Skip to main content

Full text of "The improvement of naturally cross-pollinated plants by selection in self-fertilized lines"

See other formats

Digitized by tine Internet Arciiive 

in 2011 witii funding from 

LYRASIS members and Sloan Foundation 

Bulletin 376 September, 1935 r- , 


The Improvement of Naturally Cross- 
Pollinated Plants by Selection in 
Self-Fertilized Lines 



Agricultural iExp^rtmrat i^tattnu 

Nrvu ilaurn 



His Excellency, Governor Wilbur L. Cross, ex-officio. President 

Elijah Rogers, Vice-President Southington 

Edward C. Schneider, Secretary Middletown 

William L. Slate, Treasurer New Haven 

Joseph W. Alsop Avon 

Charles G. Morris Newtown 

Albert B. Plant Branford 

Olcott F. King South Windsor 



William L. Slate, B.Sc, Director. 

Miss L. M. Brautlecht, Bookkeeper and Librarian. 

Miss Katherine M. Palmer, B.Litt., Editor. 

G. E. Graham, In Charge of Buildings and Grounds. 


E. M. Bailey, Ph.D., Chemist in Charge. 

C. E. Shepard 1 

Owen L. Nolan 

Harry J. Fisher, Ph.D. \ Assistant Chemists. 

W. T. Mathis 

David C. Walden, B.S. J 

C. W. SoDERBERG, Laboratory Assistant. 

V. L. Churchill, Sampling Agent. 

Mrs. a. B. Vosburgh, Secretary. 


H. B. VicKERY, Ph.D., Biochemist in Charge. 

Lafayette B. Mendel, Ph.D., Research Associate (Yale University). 

George W. Pucher, Ph.D., Assistant Biochemist. 


G. P. Clinton, Sc.D., Botanist in Charge. 

E. M. Stoddard. B.S., Pomologist. 

Miss Florence A. McCormick, Ph.D., Pathologist. 

A. A. DuNLAP, Ph.D., Assistant Mycologist. 

A. D. McDonnell, General Assistant. 

Mrs. W. W. Kelsey, Secretary. 


VV. E. Britton, Ph.D., D.Sc, Entomologist in Charge, State Entomologist. 

B. H. Walden, B.Agr. ] 

M. P. Zappe, B.S. 

Philip Garman, Ph.D. \ Assistant Entomologists. 

Roger B. Friend, Ph.D. ] 

Neely Turner, M.A. J 

John T. Ashworth, Deputy in Charge of Gypsy Moth Control. 

R. C. Botsford, Deputy in Charge of Mosquito Elimination. 

J. P. Johnson, B.S., Deputy in Charge of Japanese Beetle Quarantine. 

Miss Helen A. Hulse \ 

Miss Betty Scoville J 



Walter O. Filley, Forester in Charge. 

H. W. Hicock, M.F., Assistant Forester. 

J. E. Riley, Jr., M.F., In Charge of Blister Rust Control. 

Miss Pauline A. Merchant, Secretary. 

Plant Breeding. 

Donald F. Jones, Sc.D., Geneticist in Charge. 
W. Ralph Singleton, Sc.D., Assistant Geneticist. 
Lawrence C. Curtis, B.S., Assistant. 


M. F. Morgan, Ph.D., Agronomist in Charge. 
H. G. M. Jacobson, M.S., Assistant Agronomist. 
Herbert A. Lunt, Ph.D., Assistant in Forest Soils. 
Dwight B. Downs, General Assistant. 

Tobacco Substation Paul J. Anderson, Ph.D., Pathologist in Charge. 

at Windsor. T. R. Swanback, M.S., Agronomist. 

O. E. Street, Ph.D., Plant Physiologist. 
Miss Dorothy Lenaed, Secretary. 

Printing by Quinnipiack Press, Inc., New Haven, Conn. 



Introduction 653 

The Performance of Inbred Strains when Crossed 657 

Parental Characters in the Hybrids • 661 

The Results of Crossing after Inbreeding and Selection 663 

White flint corn 663 

Evergreen sweet corn 667 

Whipple Yellow sweet corn 671 

Method of Testing Inbreds 677 

Single crosses 678 

Reciprocal crosses '. 679 

Double crosses 681 

Advanced generation crosses 681 

Multiple crosses 682 

Variety-inbred crosses 683 

Single plant crosses 683 

Method of Producing Inbred Strains 685 

The Effect of a Deleterious Factor on Yield 686 

Seed Production 686 

Summary 689 

The Improvement of Naturally Cross-Pollinated 
Plants by Selection in Self-Fertilized Lines 


Donald F. Jones and W. Ralph Singleton 

The application of inbreeding to the improvement of corn has received 
an increasing amount of attention since its inception a quarter of a century 
ago. In 1932 more than 80,000 acres of corn were planted in this country 
with seed that had resulted, in one way or another, from the crossing of 
inbred plants. Although the production of hybrid corn seed seems to be 
well started in commercial practice, the development, testing and most 
advantageous use of inbred strains still present many problems. 

The first attempt to account for the increased growth immediately 
resulting from cross-fertilization may be called the physiological hypothesis 
of hybrid stimulation. In this it was thought that the bringing together 
of germinally diverse elements by crossing different lines, generated the 
growth activity commonly referred to as hybrid vigor. Conversely, the 
reduced size and slower growth rate of inbred lines was considered the 
natural result as the germinal constitution is simplified. 

It was later pointed out that hybrid vigor can be given a genetic inter- 
pretation. It was noted that many of the heritable factors which favored 
growth and reproduction are dominant in their expression. Therefore in 
heterozygous combinations it is possible to have a larger number of 
different factors and hence more favorable factors than in homozygous 
combinations. Due to linkage and other difiiculties of recombination, it 
is extremely difficult, if not impossible, to obtain all of the growth-pro- 
moting genes that exist in a species, combined in one individual. Such a 
plant, if obtained in a homozygous condition, would show no benefit from 
crossing with other individuals and no reduction when inbred. 

As a general rule, nearly every individual is benefited temporarily, in 
one way or another, by crossing with other specimens of the same or closely 
related species. This does not follow in every case since certain combina- 
tions are not compatible. Wide crosses are usually not viable and when 
offspring are produced they are not able to function properly. Hybrid 
weakness as well as hybrid vigor may result from crossing and this is 
more easily explained by the genetic interpretation than by the physio- 
logical hypothesis. 

* Part 1, Station Bulletin 266, March, 1925. 


Connecticut Experiment Station Bulletin 376 

Figure 95. Two short-eared inbreds of Gold Nugget flint and their first 
generation hybrid below. 

Introduction 655 

According to the dominant gene hypothesis, the effects of inhreeding 
are due largely, if not wholly, to the inheritance received. Species that 
are regularly cross-fertilized maintain a condition of heterozygosity which 
allows recessive factors to accumulate. Such organisms are usually reduced 
in size and rate of growth by inbreeding, and this condition is frequently 
accompanied by the production of some abnormal and undesirable indi- 
viduals. Since these are generally weak and often sterile, they automati- 
cally tend to eliminate themselves. The survivors of a severe and continued 
process of inbreeding are usually brought to a high degree of homozygosity 
and are large or small, productive or non-productive, disease-resistant or 
not, according to the specific allotment of genes they have received. 

This genetic interpretation of the effects of inbreeding and crossing 
has emphasized the importance of selection during and after the inbreeding 
process. Since the results of crossing depend upon the constitution of the 
inbred strains used, it is essential that the inbred material contain the 
qualities to be expressed in the hybrid, or in the new variety or breed 

The first report of this series of experiments on Selection in Self- 
Fe7'tilised Lines was published in 1925 as Station Bulletin 266. It dealt 
with the production of inbred strains of corn. The possibility of obtain- 
ing vigorous inbred plants was considered, along with methods of hand- 
pollination, systems of selection during the inbreeding process, criteria 
of selection, and the correlation of characters in successive generations. 
The general conclusion from this study was that selection during the 
inbreeding period was effective in establishing specific characters. 

The number of progenies grown, the number of individuals in each 
progeny, the number of plants self-pollinated, and the selection among 
the plants self-pollinated had an important bearing upon the results ob- 
tained with specific characters, such as height of plant, time of maturity, 
and resistance to disease. Selection during the inbreeding process for 
general vigor and productiveness in the inbred strain itself was of little 
avail. Good plants at the start gave both good and poor strains after 
several generations of self-fertilization. Many poor strains at the begin- 
ning were numbered among the best at the finish. Selection of the most 
vigorous and productive plants tended to perpetuate the heterozygous 
condition that existed at the start and delayed the attainment of homo- 
zygosity. The general conclusion was reached that, when productiveness 
is the principal objective, extensive selection within the inbred lines does 
not seem to be advisable. 

The importance of selection lies in choosing the most useful strains 
after uniformity and constancy have been obtained. Most effort should 
be expended in producing the largest number of homozygous strains. 
This emphasizes the importance of starting many lines and selecting the 
plants for progenitors largely at random, unless the specific qualities sought 
are evident despite hybrid vigor. 

When inbred strains have been produced, the next consideration is 
to test them to find which are the most useful for the purpose in mind. 
Having found the best strains, how can they be utilized to the greatest 


Connecticut Experiment Station 

Bulletin 376 

Figure 96. Two long-eared inbreds of Canada Yellow flint and 
their first generation hj^brid below. 

