428 The American Naturalist. [May, 
factor would be partly but not wholly set aside by proof that 
somatogenic variations are not inherited. St. Hilaire’s factor 
of the action of environment upon early stages of development 
would result in purely fortuitous variations, and, as he himself 
clearly perceived, would require Selection to give it an adap- 
tive direction. Niigeli’s factor, on the other-hand, assumes de- 
finite but not necessarily adaptive ‘ phylogenic’ variation—his 
views have been very generally misconceived on these points 
—and, as he pointed out, his factor would also require Selec- 
tion to. determine which of the definite lines of growth were 
adaptive. 
It seems necessary to thus’clearly state the relations of the 
time stages of variation to each of the five factors, in order to 
show the decisive bearings our future exact research will have 
upon them. For example, the proof that variation is either 
‘definite’ or that it is ‘adaptive’ prior to or independently of 
Selection, will constitute conclusive disproof not of Darwin’s 
theory but of Neo-Darwinism. The fate of Lamarckism, on 
the other hand, depends upon the demonstration that phylo- 
genic variation is not only ‘ definite’ and ‘ adaptive’ but that 
it is anticipated by{corresponding somatogenic variation. 
A review of recent thought upon the variation problem 
shows that these life stages are becoming generally recognized. 
I shall pass by Lamarck’s and Darwin’s factors which are so 
thoroughly understood and speak only of the other three. 
Burron’s Factor IN VARIATION. 
As regards Buffon’s factor, which is the most comprehensive 
of all, we know that Spencer and Weismann both assumed that 
the direct action of the environment was primarily a factor of 
evolution. Weismann first regarded this solely as the proto- 
zoan source of Variation, but has recently given it a wider 
play in the action{of environment upon the germ-cells as a 
cause not of definite variation but of variability. The line 
of research upon the dynamic action of environment in its 
influence upon somatogenic variation followed by Hyatt, 
Dall, and others, is {paralleled in the more recent specula- 
