THE EXPERIMENTAL STUDY OF GENETIC RELATIONSHIPS 1 39 
and the parent form. The presence of either factor suffices to give the 
capsule the triangular shape typical of Capsella. C. Heegeri, the 
recessive, has a capsule shaped like that of Camelina, so that it really 
departs from the parent form in a generic character. As Solms- 
Laubach has shown, no systematist who did not know of its mutative 
origin or of its relationship as determined by crossing would place it 
in the genus Capsella. All the plants in existence have descended 
from one individual which was found by Professor Heeger in a colony 
of Capsella Bursa-pastoris in the market-place at Landau, Germany. 
When the two species are crossed the first hybrid generation is, of 
course, uniform and like Capsella Bursa-pastoris. In the second 
generation there is one plant of C. Heegeri in 16. The 15 sister plants 
all look alike, but are in reality of three kinds. In the third generation 
seven will remain constant, four will give one plant of C, Heegeri in 4, 
and four will give one plant of C. Heegeri in 16. As far as experimental 
evidence can be applied it shows that a plant into which the character 
of C. Heegeri has been introduced by hybridization must either give 
rise to C. Heegeri not less than once in 16 times, or not at all. Most 
recessives, if crossed with the parent form, reappear in the second 
hybrid generation in a typical i : 3 ratio. If they originated ac- 
cording to the multiple factor hypothesis we should have to assume that 
factors which had been independent for countless generations could 
suddenly become indissolubly associated in the generation in w^hich the 
pure recessive appeared. There is not the least evidence that this 
takes place. Now that we have cases in which segregation in the 
second hybrid generation occurs in the ratios 15 : i and 63 : i, we 
naturally expect that some of the recessives which are supposed to 
depend upon the concurrent absence of so many characters would not 
reappear at all when recrossed with their parents. With as few as six 
independent identical factors there would be only one recessive in a 
second hybrid generation consisting of 4,096 individuals. In a case 
like that of Capsella Heegeri it would be necessary to postulate more 
characters than six to account for a non-mutative appearance so seldom 
that only one individual has ever been observed. On the whole there 
seems no reason to doubt that the sporadic appearance of recessives in 
supposedly pure lines is really due to mutation. The imaginary in- 
fluence of past hybridization is a bogey that need not bother us. 
Although recessive mutations and recombinations following 
hybridization may contribute much to specific and varietal diversity, 
