APRIL 19, 1912] 
peared wiser to discuss more in detail the 
problems of selection and variation. I 
ean not, however, close this address with- 
out referring to this very important field 
of genetics. 
No discovery in the field of breeding has 
had more effect or is more far reaching in 
its importance than the discovery of what 
have now come to be knownas Mendel’s prin- 
ciples of heredity. While, as stated in the 
beginning of this address, breeders had long 
before the rediscovery of Mendel’s papers 
come to understand that there was a segre- 
gation of characters in the F, generation 
and that it was possible to recombine in 
certain hybrids the desired characters 
from different parents, there was no defi- 
nite understanding of the underlying 
principles, and no conception of the almost 
infinite possibilities of improvement which 
the field of hybridization opened to us. 
The law of dominance, while not univer- 
sal, has explained many eases of prepo- 
tency in one generation and failure of cer- 
tain individuals to transmit the character 
in the next generation. It has explained 
many cases of latency of characters and 
may account for all such cases. 
The law of segregation has shown us 
that the splitting of characters follows a 
definite method and that we can in general 
estimate the frequency of occurrence of a 
certain desired combination, if we know 
the characters concerned to be simple unit 
characters. 
The study of hybrids has been resolved 
into a study of unit characters and their 
relation to each other. By hybridizing re- 
lated types having opposed characters and 
observing the segregations which occur in 
the later generations, we analyze the char- 
acters of each type and determine when we 
have a character pair. The researches on 
this subject by Mendel, Bateson, Daven- 
port, Castle, Punnett, Shull, Hurst, Cor- 
SCIENCE 
607 
rens, T'schermak, Hast and dozens of other 
now well-known investigations, have de- 
veloped a science of heredity of which we 
had no conception a few years ago. 
We can now study the characters pre- 
sented by the different varieties of a plant 
or of different species, which can be crossed 
with it and definitely plan the combination 
of characters desired in an ideal type, and 
can with considerable confidence estimate 
the number of plants it will be necessary 
to grow to get this combination. We now 
know in general how characters behave in 
segregation and inheritance so that we can 
go about the fixation of a desired type, 
when one is secured, in an orderly and in- 
telligent way. 
The farther the study of characters is 
carried the more we are coming to realize 
that the appearance of apparently new 
types following hybridization is due to re- 
combinations of different units which in 
their reactions give apparently new char- 
acters. As an illustration, in a study of 
pepper hybrids which I have carried on 
during the past four years it has become 
evident that the form of plant and branch- 
ing is due to three pairs of characters or 
allelomorphs; namely, first, erect or hori- 
zontal branches; second, large or small 
branches; and third, many or few branches. 
In crossing two medium-sized races, one 
with large, horizontal and few branches, 
and the other with small, erect and nu- 
merous branches, there result many new 
combinations of characters, among which 
appear some with small, horizontal and few 
branches, which gives a dwarf plant, and 
others will have a combination of large, 
erect and numerous branches, which gives a 
giant plant. These dwarfs on the one hand 
and giants on the other, appear as distinct, 
new creations, though they are very evi- 
dently merely the recombinations of al- 
ready existing unit characters, and dwarf- 
