296 
R. A. HARPER 
of a critical consciousness of the whole protoplasm question shown more 
clearly than in the frequent looseness of such assumptions. 
The assumption is that we can draw specific conclusions as to the con- 
stitution of the germ plasm from the behavior of visible characters in the 
multicellular organism, even though this behavior is so complex that it 
necessitates the assumption of several independent units in the germ plasm 
for one apparently individual character. Bateson's amoeba would have 
to be represented as of gigantic size in order to get points of origin for all 
the characters that have come out of it and been eliminated in evolution. 
There can be no question that with the intensive study of the past two 
decades the problems of heredity have been found to be vastly more com- 
plex than it was hoped they would appear when treated by the Mendelian 
method of analysis. The pangens of De Vries, at least in the earlier form 
of his theory, were regarded as relatively few in number and broadly per- 
vasive in their visible effects in the organism. We are now apparently 
more inclined to Weismann's conception of indefinitely numerous deter- 
miners in more or less fixed space relations to each other. 
The realization of the weakness of Mendelism in relation to facts as to 
the all-pervasiveness and interdependence of plant characters has led some 
of the defenders of the theory to assert that each unit factor may possibly 
influence every part of the mature plant. 
The difficulty with all such assumptions of hereditary units of whatever 
kind is more fundamental. Take the case of the serrations of the leaf of the 
common nettle. Correns told us in 1903 that entire margins and serrate 
margins were due to a factor in the germ plasm for serratures paired with 
one for entire margin, and the whole was made an example of dominance 
and segregation. It is to be noted that Correns found teeth weakly de- 
veloped in his recessives, and while so far as I am aware no one since has 
gone over the matter I am willing to predict on the basis of my studies on 
sugar and starch characters and aleurone color in corn that a whole series 
of intermediates between serrate and entire can be found and that a present 
day student instead of saying that there is one factor for toothed margin 
would say there are several or perhaps even twenty. 
If serrateness and entirety were found to be absolutely hard and fast 
categories, units in behavior, there might be something in favor of assuming 
as a working hypothesis that each was represented by an equally hard and 
fast, definitely limited section of a chromosome. But if there are all 
degrees of variation from entire to deeply serrate, the existence of a series of 
units in the germ plasm, one for each depth of serrateness, is not obviously 
suggested. The series of fluctuating variants has a unity to the human 
mind quite as natural as any one of the particular grades of serrateness. 
This is evidently felt vaguely by those who assume modifying factors and 
factors of fluctuating potency. To assume, however, that we have ex- 
plained anything or in any way contributed to clear up our knowledge of 