The Performance of Inbred Strains When Crossed 657 

advantage? The testing and utilization of inbred strains of corn is the 
subject that will be considered in this part of the series. 


The results of self-fertilizing four varieties of corn for five generations 
are described in Part I of this series. These varieties are : Burwell's 
Flint, an early maturing type of eight-rowed Canada Yellow Flint ; Gold 
Nugget, a late maturing, eight-rowed flint having large kernels ; Century 
Dent, an early maturing dent corn having broad, shallow kernels; and 
Leaming, a local strain of this well known variety of corn that is rather 
late maturing in Connecticut. About 20 selected lines were started from 
each of these four varieties. The appearance and nature of the selected 
lines are described in detail in the previous publication. 

From a study of the yields obtained in the first generation after crossing 
these inbreds in various combinations, it was at once apparent that some 
lines had a much more favorable efifect upon productivity than others. 
An attempt was made to cross each strain with all of the other strains 
of the same variety. Not all combinations were successful in producing 
enough seed to grow. The yields were averaged for each series of crosses 
in which one strain was used as one of the parents. Averages represent 
from 3 to 20 crosses. All were grown in the same field under as nearly 
similar conditions as could be obtained and were distributed at random. 

In Table 1 it will be seen that some strains produced more than twice 
as much grain as others when used in combination with other inbred 
lines out of the same variety. This is the usual result to be expected from 
crosses of a series of inbred lines. But is it possible to foretell the be- 
havior of any given strain before the cross is made? If so, is it feasible 
to eliminate certain lines without going to the trouble of making the 
crosses and testing them? 

Table 1. The Average Yield of a Series of Crosses of Inbred Strains 

Classified as Productive (P), Unproductive CU), or Intermediate 

(I), According to Their Performance before Crossing. 

Strain Strain 


Year Grown 







Burwell Flint 




47 ± 






Burwell Flint 




27 ± 






Leaming Dent 




45 ± 






Leaming Dent 




57 ± 



112-1 ■ 



Of the eight strains listed in Table 1, two were classified as unpro- 
ductive in Bulletin 266 (Figures 53 and 57). Both of these, 30-6 and 
112-3, were in the lowest yielding series of flint and of dent crosses in 
1924. Three were classified as productive: 40-7, 112-1, and 112-6 (see 
Figures 52 and 56, Bulletin 266) . The crosses in which these three strains 
took part as one of the parents were the most productive in three out 
of four cases. In no case were the highest average yields obtained from 


Connecticut Experiment Station Bulletin 376 

Figure 97. Two tapering inbreds of Learning dent and their first genera- 
tion hybrid below. 

The Performance of Inbred Strains When Crossed 659 

strains that were classified as unproductive. The opposite also holds true, 
although once in the dent, and twice in the flint, the highest or lowest 
average yield was made by a strain in the intermediate class. 

The classification of these strains as productive or unproductive was 
made after they had been self-fertilized for five generations, and before 
any crosses had been made. Of the 15 to 20 surviving strains in the 
Burwell Flint and Leaming Dent, four were selected as being the most 
productive and four as the least productive. 

Another classification was made, in which the yield of dry grain was 
only one criterion of selection. Others were: Size and general appearance 
of the plants, their freedom from disease, and the quality of the corn 
produced. They are listed and described in Part I on page 411. For 
convenience they are called "good" and "poor" lines. Some of the poor 
lines yielded a heavier weight of grain than any of the good lines, but 
were undesirable in quality and in other respects. 

Jenkins (1929) made a careful statistical study of the correlation be- 
tween the inbred parents and their hybrid oiTspring in many measurable 
characters. He found that yield in the crossbred progeny is correlated 
significantly and positively with number of days to tassel and silk, plant 
height, number of nodes, number of ears per plant, ear length, ear 
diameter, and yield of parent plants. Kiesselbach (1922) also concludes 
that "there appears to be some general correlation between productivity 
of the pure line parents and that of their hybrid offspring". The partial 
correlation between productiveness of the inbreds and of the crosses 
derived from them thus seems to be well established. 

The eight lines mentioned above were crossed with one another in the 
three possible combinations : Good X good, good X poor or the reciprocal, 
and poor X poor. The results obtained are shown in Table 2. From 
12 to 35 different crosses are represented in each combination. The yields 
are the average of all crosses grown. There is a dift'erence of 4.2 bushels 
per acre between the matings of two good and of two poor strains in 
the fliint, and 6.8 in the dent. The good by poor and reciprocal combina- 
tions give an intermediate yield as compared with the other two. 

Table 2. The Yield of First Generation Hybrids in Bushels per Acre 

Classified According to the Good or Poor Performance of the 

Parental Strains. 


Good X Good 
No. Yield 

Good X Poor 
No. Yield 

Poor X Poor 
No. Yield 

Burwell Flint 
Leaming Dent 

34 52 ± 1.2 

35 56 ± 1.4 

12 52 ± 1.3 
26 52 ± 1,5 

23 48 ± 1.3 
12 49 ± 2.1 

There are individual crosses that stand out as good or poor in contrast 
to their parentage. In every case, however, the highest yielding individual 
cross in the good X good combination is more productive than the highest 
of the poor X poor combinations. Similarly the lowest of the poor X poor 
matings is the lowest for all the crosses. The differences are not signi- 

Many of these single crosses, which combine two inbred strains, were 
again crossed, producing a first generation hybrid that is called, for con- 


Connecticut Experiment Station 

Bulletin 376 

Figure 98. Two cylindrical inbreds of Learning dent and their 
first generation hybrid below. 

No. of 
Good Strains 



Yield in Bushels 
48 53 58 63 

per Acre 
68 73 






5 3 8 4 

8 14 18 7 

2 12 12 

14 2 4 

3 1 
9 3 

7 3 
3 2 

Parental Characters in the Hybrids 661 

venience, a double cross. In Table 3 are classified the yields of 142 double 
crosses all made from inbred strains of one variety. The yields are grouped 
according to the number of strains in their pedigree that were designated 
as good before any of the crosses were made. The one combination that 
would have brought together all four of the good strains was not made, 
due to a failure in pollination. Each of the others had from none to three 
good strains in their pedigree. All of the crosses were grown in one field, 
in the same year, and under uniform conditions. Each was represented 
by several plots distributed at random throughout the field. 

Table 3. The Distribution of Yield of Double Crosses Based Upon the 
Number of Good Strains Used in Making the Crosses. 

No. of Average 

83 Crosses Yield 

27 56 ± 1.1 

61 58 ± .5 

1 ?,7 63 ± .7 

17 62 ± 1.4 

Difference and 3 6 ± 1.8 

There is a fairly consistent increase in yield from 56.2 bushels per acre, 
where all four of the parental strains were poor, to 62.1 bushels, where 
three of the parental strains were good. The difference is 6 ± 1.8 bushels 
per acre and is hardly significant. Yield alone does not indicate fully 
the superiority of crosses made from the better parental strains. There 
were noticeable differences in quality of grain as shown in freedom from 
cracking and molding. Many combinations had notable ability to stand 
erect throughout the season, to cover the ears well with husks, and had 
other outstanding qualities fully as important as the production of grain. 


The tendency of particular characters in the inbreds to be expressed 
in the crosses is shown in the accompanying illustrations. Two short- 
eared, broad-kerneled inbreds of Gold Nugget Flint, 105-18 and 105-20, 
give short broad ears in the first generation hybrid. In one of these strains 
the plant is quite short in stature, with ears starting at nodes only a few 
inches above the ground. The other strain is nearly twice as tall, the 
tip of the ears being on a level with the base of the tassels of the other 
strain. Both are out of the same variety and have not been selected for 
height. The crossed plants are taller than either parent, but not so tall 
as other crosses in the same material. (Figure 95.) 

In contrast to this is the Burwell Flint hybrid in which two rather long, 
slender inbreds, 40-7 and 40-8, are brought together. The result is a 
long, slim ear, proportionally more slender than either of the parents. In 
this cross the pericarp has a tendency to crack, an undesirable feature 
that it not noticeable in the selfed strains themselves. (Figure 96.) 


Connecticut Experiment Station Bulletin 376 

Figure 99. Two productive inbreds of Learning dent and their first gener- 
ation hybrid below. 

Results of Crossing after Inbreeding and Selection 663 

The tendency of dent corn ears to be cylindrical or tapering can be 
fairly well foretold by the appearance of the inbreds before they are 
crossed. Strains 112-9 and 112-12, out of Beardsley's Learning, each 
have a decided inclination to be narrower at the tip than at the butt. The 
first generation hybrid of these two strains has a marked tendency in this 
direction. (Figure 97.) This is in contrast to the cross of 112-1 X 112-4, 
two lines that carry their width of ear evenly throughout their length. 
One of these inbreds was the longest of all the selfed lines from this 
variety, and the hybrid is also outstanding in its length of ear. The kernels 
of both parental lines are small and nearly round on their exposed surface, 
and this feature is clearly apparent in the cross. (Figure 98.) 

