1909. 
<t71 
REAL BASIS FOR IMPROVEMENT IN CORN. 
A Study in Plant Breeding. 
Part 1. 
There appears to be a growing tendency in the 
United States to exploit plant breeding. No one 
will object to the tendency—for the subject is fas¬ 
cinating—nor will one find fault with the exploiters, 
provided they do not wander into the realms of fic¬ 
tion. It is, unfortunately, a temptation not always 
overcome, for those who are set in authority to 
teach agricultural truths, to exaggerate the benefits 
that are expected to accrue from the lines of work 
in which they are most interested. A portion of 
these exaggerations is due to laudable and honest 
enthusiasm which every man worthy of the name 
has for his own specialty. This interest may be of 
the kind that is held by the beauty of the butterfly’s 
wing and uninfluenced by a sunset, but in spite of 
its narrowness it rings true. Another portion of 
the overstatements may arise from a less commend¬ 
able desire to be one’s own press agent, albeit even 
in this case popular demand for results punctuated 
by the dollar sign, is greatly to blame. Neither of 
these causes, however, is the equal in the pernicious 
effect to that produced by the great desire of the 
lay reader to obtain startling pieces of news'. He 
would much rather read a fanciful article about a 
wheat yielding one hundred bushels to the acre, or 
a white blackberry, than an actual result of real 
worth. And when the popular taste is for such 
items, they can always be furnished in quantity 
unlimited and in quality limited only by the imagi¬ 
nation of the writer. These articles arc generally 
entrusted to the pens of persons not intimately con¬ 
nected with the work described, and who have no 
particular! reason for dealing conservatively with 
the subject. There is no 
reason why the tales of 
Baron M u ncha u s e n 
should be banished 
from a library, for the 
fictions are obvious, but 
the tales of his imita¬ 
tors upon plant-breeding 
subjects have caused the 
loss of many hard- 
earned dollars through 
t h e non-realization o f 
the alluring prospects of 
gain held out to the fol¬ 
lowers of intricate meth¬ 
ods of hybridization and 
selection. The writer 
most certainly does not 
wish to discourage an 
increase in the number 
of persons already inter¬ 
ested in the improve¬ 
ment of plants. It is a 
great field, a field where 
much has been accomplished, a field where much will 
be accomplished. But stories of roseate hue, though 
they arouse a feverish interest, will only end in dis¬ 
appointment. We must limit our dreams and get 
down to hard facts. 
Let us see just what the biological truths are, 
upon which are based the methods of corn improve¬ 
ment; then we shall be able to judge what returns 
we can reasonably expect for our work. The main 
facts will apply to other breeding work as well, al¬ 
though of course there are variations in the detailed 
application of methods. 
Ever since plant breeding has been brought to the 
front as a means of increasing yields and producing 
new qualities', distinct from the mere feeding of 
crops by means of fertilizers, we have been taught 
that by selection we could accomplish anything. Take 
the character length of ear as a concrete example. 
Variations in length are seen in every field of corn, 
no matter what the variety. If we select the longest 
ears in successive seasons we are taught that we 
shall continually improve the strain in this character. 
Even the teachers have always had an inward con¬ 
sciousness that there must be a limit to the progress 
that could be made in this way, but the thought has 
been indefinite, and has been designedly kept in 
abeyance. The man who has been so indiscreet as 
to ask where the limit of his improvement is to be 
has been judiciously steered into othw lines of 
thought. Some questions' are hard to answer, and 
it was more interesting to let him dream of the time 
he could fill an order for a bushel of corn with one 
lone ear. 
Likewise, we have been taught that by hybridizing 
two strains we could obtain any desired new char¬ 
acter or quality if we only continued on the job for 
a sufficient length of time. We have only made a 
beginning in breeding as a science, and it will be 
TTirTfiJ Bid RAL NEW-YORKER 
many years before we can predict with the accuracy 
of a chemical reaction what will take place on cross¬ 
ing two individuals; but even now we know enough 
of the laws of variation and heredity to show us 
that this point of view is all wrong. The quality of 
the grist depends upon the grain in the hopper. 
There are three and only three ways in which plants' 
can be improved, and a careful consideration of 
them shows the errors in both of the above teachings. 
1. Every plant is composed of a large number of 
characters, each of which is inherited as a unit. 
The basis and the plan by which these characters 
develop are held in the fertilized egg from which 
the seed and finally the plant results. From the 
structure within this egg-cell, each character and 
through these characters, the entire organism is 
finally self-constructed by the utilization of food 
materials from the soil and from the air. No per¬ 
manent variation occurs in a plant unless it occurs 
first in the structure of these reproductive cells. 
Such variations' arc, therefore, the only foundation 
for plant improvement, and the sole function of 
selection is to pick out the most desirable of them 
for propagation. Unfortunately for the ease of the 
task, variations due to nutrition and other causes, 
variations that affect the plant and not the reproduc¬ 
tive cells' and are, therefore, not inherited, often 
obscure the rarer, inherited variations, and thereby 
cause much unnecessary and unavailing selective 
work. 
2. The object of hybridization is to shuffle and 
recombine the unit charaeters of the parent plants. 
