68 
All of these tend to add to rather than to subtract from the varia- 
bility in inheritance. The simplest situation may be assumed to be 
that in which each character is determined by a single chromo- 
some. So let us study the possibilities in the gladiolus on that basis, 
If we take two strains or two species of gladiolus such that 
each breeds true for one of a pair of alternative characters, such as 
ruffled and plain petalled, and cross them, then the first genera- 
tion will have in its constitution the tendencies for both characters. 
If, further, this trait is determined by a single factor, or gene, 
carried in one pair of chromosomes, then the hybrid will have 
the gene for ruffled in one chromosome, for plain-petalled in its 
mate. But when this hybrid produces pollen grains and ovules, 
and reduction takes place, only one of these chromosomes can en- 
ter into the male nucleus of each pollen grain or the egg cell of the 
ovule, so that each such reproductive cell can carry either the gene 
for plain-petalled or the gene for ruffled, not both. And so, if 
this hybrid is self-fertilized (or two sister hybrid plants are 
crossed, for hybrid gladioli are frequently self-incompatible), the 
offspring of the second generation may have each two chromo- 
somes each carrying ruffled, or one ruffled and one plain-petalled 
or two plain-petalled. So there are three possibilities : ruffled, half- 
ruffled and plain-petalled. 
If we now assume that the above two strains differ further 
in two other independent traits, as hooded and open flowered, 
then considering both these traits, the possibilities in the second 
generation are: Ruffled hooded, ruffled open, ruffled half-hooded, 
half-ruffled hooded, half-ruffled open, half-ruffled half-hooded, 
plain hooded, plain open and plain half-hooded. Thus there are 
9 possibilities for two independent characters, or 32. Similarly, for 
three independent pairs of characters, there are 27, or 3° possi- 
bilities, 
Since there are 14 pairs of chromosomes in the gladiolus w€ 
may assume that our two strains may differ by as many as eas 
dependent pairs of characters. : 
_ Of course, there can be only a limited number of combina- 
tions possible with twenty-eight different chromosomes, in four- 
teen pairs. If we designate the first pair of chromosomes as 4 and 
A, the second pair as b and B, and so forth, then we get the fo 
lowing set of chromosome units in the first generation hybrids: 