In the combination 14-4 by 243, both inbred strains of Leaming from 
a source different than the 112 lines, we see the ear and kernel characters 
of one of the parental inbreds directly expressed in the hybrid. The other 
parent has kernels that are irregular in size and shape, and the ears are 
poorly filled. The plant characters of this particular inbred are good. The 
stalks are vigorous and erect ; the leaves are broad and well colored ; and 
the plants mature early. Aside from a marked susceptibility to bacterial 
wilt, this is one of the best strains of dent corn we have produced. Tlie 
other strain with the good ear characters also has a sturdy stalk, but is 
late in maturing. Their hybrid matures in satisfactory season. It is among 
the best of all the crosses so far obtained in productiveness, ability to stand 
erect and quality of grain, when grown under conditions that prevail in 
Connecticut. (Figure 99.) 


White Flint Corn 

Six varieties of eight-rowed white flint corn were used in an inbreeding 
and selection experiment. These varieties had been grown for some time 
in Connecticut and apparently were well adapted to local conditions. They 
differed considerably in length of ear, size of kernel and in uniformity 
of filling over the tips. They were all eight-rowed, with the typically 
white, corneous kernel characteristic of this class of corn. 

Twenty-five or more naturally pollinated ears were chosen out of each 
variety. After the first generation, three progenies in each line were grown 
yearly. Two plants were selected for self-pollination out of aboiit ten plants 
grown in each progeny. On the basis of plant characters, ear development 
and freedom from disease, the two best lines were noted. Whenever 
possible, the two selfed ears from the best and one from the next best 
progeny were used to continue the line the following year. Due to failure 
to secure sufficient seed such an arrangement was not possible in every 
case, but most of the lines were continued in this way for four generations. 

This system was considered to be an improvement over the method of 
selection in self-fertilized lines first proposed. Selections were not made 
until the crop matured. Each line was represented by two different pro- 


Connecticut Experiment Station 

Bulletin Z76 

genies and final elimination did not come until two of the three progenies 
had been grown for at least two years. 

After four generations of self-fertilization and selection in this way, all 
lines were pollinated by a stock of yellow dent corn regularly used as 

Figure 100. The original naturally pollinated ears of white flint from which 
inbred lines were derived. 

the pollen parent of the Canada-Learning hybrid, described in Station 
Bulletin 310. (This pollen parent is a composite of several inbred lines 
derived from Illinois and Connecticut strains of Leaming.) Jenkins and 
Brunson (1932) and Jenkins (1934) used this method of pollinating all 

Results of Crossing after Inbreeding and Selection 


inbreds by a common stock to test inbreds, and found that it had advan- 
tages. St. John (1934) obtained results which indicate that inbreds used 
as the seed parent yield less than when used as the pollen parent in the 
same combinations. According to his results, a series of inbreds, all 
pollinated by one variety — the most convenient way to make the pollina- 
tion — does not give as reliable an indication of yielding ability as when 
the crosses are made the other way. 

In the experiment here, each of the crosses of the inbred lines X 
Learning was grown in single row plots between two check rows planted 
with Canada-Learning. The check plots were similar in type to the test 
plots. Both contained flint-dent hybrids having the same general season 

Figure 101. The result of pollinating two ears in three progenies of a 
number of lines of white flint after three generations of self-fertilization. 

of ripening and about the same amount of growth. In previous tests the 
Canada-Leaming had proved to be a remarkably fast growing and high 
yielding corn, fairly uniform in type. 

A photographic record was made of the original ears and the hand- 
pollinated ears in each generation for all self-fertilized lines. Figures 100 
and 101 show the original ears of two of the six white flint varieties from 
which the lines were started, together with self-pollinated ears in the third 
generation of some of the lines that were derived from these two varieties. 
From their appearance, the original ears give no indication whatever as 
to the characters of the inbred strains derived from them. And neither 
the original ears nor the hand-pollinated ears in the third generation indi- 

666 Connecticut Experiment Station Bulletin 376 

cate clearly the results obtained when the inbreds are crossed with an 
unrelated variety. 

Based on the appearance of the growing plants and the mature ears 
when harvested after three generations of self-fertilization, 51 lines out 
of the 153 started were noted to be promising. This selection took into 
consideration general vigor of the plants as judged by the stalk growth, 
ability to stand erect, freedom from disease, total production of grain, 
and quality of the grain itself. If it had been impossible to use all of 
the inbreds, these were the ones that would have been chosen for propa- 
gation. However, all of the surviving lines, a total of 113, were crossed, 
as previously stated. 

For comparative purposes it is better to classify the inbreds during the 
same season that the crosses are grown, since one year often dififers 
markedly from another in rainfall and temperature. The inbreds and their 
crosses behave differently in different seasons. In the tests, however, 
classification of the inbreds was made in 1925, and the yield comparison 
in 1927, and again in 1928. 

As stated above, 113 of the original 153 lines were used in making 
the crosses. The rest were lost in the process of inbreeding. All of the 
crosses were divided into two groups according to their yield in com- 
parison with the two adjacent check plots. In the first class were those 
that yielded more than the average of the two adjacent checks, and in 
the second those that yielded the same or less. As previously mentioned, 
the white flint inbreds had been classified as "selected" and "not selected." 
Putting these figures into a four-part frequency distribution table we have : 

Inbreds Selected Not Selected 





Number of crosses 
yielding more than 
average of checks 

Number of crosses 
yielding less than 
average of checks 

The coefficient of association is -|- .40. Of all the crosses that were 
above the average of the adjacent check plots in yield, 55 per cent were 
made with selected inbreds and 45 per cent from those not selected 
previously as promising. Of the crosses yielding the same or less than 
the check, 34 per cent were made with the selected inbreds and 66 per 
cent with those not selected. Stated conversely, 28 per cent of the crosses 
of selected inbreds were above the average in yield, while only 14 per 
cent of the unselected were above average. According to this there is 
some evidence that desirable lines can be chosen in the third generation 
of self-fertilization. 

The lines from the six different white flint varieties were classified 
separately. All gave about the same result so that the figures from all 
are combined. 

Results of Crossing after Inbreeding and Selection 


A study was made of the parentage of the exceptionally good and poor 
combinations. In some cases as many as four replications of each cross 
were grown. A classification was made including only those crosses whose 
replications consistently yielded more or less than either adjacent check. 
The four-way distribution is as follows : 

Lines Selected Lines Not Selected 

Crosses yielding 
more than either check 





Crosses yielding 
less than either check 

The coefficient of association is -|- .21, less than in the previous classi- 
fication. Stating the results in general terms, we can say that a selection 
of one-third of the inbreds before crossing was successful in obtaining 
only about one-half of the higher yielding combinations, when all of the 
inbred lines were crossed with one distinctly different type of corn. Only 
one combination, replicated three times, yielded more than either check 
in all plots and this was not from a selected line. While there is some- 
thing to be gained by making an elimination before crossing, it is question- 
able whether or not any normal lines can be discarded before being tested. 

Evergreen Sweet Corn 

In 1921, fifty-six lines of Evergreen sweet corn were, started. These 
were selected from 200 open-pollinated ears on the basis of a germination 
test showing seeds free from fungous infection. The original 200 ears had 
been chosen because they were of desirable size and shape and the type 
of kernel was good for canning purposes. The 56 lines were continued 
by self-fertilization, three progenies being grown in each. At the time of 
pollination, the best-appearing of the progenies in each line was selected 
and five plants were self-fertilized. 

In the fourth generation, six lines were chosen^ on the basis of growth 
of stalk and ear development of the inbred plants, as the most promising 
of the entire lot. They were all sufficiently productive to be used as a 
possible seed parent in a crossing field. These six lines were used as 
pollinators. They were crossed on one another and on each of the other 
lines, with the exception of a few that were too poor to be continued. 

Table 4. 






The Effect of Self-Fertilized Lines of Evergreen Sweet Corn 
Upon the Yield of Crosses in Average Weight of 
Grain in Pounds per Plot. 

Used as 

seed parent 

and crossed 

by five others 

7.6 ±.13 
9.1 ± .39 
8.4 ± .05 

8.7 ± 26 
6A ± .20 
7.7 ± 27 

Used as 
pollen parent 
and crossed 
on five others 

8.3 ± .49 
8.1 ± .32 

7.5 ± .43 
7.8 ± .39 

8.5 ± .34 
7.8 ± .24 

Used as 
pollen parent 
and crossed 
on all others 

8.1 ±.10 
8.3 ± .12 
8.0 ± .12 
8.3 ± .11 

8.2 ± .12 
7.6 ± .09 

Connecticut Experiment Station 

Bulletin 376 

All of these crosses were grown in single row plots in two different 
fields, at Mount Carmel and at Orange. The original strain of Evergreen 
sweet corn was used as a check in every fifth row. The total yield of 
air-dry ears was weighed and corrected to the check yields. These were 
then averaged for the six inbreds used as a seed parent and as a pollen 
parent in various combinations. Results are given in Table 4. The 

•^ ^ to -^ •»-. — . 



Figure 102. One representative ear of first generation hybrid combinations 

of inbred strains of Evergreen sweet corn. The pollen parents 

are numbered across the top and the seed parents at the left. 

figures indicate that the combinations containing lines 50 and 63, either 
as a seed or a pollen parent, are among the highest in yield. The differ- 
ences are small and significant only for the highest and lowest average 
yields, but they furnish some evidence that these two lines impart pro- 
ductiveness to their hybrid offspring. 

