
point between the surface and the underground hydrology 
and serves as a window to the underground aquifer. In other 
words, the polje floor is affected by the oscillations of the 
local water table. Although the lake may be fed by both 
surface runoff and underground circulation, the draining of 
it occurs exclusively through the subterranean karst network 
(Fig. 11). The dynamics of the lower part of the basin is 
strictly linked with the seasonal cycle of water input, 
throughput, and output. The temporary lake is responsible 
for the solution of both the floor and the base of the slopes, 
leading to the planation of the bottom and to its lateral 
enlargement by marginal corrosion. The forming, standing, 
and dissipation of the water body is also the cause of a 
redistribution and leveling of the filling materials and of their 
volumetric reduction by solution and erosion via the 
subterranean circulation system. 

In the structure of the poljes it is possible to recognize a 
large variability in the thickness of the filling deposits, 
ranging from a few decimeters to more than 100 m. Swallow- 
ing cavities and springs may open at both the base of the 
slopes and inside the floor area. Of these, some are real caves, 
and others are covered dolines developed in the fillings above 
cavities in the underlying rock. From the hydrological point 
of view, some behave permanently as springs or as swallowing 
cavities, and others may invert their functioning from springs 
to swallow holes and vice versa. This last type is referred to as 
estavelle in France. 

The classification of poljes may be based on both the 
geomorphological characters and the hydrodynamics. Gams 
(1994) used these two criteria to distinguish five main types 
of poljes: (1) border poljes, (2) piedmont poljes, (3) peripheral 
poljes, (4) overflow poljes, and (5) base-level poljes. A border 
polje is located at the transition between a non-karstic and a 
karstic area. The recharge of the basin is, therefore, mainly 
due to surface streams and the escape of water through 
stream-sinks. A piedmont polje is located downslope of a 
mountain area from where a large amount of debris has been 
received, filling the depression and hosting a local aquifer. 
Here, the recharge is also in the main made by surface 

Closed Depressions 121 

streams. A peripheral polje is a depression fed by a large 
internal area of impermeable rocks with a centrifugal stream 
network. The sinks are located around the periphery of the 
inlier. An overflow polje is underlain by a belt of relatively 
impermeable rocks that act as a hydrological barrier to water 
that emerges at springs on one side of the polje floor and 
escapes via stream-sinks on the other side of the basin. A 
base-level polje is a polje with a floor cut entirely across karst 
rocks and is affected by the vertical oscillations of the water 
table; consequently, it is inundated during high-level periods. 

Based on the earlier work of Ford and Williams (1989), it 
is possible to reduce the categories to three basic types: (1) 
border poljes, (2) structural poljes, and (3) base-level poljes. 
From the geomorphological point of view, most poljes 
correspond to tectonic depressions such as graben, fault- 
angle depressions, pull-apart basins, etc. These dislocations 
often lead to contact between rocks with different perme- 
abilities, thus leading to conditions of lithological contact. 
Many structural poljes have also been described as litholo- 
gical contact poljes (Fig. 12), a situation corresponding, at 
least partially, to that of the border poljes. 

Some structural poljes have trapped hundreds of meters of 
sediments hosting local aquifers. An example is the Piano del 
Fucino in the Central Apennines (Italy) the bottom of which, 
during most of the Holocene, was occupied by a lake subject 
to numerous vertical variations in the level. In early Roman 
and modern times tunnels were excavated to drain the lake 
and to reclaim the floor for agriculture. Today, the intensive 
agriculture is supported by overexploitation of the alluvial 

FIGURE 11 A ponor (swallow hole) in the Cernicko polje (Slovenia). The 
function of the gate is to block coarse debris (in particular, tree trunks and 
branches) that could impede the discharge of the water. 

Limestone-shaley PLATEAU 


: : i dolines 
= floor of the Oe ici = ' 
S inside ee ones 
‘ ' 


pure reef-limestone 0 1 km 

FIGURE 12 The structural polje of Piano del Cansiglio in the Venetian Prealps (Italy). The polje is now inactive and hosts smaller depressions nested inside 
the main floor. This form has been described as a lithological contact polje. 

122 Coastal Caves 

Some poljes have developed along typical fluvial valleys. A 
specific example is represented by Popovo polje in Yugoslavia, 
which is 40 km long and only 1 to 5 km wide. The floor 
slopes gently throughout its length as demonstrated. by the 
course of the river Trebinjcica. Before the construction of a 
draining tunnel, the bottom was occupied from October to 
May by a seasonal lake. 

The best examples of true karst poljes can be found in 
tropical karst regions, such as in Southern China. Here, the 
floors of certain populations of large cockpits are in the 
vertical oscillation zone of the water table. The base-level 
floors of these cockpits expand by marginal corrosion, 
leading to the gradual dismantling of the ridges and to fusion 
of the floors. The resulting compound depressions represent 
true karst poljes, probably a unique type of polje created by 
karst processes only. 

See Also the Following Articles 
Soil Piping and Sinkhole Failures 


Beck, B.E (1984) Sinkholes: Their Geology, 
Environmental Impact. A.A. Balkema, Rotterdam. 

Bondesan, A., M. Meneghel, and U. Sauro (1992) Morphometric analysis of 
dolines. Int. J. Speleol. 21(1-4), 1-55. 

Cvijié, J. (1985) Karst. Geografska Monografija, Beograd. 

Ferrarese, F and U. Sauro (2001) Le doline: aspetti evolutivi di forme 

Engineering and 

carsiche emblematiche [The doline: evolution aspects of the emblematic 
karst form]. Le Grotte d'Italia V(2): 25-38. 

Ford D. and PW. Williams (1989) Karst Geomorphology and Hydrology. 
Unwin Hyman, London, 601 pp. 

Gams, I. (2000) Doline morphogenetical processes from global and local 
viewpoints. Acta Carsologica 29(2): 123-138. 

Gams, I. (1994) Types of poljes in Slovenia, their inondation and land use. 
Acta Carsologica 23: 285-300. 

Gunn, J. (1986) Solute processes and karst landforms. In: Solute Processes. 
(S.T. Trudgil, Ed.) John Wiley & Sons, New York, pp. 363-437. 

Klimchouk, A., D.C. Ford, A.N. Palmer, and W. Dreybrodt, Eds. (2000) 
Speleogenesis and Evolution of Karst Aquifer. National Speleological 
Society, Huntsville, AL. 

Miheve A. (2001) Speleogeneza divaSkega Krasa [The speleogensis of the 
Divaca Karst], in Slovenian. Zalozba ZRC, ZRC SAZU, Ljubljana 27: 

Nicod, J. (1975) Corrosion de tipe crypto-karstique dans le karst 
méditerranéen. Bull. Ass. Geogr. Fr. 428: 284-297. 

Sauro, U. (1995) Highlights on doline evolution. In Environmental Effects 
on Karst Terrains, Vol. 34 (I. Barany-Kevei, Ed.). Universitatis of 
Szegediensis, Szeged, pp. 107-121. 

SuSterig, E (1994) Classic dolines of classical sites. Acta Carsologica 
23: 123-156. 

Twidale, C.R. and J.A. Bourne (2000) Dolines of the Pleistocene dune 
calcarenite terrain of western Eyre Peninsula, South Australia: a reflection 
of underprinting? Geomorphology 33: 89-105. 

White, W.B. (1988) Geomorphology and Hydrology of Carbonate Terrains. 
Oxford University Press, London, p. 464. 

Williams, PW. (1985) Subcutaneous hydrology and the development of 
doline and cockpit karst. Zeitschrift fiir Geomorphologie, 29: 

Coastal Caves 

John E. Mylroie 
Mississippi State University 


Caves that form in coastal environments will be controlled by 
factors that separate them from caves that form in traditional 
inland settings. The first and most obvious factor will be the 
physical and chemical power of waves and saltwater acting on 
coastal rocks. Second, and less obvious, is the fact that sea 
level can change, and with that change the position of the 
coastline moves. Therefore, the positions of cave develop- 
ment by coastal processes will also move. Sea level can change 
in a variety of ways, but there are two ways that are of 
particular importance to cave formation on coasts. Sea level 
can change on a global scale, in which case it is referred to as 
a eustatic sea-level change. The most common way for this 
change to occur is by changing the amount of ice on the 
continents. During the last 1.6 million years, called the 
Pleistocene Epoch, the Earth has undergone a series of ice 
advances called glaciations (the ice ages of the popular press) 
and a series of retreats called interglacials. As ice sheets grow, 
evaporated seawater falling as snow on land is trapped as ice, 
and the sea level drops worldwide. When the ice melts as an 
interglacial occurs, sea level rises as the meltwater flows back 
into the ocean basins. A eustatic sea-level change of this type 
is glacioeustatic. Evidence indicates that the Earth has gone 
through at least 15 of these glacial cycles (and, hence, sea 
level changes) in the Pleistocene. Changes in sea level can also 
occur as isolated events at specific locations because the land 
is either subsiding or being uplifted. Such a sea-level change 
is considered a /ocal sea-level change, as only that local area is 
affected, and it is commonly caused by tectonic movements 

of the Earth. 


The most common of the coastal caves are sea caves, and they 
are found the world over. The precise scientific name is 
littoral cave, meaning a cave formed within the range of tides. 
Sea caves are caves that form by wave erosion in coastal areas 
that contain exposed bedrock. They can develop in almost 
any type of bedrock, with wave energy utilizing fractures and 
other preexisting weaknesses in the rock to quarry out voids 
by mechanical action. The chemical action of saltwater can 
also exploit rock weakness. The compression of air caused by 
water flowing forcefully into cracks and fractures in the rock 
can break rock, including rock above sea level. Sea caves can 
vary from small voids only a few meters across to very large 
chambers up to 100 m deep and wide (Fig. 1). The sea caves 
seen on coastlines today have formed rapidly, as sea level has 

FIGURE 1 Looking out a sea cave on Eleuthera Island, Bahamas. 

only been at its present elevation for perhaps 3000 to 5000 
years, following melting of the large continental glaciers 
of the last glaciation, which ended about 10,000 to 12,000 
years ago. In areas such as Norway, where the Earth’s crust 
was depressed by large masses of ice during glaciation, the 
shoreline is now rising as the crust rebounds to a stable 
position following melting of the ice. In so doing, sea caves 
formed many thousands of years ago are now found high 
above modern sea level. 

The Earth is currently in an interglacial (between glacia- 
tions), so glacial ice is at a minimum and sea level is high. 
The last interglacial occurred 131,000 to 119,000 years ago. 
During that time, the ice melted back a bit more than present 
conditions, and sea level was about 6 m higher than it is 
today. On some rocky coasts, sea caves produced at that time 
are still visible, 6 m above the ocean, if more recent erosion 
has not obliterated them. 

Sea caves have had a long history of interaction with 
people, especially wherever sailors have used the ocean on 
rocky coastlines. Sea caves were particularly favored by 
smugglers to hide stolen goods and also to hide the small, 
fast sailing ships that carried such cargo out of reach of the 
taxman. Pirates allegedly buried treasure in sea caves, but 
most sea caves are in an active erosional environment, and 
anything buried would not survive long. Pirates who chose 
old sea caves above modern sea level would have had better 
success. When sea level was lower, sea caves that had formed 
before the sea level fell were left abandoned far from shore. 
Some sea caves in this situation contain significant 
archeological remains. 

Sea caves are ubiquitous on the rocky coasts of the world. 
Fingal’s Cave in Scotland, the Blue Grotto of Capri in the 
Mediterranean, Sea Lion Cave on the coast of Oregon, and 
Arcadia Cave on the coast of Maine are well-known sea caves 
visited by tourists on a regular basis. Many organisms use 
sea caves as a refuge, particularly seals, sea lions, and other 
marine mammals, as well as birds, which roost in the ceiling 

Coastal Caves 123 

ledges above the reach of waves. From the viewpoint of cave 
exploration, sea caves are not of major interest, primarily 
because they are short in length. In areas where other types of 
caves are rare, such as in southern California, sea caves offer 
the best cave exploration option. Exploration of sea caves can 
be very dangerous for those not experienced in handling 
strong waves and currents. 


The coastal environment creates a very unusual type of cave 
when limestones are present. The interaction of freshwater 
and seawater produces a unique geochemical situation that 
allows caves to form by dissolution that are very different 
from both sea caves (made by mechanical wave action) and 
limestone caves of the interior of continents (which are under- 
ground stream conduits formed by freshwater dissolution). 
Freshwater is slightly less dense than seawater, because of 
the extra salt dissolved in seawater. Average freshwater has 
a density of 1.0 g/cm®; average seawater has a density of 
1.025 g/cm®. The difference in density is only 1 part in 40, 
but it is sufficient such that when freshwater flows toward 
the ocean inside an aquifer it floats on top of the seawater 
that has invaded the aquifer from the ocean. If it is a sharp 
boundary, the boundary between the freshwater and 
saltwater is called the Aalocline (from halo meaning “salt” and 
cline meaning “boundary”). If the boundary is broad, 
containing water of brackish salinity, it is called a mixing 
zone. The freshwater flows toward the ocean because rainfall 
infiltrates the land behind the coast, piling up in the aquifer 
until there is sufficient slope to drive the water toward the 
ocean. Where the freshwater is piled up inland, because of 
buoyancy the halocline sinks downward into the seawater, 
much like a piece of wood floating in water. Because the 
difference in density is 1 part in 40, for each 1 m that the 
freshwater piles up above sea level, it sinks 40 m into the sea 
water. As the freshwater flows down this small 1-m-high 
slope toward the ocean, its elevation above sea level decreases, 
and in buoyant response the halocline rises up toward sea 
level, a 40-cm rise for each 1 cm of elevation loss of the water 
table. At the coast, the freshwater discharges to the sea as a 
thin sheet. This configuration of freshwater over seawater is 
called the freshwater lens, because when seen in cross section 
at an island (where the water discharges to coasts on either 
side), the freshwater body is seen to have the shape of a lens, 
similar to a lens in a magnifying glass. Figure 2 shows this 
relationship, with vertical exaggeration. Understanding the 
freshwater lens is critical to successful exploitation of fresh 
groundwater in island and coastal areas. 

When the freshwater lens is formed within a limestone 
aquifer in a coastal region, a unique type of cave, called a 
flank margin cave, can develop. Seawater is usually saturated 
with CaCO3, the mineral that makes up limestone (CaCO, 
can be calcite or a slightly different form, or polymorph, 
called aragonite), and cannot dissolve limestone very well. 

124 Coastal Caves 




C= a re 








FIGURE 2 Diagram of a freshwater lens in a limestone island. The lens is drawn with 10 x vertical exaggeration. The term epikarst describes the dissolutional 

forms on the limestone land surface; autogenic recharge means all the water entering the limestone came as precipitation directly on the limestone, not as flow 

from adjacent non-limestone areas. 

Freshwater that has had a long residence time in a limestone 
aquifer is also commonly saturated with CaCO; and also 
cannot dissolve more limestone. However, because the 
freshwater and seawater became saturated with CaCO; under 
different initial conditions, when they mix they are capable 
of undergoing more dissolution, a process called mixing 
corrosion or mixing dissolution. Where the freshwater lens 
meets the seawater at the halocline, it is possible to dissolve 
out large voids, or caves, that otherwise would not be able 
to form. 

The top of the freshwater lens is also a place where waters 
can mix. In this case, the freshwater at and below the water 
table (in this case, the top of the lens) is called phreatic water, 
and the water descending from the ground surface above is 
called vadose water. It is common for both the phreatic and 
vadose waters to be saturated with respect to CaCO3, but as 
with the case of mixing seawater and freshwater, the phreatic 
and vadose waters saturated at different initial conditions so 
that when they mix the water can dissolve more CaCO3. 
Therefore, both the top and bottom of the freshwater lens are 
a favorable environment for the dissolution of CaCO . 

The top of the lens (the water table) and the bottom of the 
lens (the halocline) represent density interfaces. Organic 
particulate material transported by the vadose water flow 
from the land surface commonly floats on the top of the 
water table. Some of this organic material may then become 
water logged and work its way to the bottom of the lens, 
where it floats on denser seawater at the halocline. The decay 
of the organic matter at these interfaces creates CO), which 
dissolves in the water to make carbonic acid, which promotes 
CaCO; dissolution. In both situations, if the amount of 
organic material becomes too great, its decay will use up the 
local oxygen supply to create anoxic conditions. If the anoxic 
conditions persist, anaerobic bacteria will create H,S, which 
can later encounter water with oxygen in it to create H,;SO,, 

or sulfuric acid, a very powerful acid that can dissolve even 
more CaCO3. 

The mixing and possible anoxic conditions that promote 
dissolution can occur at the top and bottom of the lens 
throughout the lens area; however, these environments are 
superimposed on each other at the edge, or margin, of the 
freshwater lens where the top of the lens slopes down to sea 
level and the bottom of the lens rises up to sea level. Because 
the lens is thin at this point, its flow velocity is high because 
the entire lens discharge is being forced through a thinning 
wedge. The combination of increased flow and superposition 
of favorable geochemical environments for CaCO dissolu- 
tion results in large voids forming very rapidly at the margin 
of the lens. As the discharge of the lens occurs at the margin 
of the lens, at the flank of the land, the caves thus developed 
are called flank margin caves. In addition, in tropical islands 
and coasts, the limestone is commonly very young. Unlike 
the ancient limestones of the mid-continent regions, these 
young limestones have not been buried, squeezed, or greatly 
altered. The rocks still have a high degree of primary porosity, 
and water moves through them easily. Such rock is said to be 
eogenetic. True conduit flow is difficult to develop, as the rock 
has a high degree of permeability, meaning that the water has 
many flow routes to choose from. So, how do caves form in 
this setting? 

Flank margin caves are not true conduits, such as caves 
formed by sinking streams in continental interiors. They are 
instead mixing chambers. Conduit caves form by turbulent 
water flow, but flank margin caves develop in the laminar 
(nonturbulent) flow of a porous, eogenetic limestone aquifer. 
The freshwater enters the developing flank margin cave as 
diffuse flow and exits, after mixing with seawater, as diffuse 
flow. Flank margin caves form without human-sized 
entrances. As a result of their nonconduit origin, flank 
margin caves do not have long tunnels or the dendritic 

pattern found in most conduit caves. The flank margin caves 
are a series of oval rooms that tend to be extensive in the 
horizontal direction, but limited in the vertical direction, a 
result of developing in the thinning margin of the freshwater 
lens. The chambers can connect in a somewhat random 
manner, creating caves that are unpredictable in their 
pattern. Maze-like areas are common, indicating regions 
where chamber development did not go to completion. As 
the caves were growing, the mixing zone advanced into them, 
such that the back wall of the cave (the wall farthest from the 
ocean) is the youngest. Complex cross-connections between 
chambers can develop, and the caves can be quite intricate 
despite their simple mode of development. The typical flank 
margin cave consists of one or more large chambers located 
just inside the edge of the island. The cave may trend parallel 
to the coast for some distance, but it rarely penetrates very far 
inland, as its development site was restricted to the margin of 
the lens (Fig. 3). The caves form entrances when the erosion 
of the hillside that contains them breaches into the under- 
lying cave. Initially, this entrance may be a small opening, 
but through time it can enlarge as more of the outer wall of 
the cave erodes away. Throughout the Bahamas and other 
carbonate island areas, flank margin caves in all states of 
erosional destruction, from almost intact to almost entirely 



‘Total Surveyed Cuve- 9.32 kilometers! 23.7 meters depth 

Main Entrances Se 

Suunto and Tape Survey by 
the 1998 Isla de Mona Expedition 
Marc Obms Mike Lace 
Chris Beck Myk Coughlin 
Richard Gantt John Myiroie 
Erik Myiroie Joan Mylraie 
Leif Mylraie Lars Mylroie 
Jake Turin Jon Smith 
Joe Troester Pat Kambesis 
Ron Richards Ramon Carrasquillo 

Cartography by Marc J. Ohms sous) 


‘Titaade Renee 

Coastal Caves 125 

removed, can be seen. On islands such as Bermuda, where 
weathering and erosion of the limestone hills is more rapid 
than in the Bahamas, most flank margin caves have been 
entirely removed. 

Because they develop in the freshwater lens, flank margin 
caves are sensitive to sea-level change. If the sea level falls, 
the caves will be drained as the freshwater lens follows the 
sea level downward, and cave development will cease. If the 
sea level goes up, the lens will be pushed higher, the caves will 
become flooded with pure sea water from below, and 
dissolution and cave enlargement will stop. 

The Bahama Islands are tectonically stable, meaning they 
are not rising or falling because of tectonic forces. The dry 
flank margin caves that explorers enter there today resulted 
from the freshwater lens being 6m higher than today 
131,000 to 119,000 years ago, during the last interglacial. 
That sea-level highstand lasted only about 12,000 years, but 
caves with individual chambers up to 14,000 m? in volume 
developed in that time frame, indicating how rapidly this 
mixing dissolution process can occur. On stable limestone 
coasts around the world are many flank margin caves that 
developed during the last interglacial sea-level highstand. 

Isla de Mona, halfway between Puerto Rico and the 
Dominican Republic, is a small island that has been uplifted 

Connection to Cucva al lado del Faro 

Caribbean Sea 
Breakdown- 00 Dirt- a 
Teeioa Ceiling Height- @) Drop-off 
Ceiling Change- + 4 Dripline- ~--< 
Colama- 5 Stalagmite- a 
af Q Stalactite- Rimstone- <> 
Cg MBS) 3 
Fai ofthe Cathartoral Rivown Flowstone- =“ Sod: Tv 
Upper Level Passage- 2? 

FIGURE 3 Map of Cueva del Lirio on Isla de Mona, Puerto Rico, a very large flank margin cave. (Cartography by Marc Ohms.) 

126 Coastal Caves 

by tectonics. It has huge flank margin caves that formed 
almost 2 million years ago (Fig. 3). These caves are very big 
because they developed in a freshwater lens before the 
Pleistocene glaciations began and so had a longer time to 
dissolve before sea-level fluctuation caused the freshwater 
lens to change position. When the initial glaciation began 
early in the Pleistocene, these caves were drained and speleo- 
thems (e.g., stalactites and stalagmites) developed. When a 
subsequent interglacial sea-level highstand occurred, the sea 
level rose and the cave was invaded once again by the fresh- 
water lens, partially dissolving the speleothems produced 
during the dry phase. The caves were then uplifted by 
tectonics well beyond any further glacioeustatic sea-level 
changes and have been preserved for exploration today. In 
this case, local sea-level change affected only this one island. 
Some of the cave chambers are over 400,000 m? in volume, 
much wider than they are high, but with many complex 
connections with adjacent chambers. The chambers have 
ancient speleothems that are much modified by attack from 
an invading freshwater lens, as well as more modern 
speleothems that have grown since the last uplift event and 
are pristine in condition. The longest flank margin cave in 
the world, Cueva del Lirio, is located on Isla de Mona, and 
has over 19 km of survey on its map (Fig. 3). Flank margin 
caves commonly have numerous entrances, and in tropical 
settings are warm and friendly. The pleasant conditions and 
many entrances make them easy to explore. When movies 
and television dramas show pirates in caves, it seems they 
always have enough light and they move easily through large, 
open passages. Flank margin caves actually do look a bit like 
this fictional characterization. 


Blue holes are names for large, deep pits that form on islands 
and lagoons in tropical waters. They are named blue holes 
because their great depth gives them a very deep, dark blue 
color. They commonly connect to cave systems at depth. The 
name blue hole was first published in 1725 and later appeared 
on British Admiralty charts from the Bahama Islands in the 
1840s. Blue holes became popularized in the early 1970s 
when cave divers began to make the first serious investiga- 
tions of their depths. Since that time, blue holes have been the 
subject of a number of major scientific investigations, includ- 
ing the discovery in 1979 of a new class of Crustaceans. 
Blue holes are defined as “subsurface voids that are 
developed in carbonate banks and islands; are open to the 
Earth’s surface; contain tidally influenced waters of fresh, 
marine, or mixed chemistry; extend below sea level for a 
majority of their depth; and may provide access to 
submerged cave passages” (Mylroie et al. 1995, p. 225). Blue 
holes can be additionally characterized as being found in two 
settings: (1) ocean holes, which open directly into the present 
marine environment and contain marine water, usually with 

tidal flow; and (2) inland blue holes, which are isolated by 
present topography from marine conditions, open directly 
onto the land surface or into an isolated pond or lake, and 
contain tidally influenced water of a variety of chemistries 
from fresh to marine. 

While blue holes are best known from the Bahama Islands, 
they are found in a wide variety of tropical coasts and islands. 
Their origin is tied to the coast, island, and lagoon 
environments where they are found, which means that their 
development has been influenced by glacioeustatic sea-level 
changes of the Pleistocene. Blue holes commonly contain 
many stalactites and stalagmites (known as speleothems) that 
are now under water, indicating that the blue holes were 
drained by glacioeustatic sea-level lowstands, which allowed 
the speleothems to form from dripping vadose water. They 
were then flooded by the return of a higher sea level as the ice 
sheets melted on the continents at the end of the last 
glaciation. Some of these speleothems are more than 350,000 
years old, indicating that the blue holes containing them are 
very old and have undergone repetitive sea-level lowstands 
and highstands. 

The four main hypotheses regarding how blue holes form 
are (1) drowning of surface karst features such as pits and 
sinkholes, (2) phreatic dissolution along a descending 
halocline, (3) collapse of deep-seated phreatic dissolution 
voids, and (4) bank margin fracturing. Blue holes come in 
a variety of morphologies and may represent features of 
polygenetic (many origins) development, in which case a 
combination of the above hypotheses may be correct. 

Exploration of blue holes generally involves cave diving 
to great depths. Such exploration is at the leading edge of 
technology and stamina, requiring the use of mixed gases, 
long decompression stops, total darkness, the danger of silt- 
out (stirring up silt so that the way out cannot be seen), and 
tight passages. Cave diving in blue holes is extremely 
dangerous, and many cave divers and scientists have lost their 
lives trying to penetrate into the unknown. Unlike many 
other types of exploration, however, there is no substitute for 
direct human exploration. The blue hole depths and their 
caves, and the contents of those caves, cannot be viewed, 
measured, or sampled without someone going in them. 


Coastal caves are important to science as they contain 
information about present and past sea-level conditions. 
Their utilization as a habitat makes them important for 
many organisms over the breadth of the animal kingdom. 
The speleothems contained within blue holes and flank 
margin caves contain within their layers evidence of changes 
in the Earth’s climate over hundreds of thousands of years. 
While flank margin caves are generally easy to explore, sea 
caves can be dangerous for the unwary, and blue holes are 
exceptionally dangerous even for the well trained. 

Contamination of Cave Waters by Heavy Metals 127 


Bunnell, D. (1988) Sea Caves of Santa Cruz Island. McNally & Loftin, Santa 
Barbara, CA, 123 pp. 

Mylroie, J.E., J.L. Carew, and A.I. Moore (1995) Blue Holes: Definition and 
Genesis. Carbonates Evaporites 10(2): 225-233. 

Mylroie, J.E. and J.L. Carew (1995) Karst development on carbonate 
islands. In Unconformities and Porosity in Carbonate Strata (D.A. Budd, 
PM. Harris, and A. Saller, Eds.). Memoir 63, American Association of 
Petroleum Geologists, Tulsa, OK, pp. 55-76. 

Mylroie, J.E., J.W. Jenson, D. Taborosi, J.M.U. Jocson, D.T. Vann, and C. 
Wexel (2001) Karst features of Guam in terms of a general model of 
carbonate island karst. 7. Cave Karst Stud. 63(1): 9-22. 

Palmer, R. (1997) Deep into Blue Holes. Media Publishing, Nassau, Bahamas, 
188 pp. 

Vacher, H.L. (1988) Dupuit-Ghyben—Herzberg analysis of strip-island 
lenses. Geol. Soc. Am. Bull. 100: 580-591. 

Contamination of Cave 

Waters by Heavy Metals 

Dorothy J. Vesper 
West Virginia University 

Hz’ metals are ubiquitous throughout nature, including 
within caves and karst environments. Evaluating the 
accumulation and transport of metals in cave waters requires 
understanding the governing physical and chemical 
processes. While the presence of heavy metals in speleothems 
and cave deposits has been investigated in some detail, the 
general metal cycling through the karst system is less well 
known but can be inferred from analogous investigations in 
surface systems. The term heavy metals is poorly defined and 
has been used inconsistently through time and in the scien- 
tific literature. In the context of this discussion, the metals 
and metalloids discussed are those defined as potentially toxic 
by the U.S. Environmental Protection Agency and the World 
Health Organization (Table I). It should be noted that many 
of these metals, while toxic in large quantities, are essential 
nutrients in small quantities. 


Metals are omnipresent in atmospheric, marine, and terre- 
strial settings. In karst environments, they are most likely to 
be found in three primary compartments: soils, the matrix— 
fracture—conduit system, and springs (Fig. 1). Within 
conduits and caves, heavy-metal rich minerals may be 
found as speleothems, coatings, fillings, rinds, and other cave 
deposits. Additionally, both caves and springs may have 
metals present in water or associated with suspended and bed 

TABLEI Potentially Toxic Metals and Their Abbreviations 
Metal Abbreviation Metal Abbreviation 
Arsenic As Nickel Ni 
Beryllium Be Lead Pb 
Cadmium Cd Antimony Sb 
Chromium Cr Scandium Sc 

Cobalt Co Selenium Se 

Copper Cu Titanium Ti 

Iron Fe Thallium Tl 

Mercury Hg Vanadium Vv 
Molybdenum Mo Zinc Zn 
Manganese Mn Nickel Ni 

Metals may be part of the natural background or 
anthropogenic (Table II). Spectacular deposits of metal-rich 
speleothems can occur when caves exist in proximity to 
natural geologic sources. For example, Cupp—Coutunn Cave 
in Turkemenistan has speleothems rich in manganese, iron, 
lead, and zinc due to the presence of overlying bituminous 
coal and subsequent hydrothermal alternation. Mbobo 
Mkula Cave in South Africa also boasts unusual metal-rich 
speleothems thanks to the presence of overlying ore minerals 
and a sulfide-rich black shale. The host rock itself may 
contribute to the metal load. Low levels of trace metals in 
springs waters in Nevada and California have been attributed 
to the paleomarine chemistry at the time of carbonate 
deposition. Given that the groundwater feeding the springs 
was thousands of years old, the water chemistry was 
attributed to dissolution of the carbonates. Metals may also 
be also found as detrital material within the host rock. The 
manganese source for Jewel Cave in South Dakota has been 
attributed to such detritus. 

Anthropogenic sources of metals are widespread (Table II). 
Sources may be diffuse, such as emissions from fuel com- 
bustion, or more localized, such as point-source discharges 
from manufacturing facilities. Acid mine drainage (AMD) 
is a common metal source in many karst regions (the 
Appalachians, Kentucky, and Tennessee in the United States 
and in southern China). Studies of metals in AMD in 
Tennessee have shown that Fe and Mn concentrations 
decrease in water when introduced to karst systems. The 
most likely explanation is that metals precipitate as alkalinity 
increases and the pH of the combined AMD-karst solution 
rises. In cave stream/spring water, the drop in metal concen- 
trations is associated with the production of flocs of Fe and 
Mn hydroxides. There is some suggestion that the metals 
may precipitate as a metal armor on the host rock, thereby 
limiting the interaction between water and rock. Manufac- 
turing sources also contribute to metal contamination. Horse 
Cave in Kentucky and its associated aquifer and springs have 
been impacted by discharge from a metal-plating factory that 
began operations in 1970. Although some dilution occurred, 

128 Contamination of Cave Waters by Heavy Metals 

—_— Cave Deposits wieiet 
ondul Suspended 
Cave Water eee 
System Suspgndes sg sediments e@ @, © G6 


FIGURE 1 Locations for heavy metal storage in karst settings. 
TABLE II Sources of Heavy Metals 

Type of Source Examples of Specific Sources 

Natural Original marine deposition; detrital materials in 
bedrock; ore bodies; hydrothermal deposits; black 
shales; coals 

Anthropogenic Mining and mineral processing; agriculture (fertilizer, 

pesticides, preservatives, irrigation); emissions and 
solid wastes from fossil fuel combustion; sewage and 
solid wastes; manufacturing (metallurgical, 
electronic, ceramic, chemical, pharmaceutical); sports 
and military shooting; breakdown of metal alloys 
and paints 

elevated concentrations of Cr, Ni, and Cu were found in 
both cave and spring waters within the Hidden River basin. 

General Chemistry 

Once metals are introduced into karst settings, their storage 
and transport depend on physical processes, the specific 
metal chemistry, and the chemistry of the surrounding 
environment. Metal speciation influences solubility and the 
likely mode of transport through the aquifer. Metal specia- 
tion also controls the bioavailability and toxicity of the metal. 
Metal chemistry can be complex and depends on many 
competing variables. Although a brief description is provided 
herein, the reader is referred to the bibliography, which lists 
a few of the excellent texts on the topic. 

Soil and Bedrock Aquifer 

Bed necienesiisii 

Bed sediment 

In general, metals are present in three forms: mineral, 
otherwise associated with solids, or dissolved (Fig. 2). 
Mineral-bound metals exist as source materials in soils in 
bedrock, as secondary cave deposits, and as detrital material 
throughout karst systems. Heavy metals may also be asso- 
ciated with the surfaces of solids and are often associated with 
suspended and bed sediments. Metal interactions at solid 
surfaces range in intensity from exchangeable (loosely 
bound) to incorporation into the near-surface mineralogy 
(tightly bound). Metals may also be incorporated into surface 
coatings (both organic and inorganic) or attached to 
insoluble organic matter. It is difficult to distinguish between 
metals that are electrostatically bound to the surface, 
specifically adsorbed, or coprecipitated with the surface 
mineral or coating; therefore, they are often loosely referred 
to in combination as sorbed metals. Dissolved metals are 
present in water as free metals or soluble metal complexes. 
Metals may also form complexes with soluble or colloidal 
organic compounds and thereby increase their solubility 
dramatically. Copper, in particular, has a strong affinity for 
organic compounds. Metals may move between these 
compartments via chemical reactions such as dissolution, 
sorption, or desorption or via physical processes such as 
deposition and entrainment (Fig. 2). 

Two key variables that control metal solubility and 
chemical form are pH and reduction—oxidation (redox) state. 
Many metals are more soluble in acidic waters than in neutral 
waters and thus tend to precipitate in well-buffered karst 
waters. Redox reactions occur when electrons are transferred 
between metal species. For example, ferric iron (Fe**) is 

Contamination of Cave Waters by Heavy Metals 129 

Bedrock and 








Bed Sediments 

FIGURE 2 Storage compartments for heavy metals in karst settings. Arrows indicate some of the chemical reactions and physical processes that transfer metals 

between compartments. 

reduced to ferrous iron (Fe”*) by the addition of one 
negatively charged electron. In actuality, free electrons do not 
exist, and the reaction occurs by combination with another 
“half reaction” in which an electron is lost. Redox reactions 
are highly important because they control the solubility of 
many metals. While some metals are mobile in reducing 
conditions (As, Mn, Fe, Mo) others are mobile in oxidizing 
conditions and immobile in reducing environments (Zn, Cu 
Hg, due to sulfide mineralization in reduced settings). For 
many metals, both the pH and the redox conditions must be 
known to predict what species should be present. 

Redox states change spatially and through time. Spatial 
changes may occur on a microscale, and temporal changes 
may occur on a scale as short as hourly (¢.g., with storms and 
changing hydraulic conditions). The degree to which metal 
redox reactions in caves are mediated and catalyzed by micro- 
organisms is a topic of increasing interest. Microbes can be 
either electron acceptors or electron donors and can therefore 
either oxidize or reduce metals as part of their metabolic 
process. Hence, both precipitation and dissolution of metals 
can be induced by microbial action. Much of the microbial 
action may occur along redox gradients, such as those found 
at the edges of caves. 

Other factors may influence the redox state, such as intro- 
duction of oxygen during a storm and degradation of organic 
compounds. The breakdown of either natural organic matter 
or organic contaminants in the overlying soil, karst aquifer, 
or spring sediments can create locally reducing and acidic 
environments capable of solubilizing and mobilizing metals. 

Iron and Manganese 

Iron and manganese can be found in many locations in caves. 
While oxide and hydroxide forms are the most common, 
heavy-metal carbonates, phosphates, nitrates, and sulfates 
have also been identified in caves. In general, while reduced 

iron (Fe”*) is soluble in water, when it oxides to Fe** it forms 
oxide and hydroxide precipitates such as goethite (FEOOH) 
and ferrihydrite (Fe(OH)3). Limonite is a generic term that 
refers to both mineral and amorphous forms of iron oxides 
and hydroxides. Likewise, reduced manganese (Mn**) is 
soluble, but its oxidized forms are not. Fully oxidized 
Mn often forms the common cave mineral birnessite, which 
can be poorly crystallized (6-MnO,). In reality, however, the 
chemistry is far more complicated than this. The reactions 
also depend on metal concentration, solution pH, CO, 
partial pressure, and presence of organic compounds. Both 
reduced and oxidized metals can be mobilized by organic 
complexation, and reduced metals can precipitate into 
metal—carbonate minerals. 

Trace and Contaminant Metals 

Iron and manganese oxides and solids also play a key role in 
trace-metal chemistry because of their ability to scavenge 
trace metals from solution and because—given their relative 
concentrations—they often control the overall redox chem- 
istry of the solution or sediment. Iron oxides are commonly 
associated with As, Cu, Ni, Mn, and Zn; manganese oxides 
commonly contain Co, Fe, Ni, Pb, and Zn. Therefore, the 
transfer of Fe and Mn between mobile and immobile 
forms ultimately controls the trace-element chemistry, even 
if the trace elements are themselves not redox sensitive. For 
example, while precipitation of Fe and Mn oxides may 
remove trace and contaminants metals from solution, 
dissolution of the same oxides may rerelease trace metals back 
into solution. 

Studies of soils and marine and lake sediments have shown 
that Fe and Mn oxides are typically the controlling factor in 
determining trace-metal concentrations. Trace metals may 
also be associated with organic compounds (both soluble and 
insoluble) and inorganic coatings on particulates. 

130 Contamination of Cave Waters by Heavy Metals 


Schematic Scenarios 

Heavy metals are introduced into karst aquifers from surface 
runoff and spills, reactions in the soil zone, and dissolution 
of overlying geologic units and the host rock (Fig. 1). The 
metals can be either dissolved or associated with colloids and 
particulates (Fig. 2). Dissolved metals may be transported 
through the system or may change chemical form and preci- 
pitate as speleothems onto walls or onto sediments. The fate 
of the dissolved metals may depend on the type of recharge. 
If dissolved metals are introduced slowly from fractures in the 
cave ceiling they may—given the right chemical conditions 
—form speleothems. If the metals arrive as part of a cave 
stream, they may be more likely to be transported through 
the system or to form coatings on stream sediments. Metals 
introduced in particulate form either can travel through the 
system as suspended sediments and be discharged at a spring 
or can be deposited within the aquifer. 

Metals deposited within the fractures and conduits of the 
aquifer may be stored for extended periods. It is also possible, 
however, that they will either be later dissolved (minerals and 
coatings) or reentrained (sediments) and be flushed from the 
system. One of the final possible storage locations on the 
flowpath is spring sediments. Depending on spring 
morphology and hydrology, it is possible for some metals to 
accumulate in the springbed sediments. 

Speleothems and Cave Deposits 

Coatings of Fe and Mn are not uncommon in caves. 
Although far less common, extensive decorations incorporat- 
ing Fe, Mn, and trace metals do occur. Examples of 
manganese deposits exist as powders and coatings in Matts 
Black Cave in West Virginia and Rohrer’s Cave in Pennsyl- 
vania, as stream-cobble coatings in Butler Cave in Virginia, 
and as floor fills in Jewel Cave in South Dakota. Examples of 
iron-rich speleothems also exist in Rohrer’s Cave, which has 
stalactites, stalagmites, and columns of limonite. Spectacular 
Fe and Mn speleothems exist in Mbobo Mkula Cave in 
South Africa. The metals in this cave are mobilized when 
acidic water, produced in the overlying shales, infiltrates 
through chert layers in the limestone. Trace metals are often 
found in association with the Fe and Mn deposits. Samples 
from Rohrer’s Cave have been shown to contain up to 20% 
heavy metal oxides. Deposits of trace-metal speleothems also 
exist, although they are more rare. For example, malachite, a 
copper-carbonate, has been observed in cave crusts and 

Suspended and Bed Sediments 

Given the near-neutral pH of karst water and the input 
of oxygenated surface waters, neither iron nor manganese is 

likely to be present in appreciable concentrations as dissolved 
metals in water. This is one factor that distinguishes the 
transport of heavy metals from the more commonly studied 
alkaline earth metals (eg, Ca, Mg). The alkaline earth 
metals, present from the dissolution of the carbonate host 
rock, are almost solely present in the dissolved form. If 
dissolved heavy metals are present, they are likely to be 
oxidized and precipitated as cave deposits or directly onto 
particulates in the sediments. Once deposited in a cave or 
spring, sediment Fe and Mn may be influenced by small- 
scale transitions in redox state. This is analogous to metal 
behavior in lake and marine sediments where redox 
gradients, and the associated Fe and Mn chemistry, change 
over short vertical distances. 

The importance of particulate metal transport has long 
been established in surface-water systems where it has been 
shown that nearly all heavy metals are transported in asso- 
ciation with solids. Groundwater studies in the karst aquifer 
at the Oak Ridge National Laboratory have also shown that 
most of the metals are associated with colloidal or particulate 
matter. Metal and radionuclide transport in granular and 
fractured aquifers is often attributed to colloidal transport. 
Karst aquifers, however, are able to transport much larger 
sizes of particles; therefore, metal transport is not limited by 
the size of particulate to which the metal is adhered or 
incorporated. Data from springs in Kentucky and Tennessee, 
in which concentrations of digested and filtered samples were 
compared, demonstrated that Fe, Mn, and the trace elements 
were present in particulates larger than 0.45 Um. 

Storm-Enhanced Transport of Sediment-Associated 

The ability of the aquifer to transmit sediment is largely a 
function of flow velocity; therefore, the flushing of particles 
from the overlying soil, entrainment of sediments already 
within the aquifer, and the deposition of suspended 
sediments within conduits and in springs are all key physical 
processes controlling metal transport and are closely linked 
to the groundwater velocity. 

Groundwater velocity is not constant through time in 
karst aquifers. Systems that rapidly transmit recharge water 
exhibit increases in velocity during storm events. Recent 
work has shown that the sediment transport that occurs 
during storms controls the transport of heavy metals as well. 
Storm-event samples collected from springs in Kentucky and 
Tennessee demonstrated clearly that the heavy-metal concen- 
trations increased dramatically during storm conditions 
(Fig. 3). Data from different-sized storms and different types 
of springs in the same area suggest that this relationship is 
consistent and that total metal transport is episodic and 
enhanced by storms. 

While some “dissolved” (less than 0.45 um) heavy metals 
were present throughout the storm events, their concentra- 

Contamination of Cave Waters by Nonaqueous Phase Liquids 131 

8000 40 
60004 = 30 
c — 
= S =) 
2.4 2 =| 3 
2 40004 fe FAS 
OD a 2 = 
iL = S z 
D nO w 
p= xt 4 
20004 & 10 O 
0 0) 

Julian 1999 

FIGURE 3 Concentrations of heavy metals in spring water (Beaver Spring). 
Symbols indicate digested sample concentrations and represent the total 
metal transport (both dissolved and particulate components). (1999 data 
from Fort Campbell Army Base, Kentucky/Tennessee.) 

tions were relatively constant in comparison to the parti- 
culate metals. This suggests that “dissolved” metals transport 
through the aquifer occurs continuously at low concentra- 
tions while particulate metal occurs primarily during storms. 

Data from Matts Black Cave indicates that the distribu- 
tion of manganese coatings may also be influenced by the 
same physical flow processes. The coatings are much thicker on 
the ceilings (up to 10 mm) than on the walls (up to 2 mm) 
or along the stream (poorly coated), suggesting that metal 
deposition and stream erosion are in competition; therefore, 
the thickness and the distribution of metal coatings within 
caves may change temporally with storm events. 


