JUNE 14, 1912] 
hypothesis is perhaps consistent with the 
corpuscular theories of heredity but the 
necessity of its introduction arising from 
the variability of the products of hybridi- 
zation must be regarded as seriously affect- 
ing the basal conception of segregation of 
fixed unit characters. 
Perhaps the most important modifica- 
tion of Mendelian theories is in the con- 
tinuous discovery of further cases in which 
single visible characteristics ‘as conceived 
by Mendel or De Vries may be dependent 
for their realization on from two to several 
factors, or may arise in more than one way. 
Nilsson-Ehle tells us that the black color in 
oats and the red color in certain wheats 
may each be produced in different ways. 
Shull finds that two separate genes may be 
responsible for the common form of cap- 
sule in Capsella bursa-pastoris. Bateson 
and Saunders tell us that certain white 
varieties of peas and stocks when crossed 
give purple. This is because the colors in 
question are due to two factors instead of 
one. These two factors are not members of 
one allelomorphic pair but must be assumed 
to be the members of two distinct pairs, 
each from different parents. Hoariness in 
stocks is dependent on four factors, two 
for hoariness and two for flower color, not 
compounded in one unit but distributed in 
four allelomorphie pairs. Reversion is the 
reappearance of a character because of the 
reunion of the two necessary factors which 
had become separated. 
An extreme of this tendency is found in 
Tammes’s recent paper on heredity in flax 
in which she has worked with three types 
and apparently has used all possible pre- 
cautions as to control, ete. Tammes finds 
that the results of her study of seed size, 
petal size, petal color, etc., can only be 
brought into harmony with Mendelian 
ratios by assuming that each of these vis- 
ible characters is dependent on from one 
SCIENCE 
921 
to several factors in the germ plasm. For 
the length of the seed at least four factors 
must be assumed. For length and breadth 
of petal three factors must be assumed in 
one cross and at least four in others. For 
flower color three factors were found; for 
dehiscence of seed pod three or four fac- 
tors; for hoariness of seed pod alone one 
factor. Tammes finds in crossing two indi- 
viduals differing in a single visible character 
where Mendel would expect to find a mono- 
hybrid giving four combinations in the 
F, generation that the results are those 
which should be expected in the case of a 
di- or poly-hybrid. In plain terms this 
merely states that the F, generation is, as 
the older views of hybridization held, 
vastly more variable than the ordinary 
Mendelian expectation permits. 
It is probable that Tammes and the 
others who are assuming a multiplicity of 
factors as necessary to the production of a 
single visible character are giving the facts 
as they find them, but the doctrine of fixed 
unit characters represented by pangens or 
genes in the germ plasm certainly shows 
itself inadequate to account for such facts. 
These results attack the doctrine of the 
pangen at its very foundation. According 
to De Vries the multiplicity of plant and 
animal forms is due to the large number 
of combinations possible with relatively 
few unit characters. We are now given 
multiplicity in the germ plasm to account 
for apparent simplicity in the organism. 
On Johannsen’s view also each factor must 
be represented by a gene in the germ 
plasm. 
Furthermore, as noted, the unit char- 
acters are diffuse characteristics of the 
plant taken as a whole. Each hereditary 
factor responsible for this diffuse character 
may influence many parts of the plant. 
We have thus a most complex overlapping 
of functions among the factors, one char- 