Results of Crossing after Inbreeding and Selection 


Figure 102 shows selected ears from these various combinations. Each 
cross is represented by one ear. The pollen parents are shown across the 
top with the seed parents indicated at the sides. The combinations at the 
left are reciprocal crosses among the six pollinators. The ones at the 
right are six other lines crossed by the same pollinators. Each vertical 
row of ears has the same pollen parent and the horizontal rows have the 
same seed parent. The individual ears shown were selected in the field 
as the best from all of the plants grown in each plot. Most of the crosses 
were so uniform that the ear size and shape and the kernel characters 
of the specimens shown are fairly representative of the whole crop. 

These various combinations show a marked influence from certain 
parents. Lines 50 and 63 have a tendency to produce cylindrical, well- 
filled ears in nearly all crosses when used either as a seed or as a pollen 
parent. Lines 5, 9, and 90 are exceptions in that all combinations with 
these as seed parents are tapering. Line 96 has a noticeable tendency to 

Figure 103. Mature ears of the first generation hybrid of C.63 X 50, 
Evergreen sweet corn. 

produce long ears, while all of the combinations with 57, either as a seed 
or pollen parent, and 158 as a seed parent, are short. The cross of 57 X 
158 is the shortest of the entire lot. Ears from crosses having 9, 90 and 
195 are not well filled at the tips. Other differences in width and depth, 
color and texture of the kernels, are equally in evidence. 

Before the yields were calculated, the combination of 63 X 50 had 
been selected as the outstanding hybrid of the entire series largely on the 
basis of uniformity, ear size and ear type. This selection was confirmed 
by repeated trials. From the first trials, 12 of the most promising com- 
binations were grown the two following years at three different places 
in Connecticut. 

In the second trial, the 63 X 50 combinations showed the best devel- 
opment but the plants had a tendency to lodge. In the third trial they 
displayed the same weakness. But in spite of the plants being badly down, 


Connecticut Experiment Station 

Bulletin Z76 

they still maintained the best ear development. Most of the combinations 
with line 63 have shown root weakness, apparently associated with rootrot 
infection. No other combinations had the ability to produce such large, 
well-shaped ears. Therefore the 63 X 50 cross was selected as the most 
promising, and called Green Cross. 

In an attempt to eliminate the root weakness, remnant seed from three 
progenies in the second generation of lines 50 and 63 was grown, and 
61 new lines were started. These were selfed for three generations, select- 
ing plants that were erect each time. The most promising of the new 
number 50 lines were crossed with the most promising of the number 
63 lines, and compared with the original 63 X 50 combination. Out of 
13 crosses grown in 1932, five were selected for further testing in 1933. 
After two years of testing, none of the new combinations were appreciably 

Figure 104. In size of ear this Evergreen single cross, called Green Cross 
C63.50, is outstanding. 

better than the original cross. They differed slightly in stalk growth, ear 
shape and time of maturity, but all had a tendency to go down and none 
were sufficiently better to offer much hope of improvement by reselecting 
a self-fertilized line from the second generation. In a case of this kind 
Richey's (1927) method of convergent improvement offers more promise. 
The results obtained in this selection experiment emphasize the im- 
portance of subjecting the inbred lines to as many adverse conditions as 
possible in the hope of eliminating susceptibility. It might have been 
better to have used ears showing the heaviest infection on the germinating 
seeds in the beginning. More lines should have been started, even if it 
were necessary to grow fewer plants in each line and to make fewer pol- 
linations. The hybrids should also have been tested in a wide range of 
seasonal and soil conditions and with as severe disease infection as possible. 

Results of Crossing after Inbreeding and Selection 671 

It is becoming more and more apparent that these first generation hybrids 
of inbred strains, on account of their germinal uniformity, are lacking in 
adaptability to different localities and varying conditions. The very fact 
that they do especially well in some places and under certain seasonal and 
cultural conditions is an indication that they may be equally poor under 
other circumstances. 

Whipple Yellow Sweet Corn 

Ten years ago this native sweet corn was first being used extensively 
for market garden purposes in Connecticut. It had originated several 
years before as the result of natural crossing between an unnamed, large- 
eared, early, white, sweet corn and Golden Bantam, followed by selection 
in the hands of Mr. Silas S. Whipple of Norwich, Connecticut. This 
variety produces a large ear in a short time and is still the leading market 
garden corn. 

In 1924 a number of naturally pollinated ears were selected and planted 
in an ear-row trial. Individual plants were chosen for self-fertilization 
in each row. In addition, two lots of bulk seed were planted and individual 
plants self-fertilized. Altogether 106 lines of this early yellow sweet 
corn were started. Two plants were selfed in each line each year and 
one of these was used for propagation. 

After three generations of self-pollination, 32 lines had been eliminated. 
Of the 74 remaining, 23 were noted as promising either on the basis of 
stalk growth in the field or after a comparison of the mature ears. Six 
of these were selected as pollinators and crossed with all other lines. They 
were chosen largely for their ear characters with the expectation that 
any one of them could be used as a seed parent. It was assumed that 
any strain that would combine well with one of them could be used for 
a pollen parent if it were not satisfactory as a seed parent. 

Out of the 206 first generation Whipple crosses grown in 1929, nine 
were selected as the most promising. The whole series, after discarding 
some of the poorest, were again planted in 1930 and from these, four 
combinations were chosen. Only one of these was included in the pre- 
viously selected list. 

On the basis of their behavior when crossed with the six pollinators, 
a number of inbreds were selected as promising and nearly all possible 
combinations were made among these and tested from 1931 to 1934. Each 
year from three to six of the most outstanding combinations were noted. 

There was considerable variation in weather conditions from year to 
year. Also in 1932 and 1933 bacterial wilt was a serious disease and 
many of the best combinations were found to be quite susceptible. For 
these reasons no one combination is outstandingly good throughout the 
period of testing. Comparing all the selected crosses, we find that two 
inbreds, numbers 6 and 7, were represented in all six years. Two others, 
numbers 2 and 12, were represented by desirable crosses in four years ; 
and five others, 5, 9, 11, 24, and 89, were used in high yielding crosses 
in at least two years. 


Connecticut Experiment Station 

Bulletin 376 

Figure 105. Three Whipple Yellow inbreds, numbered from top to bottom, 

2, 7, 12. 

Results of Crossing after Inbreeding and Selection 


Taking these nine inbreds and comparing them with the list of inbreds 
noted as promising after three generations of selfing, but before any crosses 
had been grown, we find that all the selections except one are included. 
In other words, if we had discarded all of the inbred lines except the 
23 that were noted as desirable, we would have saved the testing of many 
combinations that later proved to be inferior. However, as this has not 
been the case in tests with other material, it cannot be relied upon as 
an effective method of procedure. 

The combinations finally selected as the most desirable were 6.2 and 7.2. 
The first is more resistant to bacterial wilt and somewhat earlier in ma- 
turing. The latter usually produces a larger number of marketable ears. 
Both have been tested for only two years and may yet be found less 
desirable than other combinations in certain seasons. But all three inbreds 
have shown up well in some combinations in all six years and therefore 
should be reasonably dependable. 

The influence of one inbred parent on the number and weight of ears 
per plant and on the amount of tillering is shown in Table 5. All crosses 
having one inbred in common, whether used as the seed parent or pollen 
parent, are grouped together. The number of crosses in each group varies 
considerably but in nearly every case there are from 25 to 40. The average 
weight and number of marketable ears are calculated for each lot. Inbred 
number 39 is outstanding in the average weight of mature ears. Number 
5 is the highest in number of ears. The differences between the high and 
low averages in both weight and number of ears are significant and show 
clearly that inbred strains have the ability to influence the. development 
of particular characters in a series of crosses with other strains. 

Table 5. The Effect of One Inbred Parent on the Weight, the Number 

OF Marketable Ears and the Number of Tillers per Plant, in a 

Series of First Generation Crosses of Inbred Strains of 

Whipple Yellow Sweet Corn. 

Weight of 

Number of 

Number of 






per Plant 

per Plant 

per Plant 


.35 ± .004 

1.13 ±.016 

1.39 ± .044 


.33 ± .004 

1.14 ± .024 

1.53 ± .040 


.33 ± .004 

1.19 ± .024 

.84 ± .036 


.32 ± .004 

1.11 ± .020 

1.10 ± .048 


.31 ± .008 

1.23 ± .032 

1.49 ± .040 


.31 ± .008 

1.00 ± .016 

1.07 ± .132 


.29 ± .008 

1.05 ± .024 

.95 ± .044 

Considerable difference also exists in the number of tillers per plant. 
Inbreds number 5 and 82 put nearly twice as many on their crosses as 
do numbers 7 and 55. Tillers are usually beneficial, especially for early 
sweet corn. (Jones, Singleton and Curtis, 1935). By means of tillers, 
corn is able to produce a large amount of foliage in a short time. Late 
maturing corn has enough leaf area on the main stalk to nourish a good 
yield of grain without the help of side branches. But nearly all early 
varieties of both field and sweet corn are dependent upon tillers for an 
adequate expanse of foliage. It might be useful to select for increased 
tillering, provided the tillers themselves do not produce ears. Nearly 
all of the best combinations of these Whipple inbreds tiller freely. 