Heavy metals are present throughout caves and karst systems 
due to both natural and anthropogenic sources. The 
solubility and transport of iron and manganese, two of the 
most common heavy metals, are controlled by pH and redox 
conditions. Although typically soluble in water in reducing 
conditions, the oxidized metals tend to precipitate as cave 
deposits or on sediments. While speleothems of Fe and Mn 
are unusual, coatings of the same are not. Many metals are 
never present in the dissolved state and are introduced, 
stored, and transported through the system as colloids or 
larger particles. Particulate metal transport is enhanced 
during storms when high groundwater velocities permit the 
metals to be entrained and suspended. The transport and 
storage of trace and contaminant metals are closely linked to 
the iron and manganese chemistry. Although some trace 
metals are not sensitive to redox conditions, their behavior is 
governed. by their association with iron and manganese so 
they are impacted by redox conditions. 


Allard, B. (1994) Groundwater. In Tiace Elements in Natural Waters (B. 
Salbu and E. Steinnes, Eds.). CRC Press, Boca Raton, FL, pp. 151-176. 

Hill, C.A. (1982) Origin of black deposits in caves. NSS Bull. 44: 15-19. 

Hill, C.A. and P. Forti (1997) Cave Minerals of the World 2nd Ed. National 
Speleological Society, Huntsville, AL. 

Horowitz, A.J. (1991) A Primer on Sediment—Trace Element Chemistry, 2nd 
ed. Lewis Publishers, Chelsea, MI. 

McCarthy, J.R and L. Shevenell (1998) Processes controlling colloid 
composition in a fractured and karstic aquifer in eastern Tennessee, USA. 
J. Hydrol, 206: 191-218. 

Northrup, D.E. and K.H. Lavoie (2001) Geomicrobiology of caves: a 
review. Geomicrobiol. J. 18: 199-222. 

Salomons, W. and U. Férstner (1984) Metals in the Hydrocycle. Springer- 
Verlag, Berlin. 

Sasowsky, I.D. and W.B. White (1993). Geochemistry of the Obey River 
Basin, north-central Tennessee: a case of acid mine drainage in a karst 
drainage system. /. Hydrol. 146: 29-48. 

Siegel, ER. (2002) Environmental Geochemistry of Potentially Toxic Metals. 
Springer-Verlag, Berlin. 

Vesper, D.J. and W. B. White (2003). Metal transport to karst springs 
during storm flow: an example from Fort Campbell, Kentucky/ 
Tennessee, U.S.A.. J. Hydrol. 276: 20-36. 

Contamination of Cave 

Waters by Nonaqueous 
Phase Liquids 

Caroline M. Loop 
The Pennsylvania State University 

ue to their unique properties, nonaqueous phase liquids 

(NAPLs) are widely used as solvents, insulators, and 
fuels. It is some of the same valuable properties, however, that 
make NAPLs so toxic and difficult to remove from soils and 
groundwater once they have been spilled. In the hetero- 
geneous karst subsurface, the rate of movement of NAPL 
contamination can vary by orders of magnitude. Tailoring 
a conceptual model to a particular NAPL release and sub- 
surface characterization is necessary to develop multiple 
working hypotheses and better guide detection and 
monitoring techniques. 


Four characteristic chemical properties that influence the 
behavior of NAPLs in the environment are solubility, density, 
vapor pressure, and viscosity. Nonaqueous phase liquids 
must, by definition, have limited solubility in water, allowing 



Contamination of Cave Waters by Heavy Metals 




shat DNAPL - 
Gag ees eee ee 

solubility in 

FIGURE 1 Diagram of important properties for characterizing organic contaminants. p = density (g/cm), m; = solubility of contaminant 7 in water, and P;= 

vapor pressure of contaminant 2. 

them to remain in a separate phase (Fig. 1). Thus, the 
dissolved, or aqueous phase, concentration of a nonaqueous 
phase liquid is moderate to low. Unfortunately, even the 
limited aqueous phase concentration is often much higher 
than the maximum concentration level established by the 
U.S. Environmental Protection Agency (Table I). Solubilities 
are typically less than 5000 mg/L for chlorinated solvents, 
less than 2000 mg/L for gasoline compounds, and less than 
1.0 mg/L for polychlorinated biphenols (PCBs). 

Nonaqueous phase liquids are divided between those that 
are less dense than water (LNAPLs) and those that are 
more dense than water (DNAPLs). Hydrocarbons, including 
benzene and ethylbenzene, are common LNAPLs, as are 
vinyl chloride and styrene. LNAPLs are often used in the 
synthesis of plastics, and as gasoline. With a density less than 
water, they will be present as a separate phase on top of 
the water table (Fig. 2). DNAPLs include some very toxic 
chemicals such as perchloroethylene (PCE), PCBs, and 
insecticides and herbicides such as lindane and atrazine. PCE 
and its breakdown products were commonly used as solvents, 
especially in dry cleaning. (Their use, however, is decreasing 
due to their highly toxic nature.) PCBs, for example, the 
Aroclor formulations, were commonly used as liquid 
insulators in capacitors and transformers. DNAPLs will sink 
below the water table and continue to migrate deeper in an 
aquifer until they meet sufficient resistance (Fig. 3). 

NAPLs can be characterized based on their vapor pressure, 
which is a measure of the tendency of the compound to 
evaporate from a pure liquid of the compound. NAPLs are 

either volatile, with a vapor pressure greater than 10~ atm, 
or semivolatile, with a vapor pressure between 10 and 
107! atm. Benzene has a high vapor pressure, as can be seen 
when gasoline fumes rise in the summer, whereas PCBs are 
almost wholly semivolatile. Because many NAPLs have a 
high vapor pressure, they had sometimes been thought to 
fully evaporate when poured on the ground. This belief 
turned out to be incorrect and, in fact, such practice signifi- 
cantly contaminated underlying soils and aquifers. As 
discussed later, vapor pressure may be helpful in locating 
NAPL contaminants in karst. 

Viscosity influences the subsurface mobility of non- 
aqueous phase liquids in that the less viscous the liquid, the 
further it can migrate into pores and fractures. Chlorinated 
solvents such as trichloroethylene (TCE) and PCE have very 
low viscosities. Creosote wood-treatment compounds behave 
as DNAPLs and are characterized by a high viscosity. Fluids 
can increase in viscosity with time, called weathering in the 
petroleum industry, as they lose more volatile components. It 
is easier for flowing water, such as might be found in a cave 
stream, to entrain less viscous fluids, and move them further 
from the source. 

Although petroleum compounds are formed and present 
in the subsurface naturally, they are rarely found in near- 
surface aquifers and soils unless they have been spilled. Other 
NAPLs, including PCBs, chlorinated solvents, and agricul- 
tural chemicals such as lindane and atrazine, have been devel- 
oped in laboratories and tailored to industrial uses. Freon 11, 
also known as trichlorofluoromethane, is one such chemical 

Contamination of Cave Waters by Nonaqueous Phase Liquids 133 

TABLEI Density, Vapor Pressure, and Solubility Values for Common NAPLs 

Compound Density (g/cm?) Vapor pressure (log atm) Solubility (mg/L) U.S. EPA MCL (mg/L) 
Toluene 0.86 1.42 500 1 
Benzene 0.88 0.9 1800 0.005 
o-Xylene 0.88 2.05 175 10 
Ethylbenzene 0.90 1.9 150 0.7 
Vinyl chloride 0.91 -0.53 2800 0.002 
Styrene 0.91 2.2 310 0.1 
Aroclor 1242 1.4 6.27 0.24 0.0005 
Trichloroethylene (TCE) 15 1.01 1100 0.005 
Carbon tetrachloride 1.6 0.82 760 0.005 
Tetrachloroethylene (PCE) 1.6 1.6 200 0.005 
Lindane 1:9: 7.08 Fd 0.0002 

Note: Equilibrium aqueous phase solubility concentrations are orders of magnitude larger than the U.S. Environmental Protection Agency's maximum 

concentration level (MCL). 










FIGURE 2 Distribution of LNAPLs in a hypothetical karst setting. (After Wolfe ez al., 1997.) 

134 Contamination of Cave Waters by Nonaqueous Phase Liquids 


GE? 0 i 0 oe 

A? re 

TA 0 Ge Ws I 



Water table 





FIGURE 3 Distribution of DNAPLs in a hypothetical karst setting. (From Wolfe et al, 1997.) 

that was used extensively until the toxic effects of spills and 
leaks was found to exceed its practical use. The distribution 
of NAPLs in karst soils and aquifers is nearly always the result 
of human misuse, whether by transportation accidents, 
leaking tanks, or improper disposal. 


Both the volume and rate of release can play a key role in 
the movement and distribution of NAPL contaminants. 
For example, a gasoline spill from a 10-gal tank will move 
through and be sorbed to soil and regolith differently than a 
slow leak from a 1000-gal tank that can build up a pressure 
head of product, which drives it further into the surface. On 
the extreme end of this scale is the gasoline tanker truck that 
turns over and releases thousands of gallons quickly onto the 
ground surface, which then flows overland and into a 
sinkhole or sinking stream where it can more quickly be 
transported into a bedrock aquifer. 


Climate can also impact the rate and volume of NAPL 
entering the karst subsurface. Karst soils are usually thin, and 
they overlie the regolith and irregular bedrock of the epikarst. 
In a wet climate three things may occur to move the 
NAPLs more quickly through an unsaturated soil and 
possibly epikarst zone. First, when water coats soil, organic, 
or regolith particles, if NAPL is spilled, it travels as a 
nonwetting fluid and will be held less securely between 
particle grains because it is, in a sense, lubricated by the 
aqueous phase. 

Second, and not inconsequentially, soil and regolith pores 
may periodically be flushed by precipitation. In this case, 
pressure head builds up from the accumulation of recent 
precipitation, and even nonaqueous phase contaminants 
can be forced to migrate to deeper pores. If a less permeable 
layer is present in the soil, the influx of rainwater may cause 
a NAPL that was temporarily pooled above the layer to 

Contamination of Cave Waters by Nonaqueous Phase Liquids 135 

migrate horizontally to an area where the less permeable 
layer is no longer present. At that point, the NAPL may 
resume its vertical flow toward the water table. While the 
NAPL is held in soil or regolith pores, it may dissolve into 
the aqueous phase, which may then potentially be flushed 
further through the system during precipitation events. 
Thus, NAPLs held in the soil or regolith can continue to 
dissolve and be long-term sources of aqueous phase 
contamination to an aquifer with which the NAPL is not 
in direct contact. 

A third influence of climate on the transport of NAPLs 
into karst aquifers is the impact on sinkhole formation 
and collapse. Sinkholes form localized catchments for 
storm runoff in karst regions. Some sinkholes are plugged 
with soil so that NAPLs held within them may migrate 
downward only slightly more quickly than in nonsinkhole 
soils, with the only difference being a comparatively larger 
volume of precipitation percolating through. On the 
other end of the spectrum, some sinkholes have open drains 
so that any runoff captured will be immediately carried 
into the subsurface. In some cases, however, a soil plug 
may itself be subject to piping failure, especially during 
large storms. When this occurs, both the soil and the 
NAPL are transported into the subsurface and leave an open 
hole for further contaminants to more quickly enter the 

The top of bedrock in karst is often an irregularly 
sculptured surface that may be undulating or may have deep 
crevices along joints and fractures separated by intermediate 
pinnacles. Water, moving downward from overlying soil 
into the epikarst, must often move laterally for substantial 
distances before finding an open fracture or shaft that will 
permit vertical movement into the unsaturated zone of the 
bedrock. Like ground water, infiltrating contaminants may 
be held for extended periods of time, pooling along the top 
of relatively impermeable bedrock. While still in the vadose 
zone, pools resting along the top of the bedrock can be 
described as microphase pools, which have collected in pores 
between regolith and other particles. The pools should be 
distinguished from continuous macrophase pools that may 
exist floating on water (LNAPLs) or below water (DNAPLs), 
but not distributed within pore space. 

In the event that the top of the limestone or dolomite 
bedrock is below the water table, LNAPL pools will most 
likely resist resting on the bedrock; rather they will tend to 
remain in the soil or regolith directly above the water table. 
Wells intersecting a few feet of LNAPL above the water table 
are not uncommon in contaminated areas. DNAPLs, in 
contrast, are able to migrate below the water table, and may 
pool on less permeable layers, such as the bedrock surface. In 
this case, water will flow in the direction of hydraulic 
gradient; however, DNAPLs often flow in the direction of 
bedrock dip. In one specific case, an aqueous phase plume 
developed in one direction, whereas the DNAPL itself 
migrated down dip in a different direction. 

Fractures, Open Drains, and Sinking Streams 

Karst bedrock is often dissected by vertical or near-vertical 
fractures and joints, many of which have been solutionally 
widened. These provide fast paths for NAPLs into the sub- 
surface aquifer once the critical height for entry into the 
fracture has been exceeded. The critical height of a DNAPL 
pool is a function of the interfacial tension between the 
NAPL and water, the wetting angle between the DNAPL and 
the solid surface in the presence of pore water, the density 
difference between the NAPL and water, and the fracture 
aperture. Both LNAPLs and DNALs will migrate in similar 
ways through fractures above the water table. One exception 
is that the extra density of the DNAPLs will cause them to 
move more quickly under gravity. Viscosity will also play 
an important role in the ability to move through a given 
aperture. Storm flow can, again, add to the pressure above 
the nonaqueous fluid, causing it to migrate further down the 

Once at the water table, LNAPLs and DNAPLs will 
behave very differently, with fractures acting as temporary 
traps for LNAPLs and preferential flowpaths for DNAPLs. 
In a scenario where NAPL is moving down a fracture, which 
further down intersects the water table, and then joins with 
a phreatic conduit, LNAPL will remain in the fracture above 
the water table, whereas DNAPL will continue to travel 
down the fracture into the conduit, as long as the critical 
DNAPL pool height is exceeded (Figs. 2 and 3). If drought 
conditions lowered the water table so that water flowed like 
a surface stream in the conduit, the same LNAPL could 
potentially travel down the fracture and float on the water in 
the free surface stream down the length of the conduit. When 
water level rises, the LNAPL will again get caught in a 
fracture or a sump and again be trapped. 

In this same scenario, there will be much less difference 
in the transport of the DNAPL with a variation in water 
level. Because it is denser than water, the DNAPL, with 
sufficient pool height, can move through the fracture and 
into the conduit. An air-filled fracture would likely offer less 
resistance than a water-filled one, yet in both cases, the 
DNAPL will not be held in the fracture indefinitely. In the 
event that the critical pool height is not exceeded, the NAPL 
will diffuse into the rock matrix over time. Once the NAPL 
is depleted, the contaminated aqueous phase will act to 
reverse the concentration gradient and dissolve into fresh- 
water undersaturated with the NAPL, moving through the 

Open drains, such as those sometimes found in sinkholes, 
as well as sinking streams, offer a very quick path to the 
underlying karst aquifer. NAPLs can be poured into the open 
drains. They can also either float in the case of LNAPLs, or 
be pushed along in the case of DNAPLs in surface streams. 
Sinking streams and open drains allow contaminants 
direct access to quick subsurface flow and can allow NAPLs 
to move at a rate of kilometers per hour. This can be very 

136 Contamination of Cave Waters by Nonaqueous Phase Liquids 

important to take into consideration in an area where a 
spring acts as the local water supply source. 


Once NAPLs have entered a karst aquifer, they can be a 
source of contamination for thousands of years, given their 
low solubility values and the difficulty often encountered 
when attempting to retrieve them. Natural degradation is 
faster for hydrocarbons than chlorinated compounds such as 
chlorinated solvents and PCBs, but for a large spill may take 
many decades. Little is known about the ability of microbes 
present in karst systems to degrade any of the NAPL 
compounds. Hence, it is important to consider the locations 
in which NAPL might be stored in the karst aquifer. 

Pools in Conduits 

Macrophase NAPL pools—free surface pools not confined to 
pore spaces—can be present in conduits. DNAPL pools 
could be present at the bottom of a flowing subsurface 
stream, whereas LNAPL pools would float on top of the 
water. Pools such as these could also be present in dry cave 
passages, but in all cases require a large volume of conta- 
minant, quickly injected to prevent either (1) seepage into 
sediments or (2) horizontal movement to a more permeable 
material. In turbulent flow, which characterizes most flow in 
subsurface streams, NAPL can be entrained or carried as an 
emulsion until the water velocity again decreases. Macro- 
phase pools have a low surface area-to-volume ratio, so disso- 
lution is limited. When subject to air, NAPL compounds 
may volatilize within the subsurface as occurs, for example, 
during seasonal water depressions. 


Fractures may intersect conduits in all directions. As 
discussed earlier, fractures extending from the roof of a 
conduit may act to indefinitely trap LNAPLs, but fractures in 
walls may also act to isolate NAPLs. DNAPLs are especially 
likely to migrate into subaqueous fractures in conduit floors 
(Fig. 3). Once they reach a depth where the critical pool 
height is no longer exceeded, they diffuse into the fracture 
matrix, from which they may act as a long-term source for 
aqueous phase contamination. 

Matrix and Vugs 

NAPLs can be stored in the bedrock matrix or in larger vugs. 
Diffusion of NAPLs into and out of the matrix occurs on a 
relatively slow scale. The quickest period of activity in the 
matrix may be the result of the flushing of pores during 
storm activity, as when the potentiometric surface changes 

slope from down to the conduit during baseflow to a mound 
above the conduit during intense storm flow. Storage in 
bedrock matrix and vugs may represent the portion of a 
contaminant spill most resistant to short-term remediation. 


NAPLs may sorb to or be present in pores between 
sediments. In the first case, the NAPL may travel on the 
sediment, ready to desorb once it reaches a less concentrated 
environment. In the second case, once the sediment pile is 
disrupted, the NAPL is again free to move with the water, or 
down dip in the case of DNAPLs. NAPL can be held in the 
sediment pile by capillary forces, or it may be more similar to 
a microphase pool. With storm movement, a fresh flush of 
contaminants may be remobilized when NAPL is associated 
with sediment piles in karst conduits. 

Springs and Caves 

In an area where springs and caves are accessible, monitoring 
should begin by sampling these locations during storm and 
baseflow conditions. Springs are good monitoring sites as 
they are often downgradient from spills or leaks, and if 
previous dye traces exist, they can help to define a ground- 
water basin and evaluate the risk to nearby populations. 
Spring sediments should be observed. Conductivity and 
temperature are inexpensive tools for estimating the timing 
of a spring’s response. In caves, one must be careful if NAPLs 
are a suspected contaminant. In one instance in the 1960s, a 
carbide lamp ignited a gasoline spill in a cave and killed 
several people, some by flames, but others by asphyxiation. 
Less dramatic is the instance in which a volatile NAPL 
produces toxic air within a cave. 


As mentioned previously, wells can intersect feet of LNAPL 
product above the water table. Fracture trace analysis can 
help to site wells in some instances, and wells can also be 
used for dye trace studies. In karst, wells must be carefully 
considered, especially with the slow diffusion into and out 
of the bedrock matrix as compared to the rate at which 
other transport processes may be operating in the aquifer. 
DNAPILs are likely more difficult to find using wells, because 
they may have migrated into deeper fractures in the aquifer, 
and they are harder to direct by altering the hydraulic 
gradient. With either type of NAPL, but especially with 
DNAPLs, one must be very careful to case wells properly in 
a contaminated area. An open borehole is an excellent way to 
transmit DNAPL to an underlying, formerly uncontaminated 

When a NAPL is spilled, it often passes through soil, which 

can be collected for sampling at a minimal expense. Soil 
sampling may be useful for verifying the type of NAPL and 
whether any former spills took place at the site. Soil sampling 
in a sinkhole is not recommended, due to possible piping 
failure. Soil gas sampling may be very useful, especially for 
LNAPLs. Contaminants can volatilize, with the resulting 
gas moving up through the soil. Depending on the season, 
the contaminant, and the subsurface configuration, soil gas 
sampling may be a helpful technique for identifying and 
locating a NAPL product. 


Due to the heterogeneous nature of the karst subsurface, the 
rate of either water or NAPL transport through karst aquifers 
is highly variable. Both the quantity and timing of NAPL 
releases are important for understanding how the pollutant 
might be trapped in the aquifer. Individual NAPL character- 
istics such as density, solubility, vapor pressure, and viscosity 
are also key to recovering a contaminant from soils or ground- 
water. Over time, NAPL held in the epikarst or matrix can 
dissolve into the aqueous phase. Aqueous concentrations can 
be toxic and persist for many years, especially in the case of 
chlorinated compounds, which are naturally degraded more 
slowly than hydrocarbons. However, NAPL from large spills 
may move through a conduit on the order of kilometers per 
hour. All information about a specific karst system, including 
dye traces, spring response, depth to and shape of the 
bedrock surface, is important for evaluating the potential of 
NAPLs to be held in and transported through the subsurface. 
The study of NAPL contamination in karst aquifers is a 
relatively new aspect of karst science, and in the future will 
certainly be enhanced by additional case studies and research. 


Black, D. FE. (1966). Howard’s cave disaster. National Speleological Society 
News 24, 242-244. 

Crawford, N. C., and C. S. Ulmer (1994). Hydrogeologic investigations of 
contaminant movement in karst aquifers in the vicinity of a train 
derailment near Lewisburg, Tennessee. Environmental Geology 23(1), 

Ewers, R. O., A. J. Duda, E. K. Estes, P. J. Idstein, and K. M. Johnson 
(1991). The transmission of light hydrocarbon contaminants in 
limestone (karst) aquifers. In Proceedings of the Third Conference on 
Hydrogeology, Ecology, Monitoring, and Management of Ground Water in 
Karst Terranes, Association of Ground Water Scientists and Engineers. 
National Ground Water Association, Dublin, OH. 

Jancin, M., and W. FE. Ebaugh (2002). Shallow lateral DNAPL migration 
within slightly dipping limestone, southwestern Kentucky. Engineering 
Geology 65, 141-149. 

Krothe, N. C., Y. Fei, M. R. McCann, and R. P. Cepko (1999). Poly- 
chlorinated biphenyl (PCB) contamination of a karst aquifer in an urban 
environment, central Indiana, USA. In Groundwater in the Urban 
Environment: Selected City Profiles (J. Chilton, ed.). A. A. Balkema, 


Cosmogenic Isotope Dating 137 

Loop, C. M., and W. B. White (2001). A conceptual model for DNAPL 
transport in karst ground water basins. Ground Water 39(1), 119-127. 

Mercer, J. W., and R. M. Cohen (1990). A review of immiscible fluids in 
the subsurface: Properties, models, characterization, and remediation. 
Journal of Contaminant Hydrology 6, 107-163. 

Pankow, J. E, and J. A. Cherry (1996). Dense Chlorinated Solvents and Other 
DNAPLs in Groundwater. Waterloo Press, Waterloo. 

Schwarzenbach, R. P, PR. M. Gschwend, and D. M. Imboden (1993). 
Environmental Organic Chemistry. John Wiley and Sons, New York. 

Wolfe, W. J., C. J. Haugh, A. Webbers, and T. H. Diehl (1997). Preliminary 
conceptual models of the occurrence, fate, and transport of chlorinated 
solvents in karst regions of Tennessee, U.S. Geological Survey Water- 
Resources Investigations Report 97-4097. U.S. Geological Survey, 
Reston, Va. 

Cosmogenic Isotope Dating 

Darryl E. Granger 
Purdue University 

Derek Fabel 
The Australian National University 


Natural curiosity prompts both cave explorers and first-time 
visitors to wonder “How old is this cave?” and “Why is it 
here?” Scientists have more specific reasons to study and date 
cave sediments. For example, geomorphologists use caves to 
learn about landscape evolution or about the sequence of 
events that shaped the rivers, hills, and valleys around us. 
Paleontologists study fossils in cave sediments to learn about 
animal and plant evolution and about the ecological 
communities that lived long ago. Paleoanthropologists study 
our ancestors’ bones that are found in caves—sometimes 
the bone are from cave dwellers, and sometimes they are 
from those who were eaten in the caves. These fossils and 
their dates help teach us about our own human origins. 
Archaeologists search for clues about human use of caves. 
Some scientists also study caves for their own sake, to learn 
about how water flows through rock and how the spectacular 
and labyrinthine underground environment is formed. 
Caves are important across so many fields of science 
because the conditions underground are so protected and 
stable that minerals, rocks, and fossils can be preserved in 
exquisite condition for millions of years. Sediments and 
fossils on the ground surface are gradually but constantly 
weathered and eroded as rain splashes, plants root and die, 
animals burrow, and ice crystals grow and melt. The 
landscape on the surface changes slowly but surely as 
hillslopes are worn down, rivers incise or fill their beds with 
sediment, and forests grow and recede. By contrast, caves are 
often found nearly pristine, with delicate minerals, fossils, 
and sediments still intact and unmolested by the destructive 
forces of nature above. Although the hill or mountain that a 
cave is formed in may change over time, the cave itself is 

138 Cosmogenic Isotope Dating 

contained in solid rock, so it can maintain its original shape 
until the entire mountain itself is eventually eroded away. 
Often it is not enough to simply find interesting sediments 
or fossils in a cave without knowing their age as well. Because 
knowing the age of a cave or its sediments is critical for 
learning about the past, several techniques for dating cave 
sediments and minerals have been developed. Each dating 
scheme has its own advantages and limitations. Some of the 
dating techniques such as paleomagnetism, uranium-series 
disequilibrium, and radiocarbon dating have become well 
established and widely used. This article concerns another, 
relatively new, dating technique that employs radioactive 
nuclides to date when sediment was brought into a cave. 


When attempting to date a particular cave or its contents, 
there are several possible techniques to consider, depending 
on the age and the particular fossils or minerals in question. 
Some of these dating techniques are relative, indicating 
whether one thing is older or younger than another but not 
the exact age of either. Other dating techniques are absolute, 
meaning they give a numeric age that does not depend on 
correlations with any other site. 

Relative dating techniques are based on one-way changes 
that occur over time, such as mineral weathering, sediment 
accumulation, or plant and animal evolution. For example, 
some caves in South Africa contain hominid fossils and 
artifacts that are very important for learning about hominid 
evolution, but they are difficult to date. However, the cave 
sediments also contain fossils of ancient antelope and other 
animals whose ages are approximately known from other, 
well-dated, sites. The presence of the same suite of fossils thus 
indicates that two caves are the same age, but does not reveal 
exactly what that age is. Paleontologists and anthropologists 
often use such faunal correlation as a relative dating tech- 
nique, so that when a fossil is dated at one site, the relative 
ages of the other sites can be determined. 

Absolute dating techniques, on the other hand, give 
numeric ages. Rather than relying on irreversible changes 
through time, absolute ages are defined using some sort of 
“clock” that operates at a known and constant rate. By far 
the most widely used and reliable clock is radioactive decay. 
To understand radioactive decay, it is helpful to first review 
the basic structure of the atomic nucleus. 

A nucleus is made of protons and neutrons. The number 
of protons in a nucleus determines to a large degree the way 
in which an atom behaves; in fact, the elements of the 
periodic table are defined by the number of protons they 
have. Sometimes two different atoms may have the same 
number of protons, but different numbers of neutrons. In 
this case, the atoms are of the same element, but they have 
different masses. These two atoms are called isotopes. For a 
given element, some isotopes are stable; that is, they remain 

unchanged over time. Other isotopes are radioactive, in 
which case the nucleus spontaneously breaks apart, losing 
mass and energy in a radioactive decay. The radioactive decay 
occurs at a rate that is constant for any given nuclide. If a 
certain amount of a radioactive isotope is contained in a 
rock, for example, then half of that amount will have decayed 
in a characteristic time called the Aalflife. Half of the 
remaining half will have decayed after two half-lives, and so 
forth ad infinitum. Here is the clock that we can use for 
dating radioactive materials. If the original amount of a 
radioactive isotope is known, then the remaining amount can 
be measured; the difference between original and remaining 
radioisotopes indicates the amount of time that radioactive 
decay has been occurring. The trick for cave scientists is to find 
a material with a known initial amount of radioactive isotopes. 


One way to date cave sediments is by determining the 
radioactive loss of so-called “cosmogenic” nuclides. (The 
term nuclide refers to atoms regardless of their element, 
whereas the word isotope always refers to atoms of the same 
element but different mass.) These cosmogenic nuclides are 
produced by cosmic rays—energetic particles coming from 
outer space that constantly bombard Earth. Although most 
of the cosmic rays are absorbed in the atmosphere, some of 
them reach the ground surface and cause nuclear reactions 
inside rocks and minerals found within a few meters of the 
surface. During these nuclear reactions, the nuclei inside the 
mineral grains are broken apart, forming lighter nuclides. By 
chance, some of the products of these reactions are radio- 
active. For example, we can consider reactions in the mineral 
quartz. Quartz has the chemical formula SiO, (i.e., it is made 
of silicon and oxygen in a ratio of 1:2). Silicon nuclei each 
have 14 protons, and most of them have 14 neutrons. The 
common silicon nucleus thus has a mass of 28, written **Si. 
An incoming cosmic ray particle will occasionally break apart 
a silicon nucleus. Ifa proton and a neutron are lost, then the 
28Si is converted into 7°Al. (Note: Aluminum nuclei have 13 
protons.) The fortunate thing for dating is that *°Al is radio- 
active, with a halflife of 700,000 years. Another reaction 
that occurs in quartz is the conversion of '°O to '°Be through 
the loss of 4 protons and 2 neutrons. Beryllium-10 is also 
radioactive, with a half-life of about 1.5 million years. These 
two different radioactive nuclides are produced in the same 
quartz grain, and are the key to dating sediment burial in caves. 
A diagram of the reaction producing *°Al is shown in Fig. 1. 

The cosmic rays also cause nuclear reactions inside our 
bodies, but very, very slowly so there is no significant health 
hazard to be concerned about. Rocks are exposed to cosmic 
rays for much longer than people, though, and over thousands 
of years the nuclear reactions add up to considerable quanti- 
ties of the cosmogenic nuclides *°Al and '°Be. Many repeated 
measurements of quartz grains on the ground surface 
have shown that *°Al is produced six times faster than '“Be. 


28Si(n, p2n)-SAl 

FIGURE 1 An example of a spallation-type nuclear reaction, in which an 
incoming cosmic ray neutron (left) impacts a 285i nucleus (center), to knock 
off two neutrons and a proton (right), making radioactive 26], Neutrons are 
indicated by the darker colored spheres, protons by the lighter colored 

Because both the production rates and the half-lives are 
known, the concentrations of *°Al and '“Be in rocks exposed 
to cosmic rays can be calculated in a straightforward manner. 
For most rocks at the ground surface, the 7°Al:!°Be ratio is 
6:1. If quartz is brought into a cave, though, then the grains 
are shielded from cosmic rays and 7°Al and !°Be are no longer 

After the quartz-bearing sediment is brought into the cave, 
radioactive decay gradually lowers the concentrations of 
both 7°Al and '°Be. Aluminum-26 decays faster than '°Be, so 
the original 7°AI:!°Be ratio decreases over time. After 700,000 
years half of the *°Al is gone, but only 27% of the '°Be has 
decayed. The original ratio of 6:1 has thus been lowered to 
about 4.2:1. The *°Al:'°Be ratio provides the radioactive 
clock that we can use to date cave sediments, with an original 
ratio of 6:1 that decreases exponentially over time. Figure 2 
shows the decay of the two nuclides, and the *°Al:!°Be ratio 
as a function of time. You may realize that as the 
concentrations of *°Al and '°Be get smaller and smaller over 
time, they become more difficult to measure. In fact, the 
practical limit to measurement usually occurs after about 5 
million years of burial. 

Aluminum-26 and '°Be in sediment can only be measured 
using a very sensitive technique called accelerator mass 
spectrometry (AMS). This is because the concentrations of the 
cosmogenic nuclides are extremely small. For example, only 
five atoms of '°Be may be produced in a gram of quartz in an 
entire year, and a sample may contain less than a million 
atoms of '"Be. AMS is capable of measuring an isotope ratio 
(e.g., \“Be:’Be or 26A1?7Al) as low as 10~!°. That is, if there 
are 10'° atoms of the common isotope *’Al, then AMS can 
detect a single atom of cosmogenic *°Al. You should realize 
that 10'° is a very big number; for example 10'° soccer balls 
would cover the Earth’s entire surface! 

Requirements for Burial Dating 

As with any dating technique, it is important to consider 
the circumstances for which dates will be reliable, and those 

Cosmogenic Isotope Dating 139 



= he) 


cS j = 

3 eS 

e 38 

O bem, | 

® . a 

oe) 2 

iS 1 

5S acute Tei, 

5 etches Ot te eee (6) 

) 1 2 3 4 5 
Time (millions of years) 

FIGURE 2 A graph of the concentrations of 276A] and Be (arbitrary units) 
and the 7°Al:!°Be ratio in quartz grains over time. The grains are washed into 
a cave with an original *°Al:'°Be ratio of 6. Because 7°Al decays faster than 
Be, the 7°Al:!°Be ratio decreases over time. The 7°Al:!°Be ratio can thus be 
used to date when the sediment was deposited in the cave. 

for which dates will be unreliable or impossible to obtain. 
Burial dating has several rather strict requirements. First, the 
sediment must have washed into the cave from outside, i.e., 
it must be allochthonous. Otherwise, there would be none of 
the cosmogenic nuclides to begin with. Second, the sediment 
must contain the mineral quartz, because that is the mineral 
for which we know the production rates of 7°Al and '°Be. 
Quartz is not always common in cave-forming bedrock, so 
even if there is allochthonous sediment in the cave it may not 
be datable. Third, the sample must be buried underground 
by at least 20 m for the technique to be reliable over millions 
of years. It is possible to date more shallowly buried samples, 
but this introduces higher uncertainties. Fourth, there are 
limitations on the burial times. The sediment cannot have 
been buried more than about 5 million years, or the ?°Al and 
'Be will no longer be detectable. Uncertainties in measuring 
6A and '°Be are usually 3-5%, making it difficult to achieve 
burial dates more precise than about 100,000 years. So the 
sediment must have been buried for at least 100,000 years. 
Finally, the sediment must come into the cave without a 
prior history of burial. If the sediment were, for instance, 
buried at the bottom of a sinkhole for a million years then 
washed into the cave, then the burial age would account 
for the total time spent buried, not just that in the cave. 
Although these uncertainties limit the application of burial 
dating somewhat, there are many situations for which the 
technique is ideal. We describe two examples below. 

Example 1: Dating Cave Sediments at the New River, 

The New River heads in the Blue Ridge Mountains of North 
Carolina, which are composed of igneous and metamorphic 

rocks that are devoid of caves. The bedrock contains 
abundant quartz, however, that the river washes downstream 

140 Cosmogenic Isotope Dating 

as sand, gravel, and boulders. As the river winds westward, it 
passes through the Valley and Ridge province of the 
Appalachian Mountains in southwestern Virginia, a region 
where valleys are often floored by prodigiously cave-forming 
limestone and dolomite bedrock. These caves are formed as 
rainwater drains underground from the mountains, 
eventually reaching either the New River or its tributaries 
and discharging as springs. 

Although on a human timescale rivers seem to be 
permanent and fixed features of the landscape, over geologic 
time rivers change quite rapidly. The mud, sand, and gravel 
that rivers carry downstream results from erosion, which 
constantly lowers the landscape upstream. If rivers are able to 
carry more sediment than is supplied by erosion of hills and 
mountains, then the rivers will flow over bare bedrock, and 
eventually cut gorges in valley floors. On the other hand, if 
rivers cannot carry all of the sediment that is supplied 
to them, then they will agerade, or fill the valley bottoms. 
Geomorphologists are very interested in what determines 
whether rivers incise or aggrade. Sometimes climate change 
will alter the amount of rainfall in an area, which in turn 
alters both the amount of water in the river and the amount 
of erosion on the hillslopes. At other times, tectonic uplift 
will cause rivers to incise. One way to help decipher the causes 
of river incision and aggradation is to date river incision 
events to correlate them with known times of climate change 
or uplift. 

The New River currently flows through a gorge that it has 
cut through the Valley and Ridge. The steep walls of the 
gorge are often made of cavernous limestone or dolomite. 
Canoeing or rafting down the river, one looks up to see caves 
that pock the cliffs overhead. Scuba divers find other caves 
submerged alongside the riverbed. It is no great leap to 
realize that many of the caves in the cliffs above were once 
submerged beneath the river, and that the caves were left high 
and dry as the river incised its modern gorge. In a thorough 
search of more than 50 caves found near the New River, 
Granger et al. (1997) found that 5 of these caves contained 
quartz gravel unmistakably derived from the Blue Ridge 
Mountains. The only way that this quartz could have entered 
these caves was by being washed in by the New River. This 
is a perfect situation for burial dating with cosmogenic 
nuclides. The caves are repositories of river sediment, 
recording the elevation of the old river level and shielding the 
sediment from cosmic rays. 

The five quartz-bearing caves along the New River yielded 
burial ages ranging from 0.29 + 0.18 million years for a cave 
only 12 m above the river, to 1.47 + 0.22 million years for a 
cave found 29 m above the river. Considering all of the caves 
together, Granger et al. (1997) showed that the New River 
has incised its gorge through the Valley and Ridge area in 
the past 2 million years, and that the river has incised at a 
rate between 20 and 30 m per million years. Although this 
information does not definitively state the reason for river 
incision, it does show that whatever caused the river to incise 

occurred recently in geologic time, at roughly the same time 
as the beginning of the ice ages. More work along the 
New River and other rivers in the area will be required to 
completely decipher the reasons for recent river incision. 

Example 2: The Development of Mammoth Cave, 

Mammoth Cave, the longest known cave in the world, has 
developed alongside the Green River in Kentucky. It is an 
example of a water table cave, or one that has developed 
nearly horizontal passages that are closely controlled by the 
level of groundwater flow. At Mammoth Cave, the water 
level is in turn controlled by the elevation of the Green River. 
Rainfall on the nearby Pennyroyal plateau quickly infiltrates 
the karst bedrock until it reaches cave passages that are filled 
or nearly filled with water. These underground streams then 
flow toward the Green River, passing beneath a sandstone- 
capped limestone plateau to discharge as springs on the 
Green River. In addition to the large recharge area that 
captures abundant rainfall, the sandstone-capped plateau is a 
major reason why Mammoth Cave is so long. The sandstone 
is a rock made of cemented quartz sand that is very resistant 
to erosion. Over time, as the Green River has cut through the 
sandstone and into the underlying limestones, cave passages 
have formed at successively lower levels. In most situations, 
the old cave passages above would be destroyed by erosion 
as new cave passages were formed below. However, at 
Mammoth Cave the sandstone is so resistant to erosion that 
the older passages have not eroded away. The sandstone forms 
ridges beneath that are preserved stacks of cave passages, with 
the oldest passages near the top and the youngest passages at 
modern river level. 

The preservation of old passages at Mammoth Cave 
provides a wonderful opportunity to study how the cave has 
developed and how the Green River has incised and aggraded 
over time. Geologist and hydrologist Art Palmer has spent 
many years carefully working out the sequence of events that 
are encrypted within Mammoth Cave’s passages. It is very 
difficult, though, to decipher the history of the cave without 
dates that can be used to tie passages together across the 
cave system and to match episodes of cave development with 
other geologic events. Burial dating with cosmogenic 
nuclides has provided a new set of dates that helps show how 
the development of the Mammoth Cave system has been 
strongly influenced by climate change and the growth of ice 
sheets across North America. Although ice sheets did not 
reach Mammoth Cave itself; they did impact the Green 
River, which alternately incised and aggraded, forming sets of 
passages beneath the ridges at Mammoth Cave. 

Burial dating works at Mammoth Cave because quartz 
pebbles from conglomerates within the sandstone upland are 
carried into the cave through sinking streams. These streams 
carry the pebbles through the cave system and into the Green 
River. When the river incises, however, new passages are 

formed at lower elevations. The old passages are no longer 
occupied by streams, so whatever quartz pebbles were being 
carried through the cave are left in place, to sit within the 
now-abandoned cave passages. These packages of quartz- 
bearing sediment can be found throughout nearly the entire 
cave system. It is important to realize that the sediments 
indicate not when the passage formed, but when the passage 
was abandoned. It is only through careful analysis of the cave 
that the abandonment of one passage can be linked to the 
growth of another. 

Granger et al. (2001) dated sediments from throughout 
the Mammoth Cave system. These samples reach ages up to 
3.5 million years old in the uppermost levels of the cave, and 
they tell an interesting story of how the cave developed over 
time. First, the dates show that the Mammoth Cave system 
is quite old. If the sediments that fill the cave are up to 3.5 
million years old, then the cave itself must be significantly 
older than that! The upper levels of the cave system 
substantially predate the ice ages, which began roughly 2.5 
million years ago. The cave, then, reveals how the landscape 
of central Kentucky responded to this major climate change. 
The initial response to climate change seems to be that 
the entire cave system filled up with sediment. Visitors to 
Mammoth Cave will notice telltale signs of sediment every- 
where, even in nooks and crannies on the passage ceilings. 
These sediments show that much of the cave was filled up 
about 2.4 million years ago, which indicates that the nearby 
Green River must be filled up as well. The landscape response 
to climate change was river aggradation, perhaps due to 
increased hillslope erosion that would have supplied more 
sediment than the river could carry. 

The next chapter in the story of Mammoth Cave is river 
incision and cave development at lower levels. Mammoth 
Cave is developed in levels, indicating that river incision was 
episodic. Granger et a/ (2001) found that these incision 
pulses correlate with large glaciations that covered most of 
eastern Canada and the northeastern United States, 
advancing as far as the northern edge of Kentucky. These 
large ice sheets completely reorganized river systems that 
were either buried beneath the ice or blocked by great ice 
dams. In fact, the modern courses of rivers such as the Ohio 
River, the Missouri River, and the northern Mississippi River 
were shaped along the edges of ice sheets. These ice sheets 
caused pulses of river incision that propagated southward 
into Kentucky, eventually causing Mammoth Cave to form 
at progressively lower levels. 


These two examples of burial dating with cosmogenic 
nuclides show only the beginning of what the dating 
technique can do. The caves along the New River proved 
ideal for burial dating because they contained quartz gravels 
washed in from the river outside. In this case, the caves were 
merely repositories of the sediment, holding it protected for 

Crustacea 141 

millions of years while the landscape changed outside. On 
the other hand, the sediments in the Mammoth Cave system 
were an integral part of how the cave was formed. The 
sediments reveal the evolution of the cave system and how 
cave development is tightly coupled to river incision and 
aggradation. In this case, Mammoth Cave was ideal because 
it was a water table cave that carried quartz from local 
bedrock. There are many more situations where geomorpho- 
logists, paleoanthropologists, and other scientists can benefit 
immensely from dating cave sediments over the past 5 
million years. 


Brain, C. K. (1981). The Hunters or the Hunted? An Introduction to 
African Cave Taphonomy. University of Chicago Press, Chicago. 

Granger, D. E., and PR. E Muzikar (2001). Dating sediment burial with 
cosmogenic nuclides: Theory, techniques, and limitations. Earth and 
Planetary Science Letters 188, 269-281. 

Granger, D. E., J. W. Kirchner, and R. C. Finkel (1997). Quaternary 
downcutting rate of the New River, Virginia, measured from differential 
decay of cosmogenic 7°Al and 'Be in cave-deposited alluvium. Geology 
25, 107-110. 

Granger, D. E., D. Fabel, and A. N. Palmer (2001). Pliocene-Pleistocene 
incision of the Green River, Kentucky, determined from radioactive 
decay of cosmogenic **Al and ‘Be in Mammoth Cave sediments. 
Geological Society of America Bulletin 113, 825-836. 

Lal, D. (1991). Cosmic ray labeling of erosion surfaces: In situ nuclide 
production rates and erosion models. Earth and Planetary Science Letters 
104, 424-439. 

Lal, D., and B. Peters (1967). Cosmic ray produced radioactivity on the 
Earth. In Handbuch der Physik (S. Flugge, ed.), Vol. 46, pp. 551-612. 
Springer-Verlag, Berlin. 

Palmer, A. N. (1981). A Geological Guide to Mammoth Cave National 
Park. Zephyrus Press, Teaneck, NJ. 

