The Theory of Evolution Bat 
hybrid. If all hybrids followed Mendel’s law, which they do 
not in fact, we should expect a resolution of the parent types in 
the second generation. Since this does not occur in most of the 
mutations of cenothera, we must infer that if the mutation is a 
hybrid it belongs to the class of stable hybrids. This is in sub- 
stance de Vries’s interpretation. If, as de Vries assumes, the 
mutant is the result of a mutating germ-cell meeting one of the 
ordinary kind and producing a stable hybrid, it might appear 
that if the hybrid were back-crossed with the parent, the off- 
spring should be again like the hybrid, but this was not the 
result found. These and other considerations show that there are 
some obscurities concerning the origin of the cenothera mutants. 
Their behavior when hybridized is also difficult to harmonize 
with other cases of discontinuous inheritance. 
Variation and Mutation in Helix 
In connection with his experiments on snails Lang has made 
some important observations on the relation of continuous 
(or fluctuating) and discontinuous variation. He has found that 
in certain colonies of Helix hortensis, in the neighborhood of 
Zurich, only two kinds of individuals exist: yellow bandless 
and the yellow five-banded. When inbreeding they produce 
only these two types again, the bandless condition dominating 
in the first generation as already explained. No intermediate 
forms appear. This is an example of strictly discontinuous 
inheritance. 
If the banded individuals in other colonies are closely 
studied, it is found that the number and character of the bands 
vary; they are broader or narrower, they may remain separate or 
become united, especially the fourth and fifth bands. Their 
color may be dark or light. An examination of colonies from 
other localities shows that all transitions exist from the most 
highly developed five-banded condition to a bandless individual. 
Experiments establish that the offspring of the same parent may 
show these extremes. Here there seems to exist a fluctuating 
variability whose extremes include the two types of discontinuous 