Connecticut Experhnent Station 

Bulletin 376 

Results of Crossing after Inbreeding and Selection 675 

In Table 6, a series of inbreds is compared with their first generation 
hybrid offspring in height of plant, number of tillers and percentage of 
smut (Ustilago Zeae) infection. In this comparison, from one to fourteen 

Table 6. The Relation Between One Inbred Parent and the Average 
OF All of its Fi Offspring in Height of Plant, Number of Tillers 
and Percentage of Smut Infection in Whipple Yellow Sweet Corn. 


Height of Plant 

Average Number 

Percent of Plants 



of Tillers per Plant 

with Smut 



Inbred Fi 

Inbred Fi 




1.7 1.7 

21 4 




.1 1.5 

65 12 




.8 1.9 

7 7 




1.3 1.6 

7 9 




.5 1.9 

20 7 




.6 1.9 

5 9 




.6 2.2 

13 10 




.02 1.1 





2.0 2.3 

23 6 

89 56 72 .9 1.5 22 6 

Correlation Coef. + .74 ± .10 + .29 ± .20 + .30 ± .19 

crosses are used in each series having the inbred stated as one parent. 
Each inbred influences strongly all three characters. The tallest inbred, 
the one with the largest number of tillers, and the one with the least smut 
infection produce crosses with the same outstanding traits. The parent- 
offspring correlations are positive in all three characters and clearly sig- 
nificant in the case of height. In tillering and smut infection, certain 
inbreds that do not show the characters themselves have the ability to 
impart them to their offspring. 

Ear characters are particularly important in sweet corn. A series of 
Whipple crosses grown in 1930 is grouped according to the inbred parents 
and averaged for the number of rows of grain on the ear, length of ear, 
breadth of ear, (average of butt and tip diameter), taper (butt diameter 
divided by tip diameter), kernel length and breadth. In all respects the 
crosses show small but noticeable differences as shown in Table 7. 

Table 7. The Effect of One Inbred Parent on Ear and Kernel Characters 
IN A Series of First Generation Crosses of Whipple Yellow Sweet Corn. 





Ear Row 

Ear Length 

Ear Breadth 

Ear Taper 

















































































Connecticut Experiment Station 

Bulletin 376 


CO ^ 

"& I- 

> s 

■J 3 

C/2 (-1 



Method of Testing Inhreds 677 

In 1931 a number of crosses were badly blown down. Notes were made 
at the end of the season as to whether the plants were generally erect or 
leaning. The majority in each lot were uniform in their performance, that 
is, all of one combination were either erect, bent at an angle, or in some 
cases flat on the ground. When most of the plants were bent at an angle 
of more than 60°, they were classified as leaning. Table 8 shows the 
behavior of leaning and erect plants with respect to number of tillers, 
height of plants, number of marketable ears and weight of ears. In every 
instance the upright crosses were slightly superior, but not significantly 
so, except in the case of tillering, to the combinations that were not erect 
at the end of the growing season. The upright plants were taller, had 
more ears, a heavier total weight of mature ears and a significantly higher 
number of tillers per plant. 

Table 8. The Relation of Erectness of Plant to the Number of Tillers, 

Height of Plant, Number of Ears and Weight of Ears Produced 

IN A Series of First Generation Crosses of 

Whipple Yellow Sweet Corn. 

of Crosses 

No. of Tillers 
per Plant 

Height of Plant 

No. of Ears 
per Plot 

Weight of Ears 

per Plot 

in Pounds 


2.0 ± .05 
1.8 ± .04 

74 ± .94 
73 ± 1.65 

20.3 ± .56 

18.7 ± .32 

7.1 ± .14 
6.8 ± .11 


The combination of two inbred strains, grown only the first year after 
crossing, is generally called a single cross. When such a cross is again 
cross-pollinated with a third inbred strain, it is referred to as a three-way 
cross, and the hybrid union of two single crosses is named a double cross. 
Instead of using first generation hybrids to produce a double cross, the 
second or later generations following a cross, have been used and such 
combinations have been described as advanced generation crosses. 

When a number of inbred strains are combined into a true-breeding 
variety by continued inter-pollination, the result is a synthetic variety, 
or a multiple strain. And when two or more of these multiple strains are 
brought together to produce a first generation hybrid, the ofifspring is 
spoken of as a multiple cross. 

The cross-pollination of a variety by a single inbred is called a variety- 
inbred, top cross, or line-variety ' cross. In some cases strains of com 
derived from a single open-pollinated ear have been used without further 
inbreeding. These strains are called, for convenience, single ear, or per- 
haps better, single plant strains. They will probably be used in crossing, 
either among themselves or with inbreds. No good name seems to suggest 
itself for these crosses, and for the present they may be called simply 
single plant crosses, or single plant top crosses. 

By the term inbred, it is generally understood that the plants have been 
self-pollinated or sib-pollinated for a sufficient number of generations to 
be uniform and fixed in their gross morphological characters. 


Connecticut Experiment Station 

Bulletin 376 

Single Crosses 

The greater uniformity of single crosses in comparison with the original 
varieties is well known. Statistical evidence for this is given in Station 
Bulletin 207. Arnold and Jenkins (1932) have compared the variability 
of single crosses with varieties, top crosses and double crosses in several 
measureable plant and ear characters. The mean coefficient of variability 
for all characters is 9.2 for single crosses, 11.2 for double crosses, 11.3 
for top crosses and 13.4 for naturally pollinated varieties. The genetic 
diversity of different crosses of inbred strains from the same original source 
is shown in Table 9. 

Figure 108. .Mature plants of Whipcross C3.7 topped and husks stripped 


Twenty inbreds of Canada Yellow Flint, selfed four years or more, 
were crossed in many different combinations and the hybrids grown with 
adjoining check rows of the original variety. Individual plot yields are 
placed in a frequency distribution in the accompanying table. It will be 
seen that the spread in yield is greater for the crosses and the variability 
of yield is also larger. This merely emphasizes the well known fact that 
some combinations are better producers than the original variety while 
others are poorer. 

Table 9. The Distribution of Yield in Bushels per Acre of a Series of 

First Generation Hybrids Compared to the Variety from 

Which the Inbreds Were Derived. 

Yield Coefficient of 

Classes 25 35 45 55 65 75 85 Total Average variability 

Variety 18 63 48 12 7 148 50.1 ± .8 19.4 ± .8 

Single Crosses 2 4 21 32 12 7 2 80 54.6 ± .9 21.0 ± 1.2 

MctJwd of Testing Inhrcds 679 

The similarity of the plants within a single cross in all measurable 
characters is also an indication of their uniform reaction to soil and sea- 
sonal conditions. Some combinations do well one year and poorly another. 
For the same reason, crosses that are productive in certain soils in a 
given latitude and under special weather conditions may be decidedly 
unproductive when one or more of these environmental factors are changed. 
Some combinations have the ability to do well in a wide range of seasons, 
soils and latitudes. A notable example of this is Golden Cross Bantam 
sweet corn. Some of the Illinois and Iowa hybrid field corns have done 
well in New Jersey and Connecticut. In certain years Connecticut field . 
corn hybrids have been productive in Ohio, but they have never shown 
any promise in the central corn-growing states. Sweet corn single crosses, 
such as Redgreen and AVhipcross, have yielded well in western Nevada 
and Idaho, in western Oregon and in western and central Wa^shington 
where the temperatures during the growing season are more nearly like 
those of southern New England. 

The adaptation of corn to climate has been discussed by Jones and 
Huntington, (1935). The authors propose for consideration the general 
rule that corn may be moved from a less favorable to a more favorable 
corn-growing region without loss of productiveness, and with a possible 
gain, provided the climate permits proper maturity. The average yields 
over a long period of years give some measure of the conditions necessary 
for the culture of corn. Very few sections are more favorable than others 
in all respects. Those parts of the country that are not too hot and dry 
during the period of maximum growth are usually too cool and wet at 
the beginning and end of the growing season. 

Genetically diverse plants respond differently to the various factors of 
soil and climate. For that reason single crosses are expected to be less 
adaptable than the varieties from which they are derived or from other 
types of crosses that permit greater genetic diversity. Strains that are 
not reduced in size or yield to complete homozygosity produce , crosses 
that are more adaptable. A reasonable degree of fixity of type in the 
characters that determine yield, and at the same time some variation in 
other characters are desirable from the standpoint of adaptability. Davis 
(1934) made crosses between inbreds that were selfed one, two, and three 
generations. The successive crosses gave approximately equal yields. There 
is the danger that inbreds which are not uniform and fixed in their type 
will vary from year to year and may lose their good qualities. With the 
object of yield alone in view, however, there seems to be little advantage 
in long continued inbreeding. 

Reciprocal Crosses 

Those who have tested many crosses know that reciprocal crosses of 
uniform inbred strains are closely similar in structural detail and in the 
time of flowering and maturing. Ashby (1932) found a difference in 
reciprocal crosses in the total amount of growth up to 40 days after 
planting. The inbreds differed in embryo size and the dry weight of the 
reciprocal crosses varied according to the initial weight of the embryo from 


Connecticut Experiment Station 

Bulletin 376 

which the plants started. When the plants are grown to full maturity 
these early variations usually disappear. St. John (1934) demonstrated 
that there is a difference in reciprocal top crosses. 

Two inbred strains of Learning corn, numbers 237 and 243, each self- 
fertilized for more than 30 generations, differ significantly in total yield 
of mature grain. Over a period of years, 243 is about twice as productive 
as 237. The plants and seeds are larger and the ears are better filled. The 
embryos are more nearly alike but here also 243 is distinctly larger. No 
determinations of embryo weights have been made. 