Tobias, P. V. (2000). The fossil hominids. In The Cenozoic of Southern 
Africa (T. C. Partridge and R. R. Maud, eds.), Vol. 40, Oxford 
Monographs on Geology and Geophysics, pp. 252-276. Oxford 
University Press, New York. 

Vrba, E. S. (1995). The fossil record of African antelopes (Mammalia, 
Bovidae) in relation to human evolution and paleoclimate. In 
Paleoclimate and Evolution, with Emphasis on Human Origins (E. S. 
Vrba, G. H. Denton, T. C. Partridge, and L. H. Burckle, eds.), pp. 
385-424. Yale University Press, New Haven, CT. 


Horton H. Hobbs III 
Wittenberg University 


Crustaceans are one of the oldest and most diverse arthro- 
pods as well as one of the most successful groups of 
invertebrates on earth with approximately 40,000 extant 
species described. They have been extremely successful in 
aquatic habitats, yet some species have become adapted on 
land as well. Their fossil record indicates that they are an 

142 Crustacea 

ancient group, having occupied the marine environment 
since the lower Cambrian period yet very early in their 
evolutionary history they invaded freshwater habitats. 
Although about 90% of the currently recognized taxa are 
widespread in marine systems, the remaining 10% are found 
in diverse inland waters and assume important roles in 
various ecosystem processes of many surface and subterra- 
nean lotic and lentic habitats. This article focuses on the 
hypogean members, specifically on those crustaceans that are 
highly adapted to dwelling in groundwater ecosystems and 
generally referred to as stygobites or stygobionts (obligate 
hypogean aquatic forms). 

In addition to being taxonomically diverse, crustaceans are 
anatomically disparate, having evolved an assortment of 
body forms accomplished by developing highly specialized 
body segments and appendages as well as by fusing various 
segments. As a group, crustaceans are bilateral, having internal 
and external segmentation, and an open hemocoel. They 
have a rigid, chitinous exoskeleton composed of a thin 
proteinaceous epicuticle and a thick multilayer procuticle 
that in many groups is hardened by small inclusions of 
calcium carbonate. Their bodies are generally divided into 
the cephalon (head), thorax, and abdomen (with the former 
two sometimes combined as the cephalothorax). The many 
jointed appendages are biramous (or secondarily uniramous) 
may occur in all regions of the body, and these arthropods 
possess paired antennules (uniramous in all crustaceans 
except malacostracans), antennae, mandibles, and maxillae. 

Crustaceans have invaded the hypogean realm, occupying 
interstitial and other groundwater habitats, including 
anchialine waters (inland ground water with subsurface 
marine connections harboring unique fauna) in karst (see 
article entitled Anchialine Caves in this volume) as well as in 
other landforms (e.g., lava). Some species of amphipod and 
isopod crustaceans have abandoned the groundwater and 
have been successful in the terrestrial hypogean environment. 
Most species dwelling in subterranean environments exhibit 
a suite of characteristic traits that are adaptive for life in such 
extreme ecosystems. Examples include reduction or loss of 
eyes and pigments, elongation of appendages, increased 
chemical and tactile sensitivity, degeneration of circadian 
rhythms, lowered fecundity and metabolic rates, and 
increased longevity and ovum volume. These embody 
behavioral, ecological, morphological, and physiological 
modifications that include both the reduction or loss of 
characters (regressive evolution) as well as the augmentation 
of others (constructive evolution). These various adaptations 
combine to generate the convergence characteristic of most 
obligate, cave-adapted organisms that is referred to as 

Crustacean taxonomy continually undergoes reevaluation 
and revision and the classification structure used herein 
(generally based on Martin and Davis, 2001) reveals five 
classes having subterranean representatives: Branchiopoda, 
Remipedia, Maxillopoda, Ostracoda, and Malacostraca (see 

Table I); a brief discussion of the classes and their hypogean 
representatives (predominantly stygobionts) is presented 

Class Branchiopoda 

Branchiopods are relatively small heterogeneous crustaceans 
that share few characteristics, including small to vestigial 
head appendages with similar mouthparts, flattened leaflike 
thoracic legs called phyllopods that usually decrease in size 
posteriorly, and a pair of spines or claws on the ultimate body 
segment. Classification of branchiopods has undergone 
numerous revisions and, of the currently recognized orders, 
only Diplostraca has subterranean members. There are about 
100 species of the suborder Cladocera (450 total species) that 
occupy the subterranean environment (Table I). 

They are known from subterranean waters (especially the 
interstitial/hyporheos) on all continents, but especially well 
in Bosnia and Herzegovina, France, Romania, Slovenia, and 
Spain. These small transparent crustaceans have a carapace 
that is laterally compressed and attached dorsally to the body 
around which it is wrapped, excluding the head. Troglo- 
morphic adaptations are very minor except in a few species 
of the genera Alona and Spinalona, evident by a lack of eyes 
and a carapace that is translucent and sparsely pigmented. 
Additionally, some have conserved a suite of primitive 
characters (e.g., setation of valve rims), which suggests that 
the protected constancy of the hyporheic has allowed for the 
survival of some old taxa. 

Class Remipedia 

The discovery of remipedes in the subterranean waters of 
Lucayan Cavern on Grand Bahama Island in 1979 presented 
a major surprise. On one hand these blind crustaceans 
possess characteristics that are very primitive (e.g., long 
homonomous body, paddle-like antennae, double ventral 
nerve cord), yet they have attributes that are traditionally 
considered advanced (e.g., maxillipeds and biramous limbs 
that are not platelike). These small ( <1—4 cm), translucent 
crustaceans lack carapace but a cephalic shield covers the 
head, appendages of which include a pair of rodlike processes 
anterior to antennules and prehensile mouthparts; the head 
and first trunk segment comprise the cephalothorax; and they 
have an elongate trunk of up to 32 unfused segments each 
bearing a pair of similar, laterally directed biramous limbs. 
They have been observed by divers primarily within 
submerged caves below the density interface between the 
overlying fresh or slightly brackish water and the underlying 
dense saltwater although one species, Speleonectes epilimnius 
Yager & Carpenter, is found above the density gradient. Most 
swim ventral side up as the result of synchronized beating of 
the trunk appendages and apparently stay below this 


Subphylum Crustacea 
Class Branchiopoda (85) 
Subclass Phyllopoda (85) 
Order Diplostraca (85) 
Suborder Cladocera (79) 
Family Daphniidae (20) 
Family Moinidae (5) 
Family Bosminidae (2) 
Family Macrothricidae (8) 
Family Eurycercidae [= Chydoridae] (44) 
Infraorder Ctenopoda (2) 
Family Sididae (2) 
Infraorder Haplopoda (1) 
Family Leptodoridae (1) 
Class Remipedia (12) 
Order Nectiopoda (12) 
Family Godzilliidae (3) 
Family Speleonectidae (9) 
Class Maxillopoda (1104+) 
Subclass Tantulocarida (1) 
Subclass Mystacocarida (<20) 
Subclass Copepoda (1,077+) 
Order Platycopioida (11) 
Order Calanoida (36) 
Order Misophrioida (17) 
Order Cyclopoida (230) 
Order Gelyelloida (2) 
Order Harpacticoida (787) 
Class Ostracoda (976+) 
Subclass Myodocopa (235) 
Order Myodocopida (8) 
Order Halocyprida (27 strictly anchialine and approximately 200 
marine interstitial species) 
Subclass Podocopa (~365) 
Order Platycopida (1) 
Order Podocopida (120+, of which approximately half are 
Subclass Mystacocarida (<20 intertidal and subtidal interstices) 

“Number = approximate number of described subterranean species. 

interface in the oxygen-poor (as low as 0.08 mg/L) saltwater 
layer. These predators likely feed in the overlying, well- 
oxygenated freshwater lens and locate their food by chemo- 
sensory means. Remipedes are small (up to 45 mm total 
length) and are commonly associated with other stygobitic 
crustaceans such as caridean shrimps, cirolanid isopods, 
haziid amphipods, mysids, ostracods, and thermosbaenaceans. 
Twelve species reside in two families containing six genera 
(Tables I and II) and are known only from anchialine caves 
in the Cape Range Peninsula in Western Australia, Bahamas, 
Canary Islands, Cuba, Turks and Caicos, and the Yucatan 
Peninsula (Mexico). 

Class Maxillopoda 

Maxillopods are mostly small crustaceans although barnacles 
(Thecostraca: Cirripedia) are conspicuous deviations. They 
have a reduced abdomen and lack a full complement of 

Crustacea 143 

Abbreviated Classification of Crustacea Inclusive Only of Those Groups Dwelling in Subterranean Habitats* 

Class Malacostraca (2211) 
Subclass Phyllocarida (1) 
Order Leptostraca (1) 
Subclass Eumalacostraca (2,210) 
Superorder Syncarida (208) 
Order Bathynellacea (187) 
Order Anaspidacea (21) 
Superorder Peracarida (1829) 
Order Spelaeogriphacea (3) 
Order Thermosbaenacea (33) 
Order Mysida (45) 
Bochusacea (1) 
Order Bochusacea (3) 
Order Amphipoda (778) 
Suborder Gammaridea (778) 
Order Isopoda (948) 
Suborder Phreatoicidea (11) 
Suborder Anthuridea (19) 
Suborder Microcerberidea (64) 
Suborder Flabellifera (141) 
Superfamily Cymothoidea (Cirolanidae—90) 
Superfamily Sphaeromatoidea (Sphaeromatidae—5 1) 
Suborder Asellota (436) 
Suborder Calabozoidea (1) 
Suborder Oniscidea (276) 
Order Tanaidacea (6) 
Order Cumacea (12) 
Superorder Eucarida (178) 
Order Decapoda (178) 
Suborder Pleocyemata (178) 
Infraorder Caridea (96) 
Infraorder Astacidea (40) 
Infraorder Anomura (2) 
Infraorder Brachyura (41) 

TABLE II Occurrence of Remipedia 

Family Genus Species Distribution 

Godzilliidae Godzilliognomus 1 Bahamas 
Godzillius 1 Turks and Caicos 
Pleomothra 1 Bahamas 

Speleonectidae — Cryptocorynetes 1 Bahamas 
Lasionectes 2 Australia, Turks and Caicos 
Speleonectes 6 Bahamas, Canary Islands, 

Cuba, Mexico (Yucatan) 

legs. Treated below are those subclasses having subterranean 
representatives: Tantulocarida, Mystacocarida, and Copepoda. 

SUBCLASS TANTULOCARIDA These small (<0.5-mm- 
long), ectoparasitic crustaceans are restricted to crustacean 
hosts. The only stygobitic tantulocarid (Stygotantulus stocki 

144 Crustacea 

Boxshall & Huys) is known from an anchialine lava pool on 
Lanzarote in the Canary Islands where it is parasitic on two 
families of harpacticoid copepods. 

SUBCLASS MYSTACOCARIDA This small group (most 
species <0.5 mm in total length; two genera with <20 de- 
scribed species and subspecies; Table I) of marine interstitial 
crustaceans is characterized by an elongate body that is 
divided into a head and a 10-segmented trunk. The carapace 
and compound eyes are lacking; the first trunk segment has 
maxilliped but is not fused to the head; and it has a telson 
bearing large, pincer-like furca. Because of the retention 
of primitive head segmentation, the lack of fusion of the 
cephalon and maxillipedal trunk segment, simplicity of 
mouth appendages, and absence of trunk compartmental- 
ization, this is likely one of the most primitive of all crus- 
taceans. (Others argue that these features may be related 
entirely to pedomorphosis and adaptation for interstitial 

Little is known concerning the life history of this group. 
Apparently eggs are laid free and likely they have up to six 
naupliar stages. Their small size and wormlike body are 
adaptations for interstitial life, using head appendages to aid 
in crawling among sand grains where they glean detritus and 
microorganisms from the surfaces of sediment particles. 
Although probably more widespread, they demonstrate a 
patchy distribution within the littoral and sublittoral sands 
in southern and western Africa, Australia, Brazil, Chile, 
southern Europe, the Gulf of Mexico, the Mediterranean 
Sea, and the east coast of the United States. Species of the 
genus Crenocheilocaris are restricted to the neotropical region, 
whereas Derocheilocaris spp. inhabit the nearctic, palearctic, 
and ethiopian regions. 

SUBCLASS COPEPODA The subclass Copepoda is a very 
large and diverse group of crustaceans (approximately 220 
families, 2300 genera, and 14,000 species) and, because they 
can attain incredibly high densities, are considered to be the 
most abundant metazoans on Earth (more individuals but 
fewer species than the insects). Lacking compound eyes and 
a carapace, the basic body plan consists of a head with well- 
developed mouthparts and antennae; six segmented thorax 
bearing swimming appendages, with the first segment fused 
to the head with maxillipeds; and a five-segmented abdomen, 
lacking appendages but including a telson. Development 
occurs within a few days to 3 weeks from fertilized eggs that 
hatch into a larval stage called a nauplius. Six naupliar 
stages are followed by six copepodid stages, the last of which 
is the adult (no additional molts). Six of the 10 orders have 
subterranean representatives, most of which demonstrate 
varying degrees of troglomorphy, and are treated next briefly 
(Table I). 

The Platycopioida order consists of one family (Platy- 
copiidae), four genera, and 11 species that have retained 
numerous primitive characters. They are considered to be 

the first order to diverge from the main lineage of copepods. 
The eight species of the genus Platycopia are known from the 
benthos in sea and coastal waters (Africa, Bahamas, northern 
Europe, Japan, United States). Sarsicopia polaris Martinez- 
Arbizu is found in muddy sediments of the Arctic Ocean, 
and two anchialine genera and species are endemic to a single 
cave in Bermuda: Antrisocopia prehensilis Fosshagen and 
Nanocopia minuta Fosshagen, one of the smallest known 

The Calanoida order contains approximately 2400 species 
in about 250 genera, yet only 36 species assigned to 22 genera 
and seven families are described from subterranean waters 
(Tables I and HI). Calanoids have biramous antennae and 
antennules that are greatly elongated, and the point of major 
body articulation occurs between the thorax and abdomen, 
which is marked by a distinct narrowing of the body. This 
primarily filter-feeding, planktonic group is geographically 
widespread, having been found in hypogean settings in 
Australia, Bahamas, Balearic Islands, Barbuda, Belize, 
Bermuda, Canary Islands, Caroline Islands, southeast China, 
Cuba, Dinaric Alps, Fiji, southern France, Galapagos Islands, 
Herzegovina, Istria, Italy, Madagascar, Mediterranean, 
Mexico, Philippines, and Russia. 

The Misophrioida order is represented by 16 genera and 
34 widely distributed copepods that are planktonic, hyper- 
benthic, and deep-sea, open-ocean, and hypogean water- 
dwellers. The subterranean species demonstrate particularly 
disjunct distributions, which are summarized in Table IV. 

Members of the Cyclopoida order are free-living, 
planktonic, and associated with various substrates in benthic 
or littoral habitats. They detect their prey with the aid of 
mechanoreceptors on their first antennae and grasp their 
food with precision with their first maxillae or, occasionally, 
with their second maxillae and maxillipeds. Most are 
voracious predators although some are parasitic. Antennules 
and uniramous antennae are moderately long (never as 
long as antennules of harpacticoids; see below); a fifth pair of 
legs is highly reduced, the sixth pair vestigial. Cyclopoid 
copepods have not been well studied in subterranean 
environments. Nearly 750 species are recognized, but 

TABLE III Occurrence of Described Stygobiont Calanoid 
Copepods in Anchialine, Cave, or Well Habitats 

Family Genus _Species/ Occurrence 

Boholiniidae 2 2 Anchialine 

Diaptomidae 8 11 Fresh water 

Epacteriscidae 3 if Anchialine; fresh water 

Fosshageniidae 1 1 Anchialine 

Pseudocyclopiidae 3 4 Anchialine; marine 

Ridgewayiidae 4 4 Anchialine 

Stephidae 1 2 Anchialine 

Total 22 36 

TABLEIV Subterranean Misophrioid Copepod Distribution 
and Habitat 

Genus Species Distribution Habitat 
Boxshallia bulboantennulata Canary Islands Anchialine 
Dimisophria cavernicola Canary Islands Anchialine 
Expansophria 4 Canary and Anchialine 
Galapagos Islands, 
Italy, Sardinia, 
Huysia bahamensis Bahamas Anchialine 
Misophria kororiensis Atlantic Anchialine 
Palpophria aestheta Canary Islands Anchialine 
Protospeleophria  lucayae Bahamas Anchialine 
Speleophria 4 Angaur Islands, Anchialine, 
Palau; Balearic cave 
Islands; Bermuda 
Speleophriopsis 2 Balearic and Anchialine 
Canary islands 
Stygomisophria _ kororiensis Koror Island, Palau —_ Anchialine 

only about 170 species and subspecies are reported from 
freshwater caves, interstitial habitats, and wells (Table I) (e.g., 
Acanthocyclops, Diacyclops, Eucyclops, Halicyclops, Idiocyclops, 
Kieferiella, Metacyclops, Speleoithona, and Speocyclops) from 
Africa, Asia, Australia, Bahamas, Cuba, Europe, Madagascar, 
and North and South America. About 60 additional species 
and subspecies are known from marine and brackish water 

The Gelyelloida order is represented by a single genus and 
two species inhabiting European freshwater karst systems. 
Gelyella droguei Rouche & Lescher-Moutoué and G. monardi 
Moeschler & Rouch are restricted to karst waters in 
Montpellier, France, and in the Swiss Jura, respectively. This 
copepod order is characterized by a distinct combination of 
gnathostomous mouthparts and unusual derived features. 

The Harpacticoida order is characterized by a body that 
is generally wormlike and cylindrical, with anterior segments 
not much larger than posterior ones; antennules and 
biramous antennae are quite short and the point of major 
body articulation occurs between the fifth and sixth thoracic 
segments; locomotory pereiopods are reduced, consistent 
with loss of swimming ability and use of these appendages 
as levers against sand grains. This group of copepods occurs 
virtually in all aquatic environments; most are benthic and 
are well suited to move through interstices feeding on detritus 
and on microorganisms (e.g., bacteria, diatoms, and proto- 
zoa). The harpacticoid copepods comprise 42 families with 
approximately 375 genera containing about 3000 species. 
At least 12 families (about 445 species) are known from 
freshwater caves, interstitial ground waters, and wells (e.g., 
Chappuisius, Elaphoidella, Parastenocaris, Spelaeocamptus, 
Stygonitocrella). Some 342 species and subspecies in 63 
genera and 13 families are found in the marine interstitial 

Crustacea 145 

habitat (e.g., Leptocaris, Nitocra, Novocrinia, Psammotopa, 

Class Ostracoda 

Ostracods are small (usually <1 mm, rarely 2 cm) crusta- 
ceans, with short, oval bodies encased within bivalved shells 
hinged dorsally. This is a very diverse group (5700 species) 
and although most are found in the marine environment (to 
depths of 7000 m) they are abundant worldwide in all 
aquatic systems. Most species are benthic, many are 
planktonic, some are commensal on echinoderms or various 
malacostracan crustacea (e.g., Sphaeromicolinae on Isopoda, 
Entocytherinae on cambarid Decapoda), and a few are 
terrestrial in moist habitats. 

Although some species are known only from caves, most 
hypogean ostracods are found in anchialine habitats, 
interstitial waters, springs, and wells. Examples of ostracods 
occupying various subterranean waters are Cavernocypris 
lindbergi Hartman (Afghanistan cave), Pseudocandona 
jeanneli (Klie) (cave in Indiana, United States), Mixta- 
candona juberthieae Danielopol (cave in southern France), 
and Pseudolimnocythere hartmanni Danielopol (well in 
Greece). These stygobionts display various troglomorphic 
adaptations including reduction or loss of eye structure, 
pigments, number of setae, and some have a very elongate or 
trapezoidal-shaped carapace. Most of the approximately 980 
subterranean species are assigned to the order Podocopida 
although the order Myodocopida is dominated by ostracods 
living in marine interstitial habitats (Bermuda, Galapagos 
Islands, Jamaica) and a coastal sea cave on Niue in the central 
Pacific (Dantya ferox Kornicker and Iliffe) (Table I). The 
order Halocyprida is represented by troglomorphic species 
dwelling in anchialine habitats, blue holes, and caves: 
Danielopolina (11 species in the Bahamas, Canary Islands, 
Cuba, Galapagos Islands, Jamaica, Yucatan Peninsula of 
Mexico, and the Cape Range Peninsula of Western 
Australia), Deeveya (7 species in the Bahamas and the Turk 
and Caicos islands), Euconchoecia (single species in Palau), 
and. Spelaeoecia (10 anchialine species in the Bahamas, 
Bermuda, Cuba, Jamaica, and the Yucatan Peninsula); 
approximately 200 are marine interstitial species. 

Of interest, the wide, irregular distribution, primitive 
nature, and troglomorphic adaptations of these taxa point to 
an extended history in a suitable cave environment. However, 
during the most recent period of Pleistocene glaciation, the 
sea level was lowered at least 100 m, resulting in coastal, 
anchialine caves becoming dry and then reinundated 
(substantiated today by the presence in these submerged 
caves of stalactites and stalagmites and other speleothems 
that were formed only in air by dripping or flowing water). 
This suggests that present-day anchialine ostracod fauna are 
recent invaders (within the past 15,000-18,000 years) and 
that they likely used an alternate, deeper habitat as refuge for 
considerable periods of time. 

146 Crustacea 

Class Malacostraca 

This diverse group (23,000 species) far excels the species 
richness of any other crustacean and is divided into three 
subclasses, of which two have species dwelling in hypogean 
environments: Phyllocarida and Eumalacostraca. The 
body fundamentally comprises a five-segmented cephalon, 
eight-segmented thorax, six-segmented abdomen (seven 
in leptostracans), and telson; the carapace may be absent, 
reduced, or may cover part or all of the thorax and even 
several abdominal segments; may have none to three pairs of 
maxillipeds; antennules and antennae are usually biramous; 
the abdomen generally bears five pairs of biramous pleopods 
and one pair of biramous uropods; and they are mostly 

SUBCLASS PHYLLOCARIDA The Subclass Phyllocarida 
is represented by the single order Leptostraca, which is 
characterized by a head with movable, articulated rostrum; 
biramous antennules and uniramous antennae; absent 
maxillipeds; phyllopodous thoracopods; a bivalved carapace 
(lacking hinge) covering the thorax; and an elongate 
abdomen consisting of seven free pleomeres plus telson. 
The order is represented by fewer than 20 species assigned to 
six genera, most of which are small (5-15 mm long), but 
one species is nearly 4cm in length. Most are epibenthic, 
are suspension feeders, and occur in low-oxygen marine 
environments. A single stygobiont, Speonebalia cannoni 
Bowman, Yager, & Iliffe (Nebaliidae), is known from two 
caves on Providenciales, Caicos Islands. 

class have head, thorax, and abdomen; up to three thoraco- 
meres are fused with the head, appendages of which are 
usually modified as maxillipeds; most groups have a well- 
developed carapace; and they possess a telson and paired 
uropods. There are three of four superorders with 
subterranean representatives. 

Superorder Syncarida This freshwater group, derived 
from marine stock, demonstrates the most primitive living 
body architecture of any eumalacostracan and many of the 
rather uniform trunk segments lack appendages. They lack a 
carapace; telson with or without furcal lobes; some 
pereiopods are biramous; and the pleopods are variable. They 
either crawl or swim and, although little is known about 
most species, some are likely omnivorous. Unlike most other 
crustaceans, which transport eggs and thus carry early embryos, 
syncarids lay their eggs or release them into the water subse- 
quent to copulation. Approximately 210 species have been 
described that are placed into two orders (Table I): Bathy- 
nellacea and Anaspidacea; 95% of these are stygobionts. 

The primitive Bathynellacea order lacks maxillipeds, is by 
far the more diverse group of syncarids, and is worldwide in 
distribution. Approximately 187 species are assigned to 56 
genera that are placed within three families: Bathynellidae 

(e.g., Bathynella chappuisi Delachaux, Switzerland;, Bathynella 
primaustraliensis Schminke, Australia), Leptobathnelidae 
(e.g., Acanthobathynella knoepffleri Coineau, Ivory Coast, 
Africa) and Parabathynellidae (e.g., Notobathynella williamsi 
Schminke, Australia; Parabathynella stygia Chappuis, Croatia, 
Slovakia). They typically inhabit freshwater interstitial media 
in epigean, cave, and well hypogean habitats in Africa, Asia, 
Australia, Europe, Japan, Madagascar, Malaysia, New 
Zealand, and North and South America. 

Order Anaspidacea (four families) has one pair of 
maxillipeds, is endemic to Australia, and species richness 
is particularly high on the island of Tasmania (numerous 
species awaiting formal descriptions). The Anaspididae 
family [e.g., Amaspides tasmaniae (Thomson)] inhabits 
various freshwater environments, including caves. However, 
most other syncarids are interstitial dwellers or live strictly 
in subsurface ground water, including caves and springs (e.g., 
the exclusively stygobitic family Psammaspididae— 
Psammaspides spp.—particularly diverse in New South 
Wales, Australia and Eucrenonaspides spp. in Tasmania). 
Family Koonungidae is found in sediment interstices, open 
water, and sinkholes (e.g., Koonunga crenarum Zeidler). 
Family Stygocarididae (historically considered a separate 
order) is represented by fewer than 10 species assigned to 
four genera: Oncostygocaris, Parastygocaris, Stygocarella, and 
Stygocaris. These tiny species inhabit interstitial waters in 
epigean and hypogean environments in New Zealand and 
South America. 

Superorder Peracarida Crustaceans placed into this 
superorder demonstrate a trend toward reduction of the 
carapace. They possess one (rarely two to three) maxillipeds; 
gills are thoracic or abdominal; unique thoracic coxal endites 
(oostegites) form a ventral brood pouch or marsupium in 
females; they lack true larval stages; the young hatch as 
mancas, a prejuvenile stage lacking the last pair of thoraco- 
pods; pleopods lack appendix interna; and the telson is 
without caudal rami. 

This highly successful group of malacostracans (approx- 
imately 11,000 species) is divided into 10 orders; all but the 
Lophogastrida have known subterranean species (Table I). 
Most are marine but many occupy freshwater and terrestrial 
habitats. Peracarids are diverse in their habits and size, 
ranging from a few millimeters to 44 cm in length and some 
are symbionts as well as stygobionts. 

The Spelaeogriphacea order is limited to fresh ground- 
water habitats in South Africa, South America, and Western 
Australia (Gondwana plates). This group was initially repre- 
sented by a single species, Spelaegriphus lepidops Gordon, 
which was described from pools and a stream in Bats Cave, 
Table Mountain (now known from a second cave), South 
Africa, where it was observed swimming swiftly using rapid 
undulations of the body. Evidence suggested that it feeds 
largely on detritus. In 1987 and again in 1998 two mono- 
typic genera were described, one from a lake in a freshwater 
cave in Brazil (Potiicoara brasiliensis Pires—now known from 

two other caves) and the other from the Millstream aquifer 
in arid northwestern Australia (Mangkurtu mityula Poore & 

These small, blind, unpigmented stygobionts possess a 
short, saddle-like carapace fused with the first thoracomere 
and, anteriorly, produced into a broadly triangular rostrum; 
they have one pair of maxillipeds; pereiopods 1 through 7 are 
simple, biramous, with shortened exopods; the exopods on 
pereiopods 1, 2, and 3 are modified for producing currents, 
and on pereiopods 4 through 7 they are modified as gills; the 
abdomen is elongated, often exceeding half of the total body 
length; and pleopods 1 through 4 are biramous and natatory 
and pleopod 5 is reduced. 

Order Thermosbaenacea is a group of small (2- to 5-mm), 
aquatic crustaceans that have a short carapace fused with the 
first thoracic somite (remaining seven thoracic segments free) 
extending posteriorly over two to three additional segments. 
In females, the carapace provides a brood pouch; there is a 
single pair of maxillipeds; pereiopods are biramous, simple, 
and lack epipods; there are two pairs of uniramous pleopods; 
the uropods are biramous; and the telson is free or forms a 
pleotelson with the last pleonite. 

At least 33 stygobitic species have been assigned to seven 
genera and four families (one additional species of Limno- 
sbaena not formally described) (Table V; see Wagner, 1994, 
for a revision of the order). They are known from anchialine 
habitats, caves, cenotes, various interstices, cold springs, 
thermal springs (45°C), and wells in fresh, to oligohaline, to 
hypersaline waters. These tethyan relicts demonstrate a very 
large geographic range in those areas once covered by the 
shallow Tethys Sea or along its former coastlines. 

Order Mysida is represented by some 1000 species that are 
widespread over all continents where they inhabit coastal and 
open ocean waters as well as continental fresh waters and 
various groundwater habitats. Some species are intertidal and 
burrow into the sand during periods of low tides; most of 
these shrimplike crustaceans swim with the aid of thoracic 
exopods and are omnivorous suspension feeders that eat 
algae, zooplankton, and suspended detritus. They range in 
length from approximately 2 mm to 8 cm and display a well- 
developed carapace covering most of the thorax. Compound 
eyes are stalked, sometimes reduced; they have one to two 

TABLEV Occurrence of Subterranean Thermosbaenaceans 

Family Genus Species Occurrence 

Halosbaenidae 3 5 Caves, marine interstitial 

Monodellidae 2 24 Anchialine, artesian wells, 
brackish wells, caves, coastal 
springs, interstitial, phreatic, 
thermal springs, wells 

Thermosbaenidae 1 1 Thermal springs 

Tulumellidae 1 3 Anchialine caves, caves, cenotes 

Total 7 33 

Crustacea 147 

TABLE VI Occurrence of Described Subterranean Mysids 
(Order Mysida) 

Family Genus Species Occurrence 

Lepidomysidae 2 10 Anchialine, caves, crab burrows, 
interstitial, phreatic, prawn 
culture field 

Mysidae 16 28 Anchialine, caves, coral reef caves, 
land crab burrows, marine caves 

Stygiomysidae 1 7 Anchialine, caves, phreatic 

Total 19 45 

pairs of maxillipeds not associated with cephalic appendages; 
the abdomen is elongated; pereiopods are biramous although 
the last pair is sometimes reduced; pleopods are reduced or 
modified; and the statocyst is usually located in each 
uropodal endopod. 

Of the two suborders, only the Mysida has stygobitic 
representatives and is divided into four families, three of 
which have stygobitic and stygophilic species (Table VI). At 
least 45 stygobitic species are recognized, most of which are 
endemics. The current distribution of the majority of these 
suggests that most colonized ground waters as a consequence 
of uplifting and stranding of their marine ancestors, which 
resulted from regressions of the Tethys and Mediterranean 
seas. Clearly other taxa have invaded ground waters more 
recently, most being stygophiles. Adaptations for subterra- 
nean existence seem to be limited to reduction or loss of body 
pigments (red pigments often retained) as well as reduction 
or loss of eyes but not eyestalks that contain endocrine organs. 

Mictaceans are stygobitic crustaceans, lacking body 
pigmentation and without visual elements in reduced 
eyestalks. The head narrows anteriorly into a triangular 
rostrum and is fused posteriorly with the first thoracomere. 
The carapace is not developed but small lateral carapace folds 
produce a head shield laterally covering bases of mouthparts; 
gills are lacking; pereiopods are simple; pleopods are reduced 
and uniramous; and uropods are biramous, with two to five 
segmented rami. 

A recent revision of the order Mictacea has resulted in 
it being monotypic (Table I). The single species, Mictocaris 
halope Bowman & Iliffe, assigned to the family Micto- 
carididae is known from four marine caves on Bermuda. This 
small species (up to 3.5 mm in total length) swims, using 
exopods of thoracopods 2 through 6, and rarely rests or walks 
on the substrate. Mouthparts suggest that it is not a predator 
but probably procures food by scraping and/or filtering fine 
particulates from the water column and benthos. 

Order Bochusacea is the most recently established pera- 
carid order, which was erected to accommodate three species 
placed in two genera belonging to the family Hirsutiidae 
(initially placed in the order Mictacea; two deep-sea benthic 
species: Hirsutia bathylis Sanders, Hessler, & Garner off 

148 Crustacea 

Surinam in northeast South America, and Hirsutia 
sandersetalia Just & Poore from Bass Strait, Australia). With 
additional material it became clear that a new order, 
Bochusaceo, should be erected to house the hirsutiids. 
Additionally, a third species, Thetispelecaris remex Gutu & 
Iliffe (Table I), from three marine/anchialine caves in the 
Bahamas, was placed in the order. Although similar to the 
mictaceans, the bochusaceans possess different pereiopod 
forms, among other features. They too are swimmers and 
filter feeders; pereiopod 1 is specialized for feeding and not 
for locomotion. 

The Amphipoda order is characterized by the absence of a 
carapace; the body is divided into head, thorax, and 
abdomen, each bearing appendages; the head bears two pairs 
of antennae; it has a single pair of maxillipeds, seven pairs of 
uniramous pereiopods, with the first, second, and often 
others modified as chelae or subchelae; periopodal coxae are 
expanded as lateral side plates; gills are thoracic; the abdomen 
consists of two regions of three segments each, an anterior 
pleon with pleopods, and posterior urosome with appen- 
dages modified as uropods; and the telson is free or fused 
with the last urosomite. This group of crustaceans is generally 
slender and laterally compressed although some are dorso- 
ventrally flattened (e.g., Heterophlias). They range in length 
from 1mm to giant deep ocean benthic forms reaching 
25 cm, 

Amphipod crustaceans are commonly found in numerous 
aquatic ecosystems around the globe where they have 
invaded freshwater, brackish, and marine environments, 
often comprising a large portion of the biomass in a habitat. 
A few species also dwell in various terrestrial ecosystems 
(Talitridae; e.g., supralittoral sandy beaches, moist forest 
litter) and they exhibit a great diversity of feeding strategies 
including carnivory, herbivory, parasitism, scavenging, and 
suspension feeding. Approximately 7000 species are assigned 
to three suborders: Gammaridea, Caprellidea, and Hyperiidea. 
Stygobiont amphipods are known only from the very large 
(>5700 species) Gammaridea suborder where some 778 tro- 
glomorphic species are assigned to approximately 150 genera 
and 29 families (Table VII). The following families demon- 
strate the richest subterranean biodiversity: Bogidiellidae, 
Crangonyctidae, Hadziidae, Melitidae, and Niphargidae. 
The amphipod suborder is widespread globally with the 
greatest taxonomic diversity of troglomorphic species occurr- 
ing in eastern and southern North America, the Caribbean/ 
West Indian region, central and southern Europe, and the 
Mediterranean region. These stygobionts evolved from 
surface ancestors, moving into hypogean ground waters from 
both freshwater and marine environments involving both 
active and passive dispersal. 

Order Isopoda is quite a diverse group with more than 
11,000 described aquatic and terrestrial species. They inhabit 
nearly all environments where most are free living yet some 
are partly (Flabellifera) and others are exclusively (Epicaridea) 
parasitic. Their feeding habits extend from herbivores to 

omnivorous scavengers, detritivores, predators, and parasites. 
They range in length from about 0.5 mm to 4.4 cm (the 
largest is the benthic Bathynomus). This diverse group of 
crustaceans is dorsoventrally flattened; they lack a carapace; 
the first thoracomere is fused with the head; they have a 
single pair of maxillipeds and seven pairs of uniramous 
pereiopods, which are modified as ambulatory, natatory, or 
prehensile limbs; pleopods are biramous and modified as 
natatory or for gas exchange (gills in aquatic taxa, pseudo- 
tracheae in terrestrial Oniscidea); and the telson is usually 
fused with pleonites 1 through 6. 

Representatives of subterranean isopods are known from 
six suborders: Phreatoicidea, Anthuridea, Microcerberidea, 
Flabellifera, Asellota, Calabozoidea, and Oniscidea (Table 
VIID) and are global in distribution. 

Suborder Phreatoicidea is the most ancient group of 
isopods and is derived from marine ancestors. These isopods 
currently are restricted mostly to fresh waters of South Africa, 
Australia, India, and New Zealand. Most are epigean; how- 
ever, stygobitic species (11 species in eight genera), known 
from caves and wells, are extant on all of these continental 
Gondwana fragments except South Africa. 

Anthuridea is principally a marine group consisting of 
four families; however, at least 19 troglomorphic species 
assigned to two families and three genera dwell in the 
interstitial sediments of anchialine habitats, bays, beaches, 
caves, and wells in the Canary Islands, the Caribbean and 
Indian Ocean islands, Indonesia, Mexico, New Zealand, and 
South America. 

The Microcerberidea consist of two families. The Atlanta- 
sellidae is represented by a single species, Atlantasellus 
cavernicolous Sket, which is restricted to anchialine caves in 
Bermuda. Clearly its recent origin is from marine stock. The 
Microcerberidae are slender, stygobionts of which approx- 
imately 63 species in six genera inhabit interstices of caves, 
marine beaches, and wells in the coastal regions of southern 
and western Africa, southeastern Asia, the Caribbean, Indian 
Ocean islands, Japan, the Mediterranean, and western North 

Suborder Flabellifera is represented by 18 families, only 
2 of which have subterranean members: Cirolanidae and 
Sphaeromatidae. The predominantly marine family Cirola- 
nidae has species dwelling in subterranean environments 
ranging from fresh water to salinities near that of seawater. 
Approximately 92 species assigned to 29 genera are currently 
recognized and range from less than 3mm in length 
(Arubolana parvioculata Notenboom) to around 33 mm 
[Speocirolana bolivari (Rioja)]. Most are blind, lack body 
pigments, are somewhat convex dorsally, and are opportu- 
nistic, benthic scavengers and predators. Their geographical 
distribution includes east Africa, the western Atlantic, and 
North America. (Cave and phreatic habitats are occupied 
by Speocirolana in Mexico and Texas, and by Antrolana 
and Cirolanides in Virginia and West Virginia, and Texas, 

Crustacea 149 

TABLE VII_ Occurrence of Subterranean Amphipods (Suborder Gammaridea) 

Family Genus Species Occurrence 

Allocrangonyctidae | 2 Caves: United States (Missouri, Oklahoma) 

Aoridae 1 1 Marine interstitial: Ile de Batz (Atlantic Ocean) 

Bogidiellidae 23 110 Caves, hyporheos, interstices of sandy beaches and sublittoral sands, cold mountain springs, wells: 

(including Artesiidae) nearly worldwide (not from continental Africa south of equator, continental Australia) 

Calliopiidae 3 3 Ground water, springs: New Zealand (North and South islands) 

Crangonyctidae 6 150 Caves, springs, wells: Holarctic region (mostly North America) 

Gammaridae 17 37 Caves, springs, wells: Canary Islands, Eurasia, Mediterranean region, North America, Solomon Islands 

Hadziidae 26 78 Caves: Australia; southern Europe; Bahamas, Caribbean, Fiji, Hawaiian islands; Mexico; United States 

Hyalellidae 2 Caves: Australia, Brazil, Venezuela 

Hyalidae 3 Subterranean freshwaters: Comores and Zanzibar islands (off coast of eastern Africa) 

Ingofiellidae 3 37 Caves, interstitial (marine), springs, wells: Africa, Caribbean and Mediterranean regions, 
South America 

Liljeborgiidae 1 1 Marine cave: Bermuda 

Lysiannasidae 1 1 Anchialine cave: Galapagos Islands 

Melitidae 23 54 Anchialine, caves, freshwater and marine interstitial and wells: Australia; Balearic, Canary, Caribbean, 
Galapagos, Hawaiian, Pacific, and Philippine islands 

Metacrangonyctidae 2 14 Hyporheos, springs, wells: Canary Islands, Mediterranean region 

Metaingolfiellidae 1 1 Well: Southern Italy 

Neoniphargidae 1 1 Shallow groundwaters: Victoria and Tasmania (?), Australia 

Niphargidae 8 221 Caves, hyporheos, interstitial, wells: Asia, Europe 

Paracrangonyctidae 2 3 Interstices, wells: Chile, Kerguélen Island (in Indian Ocean), New Zealand 

Paramelitidae 9 10 Caves, groundwater habitats: South Africa; Western Australia, southern Victoria, and Tasmania, 

Pardaliscidae 2 3 Anchialine caves: Bahamas, Galapagos, Lanzarote, Turks and Caicos 

Perthiidae 1 1 Caves: Southwestern Australia 

Phreatogammaridae 2 2 Caves, wells: South Australia, New Zealand (South Island), Spain 

Pontogeneiidae 1 2 Cave: Japan 

Plustidae 2 3 Freshwater caves and wells: Japan 

Pseudocrangonyctidae 2 11 Caves, springs, wells: Northeastern Asia 

Salentinellidae 2 14 Freshwater and brackish water caves, hyporheos, wells: Mediterranean region of northern Africa and 
southern Europe 

Sebidae 1 Artesian well: Texas (Hays County) 

Sternophysingidae Caves, springs: Southern Africa 

Talitridae 3 Caves: Corsica, Isla de la Palma (Canary Islands), Kauai (Hawaiian Islands), Sardinia 

Total 148 778 

The Sphaeromatidae is mostly a marine group with four 
genera of stygobionts represented by 43 species inhabiting 
subterranean waters in southern Europe and 8 species in hot 
springs in the southwestern United States and north-central 
Mexico. The very diverse genus Monolistra (at least 35 
species) occupies karst waters along the Dinaride and Italo- 
Dinaride systems. 

The Asellota are divided into a number of superfamilies 
(not treated herein) and nine families have subterranean 
representatives. The family Asellidae is a large benthic group 
(about 18 genera and 260 species) that is found primarily in 
caves, springs, and wells in northwestern Africa, Europe, 
Japan, and Central and North America. The Stenasellidae 
represent an ancient group that demonstrates primitive 
characters and that are blind and often orange or pink in 

body color. At least 67 are assigned to 10 genera and these 
crustaceans inhabit caves, the hyporheos, cool and thermal 
springs, and wells in much of Africa, southeastern Asia, 
eastern, southern, and western Europe, Indonesia, Malaysia, 
Mexico, and the Edwards Plateau in North America. The 
small, monotypic Stenetriidae isopod is known only from 
mixohaline waters in a cave on Curacao. The Janiridae is 
represented by 13 species assigned to five genera of marine 
origin. They are known from caves, springs, and wells in east 
Asia, western Europe, Italy, Japan, and North America 
(California). Microparasellidae are small, quite diverse (79 
species in four genera), and widespread. They are known 
from eastern Asia, Australia, the Caribbean, Indian Ocean 
islands, Japan, and the Mediterranean. The Paramunnidae is 
represented by a single species inhabiting subterranean 

150 Crustacea 

TABLE VIII Taxonomic Treatment of Subterranean Isopods 
Suborder Family Genus —_Species/subspecies 
Phreatoicidea Amphisopidae 3 5 
Nichollsidae 1 2 
Phreatoicidae 3 4 
Total 3 7 11 
Anthuridea Anthuridae 1 15 
Paranthuridae 2 4 
Total 2 3 19 
Microcerberidea —_Atlantasellidae 1 1 
Microcerberidae 6 63 
Total 2 7 64 
Flabellifera Cirolanidae 29 92 
Sphaeromatidae 4 51 
Total 2 33 143 
Asellota Asellidae 18 260 
Stenasellidae 10 67 
Stenetriidae 1 1 
Janiridae 5 13 
Microparasellidae 4 79 
Paramunnidae 1 1 
Gnathostenetroididae 3 6 
Protojaniridae 4 9 
Total 8 46 436 
Calabozoidea Calabozoidae 
Total 1 1 1 
Oniscidea Armadillidae - - 
Armadillidiidae - = 
Berytoniscidae - - 
Cylisticidae - - 
Eubelidae = = 
Ligiidae - - 
Mesoniscidae - - 
Oniscidae - - 
Philosciidae - - 
Plathyarthridae - - 
Porcellionidae - - 
Scleropactidae - - 
Scyphacidae - - 
Spelaeoniscidae - - 
Styloniscidae = - 
Trachelipidae - - 
Trichoniscidae 54 = 
Total 17 99 276 
GRAND TOTAL = 35 196 950 

waters. Although the genus Munnogonium currently contains 
seven species, only M. somersensis Kensley inhabits a marine 
cave in Bermuda and shows no obvious troglomorphic adap- 
tations. The gnathostenetroidid isopods are small crustaceans 
of marine origins and the six species distributed evenly over 
three genera demonstrate a disjunct distribution in anchialine 
caves in the Caicos and Turks islands and Bermuda (the eyed 
Stenobermuda), an anchialine cave in the Bahamas (/Veostene- 
troides), and the intertidal zone in Japan and a thermal spring 
on Ischia, a small coastal island southwest of Naples, Italy 
(Caecostenetroides). Nine species belonging to four genera are 
assigned to the Protojaniridae family and are quite separate 

from other members of the Asellota suborder. They are 
found in subterranean fresh waters in South Africa, 
Argentina (South America), southern India, islands of the 
Indian Ocean, and Namibia. 

