608 
ness and giantness are the results of the 
reaction of the different units combined. 
When we remember the large number of 
distinct characters which are presented by 
the very numerous varieties of any of our 
cultivated plants, we arrive at an under- 
standing of the possibilities of improve- 
ment which the field of hybridization af- 
fords, yet I doubt if many of us have even 
then an adequate conception of the possi- 
bilities. Possibly I may make this more 
clear by an illustration from my timothy 
breeding experiments. While the various 
characters presented by the different types 
under observation have not been carefully 
studied in inheritance, the following char- 
acters can be distinguished plainly, and 
from observations on accidental hybrids are 
known to segregate. The following is a list 
of 28 such character pairs which it is be- 
lieved will prove to be allelomorphs. 
TIMOTHY CHARACTER PAIRS 
Heads 
Long or short. 
Thick or thin. 
Dense or lax. 
Greenish or purple when young. 
Gray or tawny when ripe. 
Simple or branched. 
Erect or nodding. 
Continuous or interrupted. 
Apex blunt or pointed. 
Base blunt or attenuated. 
Seeds large or small. 
Leaves 
Long or short. 
Broad or narrow. 
Erect or reversed. 
Rolled or flat. 
Clustered at base or extending to top of 
culm. 
Culms 
Tall or short. 
Thick or thin. 
Straight or wavy. 
Erect or bent outward. 
Green or purplish. 
Many or few. 
SCIENCE 
[N. 8. Vou. XXXV. No. 903 
Nodes 
Many or few. 
Green or brown. 
Internodes long or short. 
Habit Characters 
Lodging or non-lodging. 
Rusty or rust resistant. 
Early or late season. 
It is possible that some of these char- 
acters may be expressions of the same unit, 
but in a number of cases they certainly re- 
present several different unit characters. 
For instance in length of head, height of 
culm, number of culms, and season of matur- 
ing, several different degrees are certainly 
present which are fully heritable. Doubt- 
less there are many more than 28 pairs of 
unit characters which could be distin- 
guished by careful study. If we have two 
pairs of characters, such as tall or short and 
early or late, we know that 4 homozygous 
combinations are possible. If three pairs 
are considered, 8 combinations are possible. 
Every time we add a different character 
pair we double the number of different 
combinations that are possible. Twenty- 
eight character pairs would thus give 
us aS many possible combinations as 2 
raised to the 28th power, or the astonish- 
ing number of 268,435,456. It would be 
possible then to produce this tremendous 
number of different varieties of timothy if 
there was any reason to do so, and each va- 
riety would be distinguished from any 
other variety by one distinct character and 
would reproduce true to seed. 
The task of the breeder, then, is to find 
which among these character combinations 
gives the superior plant for commercial 
cultivation. He will soon eliminate cer- 
tain characters as unimportant and con- 
centrate his attention on those qualities 
that are essential. 
It would be interesting to discuss the 
factor hypothesis, purity of germ cell, sex- 