Figure 109. Whipcross C2.12. 

Reciprocal crosses of these inbreds were made and grown in nine 
alternating rows. The average yields in bushels per acre were as follows: 
243 X 237 = 69.0 ± .11 ; and 237 X 243 = 67.4 ± 1.62. The dift'erence 
in yield is less than the probable error of the difference. The yield of 
grain sums up the plant's ability to grow better than any other one meas- 
ureable character. 

Similarly two inbred Whipple sweet corn strains, numbers 2 and 7, 
are quite different in plant growth and seed size. Reciprocal crosses com- 
pare as follows : 

Method of Testing Inbreds 681 

Days to Ear Ear Av. Weight ^"ield Thousand ears 

Average Length Row Mktbl. Lbs. per per acre 

Tassel Inches Number Ears Plot Not Mktble. Mktble. 

Whipcross C2.7 56 7.5 14-16 .34 6.7 3 10 

Whipcross C7.2 57 7.5 12-16 .31 7.2 3 12 

These figures are based on single plots of 25 plants each, grown under 
similar conditions. They show that reciprocal crosses of inbred strains 
are sufficiently alike that they may be considered the same for practical 

Although these crosses give similar results, usually the hybrid is 
made more conveniently one way than the other. Inbred strains differ 
in productiveness, in size and quality of seed and in time of flowering. 
Naturally the inbred that produces the larger yield and the better quality 
of seed is used for the seed parent, whenever possible. If practicable, the 
seed and pollen parents should silk and tassel at the same time, when 
planted on the same day. This is not always feasible but it is highly 

Double Crosses 

Double crosses and three-way crosses have considerable advantage from 
the seed production standpoint. With double crosses, both the seed parent 
and pollen parent are' vigorous and productive. Only one row of the 
pollen parent need be planted to four or more rows of the seed parent. 
With three-way crosses, the seed parent is productive provided the inbred 
pollen parent supplies sufficient pollen. Combinations of this kind are 
genetically more variable than single crosses and for that reason some- 
what more adaptable and less susceptible to injury by adverse growing 
conditions at critical periods. The maintenance of three or four inbred 
lines is a serious problem. If the single crosses are in commercial pro- 
duction, then the production of other types of crosses with them is a 
simple matter. 

Advanced Generation Crosses 

Production of seed would be simplified if the second or later genera- 
tions of a single cross could be used. Theoretically a random sample of 
F2 plants has the same proportion and composition of gametes as Fi 
plants, provided there is no selective elimination. Kiesselbach (1930) 
reported the results of advanced generation crosses to be approximately 
the same as crosses of first generation hybrids. The Burr-Leaming hybrid, 
described in the Storrs Connecticut Station Extension Bulletin 108, has 
been produced as a double cross of first and second generation hybrids. 
The yields for two years compare as follows : 









(21X20)F,X (243X237)Fi 
(21X20)F,X (243 X 237) F, 

The difference in average yield is not significant. In appearance, time 
of maturity and quality of grain, the two lots appeared to be the same. 
The seed produced on F2 plants was more variable and less attractive in 


Connecticut Experiment Station 

Bulletin 376 

Multiple Crosses 

Instead ^f using first or later generation crosses for the production of 
hybrid seed, it is possible to combine a number of inbreds to form a new 
synthetic variety, and to cross two of these. This has been done in the 
case of the Burr Leaming hybrid. Eight inbreds out of Burr White 
were combined in pairs by hand pollination. The Fi seed was mixed and 
planted in an isolated field and allowed to interpollinate. The same thing 
was done with eight Leaming inbreds. Both lots have been continued 
by natural pollination in isolated fields for several years with selection 
for certain ear characters in each lot. 

Figure 110. Increase field of Whipple inbred C2. 

In 1933 the Multiple Burr Leaming, produced by crossing these two 

composites, was compared with the original double cross made both ways. 

The results in bushels of grain per acre were as follows : 

Multiple Burr Leaming 61 ± 1.5 

Double Crossed Burr Leaming 64 ± 1.9 

Reciprocal Double Cross 67 ± 1.2 

Although the original double cross made both ways yielded more than 
the cross of the composite, the difi^erences are not significant. 

It might be expected that in later generations reduction in yield would 
be less for multiple than for double crosses. There is always a tendency 
for the corn grower to save his own seed from high yielding hybrids. In 
many cases this would be desirable provided the sacrifice in yield were 
not too great. 

Method of Testing Inbreds 683 

A comparison was made between the first and second generations of 
the Canada Leaming hybrid. This is a first generation cross of two 
composites similar to Burr Leaming. One is a Canada Fhnt type, the 
other the same Leaming as used in Burr Leaming. The first generation 
yielded 59.5 ± .05, and the second 46.0 ± .92, a decrease of 22 per cent. 

Richey (1934) compared the yields of 10 first and second generation 
double crosses. The reductions ranged from 5 to 24 per cent. In single 
crosses this percentage is usually higher. The senior writer (1924) found 
a reduction of 32 per cent between the first and the second generations 
of two long-inbred Leaming strains. The crosses were grown over a 
period of six years. With every type of cross using inbred strains in both 
parents, either singly or in combination with others, there is an appre- 
ciable loss of yielding power after the first generation. This makes it 
advisable to use only freshly crossed seed each year. 

Variety-inbred Crosses 

Lindstrom (1931) called attention to the practical value of variety by 
inbred crosses which he calls top crosses, a term borrowed from animal 
breeding. In 1917 a number of crosses of local Connecticut varieties of 
flint and dent corn were made with an inbred Leaming strain No. 243. 
These were grown with a number of single crosses, about 60 in all. Ten 
of them yielded more than 100 bushels per acre. Eight of the ten were 
variety-inbred crosses. The combination of Canada Yellow Flint variety 
by Leaming inbred 243 gave the highest yield. The grain was well 
matured and uniform in its intermediate flint-dent type. Twenty-two 
variety-inbred crosses averaged 95 zh 1.4. Under the same conditions 
49 single crosses yielded 91 ± 1.3 (Jones 1922). This method of pro- 
ducing seed is now widely used with both field and sweet corn and gives 
excellent results, especially when distinctly difl^erent types of maize are 
brought together. 

The Crosby variety of sweet corn crossed by an Evergreen inbred, 
Connecticut 77, gives a productive and fast growing corn of good can- 
ning quality. Spancross C2*, Marcross C6, Whipcross P39 and Bancross 
P39, are other variety-inbred crosses using Spanish Gold, Golden Early 
Market, Whipple Yellow, and Golden Bantam, that have outstanding 
value either in early maturing, large size of ears, prodvictiveness or quality. 
From the standpoint of ease of producing seed and the adaptability of 
this seed as compared to that of single crosses, there is much to be said 
in favor of variety-inbred crosses. 

Single Plant Crosses 

In sweet corn for canning, single plant lines have been used for field 
production where the uniformity of kernel type is more important than 
gross yield. These single plant lines are the descendants by natural pol- 
lination of an individual plant, selected on the basis of the progeny per- 

*According to this system of naming hybrids, a syllable of the varietal name is combined with 
the word "cross". The letters and figures following indicate the source of the inbred (C= Con- 
necticut, P= Purdue) and its pedigree number. One figure shows that it is a variety-inbred cross, 
two figures a single cross, etc. 


Connecticut Experiment Station 

Bulletin 376 

formance test. Davis (1934) has shown that two generation selfed lines, 
crossed with open-pollinated varieties, yield as well as those crossed with 
lines that are inbred longer, although they are undoubtedly more variable 
and more subject to change. This suggests that single plant lines can be 
used for crossing with homozygous inbreds to give high yielding hybrids 
with more uniformity and fixity of type than a variety by inbred combina- 
tion. These single plant lines may be open-pollinated, selfed, or offed for 
one or two generations, depending upon the vigor and yielding ability 
required in the seed parent, and the uniformity and constancy desired in 

Figure 111. Uniformity in plant and ear characters and in time of maturity 
is characteristic of single crosses. 

the hybrid. A simple way to find a good combination of this kind is to 
self individual plants, save part of the pollen and apply this to an inbred 
selected as a tentative pollen parent. A series of such crosses would 
indicate the most desirable combinations and these could be tested more 
thoroughly by using the mixed progeny of the once inbred line as the 
seed parent. This method is now being tried with sweet corn. 

Crosses of two, open-pollinated, single plant lines may have value and 
should be tried. The earlier work with ear-row selections, in which part 
of the rows were detasseled and crossed with the other lines, indicated 

Method of Producing Inbred Strains 685 

that individual plants had outstanding productiveness. If there were any 
way of maintaining this yielding ability from year to year without the 
expense and delay of fixing it by inbreeding, it would be a real saving. 


Up to the present time inbred strains have been produced chiefly by 
self-fertilization. It is theoretically possible (Haldane, 1931) to obtain 
a more thorough recombination of desirable genetic factors by other forms 
of inbreeding where homozygosity is more slowly approached. Brother 
and sister mating, or sib-crossing as applied to plants, ofifers a more 
convenient method of inbreeding than self-fertilization, from the hand- 
pollination standpoint. It may also have a distinct advantage in per- 
mitting freer recombination of genetic factors. 