The monotypic suborder Calabozoidea [Calabozoidae 
(Calabozoa pellucida Van Lieshout)] is known only from 
wells in the environs of the small town of Calabozo in 
northern Venezuela. This transparent, blind isopod appears 
to be restricted to phreatic waters. 

The suborder Oniscidea is generally a terrestrial group 
(e.g., Sinoniscus cavernicolous Schultz, China) characterized 
by extreme reduction of the first pair of antennae and by 
biramous pleopods modified into pseudotracheae. Most 
members of the suborder are troglobites yet some are 
stygobites; some 276 aquatic and terrestrial species assigned 
to 99 genera and 17 families are recognized. Examples of 
aquatic species are Abebaioscia troglodytes Vandel from the 
Nullarbor Plain, Australia; Haloniscus searlei Chilton, a 
scyphacid from Australia in lakes up to 159 parts per 
thousand; the styloniscid, Thailandoniscus annae Dalens, 
from Thailand; and the trichoniscid Mexiconiscus laevis 
(Rioja), which is terrestrial as a juvenile and aquatic as an 

The order Tanaidacea is known worldwide from benthic 
marine habitats and only a few species live in brackish or 
nearly fresh water. They often dwell in burrows or tubes and 
are known from all ocean depths. Approximately 850 species 
are recognized and, being small (0.5-2 cm long), many 
are suspension feeders, some detritivores, and others are 
predators. The carapace is fused with the first two thoracic 
segments; the first and second thoracopods are maxillipeds, 
with the second one being chelate; and thoracopods 3 
through 7 are simple, ambulatory pereiopods. 

Very little is known about subterranean species, mostly 
due to very minimal sampling. Currently six species, assigned 
to the families Apseudidae and Anarthruridae, are known 
from anchialine/marine caves in Bermuda, Eli Malk and 
Koror islands (Palau), and from Niue Island (South Pacific) 
(Table IX). 

The order Cumacea is represented by rather strange- 
looking crustaceans having a large, bulbous anterior end and 
elongated posterior, a carapace that is fused to and covers 

TABLE IX Occurrence of Tanaidaceans in Subterranean 

Genus Species Occurrence 
Aspeudes bermudeus Anchialine: Bermuda 

bowmani Marine: Koror Island (Palau) 

orghidani Anchialine: Bermuda 
Calozodion propinguus Anchialine, open ocean: Bermuda 
Nesotanais maclaughlinae Anchialine: Eli Malk Island (Palau) 
Pugiodactylus  agartthus Marine cave: Niue Island (South Pacific) 

TABLEX Subterranean Cumaceans 

Family Genus Species Occurrence 

Bodotriidae Cyclaspis simonae Submarine cave: Jamaica 
Diastylidae Oxyurostylis antipai Submarine cave: Jamaica 
Nannastacidae Campylaspis 2 Blue hole: Andros Island, 

Bahamas; Submarine cave: 

Cumella 8 Anchialine: Bermuda; blue hole: 
Andros Island, Bahamas; 

submarine cave: Jamaica 


the first thoracic segments, and three pairs of maxillipeds. 
Pereiopods 1 through 5 are simple and ambulatory, and pleo- 
pods are present in males, but usually absent in females. 
They are distributed worldwide and about 850 species are 
recognized. These small peracarids (0.5—2 cm in length) are 
mostly marine (a few brackish and freshwater species) and are 
generally benthic. Subterranean species are not well studied 
and only 12 stygobionts placed in four genera and three 
families are known from blue holes and submarine caves on 
Andros Island (Bahamas), Bermuda, and Jamaica (Table X). 
Superorder Eucarida 
treated herein and is characterized by crustaceans having 
a carapace covering and fused dorsally with the head and 
all thoracomeres (cephalothorax); they usually have stalked 

Eucarida is the final superorder 

compound eyes; gills are thoracic; and the telson lacks 
caudal rami. 

Order Decapoda is the only eucarid found in subterranean 
environments. This order is amazingly diverse with nearly 
10,000 species known from marine, brackish, freshwater, 
and terrestrial habitats. Although widespread in distribution, 
these crustaceans are significant subterranean players in 
temperate and tropical regions of the world dominated by 
karst or volcanic terrains and are well represented in the 
western Atlantic, Caribbean region, Central and North 
America, and numerous Pacific Ocean islands. They are 
characterized by a well-developed carapace enclosing the 
branchial chamber, are unique among decapods in possessing 
three pairs of maxillipeds, and have five pairs of functional 
uniramous or weakly biramous pereiopods (thus the deriva- 
tion of the order name). The order is split into two suborders, 
the Dendrobranchiata (about 450 species, mostly penaeid 
and sergestid shrimps) and the Pleocyemata (all remaining 
decapods as well as all stygobionts). The only dendrobran- 
chiate decapod that occupies subterranean waters is a 
stygoxenic penaeid shrimp, Penaeus indicus H. Milne 
Edwards, found in Mangapwani Cave in East Africa. The 
suborder Pleocyemata contains all of the remaining decapods 
and is geographically widespread globally. It is commonly 
divided into seven infraorders, only four of which have 
stygobitic representatives: Caridea, Astacidea, Anomura, and 
Brachyura. The three remaining pleocyemate infraorders are 
represented in subterranean environments by the stygophilic 

Crustacea 151 

TABLE XI Occurrence of Stygobitic Caridean Shrimps 

Family Genus/ Species Occurrence 

Procarididae 1 3 Anchialine 

Atyidae 15 48 Anchialine; caves; cenote; coral 
limestone pools; saline caves, 
pools, and wells; wells 

Agostocarididae 1 2 Anchialine, blue hole, cenote 

Palaemonidae 11 28 Anchialine, caves, coral rock 
pools, fissures, submarine cave, 
subterranean water, wells 

Alpheidae 6 7 Anchialine, caves, coral rock 
pools, lava pools, lava tube 
(marine), submarine cave 

Hippolytidae 6 9 Anchialine, blue holes, caves, 
cenote, pools in coral or lava, 
sea cave 

Total 40 97 

Jasus edwardsii (Hutton) from an intertidal marine cave in 
southeastern New Zealand (Palinura), by the stygoxenic 
Stenopus spinosus in a lava tube in the Canary Islands, by 
Odontozona addaia from a marine cave on the island of 
Minorca, Spain (Stenopodidea), and by a probable stygophile 
Naushonia manning Alvarez, Villalobos, & Iliffe from an 
anchialine cave on Acklins Island, Bahamas (Thalassinidea). 

Approximately 2000 species of natant decapods are 
assigned to the infraorder Caridea. These caridean shrimps 
have chelate first or second pereiopods (no chelation in the 
genus Procaris); first pleopods are somewhat reduced; and 
the pleuron of the second abdominal tergite is enlarged, 
overlapping that of the first and third. Global in distribution, 
only some 97 species in 39 genera and six families are known 
from anchialine habitats, blue holes, caves, cenotes, springs, 
and wells in some temperate but mostly tropical latitudes 
(Table XI). 

The infraorder Astacidea, which includes about 900 
species of crayfishes and chelate lobsters (>500 of those are 
crayfishes), is global in distribution (all continents except 
Africa), and the first three pairs of pereiopods are chelate. 
Crayfishes are the only astacideans that are stygobitic and, of 
these, only the superfamily Astacoidea is represented in the 
subterranean environment. Some 40 species from a single 
family (Cambaridae) have been described as obligate 
cave-dwelling species and subspecies; about 50 additional 
pigmented, stygoxenic, or stygophilic crayfish commonly 
invade subterranean waters around the globe (Astacidae, 
Cambaridae, Parastacidae). The following four genera are 
assigned stygobitic species and subspecies from Cuba, 
Mexico, and North America (primarily in the Appalachian, 
Florida Lime Sink, Interior Low Plateau, and Ozark karst 
regions): Cambarus (11 species), Orconectes (8 species and 
subspecies), Procambarus (20 species and subspecies), and the 
monotypic Troglocambarus (Table XII). Most of these are 
troglomorphic and are opportunistic omnivores. 

152 Crustacea 

TABLE XII_ North American Stygobitic Crayfish (Astacidea: Astacoidea: Cambaridae) in the United States, Cuba, and Mexico 
Genus Subgenus Species/subspecies Occurrence 
Cambarus Aviticambarus 3 United States: Alabama, Tennessee 
Erebicambarus hubrichti United States:: Missouri 
Jugicambarus 6 United States: Arkansas, Florida, Georgia, Missouri, Oklahoma 
Puncticambarus nerterius United States: West Virginia 
Orconectes Orconectes 8 United States: Alabama, Indiana, Kentucky, Missouri, Tennessee 
Procambarus Austrocambarus By Cuba; Mexico: Oaxaca, Veracruz 
Leconticambarus 1 United States: Florida 
Lonnberguis 2 United States: Florida 
Ortmannicus 11 Mexico: San Luis Potosi; United States: Florida 
Remoticambarus pecki United States: Alabama 
Troglocambarus maclanei United States: Florida 

Although diverse, the infraorder Anomura has only two 
stygobitic species assigned to two families (Table I). They 
possess a variably shaped carapace; the first pair of pereiopods 
is chelate, the third pair never, and the fifth pair reduced 
(never used for walking), and function as gill cleaners. The 
pleopods are reduced or absent. The superfamily Gala- 
theoidea is represented by two families having subterranean 
species: Aeglidae and Galatheidae. Aegla cavernicola Tiirkay is 
a poorly studied aeglid, known only from a cave in Sao Paulo, 
Brazil. The aggressive galatheid Munidopsis polymorpha 
Koelbel has received much attention (life history and ago- 
nistic, feeding, and reproductive behaviors) and is restricted 
to a single cave (lava tube) and wells (marine groundwater) 
on Lazarote, Canary Islands. 

The cephalothorax of the infraorder Brachyura (true 
crabs) is dorsoventally flattened and commonly expanded 
laterally; the abdomen is reduced, symmetrical, and flexed 
beneath the thorax; uropods are usually absent; first 
pereiopds are chelate; and the males lack third, fourth, and 
fifth pairs of pleopods. 

Approximately 7000 species are recognized, most of which 
are marine; however, freshwater and tropical terrestrial 
species occur in epigean and hypogean habitats. At least 41 
crabs assigned to 28 genera and 11 families are categorized as 
stygobionts (Table XIII). Additionally, the following families 
are represented as facultative cavernicoles in primarily 
tropical subterranean settings (mostly caves and anchialine 
habitats; numbers are approximate): Trichodactylidae (1 
genus, 2 species), Xanthidae (2, 2), Potamidae (7, 9), Potamo- 
nautidae (2, 2), Pseudothelphusidae (10, 21), Gecarcinu- 
cidae (2, 2), Parathelphusidae (1, 5), Sundathelphusidae 
(5, 5), Hydrothelphusidae (1, 1), Gecarcinidae (3, 5), and 
Grapsidae (5, 12). 


Bernasconi, R. (1994). Suisse (Schweiz/Svizzera). In Encyclopaedia 
Biospeologica, Vol. 1 (C. Juberthie and V. Decu, eds.), pp. 809-818. 
Société de Biospéologie, Moulis (France) and Bucharest. 

Botosaneanu, L., ed. (1986). Stygofauna Mundi, A Faunistic, Distributional, 
and Ecological Synthesis of the World Fauna Inhabiting Subterranean Waters 
(Including the Marine Interstitial). E. J. Brill, Leiden. 

TABLE XIII_ Occurrence of Stygobitic Brachyurans 

Family Genus Species Occurrence 

Hymenosomatidae 1 1 Caves: Indonesia 

Trichodactylidae 2 2 Cave: Chiapas and Tabasco, 

Goneplacidae 1 1 Caves: New Britain, Indonesia 

Xanthidae 1 1 Anchialine cave: Ecuador 

Potamidae 5 6 Caves: Sarawak, Borneo; Laos; 
Pulau Tioman, Malaysia; 

Pseudothelphusidae 5 9 Caves: Belize; Chiapas, Mexico; 
Columbia and Venezuela, 
South America; Guatemala 

Gecarcinucidae 5 5 Caves: Sarawak, Borneo; 

Parathelphusidae 1 1 Cave: Irian Jaya, Indonesia 

Sundathelphusidae 4 7 Caves: New Britain, Indonesia; 
Papua New Guinea; 
Philippine Islands 

Hydrothelphusidae 1 1 Cave: Madagascar 

Grapsidae 2 7 Anchialine, caves: Guam; 
Jamaica; Java, Nusa Lain 
Island, Indonesia; Niue 
Island, Polynesia; Papua New 
Guinea; Solomon Islands 

11 28 Al 

Bowman, T. E., J. Yager, and T. M. Iliffel (1985). Speonebalia cannoni, 
n. gen., n. sp., from the Caicos Islands, the first hypogean leptostracan 
(Nebaliacea: Nebaliidae). Proceedings of the Biological Society of 
Washington 98(2), 439-446. 

Boxshall, G. A., and R. Huys (1989). New tantulocarid, Stygotantulus stocki, 
parasitic on harpacticoid copepods, with an analysis of the phylogenetic 
relationships within the Maxillopoda. Journal of Crustacean Biology 9(1), 

Culver, D. C., and Wilkens, H. (2000). Critical review of the relevant 
theories of the evolution of subterranean animals. In Ecosystems of the 
World 30: Subterranean Ecosystems (H. Wilkens, D. C. Culver, and W. E 
Humphreys, eds.), pp. 381-398. Elsevier Press, Amsterdam. 

Gutu, M., and T. M. Iliffe (1989). Description of two new species of 
Tanaidacea (Crustacea) from the marine water caves of the Palau Islands 

(Pacific Ocean). Travaux du Muséum d'Histoire Naturelle Grigore Antipa 
30, 169-180. 

Gutu, M., and T. M. Iliffe (1998). Description of a new hirsutiid (n.g., 
n.sp.) and reassignment of this family from order Mictacea to the new 
order, Bochusacea (Crustacea, Peracarida). Travaux du Muséum d'Histoire 
Naturelle Grigore Antipa 40, 93-120. 

Hobbs, H. H., Jr., Hobbs HI, H. H., and Daniel, M. A. (1977). A review of 
the troglobitic decapod crustaceans of the Americas. Smithsonian 
Contributions to Zoology 244, 1-183. 

Hobbs III, H. H. (2000). Crustacea. In Ecosystems of the World 30: Sub- 
terranean Ecosystems (H. Wilkens, D. C. Culver, and W. F. Humphreys, 
eds.), pp. 95-107. Elsevier Press, Amsterdam. 

Humphreys, W. E, and Harvey, M. S., eds. (2001). Subterranean Biology in 
Australia 2000. Records of the Western Australian Museum, Supplement 
No. 64. 

Juberthie, C., and Decu, V., eds. (1994-2001). Encyclopaedia Biospeologica, 
Vols. 1-3. Société de Biospéologie, Moulis (France) and Bucharest. 

Crustacea 153 

Martin, J. W., and Davis, G. E. (2001). An updated classification of the 
recent Crustacea. Natural History Museum of Los Angeles County, Science 
Series 39, 1-124. 

Petrescu, I, T. M. Iliffe, and S. M. Sarbu (1994). Contributions to the 
knowledge of cumaceans (Crustacea) from Jamaica. II. Five new species 
of the genus Cumella. Travaux du Muséum d’Histoire Naturelle Grigore 
Antipa 34, 347-367. 

Poore, G. C. B., and W. E Humphreys (1998). First record of Spelaeo- 
griphacea from Australasia: A new genus and species from an aquifer in 
the arid Pilbara of Western Australia. Crustaceana 71(1), 721-742. 

Wagner, H. P. (1994). A monographic review of the Thermosbaenacea 
(Crustacea: Peracarida): A study on their morphology, taxonomy, phylo- 
geny and biogeography. Zoologische Verhandelingen (Leiden) 291, 1-338. 

Zinn, D. J. (1986). Mystacocarida. In Stygofauna Mundi, A Faunistic, 
Distributional, and Ecological Synthesis of the World Fauna Inhabiting 
Subterranean Waters (Including the Marine Interstitial) (L. Botosaneanu, 
Ed.), pp. 385-388. E. J. Brill, Leiden. 


Keith D. Wheeland 

The Pennsylvania State University, Retired 

AN cave database is a collection of bits of data that describe 
a cave and its environment. Modern databases have been 
designed to be accessed by computer. These databases can 
be used by the general caving community to obtain, for 
example, a list of caves for a particular geographical area; or 
cave scientists might use them to obtain answers to more 
complex questions. The use of computerized cave databases 
has improved the speed and accuracy of these answers. 


Early cave databases were published in bulletin form. One 
of the earliest of these publications that applied to an entire 
state is Pennsylvania Caves by Ralph W. Stone published in 
1930. Stone went on to edit two additional bulletins about 
Pennsylvania caves in 1932 and 1953. Stone, who was a 
charter member of the National Speleological Society (NSS), 
recognized the importance of assembling cave databases. 
In 1942, he wrote, “One of the objects of the Society is to 
catalogue all known caves, whether developed or undevel- 
oped.” In 1948, Stone became the second president of the 
NSS. Other early databases were published in bulletin form 
for caves in West Virginia by W. E. Davies in 1949 and for 
caves in Tennessee by T. C. Barr in 1961. 

With the advent of computers, some early databases were 
converted to computer-readable format. In 1984 a survey 
conducted by Jerry D. Vineyard of the Missouri Geological 
Survey revealed that five statewide computerized cave 
databases existed in the United States: Alabama, Georgia, 
Missouri, Pennsylvania, and Tennessee. Some of the first 

computerized cave databases were designed to be stored on 
Hollerith punched cards containing 80 columns of data. 
Some of these early databases have been converted to 
magnetic media that still contain the 80 columns of data. 
Others have been designed to take better advantage of the 
facilities of modern computers and database management 
systems. Today, cave databases throughout the world may be 
on paper, in a home computer, or in a larger computer. Some 
databases are available to the public online. 

The collection of data—and their subsequent conversion 
to computer-readable format—has often been initiated by an 
interested individual. Cave databases may contain a parti- 
cular type of cave, saltpeter caves, for example, but usually 
the databases contain caves for a particular geographical area. 
In some cases these smaller regional databases have been 
consolidated into larger geographical or geopolitical areas. In 
some countries (Australia, for example) this has matured to 
the point at which a national cave database exists. And in 
Austria, the cave database is a cooperative effort of 22 speleo- 
logical clubs with a combined 14,000 caves. In the United 
States, there has been very little interest in forming a national 
database. Until there is the need or the will for a national cave 
database in the United States, serious scientific use of the 
databases will have to be done on a regional basis. To pursue 
a national project in the United States, the scientist must first 
find and then contact each of the stewards of the regional 
databases. A steward is the individual or group that maintains 
the database and controls access to it. 


The International Union of Speleology (UIS) recognizes that 
sharing data across cave databases can be a problem. The 
Informatics Commission of the UIS is working on a solution 
to the problem. The proposed solution is to describe a 
file-sharing schema. A steward who receives a request from 
a scientist would provide the data using this go-between 


156 Databases 

schema. In theory, each database steward would have a 
computer program that would export data to this standard 
schema. The scientist who requested data from multiple 
database stewards would expect to receive the data in one 
standard format, albeit in multiple databases. 


Usually the database process begins with an individual 
collecting the data to form a database. If the database 
collection and maintenance process is to continue after the 
individual has lost interest or is no longer able to maintain 
the database, something else must be done. This something 
else generally involves forming an organization to perpetuate 
the database. One of the important considerations the organ- 
ization faces is the determination of a set of guidelines for 
distributing information contained in the database. Getting 
individuals to contribute data to a centralized database is 
usually dependent on the amount of trust that the individual 
has toward the database steward and the policy of 
distribution of the information. 


The caving community is protective of the information that 
has been collected and placed into cave databases. The most 
important factor that shapes this protectionism is the fact 
that caves are fragile environments. Most stewards of cave 
databases will willingly share the data when they feel that the 
person receiving the data will act responsibly. Even in 
Pennsylvania, which has a long history of publishing cave 
locations (all of Stone’s caving publications include directions 
for finding the caves), the sale of certain cave publications 
is now being restricted to persons who are presumed to 
be knowledgeable about cave conservation, safety, and 
landowner relations. Even as these publications are restricted, 
so too is the access to cave databases restricted. 

Access to most databases is restricted in some fashion. The 
stewards of the databases have devised various means of 
restricting access. In the simple case where a single individual 
controls the database, it is often that individual who 
determines what requests for information will be honored. If 
the controlling entity is an organization, this becomes more 
complex. Guidelines are often set, and the person who 
handles the requests for information uses these guidelines. If 
a particular request for information does not fit the guide- 
lines, the request can be referred to the board or responsible 
group in the organization for resolution. Table I lists the 
typical criteria that are often used—alone or in combination— 
to determine whether a request for data will be granted. The 
person to whom the data will be sent must meet one or more 
of these criteria. 

Closely related to the restriction of data access is the 
concept of “secret” caves. For various reasons, some persons 
with knowledge about a particular cave do not want the 

TABLEI Database Access Criteria 

An NSS member, or an organization within the NSS 

A conservation-related group 

A researcher 

Someone who has submitted data to the database 

A land owner or land manager 

A member of the organized caving community 

A rescue group 

Someone who needs the data for engineering work for road or dam 

A subscriber to the organization that sponsors the database 

information disseminated. Some realize, however, that 
their “secret” cave data may be of importance for scientific 
research. To acknowledge this concern, database stewards 
have implemented various means for addressing the problem. 
Some of the means employed are as follows: (1) A few 
stewards refuse to accept data for “secret” caves. (2) Some will 
accept the data with the understanding that no location data 
will be divulged. (3) Others have devised procedures that give 
veto power to the person with the “secret” cave regarding 
release of the data. 

Some databases are accessible by the public online. This 
type of database generally does not include specific cave- 
location data. Anyone needing such data can apply to the 
steward of the database for extended access. 

Some databases do not contain cave data per se, but are 
still of interest to cavers. Some examples of these are data- 
bases that include information on sinkholes, springs, mines 
data, mineral resources, cave biota, and cave-associated 


The two components of a cave database are (1) the data itself 
and (2) the component that accesses and maintains the data. 
If the database is computerized, this component is known 
as the software. The software may be custom designed (the 
Pennsylvania Cave Database is one such system) or it may 
be based on off-the-shelf software as in the Tennessee Cave 


The modern cave databases are designed to be composed of 
many separate data files all linked by some common thread. 
These databases are called relational databases. This type of 
arrangement allows for storage efficiency and relative ease of 
upgrading. A simple diagram for a relational structure is 
shown in Fig. 1. This structure can be altered to reflect the 
complexity of the caves in the database. The database is really 
a data model of the existing reality. If the contents of the 
database is a collection of simple one-entrance caves, then 

Typical data elements 

. Preferred name of cave or system 
. Length 
. Internal relief 

Geological i 


| lll | 

. Name of entrance 
. Location 

. Elevation 

. Quadrangle 

. Name of formation 
. Lithology 

. Name of publication or reference 
. Detail including date 

I I 
I I 

FIGURE 1 Example of a possible relational structure for a cave database. 

the structure of the database would be simple. If, however, 
the database contains data for caves for a whole country, then 
chances are good that the structure of the database would be 

In a relational model, the various pieces of data (data 
elements) are usually linked by some identifier (or a group 
of identifiers) that is unique to the cave. This often consists 
of a unique number, either assigned by a person or by the 
computer system. 

Because some caves have multiple entrances, it is desirable 
to associate some of the data elements with the entrance 
and not with the cave itself. A few databases allow for this 
refinement. To illustrate this point, take the example of a data 
element called /ength. The length is associated with the cave. 
The data element called /ocation is associated with an 
entrance. There are other databases that link cave survey data 
with other data. In these databases it is possible to associate 
data elements with specific survey stations. These are but a 
few of the types of associations that are possible. 

The inclusion of cave survey data and cave maps varies 
with database. Some store survey data, some store digitized 

Databases 157 

maps, while others store information describing the actual 
location of the map. 


Descriptions of cave databases are varied. There are almost as 
many cave database schemata as there are cave databases. 
One such schema is described in a publication called Cave 
Inventory and Classification Systems published by the Bureau 
of Land Management. More recent information regarding 
this schema can be found on the Internet. 

Table II lists data elements that may be found in a cave 
database. These data elements may be associated with the 
entrances, with other data elements, or with the cave itself. 
In some cases, there may be multiple occurrences of these 
data elements. A particular database may contain many of 
these data elements; however, no known cave database has all 
of these data elements. 


Some database stewards have taken the approach of using 
off-the-shelf software and tailoring the software to fit their 
needs. Other stewards have designed systems specifically for 
accessing and maintaining the cave data. One such system is 
the Pennsylvania Cave Database. It is menu driven, which 
means that is uses extensive menus. It is also code driven, 
which means that many of the data elements have been 
assigned codes. For example, each county has been assigned 
a three-digit code to represent the name of the county. The 
three-digit code is stored in the database. When the county is 
shown on the screen or printed in a report, the county name 
is displayed. This approach saves storage space, is less prone 
to errors, and in the end is easily understood by the person 
reading the output. The system was designed so that new 
codes can be added without rewriting the software. 


Many of the states that have caves already have computerized 
databases, and others, like Kentucky, are in the process 
of organizing their regional collections into one statewide 
database. Many states have their own database designs, while 
others are using models developed by other states. For 
example, the states of Georgia and Alabama have adapted the 
Tennessee model (with some modifications) for their own 
databases. Some counties of Kentucky used the Pennsylvania 
database model. Because the Pennsylvania model has been 
seeded with all the state codes and county codes, and because 
it is code driven, it could be adapted for use in any state. 

It is technically feasible to use one computer database 
model for every state database. This is unlikely to happen in 
the near future, however, because of regional preferences, 
politics, and the effort involved in converting the data. 


Dinaric Karst, Diversity in 

TABLE II Data Elements That May Be Found in a Cave Database 

Name (including alternate name/s) of the cave or entrance 
Cave or entrance identifier 

Bureau of Land Management number 

Forest number 

Monument number 

Specific physical location code 





Physiographic province 

Name of nearest town 

Drainage basin 

Open/closed status 

Type of cave/mine 

Data restriction code 

Entrance type, size, and elevation 

Geological formation and lithology 


Cave length 

Cave length source (survey, estimate, unknown, etc.) 
Internal relief (vertical distance from high point to low point) 


Flora and fauna 



The National Speleological Society and the Informatics 
Commission of the International Union of Speleology may 
be able to provide more details concerning the availability of 
cave databases for specific locations. The best way to contact 
the UIS is to search the Internet. 

National Speleological Society 

2813 Cave Avenue 

Huntsville, AL 35810-4431 USA 

Barr, T. C. (1961). Caves of Tennessee. Tennessee Dept. of Conservation and 
Commerce, Division of Geology. 64. 
Bureau of Land Management. (1983). Cave Inventory and Classification 
Systems. Bureau of Land Management, Roswell District Office, NM. 
Davies, W. E. (1949). Caverns of West Virginia. West Virginia Geological 
Survey, XIX. 

Stone, R. W. (1930). Pennsylvania Caves, Bulletin G3. Commonwealth of 
Pennsylvania Department of Internal Affairs, Topographic and Geologic 


Geomorphology including sediments 

Cultural and historical 

Speleothems and detailed features 

Air movement/barometric/chimney 
Physicochemical data 

Equipment needs to visit cave 


Damage assessment 

Potential for expansion code 

Owner's name, address, phone, and e-mail 
Access name, address, phone, and e-mail 
Access policy (visitation) 
Cave-protection-sign indicator 

Significant list indicator 

Profile description or code 

Passage pattern 

Passage density 

Cave map 

Cave map attributes (grade, class, quality, storage, and year) 
Survey data including cartographer/surveyor 

Dates (discovery, added to database, changed in database) 


Dinaric Karst, Diversity in 

Boris Sket 

Univerza v Liubljani 


The Dinarides or Dinaric Alps are a part of the Alpine system 
along the east coast of the Adriatic, the western part of the 
Balkan Peninsula in southern Europe (Fig. 1). In the extreme 
northeast of Italy and the west of Slovenia they are detached 
from the Southern Calcareous Alps; by a narrow belt of the 
so-called “isolated karst” the Dinarides cover approximately 
the southern halves of Slovenia and Croatia and the whole of 
Montenegro and western Serbia (south of Sava River). In 
northwestern Albania they extend to Hellenides. 

The Dinaric land appeared atop the Adriatic microplate 
out of the sea approximately 30 million years ago as an island 
that separated the sea of Paratethys from the still existing 

S Kras - Carso - Karst 

other engortast 
haret arene 

FIGURE 1 Position of the Dinaric karst. Mtg = Montenegro. (After Sket, 

narrow part of the Tethys Ocean. Its flora and fauna origin- 
ated and developed in comparative isolation from the rest of 
Europe. After orogenic faulting and erosion of the flysch 
layer, karstification began—most probably at the end of the 
Pliocene. It caused the origin of cave habitats, hydrographic 
changes, and splintering of many biotic distribution areas. 

Tropical or subtropical climates during the Tertiary were 
followed by intermittent glaciations in the Pleistocene. 
However, the closest extensive glacier in the region was in 
the Julian Alps in northwestern Slovenia, whereas in the 
Dinarides proper, only some of the highest peaks were 
glaciated. The dry land extended during the last glacials far 
south into today’s Adriatic Sea, enabling the biota to move 
away from their former proximity to the coldest points. It has 
been supposed that after glaciations, 10,000 years ago, the 
whole Dinaric karst, except for its highest elevations and a 
narrow coastal strip, before man’s impact, had been covered 
by extensive forests. 

Man started to colonize the territory (Sket, 1997) at least 
by the end of the Riss glacial or 65,000 years BP. With the 
onset of agriculture and animal husbandry (Mesolithic, 
10,000-4300 B.c.), more forest clearing occurred. Illyric 
tribes depended primarily on domestic animals, migratory 
grazing (transhumance), and storage of hay. Wintering of 
large herds caused higher pressure on forests’ existence. The 
Copper to Iron Ages (3000-200 B.c.) meant additional 
forest devastation for mining and the primitive metallurgy. 
In the Halstatt period, southern Slovenia was already quite 
densely inhabited. Grave constructions testify to the 
existence of bare grounds in Lika (Croatia) in 500 B.c. and at 
the coast in Bronze and Roman times. Up until 60 years ago, 
before the reforestation started to be successful, very large 
parts of the Dinarides were barren rock or nearly so. The 

Dinaric Karst, Diversity in 159 

measures that are responsible for a comparatively green karst 
sites nowadays were the introduction of the Austrian pine 
(Pinus nigra) for reforestation, a ban on goats, and a remark- 
able reduction in the human population in karst areas; the 
population moved to the towns and to developing industrial 
centers and to the fertile fields of the Pannonian Plain. 

Today's Dinaric karst extends to more than 56,000 km’; 
it is an approximately 100-km-wide and 600-km-long belt 
between the Adriatic Sea and the northeast hills and plains. 
It includes most Adriatic islands and drowned caves below 
sea level, and reaches some peaks having 2000-m elevations. 
Mountain ridges generally run parallel to the Adriatic coast 
(northwest-southeast). One of the specific characteristics 
of this karst are poljes, which are large, closed, periodically 
flooded depressions, typically without a surface in- and 

Thick carbonate layers, mainly Triassic limestones (but 
also dolomites and limestones of other ages), are highly 
porous; in the approximately 7000-km? area of the Dinaric 
part of Slovenia only, more than 4900 caves have been 
registered and investigated. 

In the outer Dinarides (toward the Adriatic) the climates 
are Mediterranean or sub-Mediterranean. The mean yearly 
temperatures are between 12° and 17°C. Precipitation is 
about 1000-2000 mm/yr, but is unevenly distributed and 
more than 50% in the cold half-year; this causes regular 
droughts in summer and washing away of soil in winter. The 
climate of the inner Dinarides is more continental and cooler 
and there is less precipitation. The nonfluctuating tempera- 
tures of underground habitats approximately equal mean 
yearly temperatures of the region, which are at 8-16°C 
higher than winter temperatures even in subtropical regions. 
The floods in rainy periods cause some temperature 
fluctuations and enrich the underground with fresh food 

The comparatively isolated position of the Dinaric 
territory along with its tropical past and very dynamic geo- 
morphologic changes allowed development of a very diverse 
flora and fauna. This is attested by the fossil molluscan 
faunas of Pliocene-Pleistocene lakes as well as by the diversity 
of survivors. The territory's position on the edge of formerly 
glaciated regions allowed it to be a refuge for thermophile 
elements during the Pleistocene. Particularly useful refuges 
are the thermally conditioned hypogean habitats, where 
animals are not exposed to unfavorable winter temperatures. 



If we except the timely remote and isolated first mention 
of a cave fish in China in 1540, a series of zoological 
discoveries in caves started in the extreme northwest part of 
the Dinarides. The amphibian Proteus anguinus was first 
mentioned by Valvasor in 1689 and scientifically described 

160 Dinaric Karst, Diversity in 

by Laurenti in 1768; however, it was first seen in a cave in 
1797. The next was the most unusual cave beetle, found by 
the local cave worker Luka Ce¢ in the famous Postojna Cave, 
described by Schmidt in 1832 as Leptodirus hochenwartit. 
These and a great number of other discoveries in the 19th 
century were in the Kras (= Carso = Karst) of Slovenia, which 
gave its name to the geomorphologic phenomenon of the 
karst. Discoveries of cave animals in other parts of the world 
(first in the Caucasus, Appalachians, Pyrenees, and New 
Zealand) occurred in the middle of the century, and in other 
parts of the Dinarides even toward its end. Thanks to such 
a history, the “classical Karst” won the appellation of the 
“cradle of speleobiology (biospeleology).” Hamann’s (1896) 
book Europaeische Hoehlenfauna, which even at that time 
already cited 400 references, was still devoted mainly to the 
cave species from Slovenia. 

The most recent census of the Dinaric obligate subterra- 
nean fauna (Sket et al, 2004) revealed more than 900 
species, 600 terrestrial troglobionts, and 330 aquatic stygo- 
bionts. If we compare this to 930 and 420 troglobionts and 
stygobionts in the whole of North America, the importance 
of such numbers is evident. Among regions of approximately 
the same size class in the world, the Dinaric karst is by far 
the richest in the obligate subterranean fauna. This does 
not seem to be a consequence of the state of investigations. 
To the troglobionts, one should add some regularly occurring 
troglophiles and trogloxenes as well as a high number of 
normal surface animals (“accidentals”) that occasionally, but 
with a higher or lower regularity, happen to occur in caves. 

Terrestrial Cave Fauna 

By far the most diverse group of troglobionts is beetles, 
represented by more than 200 species. This should not cause 
wonder, because Coleoptera are by far the richest animal 
group in the world; nevertheless, its representation under- 
ground is very biased; it is represented mainly by members of 
only two families. In addition, the other insect groups— 
which are still extremely diverse on surface—are hardly repre- 
sented underground at all. The most important in Dinaric 
caves is the family Cholevidae with its subfamily Lepto- 
dirinae (known also as Catopidae: Bathysciinae). Approx- 
imately 175 species are distributed in as many as 50 genera. 
Although this is originally an edaphic (i.e., soil-inhabiting) 
group, only a few cave-inhabiting species are not specialized 
to the cave habitat. And only few Dinaric genera are not 
endemic to this region: The troglophilous Phaneropella lesinae 
Jeannel exhibits a transadriatic distribution and its conge- 
nerics occur also far in the Asia Minor. The edaphic and 
troglophilous Bathyscia montana Schioedte may serve as an 
archetype of leptodirines: It is less than 2 mm long, egg 
shaped, with short legs and antennae. Some troglobionts are 
similarly built, but most of the others attained different 
degrees of troglomorphism. They are regularly bigger and at 
least slightly elongated. 

The higher troglomorphic species may be pholeuonoid, 
which means that they have a spindle-shaped abdomen and 
a neck-shaped prothorax. Such a species include Parapropus 
spp. in the northwestern part of the Dinarides; similarly 
shaped cave beetles may occur also in other karst regions 
(such as Pholeuon in Romania). Only in the Dinaric karst 
have the beetles attained the highest, ie., the leptodiroid, 
degree of troglomorphism. In the name-giving Leptodirus 
hochenwartii Schmidt from Slovenia and Croatia, the legs 
and antennae are extremely elongated, the “neck” is nearly 
cylindrical, and high vaulted elytrae make the abdomen egg 
shaped or nearly globular. Of similar shape are more than 25 
member species of the genus Antroherpon in the southeastern 
part of the Dinaric karst. The high number of species prob- 
ably indicates the effectiveness of such a body transforma- 
tion. Nowhere else in the world has this group of beetles 
achieved such diversity as in the Dinaric caves. 

The other beetle group in Dinaric caves is the family 
Carabidae with 12 genera and 80 troglobiotic species of its 
subfamily Trechinae. The carabid beetles in surface soil 
are predators; they are twice as big and longer legged than 
their scavenger cholevid neighbors. The northwestern 
Anophthalmus spp. or the southeastern Neotrechus spp. are, 
however, larger, more slender, and with longer appendages 
than the epigean or troglophilous Trechus spp. Some other 
troglobionts such as the Aphaenopsis spp. are even more 
troglomorphic, but this transformation never reached the 
degree it has in cholevids. The Dinaric Trechinae are also 
less exceptional; their diversity is approached by other cave 
faunas, such as those in North America. To the same family, 
but its subfamily Pterostichinae, belong also two comparatively 
large, approximately 10-mm-long species of Laemostenes; 
although also living outside caves, they are worth mentioning 
because they are present in nearly all caves in the whole 
Dinaric karst and far outside it. 

The family Pselaphidae is represented by only 20 
troglobiotic species. These are tiny beetles with very slender 
thoraxes and wide abdomens with shortened elytrae. They 
too were originally soil-dwellers. The related extensive family 
Staphylinidae, with a nearly tap shaped body with very short 
elytrae, resulted in the Dinarides in no troglobiotic species, 
but a number of species occur as troglophiles or occasional 
guests in the entrances of the caves. Examples include some 
tiny species of Atheta and some larger Glyptomerus spp. Some 
tiny weevils (Curculionidae) are also known from Dinaric 
caves. Some species of the genus Troglorrhynchus are often 
regarded as troglobionts. However, they depend on tree 
roots that are not as common and extensively developed in 
karst caves as they are in lava tubes, shallow underground. 
Troglorrhynchus spp. are in fact, as soil inhabitants, merely 
guests in caves. 

Most troglobiotic beetles in Dinarides are eyeless and 
have reduced hind wings but no one is really pigmentless; 
although translucent, most are dark brown, whereas some are 
yellowish brown in color. Although carabid beetles are more 

or less evenly distributed along the Dinaric area, the specific 
cholevids become more diverse toward the southeast of it; the 
southern Herzegovina, with neighboring parts of Croatian 
Dalmacija and Montenegro by far being the richest. 

The next richest groups of troglobionts are the spiders and 
false scorpions. The genus Troglohyphantes is very character- 
istic among spiders (Araneae). Its small representatives of 
mostly less than 4 mm in body length span their small sheet 
webs either in the forest litter, in micromammalian burrows 
(they are microcavernicole), in entrance parts of the caves, 
or deeper in them. Only the latter are mostly pigmentless 
and eyeless; there are approximately 25 troglobiotic species. 
In contrast, the always more than 5-mm-long spiders of the 
family Dysderidae hunt without webs. Among 25 species 
of seven mainly fully troglobiotic genera belongs the first 
scientifically described cave spider Stalita taenaria Schioedte. 
These are shiny brown to orange in the cephalothorax and 
legs and dirty white on the abdomen. Not less than seven 
families are represented by troglobiotic species in the 
Dinarides. The most strange might be the family Anapidae; 
its species Pseudanapis relicta Kratochvil from Montenegro 
measures only 1.4 mm in length and its relatives are very 
far away. 

Similarly numerous are false scorpions (Pseudoscor- 
piones), which, however, are less diverse, representing only 
three families, which have similar body shapes. The most 
specialized troglobiotic species are those with extremely slender 
and elongated pincers on the second pair of appendages and 
an up to 8-mm body length. They are very different from 
their relatives that inhabit forest soil. Neobisium spelaeum 
Schioedte was the first species described from caves. 

The other groups represented by troglobiotic species are 
(in descending order) millipedes (Diplopoda), woodlice 
(Isopoda: Oniscidea), snails (Gastropoda), centipedes 
(Chilopoda), harvestmen (Opiliones), and planarians 
(Turbellaria: Tricladida). Remarkable millipedes are species of 
the genus Apfelbeckia that are most probably troglophile or 
even trogloxene. At approximately 10 cm they are the largest 
European millipedes. During the hot and dry Mediterranean 
summers, they are found only at the entrances of caves, 
giving those rooms a sweetish scent because of their defense 
glands. Among snails, widely distributed are tiny species of 
Zospeum, a genus, present also in the Alps and Pyrenees; they 
are about 2 mm high, ovoid, and shells may be seen on wet 
walls or may be washed out of the earth from the bottom. 
In the southeastern parts of the Dinarides, bigger shells may 
be seen, such as those of Aegopis spelaeus A.J. Wagner, which 
can be up to 20 mm in diameter. The pigmentless species 
of terrestrial planarians from Slovenia has not yet been 

Like everywhere, bats are an important trogloxene group 
of cave inhabitants. However, very large colonies hardly exist 
anymore—nowadays, a colony of 1000 specimens represents 
a very high number. One such year-round cave species is the 
long-fingered bat [Myotis capaccinii (Bonaparte)]. The most 

Dinaric Karst, Diversity in 161 

often seen species, but usually only in small numbers, are the 
different species of horseshoe bats (Rhinolophus spp.). Bats 
are important vectors of food resources between the surface 
and the underground. Individuals of other mammal species 
may also be present deep in caves, including the rodent, the 
edible dormouse [G/is glis (Linne)], which was first scienti- 
fically described just from the beech forests of the Dinarides, 
and some martens [Martes foina (Erxleben)]. No birds enter 
the totally dark parts of Dinaric caves. However, the twilight 
areas near the entrances used to be common nesting places 
for large colonies of rock dove (Columba livia Gmelin), 
giving a number of caves the name Golobina, Golubinka, 
etc. (golob or golub is the Slovene or Croatian name for a 
pigeon or dove). Unfortunately, their populations have been 
heavily decimated or dispersed by hunters. On the other 
hand, groups of feral pigeons joined their wild relatives in 
some caves. 

Aquatic Subterranean Fauna 

As is generally true throughout the world, by far the pre- 
vailing group in Dinaric hypogean waters is the crustaceans 
(Sket, 1999). Peculiar, however, to the Dinarides is their 
richness in snails (Gastropoda); therefore, we discuss them 
first here. 