If we knew what all the results would be when these 
characters recombine wc could predict the percen¬ 
tage of the progeny of a cross that would contain 
certain/ definite combinations of characters. But 
wc do not know as yet enough concerning the inheri¬ 
tance of these characters to make predictions. We 
simply know that we can only expect to combine the 
characters actually possessed by the parents. For 
example, resistance to the disease called watermelon 
wilt was found to be a single unit character pos¬ 
sessed by the first cousin of the watermelon, the 
citron. This quality was therefore combined with 
other desirable qualities possessed by the watermelon, 
by hybridization. Here was definite basis upon 
which to work and by which tangible results' could 
be obtained. But supposing no wilt-resistant melon 
had been known, it would have been utterly futile 
to have tried to breed this character into the zvatcr- 
vielon by selection. The watermelon reproductive 
cell does not possess' this character, and on this ac¬ 
count there is no basis upon which to select. The 
kind of selectionist against whom we are speaking 
would begin by going into a field and taking seeds 
from those plants that were the least affected by 
the disease, yet he would never obtain results be¬ 
cause he would be selecting non-affected plants in¬ 
stead of resistant plants. 
3. There is a third method by which a slight and 
temporary improvement can probably be made, al¬ 
though some biologists are beginning to lose faith 
in it because the results scarcely pay for the trouble. 
This* is the selection of fluctuations. Fluctuations 
are the variations that are not due to structural 
changes in the reproductive cells, but simply to nutri¬ 
tion. They are, therefore—strictly speaking—not in¬ 
herited, but simply give temporary aid in the devel¬ 
opment of the next generation. For example, let 
us imagine two corn plants having exactly the same 
characters'; one of these plants has grown on good 
soil and is well developed, while the other has grown 
on poor soil and is weakly developed. The seed of 
the well-nurtured plant has more nutriment stored 
up in it, and the young seedling that it produces has 
a better start in life than has a seedling from the 
poorly-nurtured plant. The actual characters in¬ 
herited by the two plants are the same, but the seed¬ 
lings' from the poorly nurtured plant are handicapped, 
and are not so well able to utilize their food supply 
and produce a normal well-developed plant. This 
is also true of poorly-nourished animals. 
Let us apply these principles to corn breeding. 
Corn is a wind-pollinated plant, therefore when a 
change takes place in the reproductive cells' of any 
individual of a variety, it is quickly recombined with 
different characters in other individuals. As these 
changes take place with some frequency, what we 
call a commercial variety is actually a set of hybrids 
between individuals' possessing various characters. 
The real effect of selection is gradually to isolate 
a strain having characters that we desire, in so far 
as such characters have already been produced by 
nature. There is no question of our originating 
anything by this selection. If there are plants hav¬ 
ing undesirable characters we can reject them, pro¬ 
vided there are plants that are without these charac¬ 
ters, but is as far as we can go. In this mixture 
of types, the commercial variety, there are some 
strains that produce a greater yield than others. It 
is the aim of the line breeder to take out these types 
and discard those having less efficiency. For these 
reasons we can see how great is the importance of 
our original breeding plot. If the type desired has 
been included among the original plants it can be 
selected out and established as a variety, if it has 
been left out we can only wait for nature to pro¬ 
duce such a type. And as nature is not prodigal in 
her new productions our chances arc relatively small. 
If certain desired characters arc possessed by one 
strain, and other desirable characters are possessed 
by another strain, we then have recourse to hybridi¬ 
zation. In some of the 
progeny of a cross be¬ 
tween the two kinds, we 
will obtain our two cov- 
e t e d characters com¬ 
bined. 
As an actual example 
of the subject under dis¬ 
cussion let us again 
consider the length of 
an ear of corn. The 
Longfellow flint is a va¬ 
riety that has been un¬ 
der selection for at least 
one hundred years. 
There has been some 
improvement in length, 
for nature has made the 
necessary changes in 
the reproductive cells, 
and we now have a va¬ 
riety that averages 
about 12 inches in 
length. In the last 15 
or 20 generations, however, there has' been no 
further progress. Further natural changes may take 
place, but at present we can see no prospect of ob¬ 
taining a longer ear in the Longfellow variety. Dur¬ 
ing this time another variety—the Benton corn—has 
been produced, which in length is just what breeders 
have been trying to obtain in the Longfellow va¬ 
riety. It has an inherited ability that the Longfellow 
corn does not have—the ability to produce an ear 
averaging 15 inches in length. That which nature 
has withheld in the one case she has granted in the 
other. Selection of long ears might be continued 
in the Longfellow corn indefinitely without obtaining 
the average length of the Benton corn, yet if such 
a change in ability to produce length of ear should 
take place, it can be isolated immediately and that 
is the end of it. 
The foregoing argument shows precisely what the 
line breeder may expect. Methods with greater diver¬ 
sity of detail have been proposed for the improve¬ 
ment of corn, but the differences between them have 
to do merely with the rapidity with which results 
can be obtained, and do not affect the truth of what 
I have just said concerning the fundamental princi¬ 
ple upon which the idea of selection rests. Keen 
eyes and expert judgment combined with actual field 
tests have detected in our commercial varieties, types 
which are a great improvement over what were 
grown a generation ago. It may be that still better 
types will be detected, isolated and placed on the 
market, but in the meantime let us try to divest our 
minds of the wholly erroneous idea that merely by 
continuous selection we can make a continuous and 
unending improvement. Selections must, indeed, be 
continued, lest hybridization of these good strains 
with other poorer strains cause a deterioration in 
yield. edward m. east. 
(To be concluded next week.) 
IMPROVEMENT OF CORN BY BREEDING. Fig. 198. 
The right basket represents the average yield of two Inbred types of Learning. The left basket represents the yield of a cross between the two types. 