This method is being applied to early sweet corn as described in 
Bulletin 361. Individual plants were self -fertilized the first year and from 
each of the ears harvested six seeds were taken and wrapped in tissue 
paper. Each lot was planted singly in hills, and thinned to three stalks 
in each. Two plants in each hill were sib-pollinated, using pollen from 
each other or from the third plant, whichever tasseled at the right time. 
At harvest the two hand-pollinated ears were examined and one was 
discarded in the field. In this way each ear represented an inbred line. 
No labeling and staking were needed at any time. This method has 
produced some early maturing inbred lines of sweet corn with remarkable 
yielding ability when crossed. 

If bulk seed is used, the first generation plants should be selfed, since 
sib-mating the first time would be about the same as natural intercrossing, 
If individual open-pollinated ears are used at the start, they may be 
sib-pollinated from the beginning. After three or more generations of 
such o£f-pollination, the surviving lines may be grown in rows having 
a sufficient number of plants to give an estimate of their uniformity and 
type, and provide ample plants for crossing. 

Previous studies emphasize the importance of having the largest possible 
number of lines to select from, after some degree of uniformity and con- 
stancy has been obtained. Also there is little value in selecting during 
the inbreeding process unless for particular characters which have major 
importance. Taking the most vigorous and productive plants for pro- 
genitors in each generation usually delays the reduction to uniformity 
and constancy. By growing only a few plants each year, little selection 
is possible. Therefore the resulting inbred strains represent more nearly 
a random sample of the genetic recombinations possible and this is desired. 
The importance of selection comes after some degree of uniformity and 
fixity of type has been obtained from among the available inbred lines. 
If the desired qualities are not all recombined, then the method of con- 
vergent improvement can be used to good advantage. 

686 Connecticut Experiment Station Bulletin 376 


The following experiment was devised to determine the effect of a 
noticeably harmful recessive gene, when used in crosses. A Leaming 
line in the second generation was noted as segregating for golden plants. 
Several of these were self -fertilized and the progeny grown the following 
year. In each generation a progeny segregating for golden plants was 
selected for continuing the line, always through one normal green plant 
carrying the recessive golden gene. In the seventh generation, both green 
and golden plants were self-fertilized, and the following year, homozygous 
golden and green progenies were crossed with a third uniform inbred 
strain of the same variety. 

Having been self-pollinated for seven generations, it was assumed that 
the green and golden plants would have practically the same composition 
except for this one chlorophyll determiner. In appearance the two types 
were closely alike except that the golden plants were smaller and less 
vigorous due to poor nutrition. 

The two crosses were compared in yield by growing nine alternating 
plots, five with seed from the cross of golden X green, and four with 
seed from the cross of homozygous green X green. Both lots of crossed 
plants were green, one Gg in composition, the other GG. They were 
identical in appearance. The average yields of grain were 56 ± 1.2 bushels 
from the cross with the golden parent and 59 ± 1.2 from the cross with 
the green parent. The difference in yield of 3 ± 1.7 in favor of the 
homozygous green plants is not significant. 

Mangelsdorf found homozygous dominant plants slightly taller than 
heterozygous plants in a progeny segregating for a lethal defective seed 
character. In this case no fair comparison in yield of grain could be 
made on account of the lethal factor reducing the number of normal seeds 
on the heterozygous plants. 

Even if a significant increase in yield were obtained, one could not 
be sure that the difference was due to the specific gene being studied. 
Adjacent loci could also be in the heterozygous condition. From these 
results it may be concluded that some deleterious recessive genes ma}'' 
be present in a heterozygous condition without any serious effect upon 
the ability of the hybrid to grow and to produce. Furthermore, it indicates 
that the removal of abnormal genes of this type is not the most important 
function of inbreeding. 


Various details essential to the production of crossed corn seed were 
discussed in Bulletin 361 and need not be reviewed here. The statement 
often made, that seed corn should not be planted more than 100 miles 
from its place of production, has been shown to be a too sweeping gen- 
eralization and not in line with the facts (Jones and Huntington, 1935). 
There seems to be a distinct advantage in producing varietal seed in a 
climatic region less favorable to growth than that in which the crop is 
to be planted. However this may be, it should be realized that the pro- 
duction of naturally pollinated varietal seed is distinctly dift'erent from 

Seed Production 687 

the production of crossed corn seed from the standpoint of adaptation 
and dispersal. 

Seed of double crossed Burr-Learning has been produced in Virginia, 
Maryland, Ohio and Spain and grown in Connecticut in comparison with 
local seed. The single crosses used in producing this combination were 
all raised in Connecticut ; the seed of each parental single cross was mixed 
in one lot and some was sent to each of the above mentioned places. The 
yields of grain are given in Table 10. In most cases the yields for each 
year represent the average of five plots. The five-year average (1925 to 
1930) gives 61 ± 2.0 bushels per acre for Connecticut-, and 67 dz 2.8 for 
Virginia-grown seed. The difference of 6 ± 3.4 bushels indicates that 
there is no distinction in results wherever seed is produced for use in 
Connecticut. The time of maturity and quality of grain are apparently 

Table 10. The Yield of Double Crossed Burr-Leaming in Bushels per 

Acre, When Grown in Connecticut from Seed Produced 

IN Various Places. 

Place where seed was produced 
































*Not included in average. 

Likewise, seed of Redgreen sweet corn, a single cross, has been pro- 
duced in Indiana and Connecticut and the first generation crossed plants 
grown in both places. The average yield in pounds per plot for the two 
lots of seed grown in Connecticut three years was 9.4 ± .01 for native, 
and 10.3 ± .07 for Indiana seed, a significant difiference of .9 ±: .07 in 
favor of the Indiana-grown seed. The results are given in Table 11. 
In all yeafs and in both places the Indiana-grown seed yielded more and 
the plants grew taller than those from Connecticut-grown seed, although 
this hybrid originated in Connecticut from types that had long been grown 
there. It is a type that suffers severely from low humidity and high tem- 
peratures and is usually unproductive in the Central States. 

Table 11. The Performance of Redgreen Sweet Corn When Grown in 
Connecticut and in Indiana from Seed Produced in Both Regions. 

Yield in pounds of dry ears per plot and height of plant in inches 
Place where seed 1925 1926 1927 

was produced Yield Yield Yield Height 

Conn. Conn. Conn. Ind. Conn. Ind. 

Connecticut 9.9 9.0 9.3 8.3 88 75 

Indiana 10.4 9.7 10.8 10.7 89 83 

The change from Connecticut to Indiana is not extreme, although 
Indiana produces 10 bushels less grain corn per acre according to the 20- 
year average. To have the seed produced under as widely different climatic 

688 Connecticut Experiment Station Bulletin 376 

conditions as possible, inbred seed from the same self-pollinated ears was 
planted in Texas and in Connecticut and the cross-pollination made by 
hand. In Texas the two inbreds grew less than two feet high and pro- 
duced only a few seeds. This was largely the result of moving corn to 
a shorter growing day. In Connecticut both inbreds normally grow from 
five to six feet or more in height and produce well filled ears. The crossed 
seed from both places was grown in adjacent rows in Connecticut the 
following year. There was only enough seed of the Texas crop to plant 
one row. The weight of dry ears from the two plots was 4.8 pounds 
from the Texas-grown seed, and 3.7 pounds from the Connecticut-grown 
seed. The plants tasseled and silked on the same days and were identical 
in appearance. The average height of plants from Texas seed was 77 ± .6 
and for the Connecticut seed, 77 ± .9 inches. 

Another cross of two long inbred dent strains, 14—4 and 112-1, was 
also produced in Texas and in this State and both lots were grown here. 
There was only enough seed from Texas to grow one row in Connecticut. 
The ears were so badly eaten by birds that no yields could be taken. In 
height of plant, the averages were 80 ± 1.7 for Texas- and 84 ± .7 for 
Connecticut-grown seed. Here the difference of 4 ± 1.8 is not significant. 

All of these results indicate clearly that either there is no difference in 
the seed of hybrids grown away from the place of origin or, if there is 
any significant difference, it is in favor of seed grown in the new locality. 
In all of these cases the new region is less favorable for the production 
of corn, as shown by the 20-year average production of grain corn per 
acre. Tests should be made to see whether hybrids originating in other 
districts yield less in their native home when seed is produced in Con- 
necticut or other favorable corn growing regions. 

There are many economic reasons for producing seed in those regions 
where a good crop can be obtained every year. High yields reduce the 
cost of production. Well developed and matured seeds insure strong 
germination. When seeds of naturally pollinated varieties are raised in 
such regions, they are likely to lose productive capacity when planted 
under less favorable conditions. If this is also true, even in a less degree, 
for crossed corn, it should be known and given careful consideration. 
Crosses of inbred strains are uniform and fixed in their type and do not 
become adapted to the region in which the seed is grown as quickly as 
do open-pollinated varieties. The reasons for this are obvious. 

The evidence at hand indicates that it is possible to produce crossed 
seed wherever it can be grown to the best advantage, taking into con- 
sideration cost and quality of the matured seed. This is especially true 
of single crosses and probably also of other types of crosses. To insure 
proper maturity, quality, and other essential details, the foundation inbred 
stock seed should be grown where it is originally produced. When this 
is increased, the crosses may be made in a region most favorable to seed 
production. Seed that is one or a few generations removed from this 
foundation seed will probably be all right in most cases. Continued pro- 
duction in a different region however, will in time tend to change the 
genetic constitution of the inbred stock by natural selection, and the 
rapidity of this change will depend upon the amount of germinal varia- 
tion within it. 