With approximately 130 species, the Dinaric aquatic 
gastropods represent close to half of the world’s known 
stygobiotic snail fauna. They are also a very common 
appearance in Dinaric cave waters and their empty shells may 
accumulate in more or less pure piles some meters long 
(thanatocenoses). Nearly all of them belong to the group 
Hydrobioidea whose formal subdivision is still highly 
changing. These are snails of a discoid, conical, to nearly 
rod-shaped shell, with a simple or highly elaborated, earlike 
mouth and a smooth to strongly ribbed whirl. The mostly 
2-to 3-mm-high shells are approximately twice as big as those 
of their North American relatives, which is a difficult to 
explain curiosity. The shells of Hadziella species (from the 
northwest Dinarides) are flat; Hauffenia are mostly widely 
conical, Iglica are very slender (églica in Slavonic languages 
means “small needle”); the Lanzaia and Plagigeyeria spp. 
from the southeast exhibit very diversely ornamented, ribbed 
shells with elaborate mouths. A member of pulmonate 
snails (Pulmonata: Ancylidae) is the generalist, genetically/ 
cytologically diverse and therefore widely distributed species 
Ancylus fluviatilis O.F. Mueller with a cap-shaped shell 
without a spiral. It may be present also in caves and such a 
cave population might be clinaly (gradually) troglomorphized: 
Skin pigmentation gradually disappears and eyes become 
more and more reduced along the sinking stream in the 
Postojna-Planina Cave System. Acroloxus tetensi (KuSéer), 
with a similarly cap-shaped shell, is a stygobiotic species of 

Among crustaceans, the most numerous are copepods 
and amphipods, with approximately 60 species each. Of 

162 Dinaric Karst, Diversity in 

amphipods (Amphipoda), not less than 45 species are of the 
genus Miphargus. This is one of the most diverse genera of the 
group, spread over almost all of Europe (and in Iran and 
Arabian Peninsula in Asia). It inhabits all freshwater habitats, 
including forest ditches, brooks, caves, and interstitial waters. 
Specimens of different species may be 2 to 30 mm long and 
extremely diverse in their body shapes. However, all of these 
species are absolutely without eyes. The same variability span 
is achieved in the Niphargus species of the Dinarides. One of 
the smallest species, N. transitivus Sket, inhabits interstitial 
and karst waters in the extreme northwest. The slender and 
up to 20-mm-long WN. stygius (Schioedte), the type species of 
its genus, inhabits mainly percolation waters in Slovenia; the 
very stout and large NV. orcinus Joseph and its relatives are 
spread throughout larger cave water bodies along the Dinaric 
karst. Particularly interesting is the particularly long-legged 
N. balcanicus (Absolon) with a densely spiny neck, from 
Herzegovina. The related Niphargobates orophobata Sket 
from the epikarst zone in Slovenia has the only close relative 
in the Greek island of Kriti (Crete). Another relative of 
Niphargus, Carinurella paradoxa (Sket), which inhabits 
interstitial waters within the northwest Dinaric karst, is 
able to roll into a ball. The up to 25-mm-long and stout 
Typhlogammarus mrazeki Schaeferna is an omnipotent raptor 
that can catch a shrimp swimming or climb up a vertical wall 
with a thin layer of water. The extremely fragile Hadzia 
fragilis S. Karaman has particularly numerous relatives in 
the Mediterranean and in the Caribbean. It is a mainly 
anchihaline species but a few populations are present in 
continental freshwaters. 

Copepods (Copepoda) are represented by some generalist 
and a number of specialized species, with a particularly 
numerous number coming from the genera Diacyclops and 
Acanthocyclops. Cyclopoids are particularly richly represented 
in larger bodies of water, where they may even be accom- 
panied by the explicitly planctonic diaptomids (Calanoida: 
Diaptomidae); four Dinaric cave species is an enormous 
number for freshwater Calanoida. Harpacticoids are parti- 
cularly well represented in the epikarstic percolation waters. 
Their genera Elaphoidella with 20 and Parastenocaris with 15 
species are the leading groups and numbers are still growing. 

Isopoda are not numerous, but are very diversely 
represented. The asellids (family Asellidae) are represented by 
the troglomorphic races of Asellus aquaticus (Linne) and 
by some troglomorphic Proasellus species. Very characteristic 
is the rich group of aquatic pill-bugs Monolistra (family 
Sphaeromatidae), some of them with long spines on their 
backs. The largest isopods are cirolanids (Cirolanidae) 
Sphaeromides, distributed close to the Adriatic coast but 
without any known ecological or historical connection to 
the sea. The water fleas (Cladocera), with only three tiny 
(less than 0.5-mm-long) stygobiotic Alona species, again are 
relatively very richly represented. 

Thermosbaenacea are represented by an anchihaline 
coastal species, Monodella halophila S. Karaman, and a 

freshwater species Limnosbaena finki MeStrov & Lattinger. 
The decapod shrimp belong to the family Atyidae; they are 
very common in Dinaric caves, the most widely distributed 
“species” Troglocaris anophthalmus (Kollar) is likely to be split 
into a number of independent species. 

Also worth mentioning are 13 epizoic to parasitic species 
of Turbellaria Temnocephalida, living mainly on atyid cave 
shrimp. Besides these stygobiotic species, only one epigean 
species has been known from Europe, again from the same 

Three species of true filter feeders should be mentioned, 
because these types of animals are very rare in caves. Besides 
some troglophilic species the only stygobiotic freshwater 
sponge (Porifera: Spongillidae), Eunapius subterraneus Sket & 
Velikonja, inhabits Dinaric caves near Ogulin (Croatia); it is 
of a softer consistency than the surface species and forms 
only few gemmulae. The only stygobiotic clam (Bivalvia: 
Dreissenidae), Congeria kusceri Bole, is similar to and related 
to the well-known zebra clam; it fastens itself to rocks by 
means of byssal threads. The only stygobiotic tube worm 
(Polychaeta: Serpulidae), Marifugia cavatica Absolon et 
Hrabe, fastens itself to rocks with its less than 1 mm wide 
calcareous tubes. Particularly interesting is that generations 
of tubes may accumulate to build meter-thick tufa-like layers. 
Ecologically similar is also the only stygobiotic cnidarian 
(Hydrozoa: Bougainvilliidae), Velkovrhia enigmatica Matjasi¢ 
et Sket. A colonial species, it attaches to the substratum 
by means of stolons, with sessile medusoids still distinctly 

The only stygobiotic vertebrate in Europe is the usually 
approximately 20-cm-long cave salamander Proteus anguinus 
Laurenti. Its distribution area is holodinaric with some local- 
ities in the small Italian part of the Dinaric karst and with the 
extreme southeastern locality reportedly in Montenegro. At 
present, two subspecies are formally described, although this 
does not match its whole diversity. All proteus populations 
exhibit outer gills, only three toes on the first legs, and two 
on the hind legs. The holodinarically distributed troglo- 
morphic race P a. anguinus has a more or less colorless skin 
and an elongated head with reduced eyes, hidden below the 
skin. The nontroglomorphic P a. parkelj Sket & Arntzen is 
limited to a tiny area in the extreme southeastern part of 
Slovenia. It is very darkly (sometimes black) pigmented, has 
normally developed eyes, and varying body proportions. 
However, both races are obligate cave-dwellers and both 
come out of springs on some nights. 

In the central and southern parts of the Dinaric karst, a 
number of endemic cyprinid fishes (Pisces: Cyprinidae) with 
small distribution areas are present. Most of them are known 
as regular periodical colonizers of caves, leaving the open 
waters of poljes before the retreat of the water and appearing 
again soon after a flood. They were observed to do both 
directions of migrations actively, not because they were 
directly forced to do so by hydrological events. In ancient 
times some species were economically important; special 

constructions were placed in caves to catch them in masses 
since they were a highly prized export article of the otherwise 
poor karst regions. A number of species belong to the genus 
Phoxinellus that exhibit different degrees of scale reduction; 
P. alepidotus (Heckel) possesses only a few scales along the 
lateral line. The others are Leuciscus and small Chondrostoma 
spp.; taxonomically the most diverse species is Aulopyge 
huegelii Heckel. 


Widely Spread Taxa 

Species or species groups may inhabit very different, wide 
areas within the Dinaric karst or—in a few cases—beyond it. 
The latter may be particularly instructive, possibly explaining 
to us the noncontiguous distribution areas of even highly 
specialized cave species. Such is the case with the isopod 
Asellus aquaticus (Linne), amphipod Synurella ambulans 
(EF. Mueller), and the gastropod Ancylus fluviatilis. They 
are highly generalist species of approximately European 
distribution that penetrate underground particularly along 
sinking rivers. In the Dinaric karst, some cave populations of 
these species became troglomorphic and stygobiotic. Similar 
but less well known is the situation with some generalist 
cyclopoid copepods. 

Probably a result of a similar situation in the past is 
the recent distribution of the cave shrimp Troglocaris; 
relatives of a handful of Dinaric species are present in 
France in the West and in Georgia (Gruzija) in the East. 
Among terrestrial animals, such are the gastropods Zospeum, 
a genus reaching from the southeastern Dinarides far into the 

Holodinaric Distribution 

Also the holodinaric distribution of some taxa may be a 
consequence of multiple (i.e., polytopic and/or polychronous) 
immigration with subsequent extinction of surface popula- 
tions. In some cases, the immigration was followed by 
speciation; in other cases, the whole Dinaric area seems to 
be inhabited by a homogeneous species (or a group of still 
unrecognized sibling species). In all such cases, the species (or 
genus) is never contiguously spread over the Dinaric area; it 
is instead split into groups of populations in hydrographically 
isolated karst areas. Such a distribution exhibits the most 
unique cavernicolous representatives of their higher 
groups: the amphibian Proteus anguinus, the clam Congeria 
kusceri, the tubeworm Marifugia cavatica, and the cnidarian 
Velkovrhia enigmatica. Other cases are particular represen- 
tatives of bigger groups, e.g., one gastropod species, Zospeum 
amoenum (Frauenfeld), the spider Parastalita stygia (Joseph), 
and most probably some Niphargus species whose taxonomy 
has not yet been well studied. 

Dinaric Karst, Diversity in 163 

Merodinaric Distribution Pattern 

The northwestern elements are, e.g., the monotypic beetle 
genus Leptodirus, the flat hydrobioid snails Hadziella, and the 
monolistrine (Isopoda) subgenus Microlistra. Their approx- 
imate counterparts in the southeastern Dinaric area are 
numerous leptodiroid species of the genus Amtroherpon, 
hydrobioid snail genera Lanzaia and Plagigeyeria, and the 
isopod subgenus Pseudomonolistra. No explanation for this 
bipolarity of the continental Dinaric cave fauna has been 
presented until now. 

A paralittoral distribution is exhibited by the anchihaline 
amphipod Hadzia fragilis S. Karaman, the thermosbaenacean 
Monodella halophila, and in a big part of the belt also the 
amphipod Niphargus hebereri Schellenberg. These predom- 
inantly anchihaline species are not present in the brackish 
subterranean waters of the Kvarner Gulf, which was 
inundated by the sea only during the Pleistocene times; this 
shows us clearly that the paralittoral distribution pattern has 
a historic—not purely ecological—background. 

Smaller Distribution Areas 

Most species and subspecies, but also some genera, exhibit 
smaller distribution areas within one of the merodinaric 
areas. The most interesting fact is that these areas do not for 
the most part follow the recent hydrological divides. One 
species (or subspecies) may either cross the borders of a 
divide or may be limited to a part of the drainage area. This 
has been more properly studied for species of the isopod 
genus Monolistra and of some hydrobioid snails. It has been 
supposed that such distribution areas had been achieved in 
geologically past drainage areas (sometimes still prekarstic, on 
the surface) and maintained by competition between related 
species today. Most small distribution areas are not specifi- 
cally explainable; they may be either relics of a formerly wider 
area or results of locally limited immigrations underground. 
A number of animal species is known from only one locality, 
either a cave or a spring; the most striking is the mysid 
Troglomysis hercegovinensis Stammer from only one “lake” in 
the vast cave system of Vjetrenica in Herzegovina, in the 
well-investigated area of Popovo Polje. 


Only a negligible number of obligate subterranean species 
cross the borders of the Dinaric region. Such is the case of 
some amphipods, snails and beetles crossing the slightly 
indistinct border between the Dinaric and South Alpine 
regions, and such is the case of some copepod species, which 
may even exhibit wider distribution areas. On the other 
hand, also comparatively high is the endemism between 
regions within the Dinarides. At the moment, the only 
distribution data for administrative regions (states) that are at 
hand, definitely do not reflect the natural biogeographical 

164 Dinaric Karst, Diversity in 

units. Nevertheless, the number of endemic species is mostly 
40-60%, sometimes even much higher. Consideration of 
nominal subspecies would increase the endemicity remark- 
ably as has been studied in some large genera, while even 
a number of endemic genera exists, scattered to different 
taxonomic groupings; they are particularly numerous in the 
cholevid beetles. Only 5 of the 44 present cholevid genera 
are not endemic for the Dinaric karst and at least 23 genera 
seem to be endemic for one of the states (Slovenia, Croatia, 
or Bosnia and Herzegovina). Because political borders are 
crossing regions with homogeneous faunistic assemblages, 
the endemism degree would certainly be much higher in 
natural biogeographic provinces. 


In this highly biodiverse area, some systems have particularly 
rich cave faunas. When searching for caves with more than 
20 obligate cave inhabitants, 20 such caves or cave systems 
were traced and as many as 6 of them were in the Dinaric 
karst (Culver and Sket, 2000). The Postojna-Planina Cave 
System is, with 84 such species, the richest among them. The 
system consists of 17- and 6-km of passages connected by 
2 km of flooded corridors not yet mastered by cavers. The 
dry parts of the Postojnska jama is one of the oldest (since 
1818) and most famous tourist caves in the World, with a 
small railway in it; some parts of the system are used 
extensively by tourists, but there are still many “wild” parts 
that are rich in fauna. This is the type locality for a number 
of “first cave” animals, including the first described 
troglobiont, the beetle Leptodirus hochenwartii, the first cave 
spider Stalita taenaria Schioedte, and many others. 
Altogether nearly 60 animal species or subspecies were first 
found and described from this system, and the European 
cave salamander, Proteus anguinus, was first seen in its natural 
habitat in Crna jama in 1797. It is also a site of long-term 
ecological studies. Dry parts of the system are inhabited by 
35 species, 9 of which are beetles. There are 49 stygobiotic 
species; particularly numerous are crustaceans, snails, and 
oligochaetes. The main artery of the system is the sinking 
river Pivka, which sinks at the Postojnska jama and resurges 
in the Planinska jama; there is approximately 10 km of 
underground bed in between. A mixed fauna of stygobiotic, 
stygophile, and stygoxene species inhabits it. The last ones 
are mainly insects; aquatic larvae of some of them are present 
along the course. Populations of some stygophile species 
exhibit a clinaly (gradually) increased troglomorphy along 
the stream. These caves are in Slovenia, which is the richest 
in the world for aquatic subterranean fauna. The next four 
richest caves are also in Slovenia: Krizna jama, Logarcek, 
Sica-Krka System, and Grad. 

In Herzegovina, which is the richest area in the world 
for terrestrial fauna, evidence is growing that a number of 
additional caves will be catapulted to the status of the richest 
ones. Until now, only Vjetrenica in Popovo Polje, with more 

than 70 species was registered. This complex cave system has 
7.6 km of passages, which include a number of small streams, 
pools, and trickles of water. The surface animals are very few 
here. There is no sinking stream, but water jets after rains 
import organic debris through crevices and shafts by which 
the rich fauna is being fed. The most recent census shows 
40 species of stygobionts and 35 terrestrial troglobionts. 
Among stygobionts there are 10 species of amphipods, some 
of which are very large, and three species of decapod shrimps. 
There are nine species of beetles. Particularly noteworthy are 
the amphibious catopid beetle Hadesia vasiceki (J. Mueller) 
and the amphipod Typhlogammarus mrazeki Schaeferna, 
which occupy the hygropetric (described below). 


In an area with such a high number of faunal elements and, 
therefore, comparatively high pressure for interspecific 
competition, one can expect further ecological specialization 
of species. Therefore, besides a number of more or less 
“trivial” communities and synusiae, composed of species 
that sometimes merely by chance are locally combined, a 
small number of characteristic groupings with particular 
ecological demands or abilities exist. 

Ice caves are scattered all over the area. They are usually 
sack shaped and surrounded by forest in moderately higher 
elevations and may bear ice deposits during the whole 
year. Particularly wet soil with close to zero temperature is 
the ecological characteristic of the proximity of such a 
subterranean glacier in summer. This environment is usually 
particularly rich in troglobiotic and troglophilic Coleoptera 
(e.g., Nebria spp.). However, the highly troglomorphic 
leptodirine beetles of the genus Astagobius seem to have 
specialized to such an environment. 

The cave hygropetric is a rocky wall, usually sintered, over- 
flown by a thin layer of slipping water. A number of lepto- 
dirine beetles seem to be specialized for climbing in such a 
semiaquatic environment and collecting organic particles 
brought by water. All of them—although not closely related— 
have a pholeuonoid body shape, with very strong claws, and 
with a particularly hairy mouth apparatus. The best known 
representative of such beetles is Hadesia vasiceki J. Muller. 
These obligate inhabitants may join some nonspecialized 
semiterrestrial or semiaquatic guests using such environments 
as hunting grounds or simply as distribution paths. 

A number of aquatic and probably some terrestrial species 
are nearly limited to the percolation-water-filled, crevice 
systems, also known as epikarsts; however, such species are 
primarily found in percolation-water-fed rimestone pools or 
puddles on cave walls or bottoms. Morphologically adapted 
to such narrow places seem to be the harpacticoid copepods 
of the genera Parastenocaris and Elaphoidella in particular, 
while the tiny amphipod Niphargobates exhibits strong and 
curved legs and claws appropriate for climbing in less narrow 

Thermal waters are a trophically particularly inhospitable 
environment. They force higher metabolism intensities and, 
because they are purified as they make their way through 
greater earth depths, they offer little food and little oxygen. 
In hypothermal waters of 15—28°C on the northwestern edges 
of the Dinaric karst, some stenasellid isopods, like a thermo- 
phile race of Protelsonia hungarica Mehely and tiny gastro- 
pods Hadziella thermalis Bole, are present. In Slovenia, close 
to the Pleistocene Alpine glacier, stenasellids are present only in 
such an environment. It seems, in fact, that biota are limited 
to trophically less inhospitable springs, because the filtration 
of huge quantities of water from depths gave no results. 

Sinking streams are streams that, after some course on 
the surface, flow underground. The subterranean bed of a 
sinking stream is in fact an ecotone environment. Ecological 
conditions in them are more fluctuating, are food resources 
richer and more diverse than in “autogenous” waters. 
Some surface species may penetrate along such streams for 
shorter or longer stretches underground. Among them are 
adaptation-prone stygophile species such as Asellus aquaticus. 
Along such a subterranean bed, a rich assemblage of selected 
surface and subterranean animals is formed; the fauna 
changes its quantitative and qualitative composition with 
gradual ecological changes leaving the sinking point 
and the surface influences. Although some stygobionts, e.g., 
Troglocaris shrimp and even Proteus and some Niphargus 
spp.» find such a rich environment particularly inviting, 
others strictly avoid it. On the other hand, this is the only 
cave environment where a number of insect larvae (of Ephe- 
meroptera, Plecoptera, and Trichoptera) may be present. 


A number of threats face the Dinaric cave fauna. In the 19th 
and early 20th centuries, some animals had been caught in 
large numbers for trading purposes. The amphibian Proteus 
anguinus was reportedly a popular pet in parlor aquaria, 
while the rare cave beetles won high prices among amateur 
entomologists-collectors. Some dealers organized genuine 
chains of collectors all along the Dinaric karst. Because they 
were using baited traps they were potentially able to affect 
some populations. 

In the decades of economic development of the tradi- 
tionally poor and economically passive Dinaric karst after the 
Second World War, a number of hydrotechnical projects 
have been accomplished. Surface and underground dams 
and artificial tunnels changed the hydrology of some parts 
of the Dinaric karst remarkably. We may expect that some 
previously isolated populations-races might come in contact 
and fuse again or at least that some populations were 
“polluted” by repeated input of foreign genes. New 
competition assemblages may have formed, threatening the 
existence of some species. The hydrological regime of the 
underground in wide surroundings of those constructions 
definitely changed, changing also the living conditions (and 

Dinaric Karst, Diversity in 165 

sometimes threatening the existence) of cave populations and 
species. For example, the live deposits of the Marifugia-tufa 
with its builder and many interstitial guests are now dead, 
due to the lack of regular floods in Popovo Polje and the 
sinkcave Crnulja. This unique phenomenon will be—or 
already has been—destroyed. 

Of course, the most serious threat to the cave biota is the 
omnipresent and diverse pollution. The net-shaped under- 
ground hydrological connections make the direction of 
pollutant outflows unpredictable and their effect very wide 
reaching. In the case of the sinking river Pivka, a moderate 
amount of organic pollution may enable surface animals to 
increase their success in competitive situations to the detri- 
ment of stygobionts. In this way, pollution may indirectly 
extinguish cave faunas. However, spills of generally 
poisonous materials are not a rare event in the karst and only 
the fact that such spills are also detrimental to the precious 
resource of potable water allows us to efficiently protect the 
subterranean fauna. 

However, as early as 1920, the Section for Nature and 
Natural Monuments Conservation of the Slovene Museum 
Society wrote and published a memorandum that outlined a 
very complete and detailed nature conservation plan. Beside 
the Alpine sites and biota, one of its main subjects was 
specifically karst caves as a whole as well as the particularly 
interesting cave fauna. A paragraph about a total ban on the 
commercial exploitation of cave fauna was included, saying 
that it should only be exploited for scientific and educational 
purposes. This was—very generally—trealized in 1922. This 
early attempt at nature protection, initiated in a high degree 
by cave fauna, has been now followed by a number of species 
protection acts and other environmental legislation in all 
Dinaric countries. So, the legal background for the many- 
sided protection of this hot spot exists—but so do many 
practical hindrances for its practical implementation. 


Culver, D. C., and B. Sket (2000). Hotspots of subterranean biodiversity in 
caves and wells. 7. Cave Karst Studies 62(1), 11-17. 

Gottstein-Matoéec, S., T. Bakran-Petricioli, J. Bedek, D. Bukovec, S. 
Buzjak, M. Franiéevi¢, B. Jalzic, M. Kerovec, E. Kletetki, J. Kralj, P. 
Kruzi¢é, M. Kudini¢, M. Kuhta, N. Matoéec, R. Ozimec, T. Radja, V. 
Stamol, I. Ternjej, and N. Tvrtkovié (2001). Croatia. In Encyclopaedia 
Biospeologica, Vol. 3 (C. Juberthie and V. Decu, eds.), pp. 2237-2287. 
Moulis, Bucharest. 

Sket, B. (1997). Biotic diversity of the Dinaric karst, particularly in Slovenia: 
History of its richness, destruction, and protection, Special Publication 3, 
pp. 84-98. Conserv. Prot. Biota of Karst, Karst Water Institute. 

Sket, B. (1999). High biodiversity in hypogean waters and_ its 
endangerment—the situation in Slovenia, Dinaric karst, and Europe. 
Crustaceana 72(8), 767-779. 

Sket, B., K. Paragamian, and P. Trontelj (2004). A census of the obligate 
subterranean fauna in the Balkan Peninsula, In: H. I. Griffiths and B. 
Krystufek (eds.). Balkan Biodiversity. Pattern and Process in Europe’s 
Biodiversity Hotspot. Kluwer Academic Publishers B.V. (in print). 

Nase Jame—Our Caves 35(1). 

Valvasor, J. W. (1689). Die Ehre des Herzogthums Crain. W. M. Endtner, 

166 Diversity Patterns in the Tropics 

Diversity Patterns in the 

Louis Deharveng 
Museum National d'Histoire Naturelle de Paris 


Until the 1980s all authors who worked on cave fauna 
stressed that true subterranean animals were rare or absent in 
the tropics (Leleup, 1956). Why would animals develop 
adaptations to an environment that is, at least in the humid 
tropics, so similar to the outside environment in its weak 
thermal amplitude and high humidity? This idea was actually 
based on data collected in a few tropical caves, mainly 
from Malaysia (Batu Caves) and India (Siju Caves), where 
sampling strongly focused on the widespread guano habitats. 
As early as 1914, however, Jeannel and Racovitza mentioned 
that troglobites were rare but diversified in tropical caves. 
The first unambiguous evidence that a rich troglobitic 
community might exist in the tropics was drawn by Howarth 
(1973) from the study of Hawaiian subterranean fauna. 
Since then, the richness of troglobites in tropical caves has 
been largely confirmed in other parts of the world 
(Chapman, 1980; Humphreys, 1993; Gnaspini and Trajano, 
1994; Deharveng and Bedos, 2000). 


The intertropical belt has an ecological definition that 
roughly matches its geodesic limits (ie., between 23°27’N 
and 23°27’S; Fig. 1). Tropical climates, ranging from desert 
to hyperhumid areas, and from lowland to nival belt, can be 
characterized by a low thermal amplitude across seasons. 
This characteristic is also, along with darkness, the most 


prominent feature of a cave climate. Caves appear, therefore, 
as azonal habitats, and this may account for the frequent 
“deconnection” between their fauna and outside fauna. 

The subterranean environment encompasses various 
habitats which can be grouped in six categories according to 
the size of the voids available for the fauna and the presence 
of either freshwater, brackish water, or air in these voids: 
interstitial freshwater, interstitial anchialine, interstitial 
terrestrial (traditionally not dealt with in subterranean 
habitats, but with soil habitats), cave freshwater, cave 
anchialine, and cave terrestrial. Guano represents a special 
high-energy terrestrial habitat that is widespread in tropical 
caves and well characterized faunistically. 

All of these habitats host characteristic assemblages, but 
only limestone caves and lava tubes have been significantly 
sampled worldwide. The spectacular landscape and the 
huge caverns of tropical karsts have attracted cavers and 
biologists during the last two decades; the large amount of 
data they accumulated constitute the core of our present-day 
knowledge on subterranean tropical fauna. 


During the last two decades, a number of troglobites and 
stygobites have been discovered in most sampled tropical 
areas. In addition, evidence is growing that many guanobitic 
species, though not or poorly troglomorphic, might be 
restricted to subterranean habitats as well. Data, however, 
remain scarce and uneven. For instance, among the many 
caves sampled in Southeast Asia, very few have been 
investigated in detail (Fig. 2). 

Freshwater and anchialine interstitial habitats have been 
rather extensively sampled in the Neotropics (Iliffe, 1994), 
but poorly investigated in Southeast Asia, where only 11 
interstitial species are described compared to about 50 
cave-restricted aquatic species. 

Noninterstitial freshwater stygobites are more easily 
caught and have been relatively well studied in several 


FIGURE 1 Main documented hot spots (in orange, more than 10 obligate subterranean species) and cold spots (in blue, less than 10 obligate subterranean 
species) of tropical subterranean biodiversity. Anchialine and interstitial habitats excluded. Gray lines: limits of the tropical regions; red line: equator. Cave 
systems cited in Tables I and II as circles; others as triangles. (After Juberthie and Decu, 1994, 2001.) 


Diversity Patterns in the Tropics 

~ woo 
’ \ 
oe China y i 
a Pa 
— a > AK 
f : fe ( * 
iV nfo 
% as os 
we m4, 
“2 Os a a fe 

“We 4 fe 7 : ' "" G. 2 | 

} ° _ ‘ : ev gy 
“Ber p tes | 

“} L A ys 

9@. a . o ‘ 

e \os 13% 'e f 

} c~ 10 ) , £e— <9 “ ; 
ee @ 2g & 
~ \ \, ee ~~ 4 ia aS é © 
Bi $ t Le _— 17 | r 7 T | 
\<) ( j 2 7 
WW '. \ —_ { { e \ —— \ 
. f - ta \ 4 v . ~ Pom, 
| Nel <8 | 
<o— ne 16 £ 

1000 km e. => As om x> oS" rw + 
| > ‘fe a ; 


FIGURE 2 Main records of obligate subterranean species in Southeast Asia. Large yellow spots: best studied localities, with usually more than four obligate 
subterranean species; small black spots: isolated records, usually one or two species. 1: Tham Chiang Dao and Mae Hong Son karsts; 2: Farm caves; 3: Tham 
Phulu; 4: Ha Long Bay; 5: Vang Vieng; 6: Khammouan; 7: Ke Bang; 8: Hon Chong-Kompong Trach; 9: Phangnga; 10: Batu caves; 11: Montalban; 12: Niah 
cave; 13: Mulu; 14: Gunung Seribu; 15: Gunung Sewu; 16: Maros; 17: Batu Lubang. (After Deharveng and Bedos, 2000, updated.) 

tropical regions, such as Cuba and Thailand. Conversely, 
data are lacking for many of the major karsts of the tropics, 
such as those of central and northern Vietnam. 

A number of anchialine stygobitic species have been 
described from tropical America. In contrast, only two or 
three species are known in Southeast Asia, though several 
groups frequent in neotropical caves have been recently 
discovered there (Juberthie and Decu, 1994, 2001). 

Our knowledge of troglobites is also geographically and 
taxonomically uneven. Sarawak caves probably have the 
largest number of named troglobites, but no Collembola or 
microarthropod has ever been described from there. In 
contrast, Collembola are among the best studied troglobites 
in Thailand, with 16 described species. Guanobites, espe- 
cially small Arthropods, are poorly known, and their ecolo- 
gical status often pending. Even the giant Rhaphidophorid 
crickets of Southeast Asia, at the base of cave food webs, 
remain unnamed in most cases. 

More generally, most of the species listed in the literature 
are still undescribed. For instance, of the 21 troglobites cited 

from the SKT system (Sulawesi) by Deharveng and Bedos 
(2000), only 4 are described species. 


The unbalanced sampling of tropical caves and the uneven 
identification level of their fauna result in an immensely 
heterogeneous amount of data being available for 
deriving biodiversity patterns. Trends begin nevertheless to 

1. Species richness of tropical subterranean habitats varies 
from place to place, but is never very low or null as in 
northern temperate regions affected by glaciations. All large 
caves investigated so far in Southeast Asia hosted, for 
instance, at least a few cave-restricted species. 

2. At a continental scale, diversity appears higher for 
terrestrial fauna in the Oriental and Australian regions than 
in the Neotropics or Africa (Table I). This puzzling 
observation seems to be confirmed by recent works on 
Brazilian cave fauna, though the work might be biased by the 


in the Tropi 


“purpeyy, WH ‘eruezuey ‘Nyy ‘eneums sips ‘tsemeyng :T QS Syemeres 2ypys ‘reuruedyy WAY] SeARpEPY ? TVW ‘SORT ‘OV SereqewyeHxy ‘TWH ‘Tiserg :vug ‘wessy :ssy 




€ b Z Z 8 Ol Ol Il 91 91 91 91 91 81 0z 87 sauigosAas + sarrqo]sory, 
z 0 € € I 0 v c I 0 € I I ¢ € ZL samqosArg 
I I eprresuds 
I epooenscE, 
I aepraseuaig epodosy 
Z vappstuQ epodosy 
I I aeprnyuy epodosy 
I I I ¢€ erlueien epodessq 
I I I I vindyperg epodesaq 
I I vinuouy epodesaq 
I I epodado> 
I I I I I epodryduny —-eaoeasnd 
I T vosny [OY 
I I I I BLeTaqny, 
I v v v Z Or 9 2) SI 91 €l SI SI €l ral 17 souqoyBory, 
I I I viadoyuyo 
I aeprorunso.y eroidousurdpy 
I vlad ruapy 
I vroideuraq 
I I aeprydejasg eroidoajo> 
I Z I I aepiqeien viaidoajo= 
I aepidoie) esaidoajod, 
I I T I I I vaponelg BIDOSUT 
I depmajooIN euolUaséZ 
I I I I sepropoduey = emydiq 
I I I I I euogdAydurds 
Z I eydiowrompog 
I I I I I C I ¢ Z Z I Z Z eydiowodiqowoiy = eyoqurayjo 
c c I I I I c c € c I € € voppstuQ epodos, — vaovismd) 
I eprnrururaig epodordiq 
if I Z Z Z € Z I Z Z Z  eprusapdjog epodoydiq 
I I I I I I epyn{ epododiq 
I it I eplawoys) epodordiq 
I I epodopryy — epoderskyy 
I epruordioag 
I I I I T I epruordiossopnasg 
I I epipeadidjeg 
I I I I z I I I I I I epruorido, 
I I c c I y c v Cc € SeOUUAy: 
I I I I I I Hey —epruypery 
I BouIPOatey 
I I vOsny [OY] 
(AVS) (WAIN) (vad) 24% (TVIN) (SSV) (NVI) (Wud) avo (vanN3) (OVD (OVD (vad) (IVH) (VHD) °*¥d (TVW) (Ws) (Ins) 
PAB SDABD. endy,p seavy saavo aw BWI) aps souenSeyQ ep uoyy uoN uoisks SueqnyT Suey) yrusf yeimg urayshg 
wary vIN = wey soy eg onfig IznuMpDY = equiepy epueID ean) Wey] wey, seery = meg ueyL yen nes = LMS 

*(saxIqoSA3s pur sairqo]s0n QT UeYy) a10ur) sjods-10y JeuOTZax UTeUr ay} SuIpNpur ‘saavo [estdonN sarpnis-[Jam suTOs Jo UOTsOdusOD SNsTUNeJ pue AUsIoAIporg | ATAVL 

size of the sampled karst units. A different pattern emerges 
for stygobites, with highest richness in anchialine caves of 
tropical America and Australia, but here the sampling effort 
might explain much of the difference. 

3. Large karsts tend to be richer in troglobites than 
smaller karsts. Statistically, they should have a wider diversity 
of habitats for a comparable spatial configuration. Thus in 
Sarawak, Gua Air Jernih in the Mulu karst has much more 
cave-restricted species than the Niah cave in the small 
isolated hill of Gunung Subis (Table I). 

4. “Most diversity in caves is expressed regionally rather 
than locally” (Culver and Sket, 2000, p. 16). This point is 
directly linked to the impact of habitat fragmentation on 
speciation. The few available data suggest that this impact 
might be stronger in tropical than in temperate regions, for a 
similar spatiohistorical fragmentation of the habitat. This is 
illustrated by the example of the Paronellid Collembola of 
western Thailand (Fig. 3). Spiders seem to exhibit a similar 
pattern, with each karst unit having its endemic species 
(Deeleman-Reinhold, 1995). 

5. As a rule, troglomorphy appears to increase (a) with 
seasonality, (b) with decreasing mean annual temperature 
(ie., with altitude and latitude), and (c) with increasing 
drought. The more seasonal the area, the more troglomor- 
phic traits and more troglomorphic taxa are seen in its fauna. 
Thus, obligate subterranean beetles are few and weakly modified 
morphologically in Indonesia (two species), Malaysia (none), 
or Cuba (none). They are more numerous or more 
troglomorphic in eastern and northern Thailand (four or five 
species), or southern Brasil (two species); they finally reach 
levels of troglomorphy and diversity comparable to temperate 
Trechidae in southern China, just north of the tropical belt. 
Cave fish are mostly known in the northern part of the 
tropics and in the subtropics; they are very rare in the humid 
tropics. Similarly, New Guinea’s high-altitude caves host 
several highly modified beetles, which are absent in low- 
altitude caves of the same region. It is doubtful, however, that 
such patterns apply to all taxa. Araneae in particular may have 
similar richness in troglomorphic taxa at different altitudes. 

6. Howarth (1973) considered the cave species of Hawaii 
to be derived from surface fauna present on the island. He 
hypothesized that the absence of relicts might be charac- 
teristic of tropical caves. Hawaii, however, as a young 
archipelago, did not experience drastic ecological changes 
that would have eliminated surface fauna, hence, making cave 
species relicts. Actually, evidence is accumulating that cli- 
matic relicts are present in many tropical caves, for instance, 
in the Cape Range of western Australia (Humphreys, 1993) 
as well as in dry or strongly seasonal tropical regions (Laos, 
South Sulawesi). Several stygobites, on the other hand, are 
probable marine relicts secondarily adapted to freshwater, 
like the crab Cancrocaeca xenomorpha of Maros (South 
Sulawesi) or the isopod Cyathura (Stygocyathura) chapmani of 
the Mulu caves. More generally, the numerous monospecific 
genera described from tropical caves are indicative of phyletic 

Diversity Patterns in the Tropics 169 


FIGURE 3 The Troglopedetes (Collembola: Paronellidae) of western 
Thailand. 1: calvus; 2: centralis; 3: convergens; 4: dispersus; 5: fredstonei; 6: 
leclerci; 7: longicornis; 8: maffrei; 9: maungonensis; 10: microps; 11: 
mutltispinosus; 12: paucisetosus. 

isolation and often relictness. Relicts come in addition to the 
regional pools of nonrelictual taxa, contributing therefore to 
a local increase in cave species richness. 


The tropics are known to have the highest biodiversity on 
earth for surface ecosystems, but trends might be reversed for 
subterranean habitats. Culver and Sket (2000) compared the 
specific richness of 20 caves and wells worldwide having 
20 or more obligate subterranean species and came to the 

170 Diversity Patterns in the United States 

conclusion that “the scarcity of high diversity caves in the 
tropics is still a puzzle” (p. 16). Their study included two 
tropical caves, the SKT system in South Sulawesi, which is 
the richest spot of tropical subterranean biodiversity, and 
Bayliss lava tube in northern Australia. These caves ranked 
12 and 15 of 20 for total species richness, far behind the 
temperate cave systems of the Dinaric karst. However, 
aquatic microinvertebrates (Copepoda, Oligochaeta), which 
contribute the most to biodiversity in temperate caves, have 
not been studied in SKT nor in most tropical caves. A more 
relevant comparison based on terrestrial fauna only places 
these two caves much closer to the richest temperate caves 
(ranks 4 and 5 of 22 in Table II). They would rank even 
higher if the obligate subterranean guanobitic species, 
probably considerably more diverse in the tropics than in 
temperate areas, were considered. 

TABLEII ‘Tropical versus temperate cave biodiversity with at 
least 10 (troglobites + stygobites) species 

Ranked on 
Aquatic Terrestrial Total Terrestrial 

Europe Pyrenees 

Baget (FR)* 24 9 33 17 

Goueil di Herr (FR)* 14 12 26 14 
Europe: Italy 

Busso del Rana* 15 5 20 21 

Grotta dell’ Arena* 6 14 20 11 
Europe: Dinaric karst 

Grad (SLO)* 17 3 20 22 

Jama Logarcec (SLO)* 28 15 43 10 

Krisna Jama (SLO)* 29 16 45 8 

Sica-Krka system (SLO)* 27 7 34 18 

Sistem Postojna-Planina 48 36 84 1 


Vjetrenica Jama (Bosnia) 39 21 60 5 
Europe: Romania 

Pestera de Movile* 18 29 47 2 

Mammoth cave* 15 26 4l 3 

Shelta cave USA* 12 12 24 14 

China: Hunan 

Feihu Dong 4 16 20 8 
Tropical Australia 

Bayliss Cave* 0 24 24 4 
Southeast Asia 

Gua Air Jernih (SAR) 5 13 18 ie 

Ngalau Surat (SUM) i) 17 20 i 

SKT system (SUL) 7 21 28 5 

Kulumuzi cave (TAN) 0 10 10 16 
Tropical America 

Alambari de Cima 4 6 10 19 

cave (BRA) 
Areias System (BRA) 3 13 16 le 
Cueva Grande de 5 6 11 19 

Caguanes (CUBA) 

Abbreviations like table I. FR: France; SLO: Slovenia; * Culver and Sket 

Tropical cave fauna, therefore, differ from the richest 
temperate cave fauna by a different contribution of species 
from high- versus low-energy habitats and a different represen- 
tation of the main taxonomic groups, more than by their 
absolute richness in obligate subterranean species. 


Eleonora Trajano and Abel Perez Gonzalez provided important 
information about the biodiversity of tropical American caves. 


Chapman, P. (1980). The biology of caves in the Gunung Mulu National 
Park, Sarawak. Trans. Brit. Cave Res. Assoc. 7(3), 141-149. 

Culver, D. C., and B. Sket (2000). Hotspots of subterranean biodiversity in 
caves and wells. /, Cave Karst Studies 62(1), 11-17. 

Deeleman-Reinhold, C. L. (1995). The Ochyroceratidae of the Indo-Pacific 
region (Araneae). Raffles Bull. Zoology Suppl. 2, 1-103. 

Deharveng, L., and A. Bedos (2000).The cave fauna of Southeast Asia. 
Origin, evolution and ecology. In Ecosystems of the World: 30 Sub- 
terranean Ecosystems (H. Wilkens, D. C. Culver, and W. E. Humphreys, 
eds.), pp. 603-632. Elsevier, Amsterdam. 

Gnaspini, P, and E. Trajano (1994). Brazilian cave invertebrates, with a 
checklist of troglomorphic taxa. Revista Brasileira Entomologia 38(3/4), 

Howarth, FE. G. (1973). The cavernicolous fauna of Hawaiian lava tubes. I. 
Introduction. Pacific Insects 15, 139-151. 

Humphreys, W. F. (1993). Cave fauna in semi-arid tropical western 
Australia: A diverse relict wet forest-litter fauna. Mémoires de Biospéologie 
20, 105-110. 

lliffe, T. M. (2000). Anchialine cave ecology. In Ecosystems of the World: 
30 Subterranean Ecosystems (H. Wilkens, D. C. Culver, and W. E 
Humphreys, eds.), pp. 59-76. Elsevier, Amsterdam. 

Jeannel, R., and E. G. Racovitza (1914). Biospeologica XXXIII. 
Enumeration des grottes visitées 1911-1913 (5 éme série). Arch. Zoologie 
Expérimentale et Générale 53, 325-558. 

Juberthie, C., and V. Decu, eds. (1994). Encyclopaedia Biospeologica, Vol. 1, 
pp. 1-834. Société de Biospéologie, Moulis, France. 

Juberthie, C. and V. Decu, eds. (2001). Encyclopaedia Biospeologica, Vol. Ill, 
pp. 1375-2294. Société de Biospéologie, Moulis, France. 

Leleup, N. (1956). La faune cavernicole du Congo belge et considérations 
sur les coléopteéres reliques d’Afrique intertropicale. Ann. Mus. R. Congo 
Belge Sci. Zoologiques 46, 1-170. 

Diversity Patterns in the 
United States 

Horton H. Hobbs III 
Wittenberg University 


The scientific study of cave fauna (cavernicoles) in the 
United States had its inception in 1842 with the description 
by DeKay of the blind amblyopsid cave fish, Amblyopsis 

spelaea, from Mammoth Cave, Kentucky. By 1888 the 
number of obligate cave-dwellers, i.e., troglobites (terrestrial) 
and stygobites (aquatic species), had increased to more than 
50, as reported by Alpheus Packard in his compendium of 
North American cave fauna; numerous additional descrip- 
tions of cave-restricted organisms occurred through the 
1940s and 1950s. By 1960 Brother Nicholas listed 334 
species, in 1998 Stewart Peck estimated 1353 cave species 
(including interstitial and undescribed species and sub- 
species), and in 2000 Culver er a/. reported that a total of 973 
species and subspecies were described from caves in the 48 
contiguous states. Currently, 1121 species and subspecies are 
described (inclusive of Alaska and Hawaii) and are assigned 
to 238 genera and 112 families (Table I). Nearly all have 
evolved from the independent invasion of surface organisms 

TABLEI Summary of the Obligate Subterranean 
Phyla/Classes/Orders, Families, Genera, and 
Species/Subspecies Described from U.S. Caves 

Phylum/class/order Family Genus _Species/subspecies 
Turbellaria 4 6 28 
Oligochaeta 4 6 13 
Mollusca 5 16 29 
Acari y 13 22 
Aranaea 8 22, 98 
Opiliones 6 13 44 
Pseudoscorpiones 9 29 152 
Schizomida 1 1 
Scorpiones 1 1 1 
Copepoda 2 3 7 
Ostracoda 2 5 14 
Bathynellacea 1 1 1 
Thermosbaenacea 1 1 1 
Amphipoda 11 18 135 
Isopoda 6 15 97 
Shrimps 2 2 5 
Crayfishes 1 4 34 
Crabs 1 1 1 
Chilopoda 3 4 4 
Diplopoda 12 20 60 
Thysanura 1 1 1 
Diplura 2 3 8 
Collembola 6 12 75 
Orthoptera 1 2 3 
Dermaptera 1 1 1 
Hemiptera 2, 2 2 
Homoptera 1 1 7 
Coleoptera 7 24 261 
Diptera 1 1 1 
Osteichthyes 2 2 6 
Amphibia 1 5 9 
Totals 112 238 1,121 

Diversity Patterns in the United States 171 

into the subterranean realm and there, physical isolation and 
thus cessation of gene flow with surface congeners has led to 
speciation. In the stressed subterranean environment where 
there is perpetual darkness and extremely limited food, 
selective pressures have led to the evolution of parallel and 
convergent regressive forms (profound morphological alter- 
ations) that are characteristic of troglobites and stygobites 
worldwide. This combination of characteristics, which 
includes reduction or loss of eyes and pigments, gracilization 
and elongation of appendages, increased chemical and tactile 
sensitivity, degeneration of circadian rhythms, lowered 
fecundity and metabolic rates, and increased longevity, is 
postulated to be adaptive to life in such extreme ecosystems 
and is termed troglomorphy. 

