B. P. I.—468. 
THE IMPORTANCE OF BROAD BREEDING 
IN CORN. 
By G. N. Cotiins, Assistant Botanist. 
INTRODUCTION. 
A study of the primitive types of corn and the history of the more 
productive varieties affords many indications that the improvement 
of our varieties is being checked by the injurious effects of inbreed- 
ing and that a further advance can best be secured by more fully 
satisfying the natural requirement of cross-fertilization. The con- 
spicuous increase in yield that results from the crossing of distinct 
varieties and the rapid deterioration that follows self-pollination 
strongly support this view. Unfortunately the more recent changes in 
the methods of breeders have not been in the direction of better pro- 
vision for cross-fertilization, but toward a still closer approach to 
self-pollination, the attempt being made to apply to corn theories and 
methods derived from plants naturally adapted for self-fertilization. 
Until recently, the study of evolution and heredity has had little 
effect on the methods employed by breeders of domesticated plants 
and animals. Improved varieties were developed before Darwin’s 
time in much the same way that they have been since. Whether the 
doctrines of Darwin or those of Lamarck, Naegeli, Weismann, or 
Cope should be accepted as giving correct interpretations has seemed 
to be a purely academic question from the standpoint of the practical 
breeder. 
It is entirely different with the recently elaborated theory of De 
Vries. Converts to this new hypothesis have not been slow to claim 
that it has an important practical bearing, especially when taken in 
connection with the facts of Mendelism. Serious changes in the 
methods of breeding are being urged that are likely to have very 
injurious effects if they are generally applied to our varieties of 
Indian corn. 
The definite mechanical character of these theories and the facility 
with which they appear to explain some of the facts of heredity 
render them very convincing. There is a tendency to forget their 
limitations and overlook the fact that whatever their value with 
141—1y ; ahs) 