Summary 689 


Briefly stated, the results of Part I of this series of experiments, as 
summarized in Station Bulletin 266, were as follows: Selection during 
the inbreeding process of such specific characters as plant height, height 
of the ear from the ground, number of tillers, and resistance to disease, 
produced inbred strains that transmitted these characters to the first gener- 
ation hybrid. However, it was impossible to foretell before crossing, with 
any high degree of certainty, which inbred strain would give the largest 
grain yield when crossed. In other words, selection for the dominant 
growth genes responsible for heterosis was of little or no value. Selection 
must be based upon the progeny of the inbreds when crossed. 

The findings of Part II (this bulletin) corroborate in general the pre- 
vious conclusions. Several types of corn were used in the demonstration 
with the following results : 

1. In two sets of experiments with both Burwell's Yellow Flint and 
Leaming Dent, the inbreds were classified, before crossing, as good or 
poor in one case, and as productive, intermediate and unproductive in 
another case. In no case were the highest average yields obtained from 
strains that had been classified as unproductive. The opposite also held 
true, although once in the dent and twice in the flint, the highest or lowest 
average yield was obtained from a strain in the intermediate class. In 
different matings of good and poor inbreds, the good X good produced 
the largest yield ; the good X poor, intermediate ; and the poor X poor, 
the lowest. The differences, however, were not significant. 

2. In double crosses combining from none to three good strains, there 
was a barely significant increase in the crosses having three good strains 
over those having none. The double cross combining four good strains 
was not tested. 

3. Specific characters in the inbred parents that appeared in the F^ 
hybrid were : Short ear, long slender ear, cylindrical ear and tapering ear. 

4. In a series of white flint corn inbreds, all crossed by a multiple 
Leaming variety, selection of the best inbreds was of little avail. By 
choosing one-third of the best inbreds before crossing, we were able to 
obtain one-half of the most productive crosses. It is doubtful if any normal 
lines should be eliminated before crossing. 

5. In Evergreen sweet corn the most productive F^ hybrid had a serious 
root weakness, contributed largely by the C-63 parent. In an effort to 
correct this, remnant seed from three progenies in the second generation 
of lines 50 and 63 were grown, and 61 new lines were started. These 
were selfed for three generations, always using erect plants. The new 
50 lines were then crossed with the new 63 lines and the F^ hybrids 
tested. No hybrids were found that were materially better than the original 
cross. This emphasizes the futility of carrying on many lines after the 
plants have been selfed for two generations. It probably would have been 
better to start a great many lines and to carry on only one progeny from 
each originally selfed plant. 

6. In a series of Whipple sweet corn inbreds, 106 lines were started 
and 74 remained after three generations of self pollination. Of these, 23 

690 Connecticut Experiment Station Bulletin 376 

were noted as promising on the basis either of stalk growth or of ear 
characters. With one exception, all of the best F^ hybrids resulted from 
crossing these 23 inbreds in various combinations. As the experiment 
turned out in this one case, the rest of the inbreds could have been profit- 
ably discarded. However, these results held for this lot only and cannot 
be relied upon as an effective method of procedure. 

7. In the Whipple series, positive correlations of the inbreds and the 
F^ hybrids were found between number and weight of ears, and number 
of tillers per plant. There was also a positive correlation between number 
of tillers and yield of marketable ears. Selection of inbreds with many 
tillers is advisable provided the tillers do not produce ears. 

8. Inbreds can be tested by crossing them by other inbreds (single 
cross), by crossing an F^ hybrid X an inbred (three way cross), by cross- 
ing two Fi hybrids together (double cross), by crossing later generation 
hybrids together (advanced generation cross), by crossing an inbred by 
a variety (top-cross), or by crossing two synthetic varieties together 
(multiple cross). Doubtless other combinations could be devised. One 
of the greatest problems confronting the maize breeders today is how to 
test inbreds most advantageously after good inbreds have been secured. 
Each of the testing methods is discussed in this bulletin. 

9. A larger number of inbred lines can be produced with minimum 
effort by growing only one hill of each line. This method is described in 
the bulletin. It has proved advantageous in isolating good inbred lines 
of Spanish Gold, an extra early sweet corn. 

10. First generation hybrids heterozygous for Gg (the golden plant 
character) were grown in comparison with similar hybrids of GG com- 
position. There was no diff'erence in yield. 

11. Crossed seed of the same inbred parents was produced under widely 
different conditions of latitude and climate. No dift'erences in yield of 
the F-^ hybrids were obtained when all lots were grown in Connecticut. 


Arnold, L. E., and Jenkins, M. T. The relative variability of corn crosses and 
varieties. Jour. Agron., 24 : 868-871. 1932. 

AsHBY, E. Studies in the inheritance of physiological characters. II. Further ex- 
periments upon the basis of hybrid vigor and upon the inheritance of efficiency 
index and respiration rate in maize. Annals. Bot. 46: 1007-1032. 1932. 

Davis, R. L. Maize crossing values in second-generation lines. Jour. Agr. Res. 
48 : 339-357. 1934. 

Garber, R. J., and North, H. F. A. The relative yield of a first generation cross 
between two varieties of corn before and after selection. Jour. Agron. 23 : 647- 
651. 1931. 

Garber, R. J., and Quisenberry, K. S. Breeding corn for resistance to smut. 
(Ustilago Zeae). Jour. Agron. 17:132-140. 1925. 

Haldane, J. B. S., and Waddington, C. H. Inbreeding and linkage. Genetics, 16 : 
357-374. 1931. 

Hayes, H. K. Present-day problems of corn breeding. Jour. Agron. 18:344-363. 

Literature Cited 691 

Hayes, H. K., Brewbaker, H. E., and Immer. F. R. Double-crossed corn in Minne- 
sota. Minn. Agr. Expt. Sta. Bui. 260. 1930. 

Hayes, H. K., and McClelland, C. K. Lodging in selfed lines of maize and in 
Fi crosses. Jour. Agron. 20:1314-1317. 1928. 

Jenkins, ' M. T. Correlation studies with inbred and crossbred strains of maize. 
Jour. Agr. Res., 39:677-721. 1929. 

'- Differential resistance of inbred and crossbred strains of corn to 

drought and heat injury. Jour Agron. 24:504-506. 1932. 

Methods of estimating the performance of double-crosses in corn. 

Jour. Agron. 26 : 199-204. 1934. 
Jenkins, M. T., and Brunson, A. M. Methods of testing inbred lines of maize in 

crossbred combinations. Jour Agron. 24 : 523-530. 1932. 
Jones, D. F. The productiveness of single and double first generation corn hybrids. 

Jour. Agron. 14:242-252. 1922. 
Jones, D. F., Singleton, W. R., and Curtis, L. C. The correlation between tiller- 
ing and productiveness in sweet corn crosses. Jour. Agron. 27:138-141. 1935. 
Jorgenson, L., and Brewbaker, H. E. A comparison of selfed lines of corn and 

first generation Crosses between them. Jour. Agron. 19:819-830. 1927. 
Kiesselbach, T. a. Corn investigations. Neb. Agr. Expt. Sta. Res. Bull. 20. 1922. 
Kiesselbach, T. a. The use of advanced-generation hybrids as parents of double 

cross seed corn. Jour. Agron. 22 : 614-626. 1930. 
LiNDSTROM, E. W. Prepotency of inbred sires on commercial varieties of maize. 

Jour. Agron. 23:652-661. 1931. 
RiCHEY, F. D. Effects of selection on the yield of a cross between varieties of corn. 

U. S. Dept. Agr., Dept. Bui. 1209. 1924. 

Corn breeding. U. S. Dept. Agr., Dept. Bui. 1489. 1927. 

The convergent improvement of selfed lines of corn. Amer. Nat. 

61 : 430^49. 1927. 

RiCHEY, F. D., and Mayer, L. S. The productiveness of successive generations of 

self-fertilized lines of corn and of crosses between them. U. S. Dept. Agr., 

Dept. Bui. 1354. 1925. 
RiCHEY, F. D., and Sprague, G. F. Experiments on hybrid vigor and convergent 

improvement in corn. U. S. Dept. Agr., Tech Bui. 267. 1931. 
RiCHEY, F. D., Stringfield, G. H., and Sprague, G. F. The loss in yield that may 

be expected from planting second generation double-crossed seed corn. Jour. 

Agron. 26: 196-199. 1934. 
Smith, S. N. Response of inbred lines and crosses in maize to variations of nitrogen 

and phosphorus supplied as nutrients. Jour. Agron. 26 : 785-804. 1934. 
Sprague, H. B., and Curtis, N. Chlorophyll content as an index of the productive 

capacity of selfed lines of com and their hybrids. Jour. Agron. 25 : 709-724. 

St. John, R. R. A comparison of reciprocal top crosses in corn. Jour. Agron. 26: 

721-724. 1934. 
Stringfield, G. H., and Salter, R. M. Differential response of corn varieties to 

fertility levels and to seasons. Jour. Agr. Res. 49:991-1000. 1934. 

533 9 •■0 9 

University of