This article summarizes the ecological, taxonomic, and 
geographic patterns of biodiversity of cave-inhabiting fauna 
in the 50 United States, focusing on troglobites and stygo- 
bites (separated because the two environments have com- 
pletely different environmental characteristics and faunas), 
but not ignoring cave visitors (trogloxenes/stygoxenes) and 
“cave lovers” (troglophiles/stygophiles). Specific numbers of 
organisms known from various cave areas (see below) are 
derived mostly from a cave biota database that is available on 
the World Wide Web at No 
attempt is made herein to summarize the protozoans or other 
microbial communities of caves although clearly these are 
incredibly important in cave ecosystems as sources of food 
(primary and secondary production), in the deposition of 
minerals, and in the process of speleogenesis. Additionally, 
attention is directed toward the vulnerability as well as to the 
challenges of conservation of this fauna and_ associated 


The obvious feature of caves (see article in this volume on 
“What Are Caves?” and Hobbs, 1992) is perpetual darkness 
that results in the absence of green plants (producers of 
carbon-based molecules and thus “food” for organisms). 
Other characteristics of these food-poor cavities are low 
variances in temperature and humidity (usually near 
saturation), which make for a predictable, less variable 
environment. Evolutionarily, as an organism makes the 
transition to troglobite or stygobite, isolation is another key 
feature of that environment. These are unique combinations 
of characteristics that set caves apart from other ecosystems 
and that make this extreme environment inhospitable to 
most organisms. 

The zonal variation in biological, chemical, and physical 
properties influences the distribution and abundance of 
fauna occurring in caves. The entrance area of caves is the 
ecotone between the epigean and hypogean worlds and has 
received only minimal study. Clearly, here there is more 
diversity and greater environmental variability than in any 
other cave zone. On both horizontal and vertical scales, 

172 Diversity Patterns in the United States 

entrances provide a transition of characteristics (e.g., 
temperature, humidity, light) that may provide conditions 
for entry or survival for preadapted species or for relictual 
species that otherwise are rare or have become extinct on the 
surface due to climatic changes. They are important windows 
into the subterranean realm through which pass migrating 
trogloxenes (e.g., bats, crickets) and can be the point of entry 
of important organic material. Further into the cave but still 
within the limits of light penetration (the twilight zone), the 
influence of surface conditions is apparent and variation of 
meteorological conditions is significantly less than at the 
surface and entrance areas. The dark zone initially demon- 
strates considerable influence of surface conditions but as 
distance increases from the entrance, it grades into a much 
less variable environment that is far from “constant” but 
significantly reduced in fluctuations of such parameters as 
atmospheric and water temperatures. In the dark, deep 
interior of caves that is characterized by more environmental 
constancy, virtually no food is produced and organisms are 
thus dependent on input of carbon from the surface that 
energetically supports most cave ecosystems (plant debris, bat 
and cricket guano). Exceptions to these generalizations about 
food occur in caves where chemoautotrophic production by 
sulfur-oxidizing microbial organisms (e.g., sulfur bacteria: 
Achromatium, Beggiatoa, Thiothrix) occurs, resulting in suffi- 
cient energy to support and sustain complex cave ecosystems. 
These types of caves are dominated geochemically by reduced 
sulfur compounds and are rare occurrences (e.g., Cesspool 
Cave in Allegheny County, Virginia, and Lower Kane Cave, 
Big Horn County, Wyoming). 

Cave organisms can be separated on the basis of habitat 
and/or resource base: terrestrial riparian communities found 
on stream banks with a resource base of allochthonous 
particulate organic matter deposited by stream fluctuations; 
terrestrial transitory organic matter (dung) communities usually 

living within a few hundred meters of the surface with a 
resource base of organic matter (often fecal material) that is 
derived from the activities of animals (e.g., bats, crickets) 
moving in and out of caves; terrestrial epikarst communities 
living primarily in the network of small, air-filled cavities 
above the cave but below the surface; aquatic stream commu- 
nities living primarily in cave streams ultimately dependent 
on dissolved and particulate organic matter derived from the 
surface; aquatic phreatic communities found in the permanent 
groundwater at or below the cave itself, including the 
hyporheos; and aquatic epikarst communities living primarily 
in the network of small flooded or partially flooded cavities 
above the cave but below the surface and most easily sampled 
within the cave in drip pools. 


Within the continental United States, available and 
appropriate habitat for subterranean fauna is not continuous, 
thus cave-inhabiting organisms are found in distinct areas 
(mostly karst), some being widely distributed and isolated 
(Table I, Figs. 1 and 2). Due to these separations and a 
unique geology, history, and climate, it should not be 
surprising that there are distinct differences in regional fauna, 
variances expressed in diversity, population densities, as well 
as taxonomic groups. Isolation, due in part to folding of 
strata, has resulted in high species richness (particularly 
of troglobitic beetles) in the Appalachians, whereas in the 
Interior Lowlands this is not the case, likely due to cave 
connectivity. Most cave-adapted species are found south 
of the southern limits of the Pleistocene glacial ice sheet, 
yet some were able to survive subglacial conditions in the 
northern states (e.g., New York, Wisconsin) as well as in 
Canada (e.g., Alberta, British Columbia) and are represented 

TABLE II Comparison of Species Richness and Characteristics of 10 U.S. Karst Regions“ 

Major karst regions Number of Number of Total Number of Area of 

of United States troglobites stygobites species caves karst (km?) 
Appalachians 178 (02) 85 (01) 263 = (02) 7,441 2) 37,268 5) 
Black Hills 2 (09) 0 (10) 2 (10) 160 9) Ts212 (8) 
Driftless Area 11 (08) 2 (08) 13 (09) 615 7) 25,222 7) 
Edwards/Balcones 105 = (03) 56 (03) 161 = (03) 2,011 4) 65,586 2) 
Florida Lime Sinks 0 (10) 24 (05) 24 (06) 627 6) 27,338 6) 
Guadalupes 13. (07) 1 (09) 14 (08) 1,379 5) 43,522 4) 
Hawaii (lava) 35. (04) 6 (06) 41 (05) ? ? 
Interior Lowland Plateau 256 (01) 62 (02) 318 (01) 11,928 1) 60,612 3) 
Mother Lode 20 = (06) 3 (07) 23 ~~ (07) 179 8) 390 9) 
Ozarks 31 (05) 51 (04) 82 (04) 6,964 3) 110,125 (1) 
Total 651 290 941 31,304 377,335 

“Ranks from highest to lowest are presented in parentheses (modified from Culver and Hobbs, 2002). An additional 179 species are from other cave and 
karst regions throughout the contiguous United States and one described stygobitic amphipod occurs in Alaska, bringing the total number of described obligate 

cavernicoles in the United States to 1121. 

Diversity Patterns in the United States 173 

{ Driftl 



Edwards Plateau 
& Balcones 

FIGURE 1 Map of major karst regions treated in this article (excluding the Hawaiian Islands, which are lava). The Interior Lowland Plateau is stippled in 
order to differentiate it from the Appalachians (extend into Canada; truncated herein at Maryland—Pennsylvania border) and the Ozarks. (Modified from 
Culver et al., 2003.) 


FIGURE 2 Dot map of the distribution of caves in the United States. Each dot represents one cave. (Modified from Culver et a/., 1999.) 


Diversity Patterns in the United States 

primarily by groundwater amphipods and isopods. In the 
Driftless Area (parts of Illinois, Iowa, Minnesota, and 
Wisconsin) and other smaller karst regions above the glacial 
boundary, low biodiversity in caves is attributed to the lack 
of time for colonization and isolation due to the effects of 
glaciers. Also, the Guadalupe Mountains of New Mexico 
demonstrate depauperate cave fauna likely resulting from the 
aridity of the region. Although high densities of lava tubes 
occur in the western United States and limestone caves have 
developed in numerous smaller karst areas, this article focuses 
primarily on caves formed in soluble rock, particularly 
limestone and follows the nine geologically defined cave 
regions discussed by Culver and Hobbs (2002) (Fig. 1). 
These are the Appalachians, Black Hills, Driftless Area, 
Edwards Plateau & Balcones Escarpment, Florida Lime 
Sinks, Guadalupe Mountains, Interior Lowland Plateaus, 
Mother Lode, and the Ozarks. Also, a tenth, the Hawaiian 
region and one of volcanic origin (thus many islands are 
“young” ecologically), is treated. 

Caves of the coterminous United States are shown in 
Fig. 2 resulting from a plot of nearly 45,000 caves. Clearly 
the distribution of cavernicoles is influenced by the distribu- 

tion and abundance of cavities available for habitation. 
Indeed, although it is useful to examine the number of 
cavernicoles in states (Fig. 3) or within various karst areas 
(Fig. 4), Christman and Culver (2001) and Culver et al. 
(2003) demonstrated that the availability of habitat expressed 
by the number of caves in a region (Table II) is the best 
predictor of the biodiversity of cave-dwelling organisms. 
Also, the distribution of cavernicoles within caves is patchy 
and nonrandom and usually is associated with concentrations 
of food sources. Whereas troglobite diversity within caves is 
likely affected not only by resource availability but also by 
variety (organic plant debris, bat or cricket guano, mamma- 
lian scat), stygobites tend to be less diverse in part due to 
the preponderance of feeding generalists and the lack of 
specialization on food type. 

Class Turbellaria 

This diverse group of stygobitic flatworms (28 species) is 
assigned to four families and six genera. The largest genus, 
Sphalloplana, is represented by 16 species found in subterra- 

KA 160 9 141 
"5 140m 125 15 
) J ul 
120 5 
S 8075 
5 60 5 48 46 43 
= 40 - [| + at. 9 
15 14 
5B 207 in N98 6 6 6 6 6 6 6 6 5 4 3 3 2 2 4 7 4 
Z o- [= s I 2 Ebene eee ee eee ee KK 
ral J = < % em N 1S) S > EAA > 
g | w 
3 Om 
40 3449 
30 a TT 44 
2 ee 8 16 45 
a PW tdideccccccc: 
2 2 1 1 1 1 1 1 1 1 1 
ei Be Bee eee a a Be eee 
= N se 
S 200 vi 
is 170 
2 160 148 143 14) 
% 120 
et 80 
2 80 57 
Ea iT T 7 Bo 95 
5 3 1 4 1 MN 0 9 9 9 9 7 7 6 5s 4 4 
2 91 Tibnneas Bee = pee ae ee a eee ee ae Bee 
f= be 

FIGURE 3 Frequency histograms by state of number of described obligate subterranean species. (A) Troglobite biodiversity by state, (B) stygobite biodiversity 

by state, and (C) combined cavernicolous biodiversity by state. 

Diversity Patterns in the United States 175 


® 250 - 

= 200 4 178 A 


ty (150 5 105 

5 100 - 

= 353 

— 50> a 20 13 «11 2 

Z 0 T T HE fs = = - 
Ay A, m = N < = an Nn 
7 < @ = 6 8 > Oo m 

100 - 

20 5 

Number of Stygobites 
° 5 8 &§ 
AP a = 
LP a 2 
a 2 
HI Jfo 


GUA | - 

3 400 5 C 
E 318 
7) = 
> 300 263 
2 200 - 161 
2 100 5 
S li: 200 14 138g 
$ 0 I T - -_mie e 2 
= Ba ae MN FB HAH fv << fF 
ae < 8 oF As 5 A wg 
Major Karst Region 

FIGURE 4 Frequency histograms by major karst region of described obligate cavernicoles. (A) Troglobites, (B) stygobites, and (C) total. AP, Appalachians; BH, 
Black Hills; DL, Driftless Area; EB, Edwards Plateau & Balcones Escarpment; FLS, Florida Lime Sink; GUA, Guadalupes; HI, Hawaii; ILP, Interior Lowland 
Plateau; ML, Mother Lode, OZ, Ozarks. 

176 Diversity Patterns in the United States 

nean settings primarily in the Appalachians and Interior 
Lowlands major karst regions. 

Class Oligochaeta 

This group of aquatic worms is represented by 13 species 
belonging to six genera and four families mostly from the 
Appalachians and Interior Lowlands. They are poorly known 
from cave soils and streams as well as from interstitial waters, 
the latter habitat only rarely sampled. Of interest, 7 species of 
the genus Cambarincola (Branchiobdellidae) are restricted as 
ectosymbionts to cambarid crayfishes occupying subterra- 
nean waters. Tubificid worms are common to subterranean 
streams containing much organic material (or that are receiv- 
ing sewage effluent!), but no stygobitic forms are known. 

Class Mollusca 

Both aquatic and terrestrial snails inhabit caves, the diversity 
of stygobites being nearly six times greater than that of 
troglobites. Twenty-four stygobionts are placed in 12 genera 
in two families and three families contain four genera and 
five species of generally small, translucent troglobitic 
gastropods. Numerous species of terrestrial epigean snails are 
deposited in caves on debris transported in by floods but 
apparently are not able to maintain populations for many 
generations. Yet a few species are troglobitic (mostly from the 
Appalachians and Interior Lowlands karst regions, e.g., 
Helicodiscus barri Hubricht), feeding on guano and decaying 
allochthonous material (mainly plant) and associated 
microbes. Hydrobiid snails have been particularly successful 
in subterranean waters with 23 stygobites described primarily 
from the Appalachians, Edwards Plateau & Balcones Escarp- 
ment, Interior Lowlands, and Ozarks. Some of these snails 
are site endemics (e.g., Antrobia culveri Hubricht, known 
from a stream in a single cave in Taney County, Missouri) or 
are restricted to a single drainage (e.g., Fontigens turritella 
Hubricht, from two caves in the Greenbrier River drainage in 
Greenbrier County, West Virginia). Physa spelunca Turner & 
Clench is unique in that it is a site-endemic species restricted 
to a thermal cave stream (Lower Kane Cave; see above) 
draining into Big Horn River, Wyoming. 

Class Arachnida 

The cave-adapted arachnids are represented by a very diverse 
group of mostly terrestrial arthropods that are widely 
distributed throughout karst terrains of the United States. 
Mites (Order Acari) are small arachnids that are dominated 
by terrestrial species (11 genera and 19 species) with only 
three stygobites placed in two genera. Mites are undoubtedly 
underrepresented in these totals due primarily to inadequate 
sampling. Seven families are represented, with the rhagidiids 
being the most diverse. Although most of the known mites 
are predaceous and commonly observed moving rapidly 

among organic debris in moist areas of caves, some species 
are parasites on bats and harvestmen. Stygobites are known 
only from the Interior Lowlands (Indiana), and the greatest 
diversity of troglobites resides in the Appalachians and 
Interior Lowlands; at least one species is endemic to lava 
tubes on Hawaii. No troglobitic ticks are known but in some 
localities ticks are common, particularly in bat caves. 

Spiders (Araneae) have been quite successful in the 
terrestrial cave environment with 98 species in 22 genera and 
8 families. These troglobites are known from all the major 
karst areas except the Black Hills and the Florida Lime Sinks. 
The genus Cicurina is particularly diverse in Texas (49 
species) and Phanetta subterranea (Emerton) and Porhomma 
cavernicola (Keyserling) are widely dispersed throughout the 
Appalachians and Interior Lowlands, and less widely within 
the Ozarks. Five species assigned to five genera and three 
families are known from lava tubes in Hawaii. In addition to 
the troglobites, a significant number of facultative species 
le.g., Gaucelmus augustinus Keyserling, troglophile; Meta 
ovalis (Latreille), trogloxene; Nesticus carteri Emerton, 
troglophile] are common inhabitants of caves in the major 
karst regions. 

The Order Opiliones (harvestmen or “daddy longlegs”) are 
widely distributed throughout the karst regions of the United 
States but are not overly diverse except in the Mother Lode 
(13 species) and Edwards Plateau & Balcones Escarpment 
(13 species). A total of 44 species has been assigned to 13 
genera and six families that are scattered over the 
Appalachians, Interior Lowlands, and other lesser karst areas, 
as well as those mentioned above. The genus Banksula is 
diverse in the Mother Lode and environs (nine species) and 
Texella is represented by 11 species in the Edwards Plateau & 
Balcones Escarpment region as well as an additional three 
species in California and the Guadalupes. Large populations 
of trogloxenic harvestmen are often encountered in ceiling 
pockets or beneath ledges, out of the dessicating effects of 
air movements where they form dense, undulating mats 
(e.g., Letobunum townsendii Weeds in caves in the Edwards 
Plateau, Texas, and L. bicolor Wood in southern Ohio caves]. 
These trogloxenic species are typically scavengers often 
associated with organic matter washed into caves while the 
troglobitic forms are more commonly predators feeding on 
microarthropods, such as collembola. 

Pseudoscorpions (Order Pseudoscorpiones) are minute 
predators that have pincer-like pedipalps (as do scorpions), 
but lack the tail and sting characteristic of scorpions. Some 
are trogloxenic/troglophilic and often are associated with bat 
guano communities, whereas others are highly specialized 
troglobites. This highly diverse group of small arachnids (152 
species within 29 genera and nine families) is particularly 
abundant in Alabama caves (41 species) and also demon- 
strates high diversity in Texas (25 species), and California and 
Virginia (13 species each). They are found in all major karst 
regions except the Florida Lime Sinks, the Black Hills, and 
the Driftless Area. Two families are particularly widespread, 

Chernetidae and Chthoniidae, with the latter being 
incredibly diverse (95 species). The chernetids are dominated 
by the extensive genus Hesperochernes, which is found in the 
Appalachians, the Edwards Plateau & Balcones Escarpment, 
Interior Lowlands, Mother Lode, and the Ozarks. The 
chthoniids are heavily influenced by the genera Apochthonius 
(15 species), Kleptochthonius (31 species), and Tyrannochtho- 
nius (37 species) and are found in all the major karst regions 
except the Black Hills, Florida Lime Sinks, and the Driftless 
Area. Two species of the genus Tyrannochthonius are found in 
lava tubes on the Hawaiian Islands (Hawaii and Oahu). 

The Order Schizomida is represented by a single terrestrial 
species, Hubbardia shoshonensis Briigs & Hom, from a cave 
in Inyo County, California. Although similar to true whip 
scorpions, they have a much shorter telson as well as other 
anatomical features that separate the two orders. They reside 
in leaf litter and under stones and are predators on small 

Scorpions (Order Scorpiones) are believed to be one of the 
most ancient terrestrial arthropods and the most primitive 
arachnid. All are terrestrial predators characterized by two 
large pedipaps terminating in chelae as well as by the stinging 
apparatus derived from the telson and bearing a sharp barb 
called the aculeus. Although numerous pigmented species are 
found in U.S. caves, only one species is troglobitic: Uroctonus 
grahami Gertsch & Soleglad, known only from Shasta 
County, California. At least 11 species are troglobitic to the 
south in Mexican caves. 

Subphylum Crustacea 

CLASS MAXILLOPODA The Subclass Copepoda (see 
Crustacea article, this volume) is not well sampled in the cave 
environment in the United States and this is reflected in 
the low species richness reported. Currently two families 
containing three genera and seven species and subspecies of 
these crustaceans are known from the Edwards Plateau & 
Balcones Escarpment in Texas, the Interior Lowlands in 
Illinois, Indiana, Kentucky, and Tennessee, and from Orange 
County, North Carolina. One obligate parasitic copepod, 
Cauloxenus stygius Cope, is associated with the Northern 
Cavefish, Amblyopsis spelaea, in southern Indiana caves. 

Like the copepods, the Class Ostracoda has not been 
studied thoroughly in U.S. caves. Even so, 14 species 
belonging to five genera and two families are stygobites in the 
Appalachians, Edwards Plateau & Balcones Escarpment, 
Florida Lime Sinks, Interior Lowlands, and the Ozarks. 
Twelve species of the family Entocytheridae are commensal 
on cambarid crayfishes inhabiting caves and two species of 
Cyprididae are free living in cave streams and lakes. 

CLASS MALACOSTRACA The Order Bathynellacea (see 
Crustacea article, this volume) is characterized by lacking 
maxillipeds and a carapace. These small crustaceans are not 
well studied, particularly in the United States, where only 

Diversity Patterns in the United States 177 

one species, [berobathynella bowmani (Delamare Deboutte- 
ville, Coineau, & Serban), is known from two localities in 
Dickens and San Saba counties, Texas. These reclusive 
organisms are interstitial or reside in groundwater, those 
habitats not having been adequately sampled. 

The Order Thermosbaenacea (see Crustacea article, this 
volume) is represented by a single species, Monodella texana 
Maguire, which is known from a cave and wells in Bexar, 
Hays, and Uvalde counties, Texas. This is the sole North 
American representative of the Thermosbaenacea and is 
endemic to the Edwards Aquifer. 

The Order Amphipoda (see Crustacea article, this volume) 
is quite diverse with 134 stygobitic species and subspecies in 
18 genera and 10 families. An additional species is a talitrid 
troglobite, Spelaeorchestia koloana Bousfield & Howarth, 
which is a rare endemic that feeds on plant roots and organic 
debris in a few caves on the island of Kauai, Hawaiian 
Islands. Most amphipods are scavenger/predators and are 
laterally compressed, giving them a shrimplike appearance. 
Amphipods occur in all major karst regions except the Black 
Hills, Driftless Area, and Guadalupes. Crangonyctidae is the 
most diverse and widely distributed family, with the genera 
Bactrurus and. Crangonyx housing 7 and 12 stygobites, respec- 
tively, yet the genus Stygobromus (98 species) is approximately 
five times as diverse as those two genera combined. In 
addition to most of these being found in unglaciated regions, 
S. quatsinensis Holsinger & Shaw is endemic to caves and 
springs north of the southern limits of Pleistocene glaciation 
in British Columbia, Canada, and southeastern Alaska. 
Also, eyed and pigmented stygophilic amphipods commonly 
inhabit cave and karst spring waters of the contiguous states 
and are major players in the trophic dynamics of some 
systems [e.g., Gammarus spp. and Hyalella azteca (Saussure)]. 

The Order Isopoda (see Crustacea article, this volume) is 
represented in caves by 78 stygobites in 11 genera and four 
families. The family Asellidae is immense with all but eight 
species of the order assigned to this family of dorsoventrally 
flattened crustaceans. Geographically, these stygobites are 
widespread, absent only from the Black Hills, Driftless Area, 
Guadalupes, and the Mother Lode major karst regions. The 
genus Caecidotea (Fig. 5) is represented by 56 species and 
subspecies and widely dispersed over these regions, whereas 
other species are much more restricted [e.g., Remasellus 
parvus (Steeves) in Florida and Salmasellus howarthi Lewis in 
Washington]. The genus Lirceus (also an asellid) has two 
stygobitic species from southwestern Virginia (L. culveri 
Estes & Holsinger and L. usdagalun Holsinger &Bowman 
from Scott and Lee counties, respectively) and numerous 
stygophilic species (e.g., L. fontinalis Rafinesque) that occupy 
springs and caves. The predominantly marine family 
Cirolanidae is represented in the United States by three 
dome-shaped and elliptical stygobitic species, Cirolanides 
texensis Benedict and Speocirolana hardeni Bowman, both 
from Texas, and the widely disjunct Antrolana lira Bowman 
from Virginia and West Virginia (Appalachians). 

178 Diversity Patterns in the United States 

FIGURE 5 The aquatic isopod, Caecidotea stygia Packard, from Mammoth 
Cave, Edmonson County, Kentucky. 

The troglobitic isopods are represented by 19 species and 
subspecies in four genera. The 15 species in the contiguous 
United States belong to three genera (Amergoniscus, 
Brackenridgia, Miktoniscus) within the Trichoniscidae and 
these “pillbugs” are found in the Appalachians and Interior 
Lowlands as well as in Oklahoma, Oregon, and Texas. Four 
species of the genus Hawaiioscia are endemic to the Hawaiian 
Islands of Kauai, Maui, Molokai, and Oahu. 

Stygobitic members of the Order Decapoda (see Crustacea 
article, this volume) are limited to 34 crayfishes, five shrimps, 
and one crab. Crayfishes belong to the family Cambaridae: 
11 species of Cambarus in the Appalachians, Interior Lowlands, 
and the Ozarks; 8 species and subspecies of Orconectes 
(Fig. 6) in the Interior Lowlands and Ozarks; 14 species and 
subspecies of Procambarus in the Florida Lime Sinks and 
Interior Lowlands; and a single species of Troglocambarus in 
the Florida Lime Sinks. Shrimps belong to two families: 
Atyidae, 2 species of Palaemonias in the Interior Lowlands, 
and Palaemonidae, three species of Palaemonetes in the 
Florida Lime Sinks and Edwards Plateau & Balcones 
Escarpment. The single crab (Grapsidae; Hemigrapsus 
oregonesis (Dana) (formerly Hemigrapsus estellinensis Creed) 
is endemic, and probably extinct, to a deep artesian salt 
spring in Hall County, Texas. Approximately 50 stygophilic/ 

FIGURE 6 Orconectes (Orconectes) inermis testii (Hay) from Mayfield’s Cave, 

Monroe County, Indiana (internal mark on second abdominal segment is 

used in population studies). 

stygoxenic crayfishes utilize caves and, of special note and 
concern, is the impact of introduced epigean species that 
utilize cave streams [e.g., Procambarus (Scapulicambarus) clarkii 
(Girard) in a cave in San Diego County, California, and likely 
introduced into the area from the southern Gulf states]. 

Troglobitic centipedes, Class Chilopoda, are rare in caves, 
fast moving, usually small, slender, and are voracious 
predators. They are found in organic debris, on clay, and 
burrowing in silt. Four genera and species from three families 
are known from the Guadalupes and the Appalachians as well 
as from northcentral California and the Edwards Plateau & 
Balcones Escarpment. 

Troglobitic millipedes, Class Diplopoda, are slow-moving 
grazers, feeding on organic detritus and associated microbes. 
Typically found in food-poor caves, some 60 species and 
subspecies in 20 genera are known primarily from the 
Appalachians, Driftless Area, Guadalupes, Interior Lowlands, 
and Edwards Plateau & Balcones Escarpment. The largest 
genus, Pseudotremia, a member of the Cleidogonidae, has 29 
troglobitic species in Appalachian and Interior Lowland 
caves. Numerous troglophiles are noted in these karst regions 
as well (Fig. 7). 

CLASS HEXAPODA The Class Hexapoda is restricted to 
the primitive orders of insects lacking wings (apterygote) 
and, in caves, is represented by three orders: Thysanura, 
Diplura, and Collembola. 

Troglobitic thysanurans (bristletails) are unusual in caves 
and only a single species, Texoreddellia texensis (Ulrich), a 
member of the Nicoletiidae family, is known from caves in 
18 counties in Texas. Trogloxenic species (e.g., Pedetontus sp.) 
are not uncommon in entrance areas. 

Troglobitic dip/urans are small and lack a median caudal 
filament but possess two cerci (“tails”). Eight species are 
assigned to three genera and two families, with the campo- 
deid Litocampa being the most diverse genus with six species. 
This poorly studied group is known only from the 

FIGURE7 Unidentified troglophilic, polydesmid diplopod from Bat Cave, 
Alachua County, Florida. 

Appalachians, Interior Lowlands, the Edwards Plateau & 
Balcones Escarpment, and a single species from Lincoln 
County, Nevada (Condeicampa langei Ferguson). These are 
observed on wet flowstone, cricket guano, and on silty 

Troglobitic springtails (Order Collembola) are small, 
jumping insects that can occur in very high densities, 
particularly where organic matter has accumulated. Often 
they are observed moving about on the surface of small drip 
pools as well as on larger stream pools. Not only do they 
provide food for many small predators but they also graze on 
bacteria and fungi. They are a diverse group with 75 troglo- 
bitic species assigned to 12 genera and are known from all the 
major karst regions, except the Florida Lime Sinks, as well as 
other minor karst areas. Entomobryidae is the largest family, 
with the greatest diversity in the genus Pseudosinella (25 
species); Sinella has 13 species. Fifteen species make up 
the widely distributed genus Arrhopalites (Sminthuridae). 
Numerous troglophilic species [e.g., Sinella cavernarum 
(Packard)] contribute to the terrestrial communities of many 
caves. Six troglobitic species assigned to three genera and two 
families are found in lava tubes on the islands of Hawaii, 
Maui, and Oahu. 

CLASS INSECTA The Class Insecta is characterized by 
those insects possessing wings (pterygote) or secondarily with 
reduced wings and is represented in caves by troglobitic/ 
stygobitic members of five orders. Troglobitic members of the 
Order Orthoptera are known only from Hawaiian lava tubes. 
Three species of flightless, mute crickets are assigned to two 
genera and belong to the family Gryllidae. Caconemobius 
varius Gurney & Rentz is known from five caves on Hawaii 
Island and displays small eyes, reduced body pigment, and a 
translucent exoskeleton. It is omnivorous or perhaps a 
scavenger and appears to be most abundant in low, wet caves. 
C. howarthi Gurney & Rentz has been collected from four 

Diversity Patterns in the United States 179 

caves on the island of Maui. It has been observed only in the 
dark zone of caves and appears to be more cave adapted than 
C. varius. Thaumatogryllus cavicola Gurney & Rentz is 
known from approximately 10 small, generally shallow caves 
on the island of Hawaii. In some of these lava tubes it has 
been observed on dangling roots of grasses and trees that have 
penetrated the ceilings in characteristically wet areas. In the 
contiguous states, pigmented rhaphidophorid crickets often 
are very numerous in caves, particularly within the first 
100 to 200 m of the entrance, and these trogloxenes are 
represented by species of the genera Ceuthophilus, Hade- 
noecus, and Euhadenoecus. These crickets not only migrate to 
the surface to forage for food during spring, summer, and fall 
(and thus provide food in the form of guano to caves), but 
they also use caves as refugia during cold periods. Some 
crickets (e.g., Ceuthophilus silvestris Bruner in small caves in 
glaciated Wyandot County, Ohio) interact with caves only 
seasonally, entering them solely during the winter months. 

A single endemic, troglobite (Anisolabis howarthi Brindle) 
from Hawaii is the only known member of the Order 
Dermaptera (earwigs) to inhabit caves. Characterized by the 
heavily sclerotized posterior forceps (cerci), they use these 
for predation and are considered to be scavenging omnivores 
as well. 

Only two troglobitic “bugs” (Order Hemiptera— 
Heteroptera) are known, both endemic to lava tubes on the 
island of Hawaii. The thread-legged bug, Nesidiolestes ana 
Gagne & Howarth (Reduviidae), is a predator of arthropods, 
and the lava tube water-treader, Speleovelia aaa Gagne & 
Howarth, sucks on rotting fluids of deceased arthropods. 

The Order Homoptera is represented by seven troglobitic 
planthoppers endemic to the islands of Hawaii, Maui, and 
Molokai. All belong to the Cixtidae family and are placed in 
the genus Oliarus. Often they are observed on roots that have 
penetrated ceiling cracks of lava tubes. 

The Order Coleoptera has been incredibly successful in 
caves with 261 species in 21 genera and five families. Only 
three species are aquatic, two of which are dryopids (Oregon 
and Texas) and one a dytiscidae from Texas (Haideoporus 
texanus Young and Longley, the only stygobitic predaceous 
diving beetle known in the United States). Beetles are 
major components of cave communities particularly in 
the Appalachians and Interior Lowlands. Troglobites are 
dominated by the carabids, particularly the genus Pseuda- 
nophthalmus that is inordinately abundant in Kentucky and 
Texas caves, 62 and 67 species, respectively. Two carabids are 
endemic to lava tubes on the Hawaiian island of Maui. 
Troglobitic ground beetles are represented by the genus 
Rhadine (17 species), which is restricted to Texas, particularly 
dense in the Edwards Plateau & Balcones Escarpment karst 
region. In certain Texas caves, Rhadine s. subterraneanea (Van 
Dyke) and R. noctivaga Barr feed exclusively on cave cricket 
eggs (Ceuthophilus spp.) that have been buried in fine- 
grained, calcareous deposits. This feeding behavior also has 
been demonstrated in Mammoth and other caves in the 

180 Diversity Patterns in the United States 

Pennyroyal Plateau of western Kentucky (Interior Lowlands) 
for the carabid Neaphaenops tellkampfii (Erichson) preying 
on eggs of the cricket Hadenoecus subterraneus, and on the 
Cumberland Plateau of eastern Kentucky the cricket is 
Hadenoecus cumberlandicus and its predator is Darlingtonea 
kentuckensis Valentine. Leiodids (Ptomaphagous spp.) and 
pselaphids (predominantly Batrisodes spp.) make up the 
remaining troglobites. Numerous troglophiles (e.g., 
Carabidae, Leiodidae, Leptodiridae, Ptinidae, Staphylinidae, 
Tenebrionidae) also are common members of cave commu- 
nities. In addition to predators, the cavernicolous beetles are 
scavengers and opportunistic saprophiles. 

A single, widely distributed troglobitic fly (Order Diptera) 
is known from caves in the Appalachians, the Interior 
Lowlands, and the Ozarks: Spelobia tenebrarum (Aldrich), 
Sphaeroceridae. A variety of flies utilize caves periodically 
and include various culicids (e.g., Culex pipiens Linnaeus), 
heleomyzids [e.g., Amoebaleria defessa (Osten-Sacken)], 
mycetophilids (e.g., Macrocera nobilis), phorids [e.g., 
Megaselia cavernicola (Brues)], and sphaerocerids [e.g., 
Leptocera tenebrarum (Aldrich)]. 

Additional insects are observed in caves but are either 
troglo/stygoxenes or troglo/stygophiles and include lepidop- 
terans [e.g., Scoliopteryx libatrix (Linnaeus)], psocopterans 
(e.g., Psyllipsocus ramburi Sélys-Longchamps), and siphonap- 
terans [e.g., Myodopsylla insignis (Rothschild)]. A particularly 
disturbing example of surface insects entering the sub- 
terranean world is the invasion of central Texas caves during 
the summer months by the trogloxenic, exotic red fire ant, 
Solenopsis invicta (Hymenoptea). 


A number of vertebrates spend varying times of their life in 
subterranean environments but most are not considered 
cavernicoles (e.g., some snakes and mammals are adapted to 
a fossorial mode of life in the soil). Other vertebrates visit 
caves for shelter during times of unfavorable surface environ- 
mental conditions (e.g., amphibians, snakes), to avoid preda- 
tion (e.g., pack rats, birds), or for breeding purposes (e.g., birds, 
bats). Generally these cave-dwellers are classified as acciden- 
tals or as trogloxenes/stygoxenes; fishes and salamanders 
are the only vertebrates in which troglophilic/stygophilic or 
troglobitic/stygobitic species have evolved. 

Class Osteichthyes 

Approximately 85 stygobitic fishes are known worldwide, 
however only six cave-obligate species are recognized from 
North American subterranean waters and belong to one of 
two families, Amblyopsidae and Ictaluridae. The amblyopsid 
cave fishes are stygobitic except for one stygophilic species, 
Chologaster agassizii Putnam, which inhabits caves and 
springs of Illinois, Kentucky, and Tennessee yet migrates to 
the surface at night to exploit food resources. Four additional 

FIGURE 8 The blind cave fish Amblyopsis spelaea DeKay from Mammoth 
Cave, Edmonson County, Kentucky. 

species are confined to subterranean waters: Amblyopsis rosae 
(Eigenmann) from the Ozarks, Amblyopsis spelaea (Fig. 8) 
from the Interior Lowlands, Speoplatyrhinus poulsoni Cooper 
& Kuehne from a single cave in the Interior Lowlands, and 
Typhlichthys subterraneus Girard from the Interior Lowlands 
and the Ozarks. The ictalurid subterranean representatives, 
Satan eurystomus Hubbs & Bailey and Trogloglanis pattersoni 
Eigenmann, are known only from deep phreatic waters of the 
Edwards Aquifer in Bexar County, southwest Texas, and have 
never been observed in accessible caves. Members of both of 
these families are predatory feeders primarily of crustaceans 
except the toothless 7’ pattersoni, which is likely a grazer 
taking up detritus. 

Other fishes occupy subterranean waters (some regularly 
demonstrate migratory patterns) and some of the families 
represented are the following: Anguillidae [e.g., Anguilla 
rostrata (LeSueur)], Centrarchidae (Lepomis spp.), Cottidae 
(Cottus spp.), Cyprinidae (Notropis spp.), Ictaluridae 
Uctalurus catus (Linnaeus)], Mugilidae (Mugil cephalus 
Linnaeus), and except for Cottus sp., show no troglomorphic 
adaptations for living in this environment. 

Class Amphibia 

Oddly, no stygophilic or stygobitic frogs are known from 
caves (possibly because they cannot complete their life 
history in the subterranean environment), but they certainly 
utilize caves (e.g., Rana palustris LeConte), primarily to 
retreat from surface climatic extremes. Globally, only 11 
stygobitic salamanders are known from hypogean waters and 
are assigned to two families, Proteidae and Plethodontidae, 
the former restricted to limestone caves and karst springs in 
Bosnia-Herzegovina, Croatia, Italy, and Slovenia. The 
plethodontids are represented by nine species and subspecies 
in North America (Fig. 9) and are known from the 
Appalachians, Edwards Plateau & Balcones Escarpment, 
Interior Lowlands Plateau, and the Ozarks as well as from 
caves in the Dougherty Plain of southwestern Georgia and 
adjacent panhandle of Florida (Tables II and III). Numerous 
additional species of plethodontid salamanders are stygo- 

FIGURE9 Adult Zyphlotriton spelaeus Stejneger from Copperhead Cave, 
Newton County, Arkansas. (Photograph courtesy of Rob Payn.) 

TABLE III Stygobitic Salamanders of the United States 
Genus Species/ State Major karst region 
Eurycea tridentifera Texas Edwards Plateua & 
Balcones Escarpment 
Gyrinophilus  gulolineatus "Tennessee Appalachians 
palleucus Tennessee Interior Lowland Plateau 
palleucus Alabama, Appalachians, Interior 
palleucus Georgia Lowland Plateau 

subterraneus West Virginia Appalachians 

Haideotriton  wallacei Georiga, Dougherty Plain, NE 
Florida Florida Panhandle 
Typhlomolge — rathbuni Texas Edwards Plateau & Balcones 
robusta Texas Edwards Plateau & Balcones 
Typhlotriton — spelaeus Arkansas, Ozarks 
(Fig. 9) Kansas 

Diversity Patterns in the United States 181 

philic [e.g., Pseudotriton ruber (Latreille)] and troglophilic 
le.g., Eurycea lucifuga (Rafinesque)]. 

Class Reptilia 

Although there are no known obligate reptiles from caves in 
the United States, these habitats are used often by snakes for 
short durations or even significant amounts of time. Most 
tend to remain near entrances and use the cave as a refuge 
from extreme surface conditions. Some snakes prey on bats 
during their flights out of the cave [e.g., the colubrid Elaphe 
obsoleata (Say)] and certain lizards are observed in the 
entrance areas of caves as well [e.g., the geckonid Coleonyx 
variegatus brevis (Stejneger) in Texas]. In Florida the 
American alligator [Alligator mississipiensis (Daudin)] and the 
cooter turtle [Pseudemys floridana (LeConte)] are seen in the 
entrances to many submerged caves and will venture short 
distances into these cavities that serve often as the source of 
large karst springs. 

Class Aves 

The most noted of all birds that utilize caves is probably the 
troglophilic oilbird of Cueva del Guacharo in Venezuela, 
South America (Steatornis caripensis Humboldt), for 
centuries prized by locals as a source of cooking and lighting 
oil. Although the eyes of this bird are large and well 
developed for use in dim light, they utilize echolocation (as 
do bats) when in the complete darkness of caves. In the 
United States, no such highly specialized species occur in 
caves yet swallows (e.g., Hirundo spp.), the canyon wren 
[Catherpes maxicanus (Swainson)], and the eastern phoebe 
[Sayornis phoebe (Latham)] are common inhabitants of cave 
entrances. Various species inhabit the twilight zone of caves, 
including the eastern (common) screech-owl [Ozus asio 
(Linnaeus)} and the turkey vulture [Cathartes aura 
(Linnaeus)] nests in this semidark zone of caves. 

Class Mammalia 

The occurrence of mammals in caves is quite common 
although none are obligate and most are temporary residents/ 
visitors, including numerous rodents such as pack rats 
(Neotoma spp.), mice (Peromyscus spp.), beaver (Castor 
canadensis Kuhl), the wood chuck, Marmota monax 
(Linnaeus), as well as carnivores such as bears (Ursus spp.) 
and raccoons [Procyon lotor (Linnaeus)]. Clearly no bats are 
obligate cave-dwellers yet 30 of the 53 species and subspecies 
found in the United States occupy caves at least occasionally. 
They are assigned to four families (Table IV) and most are 
insectivorous (e.g., the endangered Indiana Bat, Myotis 
sodalis Miller and Allen), but some feed on fruits, pollen, and 
nectar (e.g., Mexican long-tongued bat, Choeronycteris 
mexicana Tschudi). During the winter most hibernate in 
caves (or mines) often in large, dense clusters of up to several 

182 Diversity Patterns in the United States 

TABLEIV Cave-Dwelling Bats of the United States 

Family Genus Species/subspecies Common name Status* 
Molossidae Tadarida brasiliensis Brazilian free-tailed bat NT 
Mormoopidae Mormoops megalophylla Ghost-faced bat NT 
Phyllostomidae Choeronycteris mexicana Mexican long-tongued Bat OSC 
Leptonycteris curasoae yerbabuenae Lesser long-nosed bat E 
Leptonycteris nivalis Greater long-nosed bat E 
Vespertilionidae Antrozous pallidus Pallid bat NT 
Eptesicus fuscus Big brown bat NT 
Idionycteris plyllotis Allen’s big-eared bat OSC 
Lasiurus borealis Eastern red bat NT 
Lasiurus cinereus Hoary bat NT 
Lasiurus seminolus Seminole bat NT 
Myotis auriculus Southwestern bat NT 
Myotis austroriparius Southeastern bat OSC 
Myotis evotis Western long-eared bat OSC 
Myotis grisescens Gray bat E 
Myotis leibii Eastern small-footed bat OSC 
Myotis lucifugus Little brown bat NT 
Myotis septentrionalis Northern long-eared bat NT 
Myotis sodalis Indiana bat E 
Myotis thysanodes Fringed bat OSC 
Myotis velifer Cave bat OSC 
Myotis volans Long-legged bat OSC 
Myotis yumanensis Yuma bat OSC 
Pipistrellus hesperus Western pipistrelle bat NT 
Pipistrellus subflavus Eastern pipistrelle bat NT 
Corynorhinus rafinesquii Rafinesque’s big-eared bat OSC 
Corynorhinus townsendii ingens Ozark big-eared bat E 
Corynorhinus t. pallescens Western big-eared bat OSC 
Corynorhinus t. townsendii Townsend’s Big-eared Bat OSC 
Corynorhinus t. virginianus Virginia Big-eared Bat E 

“E, endangered; NT, no threat; OSC, of special concern. 

thousand individuals and the summers are spent in trees 
or buildings. A few species (e.g., gray bat, Myotis grisescens 
Howell) may live in caves throughout the year, although 
different ones are utilized in winter and summer. Some 
species, like the Brazilian free-tailed bat [ Tadarida brasiliensis 
(Saussure)], occupy caves in very large numbers and contrib- 
ute an immense amount of organic material (guano, i.e., bat 
droppings) to the cave ecosystem. During the summer 
approximately 20,000,000 individuals of this particular 
species occupy a single cave near San Antonio, Texas, and, at 
that time, the population represents the largest concentration 
of mammals in the world. Guano produced by bats living 
in caves not only supports immense communities of guano- 
philes (e.g., bat fleas, dermestid beetle larvae, gnats, and 
pseudoscorpions), but also this nitrogen-rich material has 
been mined for use as fertilizer as well as for producing saltpeter 
(potassium nitrate), which is an ingredient of gunpowder. 


The biodiversity of cavernicoles in the United States is 

summarized in Table I and their distribution patterns are 
demonstrated in Tables II and II and Figs. 3 and 4. Clearly 

the more mobile troglobites have been very successful, with 
more than twice their numbers invading subterranean 
habitats when compared to stygobites. Also, insects (359 
species and subspecies including apterygous and pterygous 
forms; 32% of all obligate species), arachnids (319; 28%), 
and crustaceans (294; 26%) dominate the species richness 
summaries, with terrestrial arachnids making up nearly one- 
third of all subterranean species. Much variation is shown 
among major karst regions, with those to the north (Black 
Hills and Driftless Area) having few obligate species. The 
number of stygobites ranged from zero (Black Hills) to 85 
(Appalachians), and the number of troglobites varied from 
zero (Florida Lime Sinks) to 256 (Interior Lowlands), and 
even at the generic level, overlap among regions is low. Karst 
regions with the greatest total biodiversity are the Interior 
Lowlands, Appalachians, and the Edwards Plateau & 
Balcones Escarpment (Fig. 4C), although troglobites show 
somewhat different patterns than do stygobites (Figs. 4A and 
B). States with the highest total species richness are Texas, 
Tennessee, Alabama, Kentucky, and Virginia. Stygobites show 
a different pattern: Texas, Virginia, Tennessee, Missouri, and 
West Virginia (Fig. 3). Most assuredly “hot spots” do exist 

where concentrations of fauna occur and northeastern 

Alabama is one of these centers of biodiversity for troglobites 
and stygobites. Of particular note, Jackson County has 1526 
caves and 66 obligate cavernicoles, which translates into 
more than three times the number of caves and nearly twice 
as many species as any other Alabama county. This shows the 
strong relationship between the number of caves observed 
and the number of species reported (7? = 0.81). 

Troglobites and stygobites make up slightly more than 
50% of the imperiled U.S. fauna that is tracked in the central 
databases of the Natural Heritage Program. There are far too 
many potential (and realized) disturbances to discuss herein 
that threaten these out-of-sight organisms and the reader is 
referred to Elliott (2000) for an excellent review of them. 
Suffice it to say that the ultimate long-term survival of 
subterranean karst communities depends on appropriate 
management and protection of the cave, the groundwater, 
and the entire catchment area. 

It becomes clear that these subterranean species are 
geographically concentrated in a small percentage of the 
landscape, with more than 50% of cave-inhabiting species 
occurring in less than 1% of the land. Hence, it is much 
easier to preserve a large percentage of at-risk species by 
focusing habitat conservation efforts in those areas of high 
concentrations of obligate cave fauna, or “hot spots.” Protect- 
ing and conserving karst habitats and their biodiversity is a 
challenging but most important task for modern and future 
speleologists. The subterranean biodiversity of the United 
States is globally significant but highly vulnerable. 


Christman, M. C., and D. C. Culver (2001). The relationship between cave 
biodiversity and available habitat. J. Biogeography 28, 367-380. 

Culver, D. C., M. C. Christman, W. R. Elliott, H. H. Hobbs III, and J. R. 
Reddell (2003). The North American obligate cave fauna: Regional 
patterns. Biodivers. Conser. 12(3) 441-448. 

Culver, D. C., and H. H. Hobbs III (2002). Patterns of species richness in 
the Florida stygobitic fauna. In Hydrogeology and Biology of Post-Paleozoic 
Carbonate Aquifers (J. B. Martin, C. M. Wicks, and I. D. Sasowsky, eds.), 
Special Publication 7, pp. 60-63. Karst Waters Institute. 

Culver, D. C., H. H. Hobbs III, M. C. Christman, and L. L. Master (1999). 
Distribution map of caves and cave animals in the United States. /. Cave 
Karst Studies 61(3), 139-140. 

Culver, D. C., H. H. Hobbs II, and J. E. Mylroie (1999). Alabama: A 
subterranean biodiversity hotspot. /. Alabama Acad. Sci. 70(3), 97-104. 

Culver, D. C., L. L. Master, M. C. Christman, and H. H. Hobbs III (2000). 
Obligate cave fauna of the 48 contiguous United States. Conserv. Biol. 
14(2), 386-401. 

Elliott, W. R. (2000). Conservation of the North American cave and karst 
biota. In Ecosystems of the World 30: Subterranean Ecosystems (H. Wilkins, 
D. C. Culver, and W. E Humphreys, eds.), pp. 665-689. Elsevier Press, 

Harvey, M. J., J. S. Altenbach, and T. L. Best (1999). Bats of the United 
States. Arkansas Game & Fish Commission, U.S. Fish and Wildlife 
Service, Asheville. 

Hobbs III, H. H. (1992). Caves and springs. Chap. 3 in Biodiversity of the 
Southeastern United States: Aquatic Communities (C. T. Hackney, S. M. 
Adams, and W. H. Martin, eds), pp. 59-131. John Wiley and Sons, New 

Nicholas, Bro. G. (1960). Checklist of macroscopic troglobitic organisms of 
the United States. Amer. Midl. Nat. 64, 123-160. 

Diversity Patterns in Australia 183 

Packard, A. S. (1888). The cave fauna of North America, with remarks on 
the anatomy of the brain and the origin of the blind species. Mem. Nat. 
Acad. Sci. 4, 1-156. 

Peck, S. B. (1998). A summary of diversity and distribution of the obligate 
cave-inhabiting faunas of the United States and Canada. J. Cave Karst 
Studies 60(1), 18-26. 

Diversity Patterns in 

William E Humphreys 

Western Australian Museum, Australia 


That Australia, and tropical areas worldwide, have diverse 
subterranean fauna has not been long recognized. Until recent 
decades, Australia was thought to be deficient in overtly cave- 
adapted (troglomorphic) animals. This circumstance was 
considered to have resulted from a number of causes: (1) the 
relative sparsity of carbonate rocks in Australia, as found in 
other Gondwanan fragments, compared with the world 
average (Fig. 1); (2) the general aridity of the continent—it is 
the most arid inhabited continent, two-thirds of which 
receives less than 500 mm of rain annually—generally result- 
ing in both dry caves and the low input of food energy into 
the underground voids; (3) the global lack of cave-adapted 

IV Climate Zone 

@* Hard-rock Karst Area 

“® = Soflt-rock Karst Area 

2 Volcano-karst Area 
Calcrete Aquifers 

{c} KGG 3.2002 

FIGURE 1 Karst areas of Australia and the bioclimatic zones: II, tropical; 
II, subtropical dry; IV, transitional zone with winter rain; V, warm 
temperate; II-IV warm temperate/tropical transition zone. (After Hamilton- 

Smith and Eberhard, 2000. Graphic by K. G. Grimes.) 

184 Diversity Patterns in Australia 

animals in tropical areas; and (4) the lack of widespread 
glaciation, which was perceived to be the main driving force 
driving the evolution of troglobites in the Northern 
Hemisphere, then the focus of biospeleological research. 
Concomitantly, there was perceived to be a high proportion 
of animals found only in caves but not specialized for cave 
life, that is, lacking overt troglomorphisms. Although not 
articulated, these arguments would have applied also to 
stygofauna, the inhabitants of underground waters in both 
karstic and alluvial aquifers. 

Understanding the biogeography of an area is reliant on 
having a broad spatial and taxonomic sample of the biota, a 
comprehensive taxonomy, a well-developed systematic and 
paleoclimate framework, and a fully developed geographical 
understanding (especially of paleodrainage and plate 
tectonics). There are serious deficiencies in information on 
most of these fields of endeavor in Australia. The taxonomic 
and systematic framework is very patchy and many groups of 
interest to hypogean questions remain largely unstudied 
(e.g., Thysanura, Collembola, Diplura, Oligochaeta) or are 
just beginning to be studied so it is still too early for them to 
contribute in detail to biogeographical understanding (e.g., 
many higher taxa in Oligochaeta, Copepoda, Ostracoda, 
Amphipoda, Diplura, Gastropoda). Hence, the focus here 
will be on some higher taxa for which there is more adequate 
information, and on some systems, such as the groundwater 
inhabitants of the smaller voids (mesovoids), for which there 
is a useful body of data. 

During the last two decades of the 20th century, more 
focused, as well as more widespread, exploration of caves and 
later groundwater has shown that the Australian tropics and 
arid zones contain especially rich subterranean fauna. 
However, no area of Australia has been well studied for its 
hypogean life, the distribution of the effort has been very 
uneven across the country, and many areas remain effectively 
unexplored for cave fauna. Detailed examination of 
subterranean biology in Australia is sparse and studies have 
been largely restricted to faunal surveys. Prominent karst 
areas, such as the Barkley and Wiso regions, have barely been 
examined because of their remoteness from population 
centers. Other remote areas, such as the Nullarbor, in which 
there has been a long history of cave research, have proved to 
have sparse hypogean assemblages, especially among the 
stygofauna. Even within those relatively well surveyed areas, 
the taxonomic effort is seriously underdeveloped. For 
example, in one compilation, 63% of the stygofauna from 
New South Wales was undescribed. Where species have been 
described, there are many oddities, not yet well placed within 
their lineage and thus contributing poorly to understanding 
the biogeography of the Australian hypogean biota. 


In contrast to the widespread glaciation that directly 
influenced many of the classical karst areas in the Northern 

Hemisphere, Australia has not been subjected to continent- 
wide glaciation since the Permian. The biogeography of 
the hypogean fauna of Australia has been influenced by the 
continent's past connections with Pangaea and Gondwana, 
as well as having formed the eastern seaboard of the 
Tethys Ocean during the Mesozoic. Australia is a fragment of 
Gondwana together with Africa, India, Madagascar, South 
America, and Antarctica. Gondwana itself fragmented and 
Eastern Gondwana (India, Antarctica and Australasia) became 
isolated from South America and Africa by 133 million years 
ago. By the Upper Cretaceous (ca. 80 million years ago), 
Australia was joined only to Antarctica and it formed the 
eastern seaboard of Tethys. These lands shared a Gondwanan 
flora and fauna, and when the final separation between them 
occurred (45 million years ago), both lands were well watered 
and supported cool temperate and subtropical forests. 

The separation of Australia from Antarctica, and_ its 
resulting rapid northward drift toward southeast Asia, has 
been the most significant factor that has shaped the 
Australian subterranean fauna in the Tertiary. It resulted in 
the formation of the Southern Ocean seaway and the devel- 
opment of the circum-Antarctic ocean currents and winds 
that markedly altered the climate of the Southern Hemi- 
sphere, causing Australia to become much drier. The 
formation of the Antarctic ice cap 15 million years ago saw 
the beginning of a series of marked climatic fluctuations that 
have greatly stressed the Australian (and other Gondwanan) 
flora and fauna. Warm and wet interglacial periods alternated 
with very dry, cool, and windy glacial stages, but only a small 
area of the Eastern Highlands and Tasmania were subject to 
extensive ice cover. These cyclic fluctuations, superimposed 
on a generally increasing and spreading aridity, provided 
conditions under which subterranean refugia played an 
important role. 

Shield Regions and the Cretaceous Marine 

Australia has several major shield regions—parts of the 
Earth’s crust little deformed for a prolonged period—that 
have been emergent since the Paleozoic. The largest is the 
Western Shield, which includes the Pilbara and Yigarn 
cratons. These stable, truly continental areas of Australia have 
a nonmarine, presumably freshwater history extending 
through several geological eras. The Cretaceous marine inun- 
dation, at ca. 120 Ma, would have eliminated nonmarine life 
in the submerged areas (Fig. 2) and only 56% of the current 
land area of the continent remained above sea level. This 
has important implications for lineages with poor dispersal 
ability, as is typical of subterranean fauna. The distribution of 
ancient lineages, both epigean and subterranean, may be 
expected to reflect this marine incursion in two ways. First, 
ancient terrestrial and freshwater lineages may have survived 
on these continually emergent landmasses. Second, marine 
ancestors may have become stranded along the shores as the 

FIGURE 2 Deep history events that have influenced the biogeography of 
Australian subterranean faunas (see text). The shaded continental areas have 

not been covered by the oceans since the Paleozoic. 

Cretaceous seas retreated and today may be represented as 
relictual marine lineages now far inland. 

Caves and other subterranean habitats can remain as 
relatively stable environments over long periods of time 
because they are well insulated from the climatic perturba- 
tions that profoundly affect surface environments and surface 
animals. There, a number of ancient geographical and phylo- 
genetic relictual groups have survived (Spelaeogriphacea, 
Remipedia, Thermosbaenacea, etc.) (Fig. 3). Owing to their 
limited potential for dispersal, their present distributions 
may contain a great deal of information about past 
geography and climates. The ghost of Cretaceous and earlier 
marine transgressions is probably reflected in the distribution 
of phreatoicideans, an ancient group of isopods, in both their 
epigean and subterranean forms, the latter being restricted to 
the tropics, and it has been well documented in the 
crangonyctoid amphipods (see Box 1). 

In this respect aquatic subterranean fauna hold a special 
significance because, unlike terrestrial troglobites, the aquatic 
troglobite fauna (stygobites) contain many relict species that 
are only distantly related to surface forms. These lineages 
provide the most compelling evidence that the distribution 
of some relict fauna occurred through rafting on tectonic 
plates moved by seafloor spreading. Recently a number of 
notable discoveries of such relict fauna have been made 
in Australia whose geographical distribution and lifestyles 
suggest origins variously in Pangaea, Gondwana, Eastern 
Gondwana, and Tethys. 

Cave Atmosphere 

The latitudinal position and general aridity of Australia make 
cave atmosphere a significant biogeographic determinant 
in Australia. Cave environments have traditionally been 

Diversity Patterns in Australia 185 

separated into different zones—the entrance, twilight, 
transition, and deep zones—with characteristics related to 
the remoteness from the surface environment, such as more 
stable temperature and humidity and reduced light and food 
energy input. On the basis of research in the Undara lava 
tube, Howarth and colleagues developed the concept of a 
fifth zone, the stagnant-air zone, which is characterized by 
elevated carbon dioxide and depressed oxygen levels. Only in 
such areas were highly troglomorphic species found in cave 
passages. However, in other tropical areas, such as arid Cape 
Range, highly troglomorphic species occur in caves that have 
unremarkable concentrations of oxygen and carbon dioxide, 
some even occurring in sunlight near cave entrances, but 
only where the air is saturated, or nearly saturated, with water 

Howarth also addressed the importance of water content 
in the cave atmosphere, largely from his Australian studies. 
Both tropical and temperate cave systems lose water when 
the outside air temperature (strictly, the outside water vapor 
pressure) drops below that in the cave. In the tropics, where 
average seasonal temperature differences are less than in 
temperate regions, caves tend to be warmer than the surface 
air at night and cooler during the day. Even if both air 
masses are saturated with water, the cave will tend to dry out 
as water vapor leaves the cave along the vapor pressure 
gradient—the so-called “tropical winter effect.” 

Owing to widespread aridity, this concept has particular 
relevance to Australia and also in tropical areas where the 
general form of many caves (giant grikes, small and shallow 
caves) and low subterranean water supply make them 
vulnerable to drying. Within this context the extent of the 
deeper cave zones (transition and deep) will fluctuate as the 
boundary of threshold humidity levels migrates with the 
changing atmospheric conditions further into or out of the 
cave. Such changes occur in ecological time, associated with 
daily and seasonal fluctuations in air density and humidity, 
and through evolutionary time, in response to climatic cycles 
and long-term climatic trends. Such changes should have 
little effect on groundwater or on troglobites in deep caves, 
which are extensive enough to contain the entire change. 
But, in the shallow caves common in the Australian tropics, 
such changes are likely to cause large areas of cave systems to 
dry out. Such processes may lead to the extinction of certain 
cave fauna, or impede movement through the epikarst and 
thus could promote speciation between different karst areas. 
The high diversity of Schizomida and of the paradoxosomatid 
millipede Stygiochiropus in arid Cape Range are candidates 
for such analysis. 

Humid caves within the arid zone have permitted the 
survival of a diverse troglobitic fauna in arid Cape Range, the 
affinities of which lie with the inhabitants on the floor of 
the rainforest, both temperate and tropical, habitats now 
thousands of kilometers distant. While the fauna is now 
geographically relict, the driving force resulting in the initial 
invasion of the caves is unknown-species may have estab- 


Subterranean animals, clockwise from upper left: 1, 7jirtudessus eberhardi (Dytiscidae), one of 50 species of blind diving beetles from calcretes 
aquifers in the Australian arid zone; 2, unnamed blind philosciid isopod; 3, head of Ophisternon candidum (Symbranchiformes), one of two Australian cave 
fish; 4, the phreatoicidean isopod Phreatoicoides gracilis; 5, Pygolabis humphreysi, from Ethel Gorge calcrete belongs to a family of flabelliferan isopods, the 
Tainisopidae, known only from groundwater in Kimberley and Pilbara regions of Western Australia; 6, Mangkurtu mityula (Spelaeogriphacea), a subterranean 
family that is known from only two locations in each of Australia, Africa, and Brazil; 7, Draculoides vinei (Schizomida), one of seven species of micro- 
whipscorpions known from Cape Range; 8, Ngamarlanguia luisae (Gryllidae: Nemobiinae) from Cape Range, the only troglobitic cricket in Australia. 
(Photographs by Douglas Elford, Western Australian Museum, except 1, W. E Humphreys from a painting by Elyse O’Grady; and 4, from GDF Wilson 
Australian Museum.) 


Australia is a major center of amphipod diversity and 
much of this diversity is represented by stygal species. 
Unexpectedly, they are diverse in the tropic areas and in 
the arid center of Australia, in typical karst and in ground- 
water calcrete aquifers. They belong to a number of higher 
taxa including the crangonyctoids (Paramelitidae, 
Perthiidae, Neoniphargidae), hadziods (Fig. 6) (Melitidae, 
Hadziidae), Ceinidae, Bogidiellidae, and Eusiridae, but 
there is scant knowledge of their distribution and 
diversity. About 65 described species occur in these eight 
families, with much of the diversity occurring in ground- 
waters of the arid region (Bradbury and Williams, 1997). 
Whereas some families appear to be restricted to the moist 
temperate southeast and southwest of the continent 
(Eusiridae and Neoniphargidae), others are much more 
widespread and encompass parts of the arid zone and 
tropical areas (Melitidae, Paramelitidae). Other families 
are more restricted and Bogidiellidae are known only from 
the northwest, whereas Perthidae and Ceinidae occur in 
the southwest and south, respectively. Bogidiellid, 
melitids, and hadziids are known from the anchialine 
waters of the northwest, especially Cape Range and 
Barrow Island. Notably, while taxa in southern areas 
comprise both stygal and epigean species, northern taxa, in 
the arid tropic and subtropics, comprise only stygal taxa. 

A clear relationship can be seen between the Cretaceous 
marine transgressions and the distribution of amphipods. 
Melitoid taxa occur near the shorelines of areas that 
have been transgressed, while in those areas that have not 
been transgressed, crangonyctoid taxa and niphargiids are 
found. Paramelitids are diverse and abundant in Tasmania 
but seemingly sparse in New South Wales, where neo- 
niphargids are diverse. More comprehensive collecting in 
Western Australia suggests that family distributions may 
be circumscribed (Fig. 4). 

lished in caves coincident with the onset of aridity to escape 
the surface drying, or they may have established in caves 
seeking resources unrelated to the onset of aridity. This ques- 
tion cannot be resolved for Cape Range because the aridity 
has been sufficiently intense to extirpate entirely close 
relatives at the surface. Other tropical areas, such as North 
Queensland, offer greater prospect of resolving such issues 
because contemporary lineages occur with surface and caver- 
nicolous species exhibiting various degrees of troglomorphy. 

Another area where a resolution of the causes of colon- 
ization of the hypogean environment may be resolved is in 
the groundwater calcrete deposits (see Box 2) in the arid 

Diversity Patterns in Australia 187 

FIGURE 4 The distribution of groundwater calcretes in Western Australia, 
which occur throughout the arid land north of 29°S. Most occur 

immediately upstream of salt lakes (playas) within paleodrainage channels 
(dotted lines). 

zone. There, many different lineages of diving beetles 
(Dytiscidae) have invaded the groundwater and become eye- 
less and flightless (Fig. 3). Each calcrete body has a unique 
dytiscid assemblage, there is almost no overlap in species 
between different calcretes, and speciation appeared to have 
taken place in situ because species pairs (one large and one 
small) are common among the 50 stygal species in the arid 
zone. Molecular studies suggest that numerous lineages 
invaded the calcrete aquifers during the constrained time 
period, which suggests that it occurred in response to a 
widespread factor, such as might be expected from spreading 


188 Diversity Patterns in Australia 

Groundwater Calcretes 

The long period of emergence and the ensuing erosion 
down to the Archaean basement has resulted in classical 
karst terrain being absent from the Western Shield. How- 
ever, thin carbonate deposits are widespread throughout 
the arid zone and are well developed as groundwater 
(valley) calcretes (hereafter termed calcrete) which occur 
widely in Australia but in isolated, though sometimes 
extensive, pockets usually associated with palaeodrainage 
lines (Fig. 5). Calcretes are carbonate deposits forming 
from groundwater near the water table in arid lands 
as a result of concentration processes by near-surface 
evaporation. They occur forming immediately upstream 
of salt lakes (playas), chains of which form such a 
prominent part of the landscape in the more arid parts of 
Australia. The playas are the surface manifestation of 
palaeodrainage channels incised into Precambrian 
basement rocks by rivers that largely stopped flowing 
when the climate changed from humid to arid in 
the Palaeocene. Hence, the palaeovalleys predate the 
fragmentation of Gondwana. 

Calcretes are especially important in the Australian 
context as they form in arid climates (annual rainfall 
<200 mm) with high potential evaporation (>3000 mm 
per year). Although quite thin (10-20 m thick) the 
groundwater calcretes often develop typical karst features 
and within them. Groundwater salinity may vary 
markedly owing the episodic recharge characteristic of the 
arid zone. 

Because they are deposited at intervals from the 
groundwater flow, the scalcrete masses are separated by 
habitat—Tertiary valley-fills, largely clays, and saltlakes— 
that is unsuitable for stygofauna. Consequently, they 
form isolated karst areas along the numerous major 
palaeovalleys, some of which date from the Permian. The 
sediments filling the palaeochannels are mostly Eocene 

Energy Supply 

Energy enters subterranean systems largely mediated by 
water, animals, and plants. Because these elements them- 
selves are not uniformly distributed across Australia they 
have the potential to influence Australian cave biogeography. 
The carriage of organic matter in surface water is strongly 
affected by seasonal rainfall and plant growth. The episodic 
rainfall, characteristic of the arid zone, means that some areas 
potentially have unpredictable energy supplies. 

Plants provide the raw material that is transported by 
water into the subterranean realm, but they also directly 
transport energy into hypogean habitats by means of sap 
transport within the roots and by root growth. Roots, 
especially tree roots, were identified as an important and 

FIGURE 5 The distribution of subterranean amphipod and isopod taxa 
in relation to long emergent land areas in Australia. The phreatoicidean 
isopods and crangonyctoid amphipods are ancient continental lineages, 
whereas the melitid amphipods have a more recent marine ancestry. 

or later but the age of the calcretes is unknown. The 
extensive alluvial fan calcretes and some of the river valley 
calcretes formed in the Oligocene ((37—30 Ma) may have 
following the onset of the continental aridity. Many of the 
calcrete areas, especially those north of 31°S, are being 
actively deposited and the others have probably been 
remobilized and redeposited, attributes that make the 
dating of calcrete deposits using standard radiometric 
methods problematic. However, a molecular phylogeny of 
the diverse diving beetle fauna, the numerous species of 
which are each restricted to a single major calcrete area 
(Leys et al., in press), indicate that the calcretes have been 
present for at least 5-8 million years. 

reliable source of energy for troglobitic cixiid and meenoplid 
fulgoroid Homoptera. These occur in the lava tubes of 
tropical North Queensland and similar fauna are found in 
karst across the tropics, into the Kimberley and down the 
arid west coast, to the south of Cape Range. Tree roots are 
also utilized by cockroaches throughout the country (e.g., in 
the Nullarbor, Trogloblatella nullarborensis). 

Tree rootmats also represent a reliable food supply for 
elements of the rich communities of aquatic invertebrates, 
including some exhibiting troglomorphisms, occurring in 
some shallow stream caves of western Australia. They provide 
habitat, and probably food, for stygofauna in the Nullarbor, 
in calcrete aquifers of the Western Shield, and in anchialine 
caves in Cape Range and Christmas Island where they 

are associated with a diverse fauna largely comprising 

Roots, like guano, often provide copious quantities of 
energy to cave communities, which may be quite diverse. 
Roots in both the Undara lava tube, Queensland, and the 
Tamala Limestone of western Australia, support diverse 
cave communities. However, whereas the former contain 
numerous highly troglomorphic species, the latter has few 
stygomorphic species, many being indistinguishable for 
surface species. 

Animals may transport energy into cave systems and 
deposit it there as excreta, exuvia, carcasses, and eggs. In 
Australia such troloxenic agents exhibit marked latitudinal 
differences. In the south rhaphidophorid crickets are the 
most conspicuous trogloxenic agents, whereas bats, while 
not diverse, are locally abundant where they form breeding 
colonies. In the tropics bats are widespread, diverse, and 
important producers of guano, as, to a lesser degree, are 
swiftlets in more humid areas. 

Guano is usually intermittently distributed in both space 
and time because it is dependent on the biology of the birds 
and bats. In consequence, the cave communities associated 
with guano are highly specialized and differ markedly from 
the cave fauna not dependent on guano. Markedly troglo- 
morphic species are not commonly found in the energy-rich, 
but temporally unstable, guano communities. 


Stygofauna are discussed in the context of Crustacea, which 
comprise the overwhelmingly majority of stygofauna, but the 
Dytiscidae example above introduced the insect component. 

The magnitude of the biodiversity present in subterranean 
waters globally has only recently been given prominence. 
Australia, especially the northwestern and southeastern parts, 
has unexpectedly come to the attention of stygobiologists 
and systematists on account of its diverse regional ground- 
water fauna (stygofauna). Recently, these have been deter- 
mined to include a number of higher order taxa variously 
new to science (for instance, an undescribed family of 
flabelliferan isopod), new to the Southern Hemisphere 
(Thermosbaenacea, Remipedia, Epacteriscidae), or new to 
Australia (Spelaeogriphacea, Pseudocyclopiidae). Many of 
these taxa occur near coastal and anchialine waters and are 
interpreted as comprising a relictual tethyan fauna. Several of 
these lineages have congeneric species, which are known 
elsewhere only from subterranean waters on either side of 
the North Atlantic—the northern Caribbean region and the 
Balearic and Canary archipelagos (see Box 3). 


The Syncarida are crustaceans now entirely of inland waters. 
The Anaspidacea are confined to Australia, New Zealand, 
and southern South America. In southeastern Australia they 

Diversity Patterns in Australia 189 

are often large and mostly surface living, although several 
stygomorphic species occur in caves streams and ground- 
water, and an undescribed family has been reported that is 
restricted to caves. In contrast, both families of Bathynellacea 
have a global distribution, often even at the generic level, and 
are widespread in Australia. Bathynella (Bathynellidae) is 
found from Victoria to the Kimberley and elsewhere the 
genus occurs globally. Genera within the Parabathynellidae 
known from Australia exhibit different regional affinities. 
Chiliobathynella and Atopobathynella are known from 
Chile and southeastern Australia, while the latter is also 
found throughout northwestern Australia, including Barrow 
Island and Cape Range, and the arid paleodrainage channels 
of the arid center. Notobathynella is found across Australia 
and New Zealand, while Hexabathynella, from the eastern 
Australian seaboard, has a more global distribution, being 
found in New Zealand, southern Europe, Madagascar, 
and South America. Bathynellacea are small stygobites, 
mostly inhabitants of interstitial freshwater environments, 
although an undescribed genus of large, free-swimming para- 
bathynellid occurs in brackish water (<G000 mg L™' TDS) 
in the Carey paleodrainage systems of the arid zone, where 
it is associated with a number of maritime copepods 
lineages such as Ameiridae (Harpacticoida) and Hatlicyclops 


Remarkably little work has been conducted on nonmarine 
copepods in Australia. Recent work on groundwater 
copepods, largely from groundwater calcretes of the Western 
Shield, and the near coastal, especially anchialine systems of 
the northwest, has revealed higher taxa not previously 
described from Australia, in some cases even from the 
Southern Hemisphere. 

Numerous new species of copepods are being described 
from Australian groundwaters, largely from the Yilgarn area 
of the Western Shield including five new genera of Cyclo- 
poida and Harpacticoida, and several genera are reported for 
the first time from Australia [NVitocrella Ameiridae (Eurasia), 
Parapseudoleptomesochra (global), Haifameira Ameiridae 
(depth of Mediterranean Sea), and the family Parasteno- 
carididae (Pangaea, freshwater)]. The broader distribution of 
these lineages within Australia awaits investigation. 

The occurrence of near-marine lineages (e.g., Halicyclops) 
in the center of the Western Shield alongside lineages consid- 
ered to be ancient freshwater lineages (Parastenocaris: 
Parastenocarididae) is notable. It may reflect both the salinity 
stratified, often hypersaline groundwater in these paleo- 
drainage systems, as well as the ancient origins of the fauna. 
Mesocyclops has a mostly tropical distribution; Metacylcops 
(trispinosus group) and Goniocyclops have an Eastern 
Gondwanan distribution; and the limits to the distributions 
of newly described genera of Ameiridae, Canthocamptidae, 
and Cyclopinae await confirmation. 


Anchialine (or anchihaline) habitats comprise near-coastal 
mixohaline waters, usually with little or no exposure 
to open air and always with more or less extensive 
subterranean connections to the sea. They typically show 
salinity stratification and may usefully be considered to be 
groundwater estuaries. They typically occur in volcanic or 
limestone bedrock and show noticeable marine as well as 
terrestrial influences. The water column is permanently 
stratified with a sharp thermohalocline separating a 
surface layer of fresh or brackish water from a warmer 
marine, oligoxic water mass occupying the deeper reaches. 
They have a significant amount of autochthonous prim- 
ary production, via a sulfide-based chemoautotrophic 
bacterial flora, as well as receiving advected organic matter 
from adjacent marine or terrestrial epigean ecosystems. 
Anchialine habitats are mostly found in arid coastal areas 
and are circum-globally distributed in tropical/subtropical 

Anchialine habitats support specialized subterranean 
fauna (Fig. 6), predominantly crustaceans representing 
biogeographic and/or phylogenetic relicts. These special- 
ized anchialine endemics are largely restricted to the 
oligoxic reaches of the water column below the thermo- 
halocline. The structure of these assemblages is highly 
predictable and, remarkably, however remote an 
anchialine habitat, this predictability frequently extends 
to the generic composition. 

In continental Australia anchialine systems occur adja- 
cent to the North West Shelf (Cape Range and Barrow 
Island), and on Christmas Island (Indian Ocean), an 
isolated seamount 360 km south of Java but separated 
from it by the Java Trench. 

Cape Range supports such a fauna comprising atyids, 
thermosbaenaceans, hadziid amphipods, cirolanid iso- 
pods, remipeds, thaumatocypridid ostracods, and an array 
of copepods such as epacteriscid and pseudocyclopiid 
calanoids, and speleophriid misophrioids. Some are the 

Stygal animals from Australian anchialine waters. Clockwise from upper left: Milyeringa veritas (Eleotridae); Lasionectes exleyi (class 

Remipedia); Stygiocaris stylifera (Decapoda: Atyidae); Liagoceradocus branchialis (Hadziidae). 

Anchialine Habitats—Tethyan Relicts—cont’d 

only known representatives of higher taxa in the Southern 
Hemisphere (Class Remipedia; Orders Thermosbaenacea, 
Misophrioida), and several genera are known elsewhere 
from anchialine systems on either side of the North 
Atlantic (Lasionectes, Halosbaena, Speleophria). The poor 
dispersal abilities of these stygal lineages and their close fit 
with the areas covered by the sea in the late Mesozoic 
suggests that their present distributions could have 

resulted from vicariance by plate tectonics (Fig. 2). 

Anchialine systems on oceanic islands support a 
different group of fauna, but the structure of these 
assemblages is similarly predictable, even between oceans. 
Christmas Island is a seamount and supports an 
anchialine fauna characterized by the stygobitic shrimp 


Ostracods recorded from Australian inland waters are 
mainly from the families Limnocytheridae, Ilyocypridae, and 
Cyprididae. In the Murchison, ostracods from the families 
Candonidae, Cyprididae, and Limnocytheridae have been 
recorded in open groundwater but stygophilic species occur 
only in the Limnocytheridae and Candonidae, the latter 
including the globally widespread genus Candonopsis 
(subfamily Candoninae), which occurs widely and in a wide 
variety and age of substrates. Species are known from Pleisto- 
cene syngenetic dune karst (Tamala Limestone), several 
species from Tertiary (probably Miocene) groundwater 
calcretes on the Western Shield, and from the Kimberley 
(Devonian Reef Limestone). In Europe there are only a few, 
mostly hypogean species that are considered to be Tertiary 
relicts with surface relatives today occurring in tropical and 
subtropical surface waters; they are especially diverse in 
Africa. The subfamily Candoninae (family) Candonidae are 
common elements of stygofauna globally but recent finds 
from the Pilbara describe about 25% of the world’s genera 
but these are more closely related to the South American and 
African Candoninae than to European ones. 

The thaumatocypridid genus Danielopolina, previously 
unreported in the Southern Hemisphere, occurs as a tethyan 
element in the anchialine system at Cape Range. Fossils in 
marine cave facies in the Czech Republic suggest that this 
lineage was already inhabiting marine caves in the Jurassic. 


Phreatoicidean isopods (Fig. 3) have a Gondwanan distri- 
bution and occur widely across southern Australia (and in 
tropical Arhnemland) in surface habitats that have perma- 
nent water—generally surface expressions of groundwater— 
usually as cryptic epigean species. Their distribution is 

Diversity Patterns in Australia 191 

Procaris (Decapoda), which belongs to the primitive, 
highly aberrant, family Procarididae that appears globally 
to be restricted to anchialine caves. This family has been 
reported elsewhere only from other isolated seamounts, 
Bermuda and Ascension Island in the Atlantic Ocean, and 
Hawaii in the Pacific. In each case, as with Christmas 
Island, the procaridids are associated with alpheid, 
hippolytid, and atyid shrimp. These cooccurrences of two 
primitive and presumably ancient caridean families 
support the contention that crevicular habitats have 
served as faunal refuges for long periods of time. There is 
no coherent theory as to their distribution to remote 
seamounts such as Christmas Island. 

strongly associated with the areas of the continent not sub- 
merged by Cretaceous seas. About 59 species in 23 genera are 
described from Australia, of which 10 species in eight genera 
are hypogean (cavernicolous or spring emergents). They are 
under active revision and numerous taxa are being described. 
Five hypogean species occur on the Precambrian “western 
shield” and the family Hypsimetopodidae is represented 
with a genus each on the Pilbara (Pilbarophreatoicus) and 
the Yilgarn (Hyperoedesipus) regions (separate cratons of the 
western shield). These are closely related to the hypogean 
genus lVichollsia found in the Ganges Valley of India, suggest- 
ing they were hypogean prior to the separation of Greater 
Northern India from the western shore of Australia. 
Crenisopus, a stygobitic genus occurring in sandstone 
aquifer in the Kimberley, is the link between African and 
Australasian lineages of phreatoicideans. The genus is basal to 
the Phreatoicidae, suggesting divergence after they entered 
freshwater but prior to the fragmentation of East Gondwana 
during the Mesozoic era. 

The flabelliferan isopod family, Tainisopidae, endemic to 
northwestern Australia, occurs in the exposed and greatly 
fragmented Devonian Reef system throughout the western 
Kimberley as well as in remote outcrops of this fossil reef 
in northeastern Kimberley. A second clade of this family 
(Fig. 3) inhabits groundwater calcretes in the Pilbara, from 
which the Kimberley was separated by the Cretaceous marine 
incursions. The location and distribution of this family is 
indicative of ancient origins but its sister relationship has yet 
to be established. Cladistic analysis suggests that this family 
is related to the cosmopolitan marine Limnoriidae and 
Sphaeromatidae, but at a basal level, suggesting it is much 
older than the more derived families like the Cirolanidae. 
Among Asellota, the Janiridae occur widely across southern 
Australia and Tasmania and the genus Heterias, which also 
occurs in New Zealand, is likely to be great; Protojaniridae 
are known from the Northern Territory. 

192 Diversity Patterns in Australia 

Terrestrial isopods (Oniscidea) are a prominent com- 
ponent of cave fauna throughout Australia, as elsewhere in 
the world, yet there are few described highly troglomorphic 
Oniscidea (Fig. 3). Their distribution seems to reflect the 
general aridity that developed following the separation of 
Australia from Antarctica, rather than to suggest more 
ancient relictual distributions. So, in humid Tasmania the 
Styloniscidae are a prominent component of cave fauna, as 
they are in the wet forest of the surface, but only one species 
is troglomorphic. Armadillidae, Ligiidae, and Scyphacidae 
are also common in Tasmanian caves but none is troglo- 
morphic. On the mainland Olibrinidae, Philosciidae, and 
Armadillidae are prominent among cave fauna. In the 
drier areas of Australia, where armadillidians are such a 
prominent part of the surface fauna, they appear in caves 
more frequently, and a few have overt troglomorphies. 
These troglobites are known from the Nullarbor, North 
Queensland (Chillagoe), Cape Range, and Kimberley. The 
troglobitic Philosciidae and Oniscidae from Cape Range and 
the trogloxene Platyarthridae remain undescribed. 

A single species of Haloniscus, an aquatic oniscidean 
isopod, is known from salt lakes (playas) across southern 
Australia. Numerous stygobitic species occur in groundwater 
calcrete deposits of the Yilgarn region of the Western Shield, 
a long emergent landmass in arid Australia, sometimes in 
saline waters. The occurrence of a congeneric species in anchia- 
line waters in New Caledonia suggests a considerable age for 
the genus and remarkable morphological conservation— 
New Caledonia separated from the Australian plate in the 
Late Cretaceous, about 74 Ma. Other stygobitic oniscideans, 
belonging to the Philosciidae and probably other families, are 
widespread in Australia but are largely unknown. 


This order of stygal crustaceans has populations known only 
from two species in separate lacustrine calcrete deposits in the 
Fortescue Valley, a major ancient paleovalley of the northern 
Pilbara region of Australia. All extant spelaeogriphaceans 
occur with very circumscribed distributions in subterranean 
freshwater habitats on Gondwanan fragments known from 
two locations in each of Africa (Table Mountain, South 
Africa), South America (western Mato Grosso, Brazil), and 
Australia. The supposition of a Gondwanan origin is refuted 
by the fossil record. A marine fossil from a shallow marine 
sediment of a laurentian plate was of Carboniferous age in 
Canada, while modern-looking fossil spelaeogriphaceans 
occur in lacustrine deposits of the Jurassic of China and from 
lower Cretaceous freshwater deposits in Spain. All living 
spelaeogriphaceans occur in or above geological contexts 
that are earliest Cretaceous or older. Their broad occurrence 
suggests a Pangaean origin. The colonization of Gondwanan 
freshwater is likely to have occurred after the retreat of 
the Gondwanan ice sheet (after 320 Ma) and prior to the 
dissolution of Gondwana (142-127 Ma). 


Atyid shrimps (Decapoda) are widespread in surface waters 
throughout the tropics and they appear as stygobitic species 
in caves and groundwaters where they are represented by 
four short-range endemic genera. Stygal species occur widely 
across northern Australia, into the arid zone (Canning 
Basin), and in anchialine and freshwater systems of north- 
western Australia (Stygiocaris; Fig. 6) and Christmas Island. 
They may have colonized Australia from Asia via the 
Indonesian archipelago, but their presence in caves of the 
Canning Basin and the apparent Madagascan affinities of 
some genera suggest a more ancient origin. 


The chelicerates globally comprise a biodiverse component 
of cave communities and they are represented in the 
Australian cave fauna by the orders Acarina, Amblypygida, 
Opilionida, Pseudoscorpionida, Schizomida, Scorpionida, 
and Araneae. 


The mite family Pediculochelidae (Acariformes) was first 
recorded in Australia from a dry cave in Cape Range where 
a specimen was attributed to Paralycus lavoipierrei that is 
described from California. Tiramideopsis (Mideopsidae) 
occurs in the Millstream aquifer, a genus previously known 
from similar habitats of India and suggesting ancient links 
(cf. Phreatoicidea). Generally, the poorly known meso- 
stigmatid mite fauna of Australian caves does not appear 
to constitute a distinctive cave fauna or exhibit any of the 
morphological characteristics of deep-cave arthropods. 


Troglophilic species of Charon are found on Christmas Island 
and in the Northern Territory. 


Cavernicolous species of Triaenonychidae in Tasmania 
and New South Wales species often show depigmentation, 
attenuation of pedipalps and legs, a reduction in (but not loss 
of) eye size, reduced sclerotization, and other troglomorphic 
features. The cave fauna of Tasmania, unlike continental 
Australia, have distribution patterns more like those of the 
other periglacial areas of the world in which profound 
environmental changes were associated with Cainozoic 
glacial cycles. The distribution of the opilionid genus 
Hickmanoxyomma, which is exclusively cavernicolous, 
appears to have resulted from the ablation of surface 
forms—in the south and east of Tasmania, where the effect 
of glaciation was most intense, and the occurrence of some 

sympatric species suggests that there may have been multiple 
phases of cave invasion. In contrast, in the coastal lowlands, 
to the north and northeast, where periglacial conditions were 
less extreme, surface-dwelling species of Hickmanoxyomma 
are present. A cavernicolous assamid with reduced eyes, but 
not strongly troglomorphic, and the strongly troglomorphic 
Glennhuntia glennhunti (“Phalangodidae”) from arid Cape 
Range are both probably rainforest relicts although the wider 
affinities of both families are unknown. 


The worldwide family Chthoniidae is most commonly 
represented among troglomorphic species in Australia. 
The genera Tyrannochthonius, Pseudotyrannochthonius, and 
Austrochthonius are widespread with cave populations in 
eastern and western parts of the continent. Syarinidae, which 
occur in the rainforests of Africa, Asia, and the Americas, 
occur widely in Australia and as a troglophile in Cape Range. 
The Hyidae, known from India, Madagascar, and southeast 
Australia, are represented in Australia in the Kimberley and 
by the markedly troglomorphic Hyella from arid Cape Range. 


Schizomids are essentially a tropical forest element that occur 
across the top of the continent, as far south as the humid 
caves in the arid Cape Range. The latter contains six troglo- 
bitic species in the genera Draculoides (Fig. 3) and Bamazomus. 
(Only five other troglobitic species, in the genera Bamazomus 
and Apozomus, are known from the rest of Australia.) 
Draculoides is endemic to Cape Range and Barrow Island. 


A troglobitic scorpion (Liocheles: Liochelidae) occurs on 
rainforest-covered Christmas Island, Indian Ocean. An 
unknown species in a new genus of troglobitic scorpion 
occurs on the arid Barrow Island and shares morphological 
features of the families Urodacidae (species in the genus 
Urodacus and endemic to Australia) and Heteroscorpionidae 
Madagascar). Only one other troglobitic scorpion is known 

(species in the Heteroscorpion, endemic to 

outside the Americas, from Sarawak (Malaysia). 


Troglodiplura, which has South American affinities, is the 
only troglobitic mygalomorph spider in Australia, and occurs 
in caves in the arid Nullarbor region. Like the cockroach 
Trogloblatella, it is heavily sclerotized, suggesting a more 
drying atmosphere than generally associated with troglo- 
morphic animals found elsewhere, such as in the arid zone 
caves of Cape Range, which have affinities with rainforest 
floor communities. The primitive araneomorph (true) spider 

Diversity Patterns in Australia 193 

Hickmania troglodytes from Tasmania is a troglophilic Austro- 
chilidae, a family that also occurs in Chile and Argentina. 

Large lycosoid spiders occurs widely in the arid areas, 
one of which, Bengalla bertmaini (Tengellidae), is highly 
troglomorphic in Cape Range, lacking eyes and pigment. 

Symphytognathidae occur as troglobitic elements in 
the tropical caves of Cape Range and Northern Territory 
(Katherine) as Anapistula, found as epigean elements in the 
wet tropics of Australia, Malaysia, and Indonesia. 

Filistatidae occur throughout Australia (Wandella) but 
the monotypic Yardiella from Cape Range has relatives in 
northeast India and the family has a generally Gondwanan 

Among the Pholcidae Wugigarra occurs along the eastern 
seaboard and the southeast of the continent while the western 
three-quarters of the continent contains old elements of the 
pholcid fauna. If the distribution of the genus were restricted 
by current ecological conditions, then the genus would be 
expected to be found in the west and other refugia, but this 
apparently is not the case. This distribution may be due to 
the marine subdivision of the continent by the Cretaceous 
marine transgression. The genus Trichocyclus occurs as a 
cavernicolous element throughout much of the rest of the 
continent from the Nullarbor to the tropics. 


Among the Nicoletiidae, Trinemura is represented in caves 
in the west, while Metrinura is found in the caves of the 


The composition of collembolan fauna changes between 
the south and north of Australia. Caves in the south of the 
continent contain up to five genera of troglobitic collembola, 
while those in tropical areas have only two genera. The 
genera Adelphoderia and Arrhopalites are not recorded as 
troglobites in tropical caves, but because the former is 
known from both temperate and tropical rainforest litter 
(Greenslade, personal communication), it seems likely to 
occur in tropical caves. This apparent trend in diversity may 
well reflect the greater sampling effort in southeast Australia. 
Oncopodura occurs in southeast Australia and in the northern 

Planthoppers: Relicts or Invaders? 

There is continuing debate as to whether cave fauna result 
from active colonization or occur as relicts as a result of the 
extirpation of surface populations by adverse conditions (e.g., 
glaciation, aridity). The cave fauna on arid Cape Range are 
clearly relictual in that they are now remote from the humid 

194 Diversity Patterns in Australia 

forest from which the fauna were sourced. However, the 
aridity is sufficiently intense to have obliterated all close 
surface relatives and so the process by which it became 
relictual cannot be resolved. By contrast, in a grossly similar 
fauna in Far North Queensland, the troglobitic cixiid and 
meenoplid planthoppers have some members with surface 
relative and many intermediate forms. These lineages show 
many reductive, but no progressive trends, and this has been 
interpreted as support for the active colonization of the 
subterranean realm, rather than as a process of relictualiza- 
tion (Hoch and Howarth, 1989). 

In North Queensland seven evolutionary lines of 
planthoppers (Fulgoroidea) of the families Cixiidae (genera 
Solonaima, Undarana, Oliarus) and Meenoplidae (Phaconeura, 
a continent wide genus) are found. Solonaima (Cixiidae) 
exhibits four independent invasions of the caves and shows 
a full range of adaptations to cave life, from epigean to 
troglobitic, together with intermediate stages. This lineage 
provides an excellent model for the stepwise evolution of cave 
forms and the reconstruction of the historic process of cave 
adaptation—the loss of eyes and pigmentation, reduction of 
wings and tegmina, and increased phenotypic variation, such 
as wing venation, even within same species, suggesting a relax- 
ation of selection pressure. To Hoch and Howarth (1989) this 
suggested that there had been fragmentation of the rainforest 
owing to the drying climate during the Miocene. This model, 
argued on other evidence, has also been suggested for the arid 
Cape Range region on the west coast of the continent. 


Cockroaches represent a widespread and common element 
of many Australian caves, particularly those where the 
predominant energy source is guano from bats or swiftlets 
where Paratemnopteryx and related genera (Gislenia, 
Shawella) are prominent. Paratemnopteryx stonei exhibits 
significant morphological variation in seven tropical caves 
spread over a 150-km distance in North Queensland, 
such variation being consistent with molecular variation 
(Slaney and Weinstein, 1997). The genus Neotemnopteryx is 
widespread on the east coast and is represented by 14 species, 
of which five species are cavernicolous, but troglobitic species 
occur in the Nullarbor and the southwest coast represented 
by the troglobitic, respectively N. wynnei and N. douglasi. In 
the Nullarbor, where caves are relatively dry, the large, eyeless 
but highly sclerotized Trogloblatella nullarborensis is found. In 
contrast, the Nocticolidae occur widely in the Old World 
tropics and a number of cave species occur throughout the 
Australian tropics, down to arid Cape Range where Nocticola 
flabella is found, the world’s most troglomorphic cockroach, 
which is distinguished by its pale, fragile, translucent 
appearance. In contrast, a more robust monotypic troglobite, 
Metanocticola, is found on Christmas Island. The genus 
Nocticola also occurs in the Philippines, Vietnam, Ethiopia, 
South Africa, and Madagascar. 


Many cave crickets (Rhapdidophoridae), which occur in cave 
and bush habitats across southern Australia, are trogloxenes, 
like some bats. During the day these moisture-loving insects 
tend to congregate in relatively cool, moist, and still air 
to avoid desiccation. In the evening, part of the cricket 
population moves outside the cave entrance to feed but they 
return underground before dawn and so transport organic 
matter into the cave. Rhaphidophoridae have a disjunct 
global distribution in the temperate zones of both hemi- 
spheres. The Macropathinae are considered to be the basal 
group and these have a circum-Antarctic distribution, 
suggesting a Gondwanan origin. Generic diversity is much 
greater in Australia and New Zealand than elsewhere. Four 
genera are restricted to Australian temperate zones and a 
further three genera to Tasmania itself: The remaining three 
subfamilies inhabit the Boreal zone, suggesting vicariance 
owing to the Mesozoic dissolution of Pangaea. 

In contrast, the only truly troglobitic cricket in Australia is 
the pigmy cricket Ngamalanguia (Nemobiinae: Gryliiidae) 
(Fig. 3), a genus endemic to Cape Range that lacks eyes, 
ocelli, tegmina, wings and auditory tympana, is pale, and has 
exceptionally long antennae. 


Globally, beetles are by far the most intensively studied cave 
animals. Chief among them are the trechine carabid beetles, 
of which more than 2000 species have been described. Of 
these, more than 1000 species are troglomorphic, inhabiting 
caves from periglacial areas of Australia and New Zealand (25 
species), eastern Palearctic (ca. 250 species), western 
Palearctic (ca. 600 species), and Nearctic and Neotropical 
(ca. 200 species). 

Unlike mainland Australia, Tasmanian caves support a 
distinctive cave fauna of carabid beetles from the tribes 
Trechini (a strongly hydrophilous group forming a dominant 
element of cave fauna of the periglacial areas of Europe, 
North America, New Zealand, and Japan) and Zolini 
(confined to Australasia) each containing two genera with 
troglobitic species. In the periglacial areas of Tasmania, 
vicariant patterns similar to those for opilionids may be 
deduced for the trechine and zoline carabid beetles, which 
form such a prominent past of the Tasmanian cave fauna. 
Harpalinae, a globally widespread and predominantly 
phytophagous group, typical of dry country, are considered 
unsuitable for cave colonization, and yet many genera are 
represented in caves in Australia. Two genera of the Calleidini 
occur in guano caves in Australia, which suggests, because 
these beetles are typically arboreal, the possibility of a reversal 
from the arboreal habit typical of this tribe, to an edaphic or 
subterranean life. 

Although the Cholevidae is well represented in the 

more humid parts of Australia, the tribe Leptodirinae 

(Bathysciinae), which comprises the predominant component 
of the rich cholevid beetle fauna of the Northern Hemi- 
sphere, is entirely missing from Australia and the rest of the 
Southern Hemisphere. In the Snowy Mountains area of 
the mainland, where periglacial conditions also persisted, 
is found the only troglomorphic psydrinid beetle known 
globally. Numerous other families of beetles occur in caves 
throughout Australia, in both the humid and arid areas, 
but most seem to be accidentals. The Australian troglobitic 
fauna, especially those that associate with periglacial areas, 
differ from those in the Northern Hemisphere, owing to the 
composition of the surface fauna, rather than due to different 
evolutionary trends. 


Caves in the wet-dry (monsoonal) tropics commonly provide 
refuge to vertebrates during the dry season and clearly this 
temporary habitation has an impact on the trophic relations 
of these caves. Among them are tree frogs (e.g., Litoria 
caerula), which are also abundant in uncapped boreholes, 
and fish, such as the common eel-tail catfish, Neosilurus 
hyrtlii and the spangled perch or grunter, Leiopotherapon 
unicolor. In the dry season, the fish may survive in caves and 
underground water systems and from there they would 
contribute to the repopulation of the seasonally inundated 

Australia has only two highly troglomorphic fishes 
which are sympatric where they occur at Cape Range. The 
blind gudgeon, Milyeringa veritas (Eleotridae) (Fig. 6), is of 
unknown affinity but inhabits water ranging from seawater 
to freshwater in a largely anchialine system in Cape Range. 
Swamp eels (Synbranchidae) are represented in Australia by 
two species of Ophisternon, of which O. candidum is a highly 
troglomorphic species (Fig. 3). The genus occurs widely in 
the coastal wetlands of the Indo-Malayan region, with one 
other troglomorphic species inhabiting caves in Quintana 
Roo, Mexico, a distribution suggesting a tethyan origin. 

Snakes are commonly seen in caves, especially in the 
tropical regions where they predate bats (e.g., the banded cat- 
snake Boiga fusca ornata). The blindsnake, Ramphotyphlops 
longissimus, from the Barrow Island karst has apparent 
troglomorphies and may represent the first troglobitic reptile. 

Birds are rarely represented in Australian caves other than 
as superficial components inhabiting cave openings. The 
exceptions are swiftlets (Collocalia species) that build their 
nests in the dark zone, on smooth concave walls high above 
the cave floor in some tropical caves in Far North 
Queensland and Christmas Island (Indian Ocean). The nests 
of some species are intensively harvested for the gourmet 
delicacy “birds’ nest soup” in Southeast Asia and India. The 
Christmas Island glossy swiftlet (Collocalia esculenta natalis) 
is endemic to Christmas Island where, in the absence of cave 
bats, they are the prime source of guano in caves. A number 
of other species of Collocalia occur in the Indian Ocean, 

Diversity Patterns in Australia 195 

Southeast Asia, and Queensland, mostly nesting in caves. 
The nests detach from the cave walls in dry air, a factor that 
may account for their absence from the drier tropical areas, 
such as the Kimberley. The various subspecies inhabit few of 
the caves available, being known from only five caves on 
Christmas Island, whereas the white-rumped_ swiftlet 
(Collocalia spodiopygus chillagoensis) occurs in less than 10% 
of approximately 400 caves at Chillagoe in Queensland. 
Bats comprise nearly a third of the Australian mammalian 
fauna. Seven families of bats, comprising about 30% of the 
Australian bat fauna, are found in caves. The 17 species of 
cave-dwelling bats in Australia are largely restricted to the 
tropics and encompass insectivorous and vertebrate predators 
(ghost bats, Macroderma gigas) and frugivorous bats. Six 
species are restricted to the Cape York peninsula and 11 
species occur across the northern part of the continent, 2 of 
them extending in the west coast to the arid Pilbara region. 
Only 4 species are restricted largely to the center of the 
continent, two being restricted to the western plains of 

Queensland and New South Wales. 


In a global context, the most striking features of the sub- 
terranean fauna of Australia are (1) the apparent age of 
the lineages present in subterranean environments and (2) 
the high proportion of geographic relicts present in the 
subterranean systems that are widely separated from their 
near relatives. Although much remains to be done to 
establish consistent patterns, numerous independent 
examples suggest similar processes but at a range of spatial 
and temporal scales. 

In the southeast there is evidence that Pleistocene glacia- 
tion influenced the cave fauna. But, over most of mainland 
Australia, the overwhelming influence seems to have been 
relict distributions resulting from increasing aridity during 
the Tertiary, particularly in the Miocene. Numerous terre- 
strial and aquatic lineages have affinities with Gondwana, 
or with Western Gondwana, often at the generic level. In 
terrestrial lineages, these are commonly associated with 
rainforests. Numerous crustaceans, often lineages entirely 
comprising stygal species, and even a fish lineage, have 
distributions throughout the area of the former Tethys ocean. 
Many lineages from northwestern Australian anchialine 
waters comprise species congeneric with those inhabiting 
caves on either side of the North Atlantic. 


I thank M. S. Harvey, I. Karanovic, T. Karanovic, and G. D. 
E Wilson for their comments and unpublished information. 

Bradbury, J. H., and W. D. Williams. (1997). Amphipod (Crustacea) 
diversity in underground waters in Australia: an Aladdin’s Cave. Memoirs 

of the Museum of Victoria, 56: 513-519. 

196 Diversity Patterns in Europe 

Hamilton-Smith, E., and S. Eberhard (2000). Conservation of cave com- 
munities in Australia. In Ecosystems of the World, Vol. 30. Subterranean 
Ecosystems (H. Wilkens, D. C. Culver, and W. E Humphreys, eds.), 
pp. 647-664. Elsevier, Amsterdam. 

Hoch, H., and E G. Howarth. (1989). The evolution of cave-adapted cixiid 
planthoppers in volcanic and limestone caves in North Queensland, 
Australia (Homoptera: Fulgoroidea). Memoires de Biospéologie, 16:17-24. 

Howarth, F. G. (1987). The evolution of non-relictual tropical troglobites. 
International Journal of Speleology, 16: 1-16. 

Howarth, FE. G., and E D. Stone. (1990). Elevated carbon dioxide levels in 
Bayliss Cave, Australia: implications for the evolution of obligate cave 
species. Pacific Science, 44: 207-218. 

Humphreys, W. E (1999). Relict stygofaunas living in sea salt, karst and 
calcrete habitats in arid northwestern Australia contain many ancient 
lineages. In The Other 99%. The Conservation and Biodiversity of 
Invertebrates (W. Ponder and D. Lunney, eds.), pp. 219-227. 
Transactions of the Royal Zoological Society of New South Wales, 
Mosman 2088. 

Humphreys, W. FE (2000). The hypogean fauna of the Cape Range 
peninsula and Barrow Island, northwestern Australia. In Ecosystems of the 
World, Vol. 30. Subterranean Ecosystems (H. Wilkens, D. C. Culver, and 
W. F. Humphreys, eds.), pp. 581-601. Elsevier, Amsterdam. 

Humphreys, W. E, and M. S. Harvey, eds. (2001). Subterranean biology in 
Australia 2000. Records of the Western Australian Museum, Suppl. No. 64. 

Leys, R., C. H. S. Watts, S. J. B. Cooper, and W. EF. Humphreys. (In press). 
Evolution of subterranean diving beetles (Coleoptera: Dytiscidae: 
Hydroporini, Bidessini) in the arid zone of Australia. Evolution. 

Slaney, D. P, and P. Weinstein. (1997). Geographical variation in the 
tropical cave cockroach Paratemnopteryx stonei Roth (Blattellidae) in 
North Queensland, Australia. International Journal of Speleology, 25: 

Diversity Patterns in Europe 

Janine Gibert 
Université Lyon, I 

David C. Culver 

American University 


To an extent that is unusual in most branches of zoology 
systematics, Europe is a both a hot spot of subterranean 
biodiversity and a hot spot of research into subterranean 
biology, both historically and at present. 

The scientific study of cave life can be traced back to 
Johann von Valvasor’s comments in 1689 on the European 
cave salamander Proteus anguinus. The only stygobitic 
salamander in European, Proteus reaches a length of more 
than 25 cm, making perhaps the largest stygobiont known 
anywhere. It occurs throughout the Dinaric Mountains in 
northeast Italy, Slovenia, Croatia, Bosnia, and Herzegovina. 
During the late 18th century and much of the 19th century, 
living Proteus were collected and delivered to many scientists 
throughout Europe. It was this animal more than any other 
cave animal that played a formative role in the emerging 
theories of evolution of Lamarck and Darwin. The first 

invertebrate was described in 1832, also from Slovenia, as 
Leptodirus hochenwartii, a bizarre appearing beetle with an 
enlarged abdomen and long spindly appendages. 

Besides the caves of the Dinaric region, the cave fauna of 
the French and Spanish Pyrenees began to attract attention, 
and the Pyrenean fauna began to be described by the 
mid-19th century. In 1907, the Romanian zoologist E. G. 
Racovitza published the enormously influential “Essai sur les 
problémes biospéologiques,” which set the agenda for 
biospeleological research in the coming decades. Together 
with the French entomologist René Jeannel, as well as Pierre- 
Alfred Chappuis and Louis Fage, he established in 1907, an 
association named Biospeologica. This association had three 
objectives: (1) to explore caves and look for subterranean 
species, (2) to obtain identifications and descriptions from 
specialists for all material sampled, and (3) to publish results 
in the Mémoires de Biospeologica. Ultimately, Biospeologica 
was responsible for the inventory of the fauna of more than 
1500 caves, mostly in Europe. More than 50 monographic 
treatments of the taxonomy and distribution of European 
cave fauna were published between 1907 and 1962. 

Even with more than a century and a half of description 
and cataloging of the European subterranean fauna, both 
species descriptions and inventories are far from complete. At 
present, several large-scale inventory projects of European 
fauna are ongoing. The most important of these is 
PASCALIS (Protocol for the Assessment and Conservation 
of Aquatic Life in the Subsurface), which has developed 
common protocols for the comparison of subterranean 
aquatic species diversity at six sites in five countries (France, 
Spain, Belgium, Italy, and Slovenia). Individual country 
assessments of subterranean biodiversity are active as well, 
and at the beginning of the project the most advanced. of 
these was in Italy, where there were more than 6000 records 
for 899 subterranean species. 


As of 2000, approximately 5000 obligate subterranean 
aquatic (stygobitic) and terrestrial (troglobitic) species from 
Europe had been described. By contrast, 1200 have been 
described from Asia, 500 from Africa, and 1000 from North 
America. The dominance of Europe in known subterranean 
species is the result of several factors. First, Europe has been 
better studied than the other continents. This is particularly 
evident in the status of biodiversity assessment in noncave 
subterranean habitats. About half of the stygobitic species 
known from Europe are not primarily cave-dwellers, but 
rather live in other subterranean habitats such as the 
underflow of streams. In North America, these habitats are 
little studied and account for less than a fifth of the known 
stygobionts. The terrestrial equivalent, the M.S.S. (milieu 
superficiel souterrain), accounts for half of the subterranean 
species richness in Italy and this habitat is unsampled in 

North America. In the cases of Asia and Africa, large geo- 
graphic areas have been little studied with respect to 
subterranean species, both cave and noncave. 

Second, subterranean biodiversity in Europe is actually 
higher than on other continents. There is some empirical 
evidence to support this. On a worldwide basis, there are 20 
caves and wells that are known to have a total of 20 or more 
stygobionts and troglobionts. Of these, 13 are in Europe, 3 
are in North America, and the rest are scattered elsewhere. In 
several regions, especially the western Balkans (northeast 
Italy, Slovenia, Croatia, Bosnia, Herzegovina, and Yugoslavia) 
and the Pyrenees (France and Spain), even the casual observer 
can note the large number of stygobionts and troglobionts 
present, relative to caves in other continents. A possible 
explanation for the increased diversity of stygobionts in the 
western Balkans is the complex biological and geological 
history of the Dinaric mountains. The amount of available 
subterranean habitat in these mountains is large and has had 
a long history of cave development and evolution. In 
addition, the complex history of the Mediterranean Sea, 
including the fact that it dried up about 6 million years ago 
(the Messinian crisis), may have resulted in a greater invasion 
rate of the subterranean realm from marine waters. 
Moreover, multiple invasions may have occurred from 
surface continental waters. In the case of terrestrial species, it 
is possible that invasion of European caves was enhanced 
during interglacials of the Pleistocene relative to North 
America. This is because mountain ranges in Europe are 
largely east-west oriented, whereas mountain ranges in 
North America are largely north-south oriented. Thus, in 
North America, species could escape hot summers either by 
migrating north or by invading caves. In Europe, the orienta- 
tion of the mountains reduced the migration potential, 
perhaps resulting in increased invasion rates of caves and 
other subterranean habitats. 


Subterranean diversity patterns in Europe are diverse, 
complex, and not always well documented. The distribution 
of the biota is certainly controlled by a combination of 
factors, including geological history, physicochemical 
variables, aquifer structure, connectivity between regions, 
and biological interactions. Moreover, biodiversity patterns 
have to be considered hierarchically. Local (a) and regional 
(y) diversity is better known than the increases in diversity 
over spatial scales (B), although it is likely that B-diversity is 
likely to be much more important numerically. Put another 
way, the diversity in a single cave is almost always much less 
than the overall diversity of caves even within a small area as 
a result of reduced opportunities for migration and invasion. 

The taxonomic diversity of the European fauna, with 
more than 5000 stygobionts and troglobionts, is too 
extensive and complex to be reviewed here, but several 

Diversity Patterns in Europe 197 

highlights are especially noteworthy. The first of these are the 
beetles in the terrestrial habitats. Two families of beetles 
dominate in temperate zone caves: the Carabidae and the 
Cholevidae. Both of these families reach their zenith of 
subterranean species richness in Europe. Among the 
Carabidae, there are seven genera with 20 or more species: 
Anophthalmus, Aphaenops, Duvalius, Geotrechus, Hydraphae- 
nops, Neotrechus, and Orotrechus. The genus Duvatlius is 
especially speciose, with more than 250 species. In contrast, 
there is only one carabid genus in North America with more 
than 20 species—Pseudanophthalmus—although it has more 
than 200 species. Among the Cholevidae, a total of 23 genera 
are known from the Pyrenees alone, two of which (Speocharis 
and Speonomus) contain more than 35 species. For example, 
the number of species and the distribution patterns of the 
two closely allied genera Bathysciola and Speonomus are very 
different (Fig. 1). Only 11 species of Bathysciola inhabit soil 
layers in large areas, whereas 39 species of Speonomus have 
strikingly restricted distributions. In the Pyrenees, a total of 
41 genera are known, with Antroherpon containing 26 species. 
Once again the contrast with North America is instructive. 
There are only two genera and Ptomaphagus has 19 species. 

Among aquatic species, two groups are especially 
noteworthy. One is the amphipod family Niphargidae. There 
are four genera and Mphargus alone has more than 200 
species. Taxonomically, the diversity of the Niphargidae is 
rivaled by that of the Crangonyctidae in North America. 
Especially speciose is the genus Stygobromus, with nearly 200 
species. However, the crangonyctids seem to be less diverse 
ecologically than the niphargids, both in terms of morpholo- 
gical variation and habitat variation. The final group worthy 
of special note are the mollusks in the Dinaric Mountains. In 
the Slovenian part of the Dinaric Mountains, for example, 
there are 37 aquatic obligate cave snails, one aquatic cave 
clam, and 11 species of terrestrial obligate cave snails. The 
entire Dinaric Mountain region has several times that many 
species. Surface freshwater in this region is also extra- 
ordinarily rich in mollusk diversity, especially among 
aquatic species. 


While the inventory of known subterranean species is far 
from complete, and the summary of this information even 
less so, existing data provide some important information of 
about diversity patterns. The first general pattern is that there 
is a gradient in species richness with diversity dropping off to 
the north. A comparison of Italy and the United Kingdom, 
the two European countries with the most thorough inventory, 
shows this in a striking manner. In the United Kingdom, 
a total of 10 troglobionts and 16 stygobionts have been 
described. In Italy, a total of 265 stygobionts, 321 troglo- 
bionts, and an additional 317 troglobionts from the M.S.S. 
have been described. Of course, part of this difference is a 
result of the covering of the United Kingdom by Pleistocene 

198 Diversity Patterns in Europe 

Ratio of Troglobites to Stygobites 

TB/SB Ratio 

[| No Data 

[ES 000-034 
Gd 060-088 
HM os9- 130 
HM 31-199 

FIGURE 1 Ratio of troglobionts (TB) to stygobionts (SB) for European countries for which data are available. 

ice sheets, but undoubtedly other factors are at work to 
produce this diversity gradient. As one moves north, mean 
annual temperature declines and, all other things being 
equal, so does net primary productivity. While little or no 
primary production occurs in subterranean habitats, they are 
dependent on surface primary production for food. Thus, 
food availability declines with increasing latitude. 

The central and southern European—Mediterranean region 
presents a taxonomic diversity and species richness higher 
than other parts of Europe. In addition to the Coleoptera, 
Amphipoda, and Gastropoda mentioned earlier, many other 
groups reach their highest diversity in this region, including 
isopods, spiders, and springtails. The distribution of species 
largely reflects the effect of historical circumstances on 
processes such as colonization, extinction, and species. As 
pointed out by Holsinger, this region is characterized by 
extensive karst areas, temperate to Mediterranean climates, 
extensive shallow embayments in the Tertiary (with the 

Mediterranean salinity crisis in the Miocene), absence of 
Pleistocene glaciation, marine interstitial and anchihaline 
habitats (see article on anchihaline communities) common in 
coastal areas of the Mediterranean, and abundant freshwater 
subterranean habitats. 

The second pattern that is evident is that the ratio of 
troglobionts to stygobionts declines with increasing latitude 
(Fig. 2). The reduction in troglobionts may be largely the 
result of Pleistocene effects. No subterranean terrestrial 
habitats existed, as far as we know, within the ice sheets. On 
the other hand, some subterranean aquatic species may have 
survived the Pleistocene in running water habitats under- 
neath the glacial ice. Additionally, it may be that recoloniza- 
tion has been more rapid among aquatic species (for 
example, through fluvial corridors and due to the hydrolo- 
gical continuum and the connectivity between aquifers). 
Stygobionts can be found in almost all parts of Europe with 
suitable groundwater habitats, even though these habitats are 

Diversity Patterns in Europe 199 


Number of species 




25 "30 
Number of 0.2 * 0.2 degrees squares 

FIGURE 2 (a) Schematic distribution patterns of species in two closely allied genera of Cholevidae (Coleoptera) in the Pyrenees. The 11 species of Bathysciola 
inhabit soil layers, and the 39 species of Speonomus are troglobionts. Known sites of occurrence are encircled by ovals; single-site species are indicated by black 
points. (b) Cumulative plot of species and subspecies richness of Bathysciola and Speonomus in the Pyrenees. 

often not cave habitats, but rather other subterranean 
habitats such as the underflow of streams. The pattern in 
Fig. 2 includes these noncave stygobionts. 

The third pattern is that of the subterranean fauna that 
result from habitat differences on a more local scale. This is 
particularly evident in the differences between the aquatic 
cave fauna (karstic aquifers) and the subterranean fauna of 
gravel and sand aquifers. The obvious physical difference is in 
the size of the available space for fauna. For example, 
comparison of subterranean communities of the Vidourle 
karst (south of France) and the neighboring alluvial aquifer 
bordering the surface stream of the Vidourle, some hundreds 
meters downstream of the resurgences, reveals important 
differences (Table I). Karstic communities are more diverse 
(49 species with 20 stygobionts) than interstitial commu- 
nities (19 species with 12 stygobionts). The large voids in 

karstic channels enable a large size range of organisms to 
occur while very tiny spaces in sandy aquifers represent a 
barrier to larger animals such as the amphipod Niphargus 
virei and the decapod atyid Troglocaris inermis, which are 
limited to the karstic aquifer. A similar pattern may exist 
in subterranean terrestrial communities. There are noncave 
subterranean habitats in superficial subsurface layers such as 
scree slopes and collectively termed the milieu souterrain 
superficiel that may also harbor small species, but this pattern 
has not been investigated. 


The current incomplete state of knowledge of groundwater 
biodiversity is a major constraint on successful implementa- 

200 Diversity Patterns in Europe 

TABLEI Comparison of Vidourle karst communities and neighbouring alluvial aquifer communities (Juberthie and Juberthie- 

Jupeau 1975) 

Karstic aquifer 
Vidourle karst 

Porous aquifer 
Neighbouring alluvial aquifer 



Crustacea: Ostracoda 

Crustacea: Calanoida 

Crustacea: Cyclopoida 

Crustacea: Harpacticoida 

Crustacea: Syncardia 

Crustacea: Isopoda 

Crustacea: Amphipoda 

Crustacea: Decapoda 
Total species 

Total Stygobionts 

Monhystrella plectides 
Tripyla tenuis 

Tripyla filicaudata 
Anatonchus sp. 
Mononchus mylanchulus 
Mononchus iotonchus 
Mononchus anatonchus 
Nygolaimus stagnalis 
Dorylaimus sp. 
Thornenema loevicapitatum 
Labronema sp. 

Pugentus sp. 

Oxydirus oxycephaloides 


Moitessieria rollandiana rhodani 
Hauffenia minuta globulina 
Bythynella reyniesi 

Bythiospeum diaphanum 
Theodoxius fluviatilis 

Mixocandona n.sp. 
Pseudocandona n.sp. 
Spelacodiaptomus rouchi 
Eucyclops serrulatus 
Acanthocyclops venustus 
Acanthocyclops viridis 
Paracyclops fimbriatus 
Diacyclops languidoides 
Mesocyclops delamarei 
Mesocyclops albidus 

Ceuthonectes sp. 

Elaphoidella gracilis 
Elapoidella leruthi meridionalis 
Nitocrella sp. 

Attheyella crassa 

Attheyella wulmeri 

Bryocamtus pygmaeus 
Limocamptus echinatus 

Bathynella sp. 

Proasellus cavaticus 

Microcharon sp. 
Faucheria faucheria 

Niphargus virei 
Niphargus kochianus 

Troglocaris inermis 

Totonchus acutus 

Cephalobus sp. 

Moitessieria rollandiana rollandiana 

Physa acuta 

Acanthocyclops venustus 

Paracyclops fimbriatus 

Tropocyclops prrasinus 
Elapoidella gracilis 

Nitocrella sp. 

B ryocamtus pygmaeus 

Nitocra hibernica 

Bathynella sp. 
Parabathynella sp. 

Proasellus coxalis 
Microcharon sp. 
Faucheria faucheria 

Niphargus gallicus 
Niphargus pachypus 
Niphargopsis legeri 
Salentinella juberthieae 


tion of its conservation. The PASCALIS project (2002-2004) 
aims at improving the scientific knowledge of groundwater 
biodiversity in Europe. The main goal of this project is 
to establish a rigorous and detailed protocol for assessing 
groundwater biodiversity and to develop operational tools for 
its conservation. 

The PASCALIS project will provide a toolbox that 
includes several validated methods for (1) determining the 
reliability of patterns of regional biodiversity revealed by 
mapping of existing data, (2) obtaining by means of a 
standardized field sampling method an unbiased estimate of 
groundwater biodiversity in regions for which no data exist, 
and (3) predicting overall species richness based on 
biodiversity indicators in regions with incomplete data sets. 

These considerations should lead to the development of a 
common database and mapping process that allows 
researchers to zoom in on subterranean landscapes in order to 
consider groundwater biodiversity at different spatial scales 
(European, country, region, and local scales). For example, 
zooming on a particular region could reveal a distribution 
pattern of biodiversity at the regional scale and emphasize 
differences between different groundwater systems and eco- 
logical units. The ultimate goal of the project is to propose a 
specific action plan for the conservation of groundwater bio- 
diversity at a European level by identifying priority regions 
for conservation, by identifying the spatial scale of relevance 
for conserving biodiversity within these regions, and by 
formulating a series of appropriate measures for maintaining 

their biodiversity. 


This work was supported in part by the European project 
PASCALIS No. EVK2-CT-2001-002121. 

Diversity Patterns in Europe 201 


Culver, D. C., L. Deharveng, J. Gibert, and I. D. Sasowsky (eds.). (2001). 
Mapping Subterranean Biodiversity. Cartographie de la Biodiversité 
Souterraine, Special Publication 6. Karst Waters Institute, Charles Town, 

Culver, D. C., and B. Sket (2000). Hotspots of subterranean biodiversity in 
caves and wells. 7. Cave Karst Studies 62, 11-17. 

Deharveng, L., H. Dalens, D. Drugmand, J.C. Simon-Benito, M.M. Gama, 
P. de Sousa, C. Gers, and A. Bedors (2000). Endemism mapping and 
biodiversity conservation in western Europe: An arthropod perspective. 
Belgian J. Entomology 3, 59-75. 

Gibert, J. (2001). Protocols for the assessment and conservation of aquatic 
life in the subsurface (PASCALIS): A European project. In Mapping 
Subterranean Biodiversity. Cartographie de la Biodiversité Souterraine (D. 
C. Culver, L. Deharveng, J. Gibert, and I. D. Sasowsky, eds.), Special 
Publication 6, pp. 19-21. Karst Waters Institute, Charles Town, WV. 

Gibert, J., and L. Deharveng (2002). Subterranean ecosystems: A truncated 
function biodiversity. Bioscience 52, 473-481. 

Holsinger, J. R. (1993). Biodiversity of subterranean amphipod crustaceans: 
Global patterns and zoogeographical implications. /. Natural History 27, 

Juberthie, C., and V. Decu (1994-2002). Encyclopaedia Biospeologica, Vols. 
1-3. Societé de Biospeologie, Moulis, France. 

Juberthie, C., and L. Juberthie-Jupeau (1975). La réserve biologique du 
laboratoire souterrain du C.N.R.S. a Sauve (Gard). Annales Spéléologie 
30, 539-551. 

Sket, B. (1999). High biodiversity ion hypogean waters and its 
endangerment—the situation in Slovenia, Dinaric Karst, and Europe. 
Crustaceana 15, 125-139. 

Stoch, E (2001). Mapping subterranean biodiversity: Structure of the 
database and mapping software SKMAP and report of status for Italy. In 
Mapping Subterranean Biodiversity. Cartographie de la Biodiversité 
Souterraine (D. C. Culver, L. Deharveng, J. Gibert, and I. D. Sasowsky, 
eds.), Special Publication 6, pp. 29-35. Karst Waters Institute, Charles 
Town, WV. 

Early Humans in the 
Mammoth Cave Area 

Patty Jo Watson 
Washington University, St. Louis 


The longest cave in the world (more than 560 km, and still 
being explored and mapped) is the Mammoth Cave System 
in west central Kentucky in the United States. Beginning 
about 4000 years ago, the people living in the Mammoth 
Cave region began exploring some portions of this immense 
subterranean labyrinth. Use by indigenous peoples of parts 
of the Mammoth Cave System continued until about 2000 
years ago. The materials discarded or lost underground many 
millennia in the past preserve detailed information, not 
only about prehistoric activities in the cave, but also about 
subsistence practices and other lifeways above ground. 


The Mammoth Cave System was not the only cave entered 
by aboriginal inhabitants of the Americas. To the contrary, 
any cave opening large enough for a person to squeeze into 
was probably explored and well known. The earliest trip yet 
documented into the dark zone of any cave in the Western 
Hemisphere is that of a 45-year-old man whose skeletal 
remains were found in Hourglass Cave, high in the Southern 
Rocky Mountains (Mosch and Watson, 1997). In Eastern 
North America, the oldest archaeological record of dark zone 
cave exploration dates to 4500 years ago when eight or nine 
people made one or two trips far back inside a 11-km-long 

cave in what is now northern Tennessee. They left a thin trail 
of charcoal from their cane torches, and—in the damp clay 
floor of one passageway—the impressions of their feet 
(Crothers et al., 2002). 

For more than 4000 years, from 4500 years ago to the time 
of European contact, indigenous peoples of the Americas 
entered caves, explored them, and used them for a variety of 
purposes: as storage places, as sources of valuable minerals or 
of magical powers, as shrines, as cemeteries, or as places to 
contact the spirit world. The archaeology of the Mammoth 
Cave System illustrates several of those cultural activities for 
the two millennia between 4000 and 2000 years ago. 


Like many other regions of the eastern United States, west 
central Kentucky was first inhabited by Paleoindian 
hunter-gatherers, who arrived approximately 11,000 years 
ago. Archaeological evidence for this time period (ca. 11,000 
to 10,000 years before present) is sparse, but data from here 
and elsewhere in the midcontinent indicates to archaeologists 
that Paleoindian communities were small and were dispersed 
fairly widely across the landscape. 

In archaeological terminology, the Paleoindian period in 
Eastern North America is followed by the Archaic period 
(ca. 10,000 to 3000 years ago), which is succeeded by the 
Woodland period (3000 to 1100 years ago), and—in some 
places—the Mississippian period (1100 to 500 years ago). 
The explorers who left their footprints in that Tennessee cave 
4500 years ago are spoken of by archaeologists as Archaic 
people, and so are the earliest explorers of the Mammoth 
Cave System. There is also archaeological evidence of Archaic 
cavers in Fisher Ridge Cave just east of Mammoth Cave 
National Park, as well as Lee Cave and Bluff Cave (both 
within Mammoth Cave National Park but not known to be 


204 Early Humans in the Mammoth Cave Area 

connected to the Mammoth Cave System) and Adair Glyph 
Cave (some distance north of Mammoth Cave). 

During most of the Archaic period, archaeobotanical 
evidence shows that all societies in the Eastern Woodlands of 
North America hunted, fished, and gathered the abundant 
resources of the forests where they lived. About 3000 years 
ago, however, at the end of the Archaic and beginning of the 
Early Woodland period, some local groups had become 
farmers growing several species of small-seeded plants: 
sunflower, a relative of sunflower called sumpweed or 
marshelder, goosefoot (also called lambsquarters), maygrass 
(also known as canary grass), bottle gourd, and a gourdlike 
form of squash (Smith, 2002). These Early Woodland folk 
were among the first agriculturists in Eastern North America, 
and some of them were also highly skilled cavers who 
confidently made their way through many kilometers of 
passages in what is now known to be the world’s longest 
cave, the Mammoth Cave System. Beginning in the 1950s, 
archaeologists and speleologists have studied the remains left 
by these early cavers to discover where they went and what 
they did underground and also to investigate the evidence 
they left of their early farming economy. 


Dry cave passages are excellent repositories of archaeological 
and paleontological remains because they lack factors of 
weathering and decay that quite rapidly alter or destroy 
all perishable materials aboveground. Some of the earliest 
Euroamerican explorers of the Mammoth Cave System noted 
well-preserved traces of prior human presence there. When 
the National Park Service acquired the land above Mammoth 
Cave, Salts Cave, and a series of other caves (known since 
1972 to form a single cave system; see Brucker and Watson, 
1987), archaeologist Douglas Schwartz at the University of 
Kentucky was asked to prepare a series of reports on 
Mammoth Cave area archaeology. Subsequently, other inves- 
tigators affiliated with the Cave Research Foundation and 
the National Park Service took up this work, which is still 
continuing (Crothers, 2001; Crothers e¢ al, 2002; Watson, 
1969, 1997). Results of this research are summarized next. 


Several dozen radiocarbon dates have been obtained on a 
wide variety of material from various parts of the prehistoric 
archaeological record in the Mammoth Cave System 
(Crothers et al, 2002, and Table 1 and Fig. 1 therein). As 
already noted, the dates span some 2000 years, from a little 
before 4000 years ago to a little after 2000 years ago. Most of 
them, however, cluster around 2500 years before present, 

plus or minus a few hundred years. Thus, two archaeological 
periods are represented: Late Archaic and Early Woodland. 
Both Archaic and Early Woodland cavers explored several 
kilometers of passages in both the upper and lower levels of 
Mammoth Cave and Salts Cave, but the Early Woodland 
record is much more abundant and much more extensive 
than is the Archaic one. 

Archaeological Evidence 

The most prominent and pervasive category of prehistoric 
material in dark zone cave passages is torch and campfire 
debris. Everywhere the ancient cavers traveled, they left a 
scatter of cane and dried weed stalks, as well as charred 
fragments of torch materials. Charcoal smudges left by 
torches on cave walls, ceilings, and breakdown boulders are 
also commonly visible. River cane seems to have been the 
preferred torch material, but dried stems of goldenrod and 
false foxglove were also used. In addition to the torch 
remains, fragmentary bindings and cordage made from the 
inner bark of certain trees are strewn along the passage floors. 
Less abundant items include wooden and gourd containers, 
mussel shell scrapers or spoons, basketry and vegetal fiber 
artifacts (including remains of bags, cordage, and moccasin- 
like footwear), digging sticks, scaling or climbing poles, 
imprints of human feet in dust or damp clay floored passages, 
dried human excrement (paleofeces), and the physical 
remains (“mummies,” actually simply desiccated bodies) of 
two of the ancient cavers themselves. 

In most places within the dry cave passages entered by the 
Early Woodland people there is evidence that they removed 
cave minerals from the walls, breakdown boulders, ceilings, 
and cave sediments. These minerals consist of various sulfate 
compounds, most prominently gypsum (hydrous calcium 
sulfate) and mirabilite (hydrous sodium sulfate), with epso- 
mite (hydrous magnesium sulfate) probably also included, 
although it is much less abundant in the cave than are 
gypsum and mirabilite. 

Archaeological excavation in the entrance areas of both 
Mammoth Cave and Salts Cave yielded a considerable quan- 
tity of charred plant material and animal bone (deer, turkey, 
raccoon, opossum, squirrel, rabbit, turtle, fish, rodents), as 
well as tools made of bone, chipped stone, and ground stone. 
The Salts Cave excavations also recovered a surprising 
quantity of human bone (Robbins, 1997). 

Physical remains of prehistoric people are known from a 
few cave sites where they died by accident (e.g., an adult man 
in Mammoth Cave and a young boy in Salts Cave) or were 
purposely buried (Crothers et a/., 2002). Use of sinkholes 
and vertical shafts as mortuary facilities, into which the 
dead were lowered or dropped, was much more common 
than cave interments across the Midwest, Midsouth, and 


The somewhat scant evidence dating to the earliest human 
presence in Mammoth Cave, Salts Cave, and Lee Cave, as 
well as the Tennessee footprint cave, indicates that Archaic 
people who ventured into these underground locales seem to 
have been exploring rather than working or spending long 
periods of time there. Adair Glyph cave is quite different, 
however (DiBlasi, 1996). Here the ancient people used sticks 
or their fingers on a mud-floored passage to make a dense 
series of curvilinear and rectilinear markings whose meanings 
are unknown to us, but whose creation must have taken 
considerable time and thought, and probably more than one 
trip into the cave. At any rate, it is obvious that those 
responsible for the markings were not simply exploring this 
room, which is about a kilometer from the entrance. At 
Fisher Ridge Cave, the archaeological evidence primarily 
indicates exploration, but there is a lattice or checkerboard 
design (scratched onto a large breakdown boulder) that was 
present when the first Euroamerican cavers entered that 
passage during the early 1980s. Finally, a few geometric and 
representational figures (some scratched, some drawn with 
charcoal) documented in both Salts Cave and Mammoth 
Cave are probably prehistoric, either late Archaic or Early 
Woodland, or both (Crothers, 2001; DiBlasi, 1996). 

In Wyandotte Cave, Indiana, and 3rd Unnamed Cave, 
Tennessee, late to terminal Archaic people quarried and 
worked chert derived from underground sources. At 3rd 
Unnamed Cave they also engraved rectilinear, curvilinear, 
and representational figures onto the limestone ceilings, 
walls, and breakdown boulders. Some are reminiscent of the 
mud images in Adair Glyph Cave. 

Thus, it now appears that the first people to enter dark 
zones of several big caves in Eastern North America were 
indeed exploring these complexes, but they were also 
obtaining chert from some of them, and in several they were 
also probably carrying out ritual activities. 

During the subsequent Woodland period (3000 to 1200 
years ago), caves were important features for many 
communities across the Midsouth and Southeast as 
ceremonial and mortuary foci (Crothers et a/., 2002). Dozens 
of examples are known in Kentucky, Tennessee, the Virginias, 
Alabama, Georgia, Florida, and Texas, but the prehistoric 
archaeological record in the Mammoth Cave System ceases 
2000 years ago for reasons not yet clearly understood. For 
the centuries between about 2800 and 2300 years before 
present, however, Early Woodland cavers spent considerable 
amounts of time in both Salts Cave and Mammoth Cave 
mining sulfate minerals. They were obviously very familiar 
with several kilometers of complicated passage systems in 
each of these caves and were quite capable of navigating their 
way through multiple levels of walk