The Use of Graphs in Paleontology. 407 
from one stratigraphical level, namely the present. The relation- 
ship of his material to that of the paleontologist is evident from what 
has just been stated. It is like a cross-section of lineages which 
stretch back into the, to him, inaccessible past. 
Mendelian experiments of necessity deal only with (differentiating) 
characters belonging to homologous parts, e.g. the end of a pod, 
pointed or rounded; the edge of a leaf, much or little serrated ; 
the epidermis of a plant, smooth or hairy. They deal, therefore, not 
with separate structural units as such, but with certain stages in 
the changes which they undergo. Thus, the end of a pod is not only 
either pointed or rounded, every gradation between these two 
extremes may be found, provided all pods are taken into account. 
The fact that such gradations do occur to-day indicate that pointed- 
ness and roundness have not arisen suddenly but gradually during 
past ages. So, likewise, with the edge of the leaf, the epidermis of a 
plant, and many other homologous parts which have supplied 
material for experimentation. 
Where such gradation exists, the experiments appear to deal with 
the extremes of the series which correspond with such percentages, 
as, say, 0 and 95 or 10 and 70 of the scheme explained in the earlier 
part of this paper. If this be the case, some, at least, of the 
Mendelian unit characters are merely stages in the serial changes 
of a structural unit. 
The fact that structural unit A at stage O or 10, for example, 
can be removed and replaced by the same structural unit at stage 
95 or 79 without producing any corresponding change in the 
elements B, C, D, etc., furnishes yet further confirmation of the 
view that serial changes in one element are independent of those 
in another. 
The terms structural unit and element used so frequently above 
seem at first sight to be synonymous with Osborn’s unit character 
or biocharacter.'' In so far as he defines this as a separable unit in 
ontogeny and phylogeny, the two are the same. But in his definition 
he appears to regard the bivcharacter as being the same as a 
Mendelian unit character. Now the unit in ontogeny and phylogeny 
is something which changes continuously. The Mendelian character, 
on the contrary, is something which is quite fixed, and, within the 
limits of experimentation, unchangeable. If it were not so the 
experimenter would lose his bearings in the multiplicity of his filial 
generations. The view adopted here therefore is that the Mendelian 
unit character is not merely a structural unit, but a structural 
unit at a certain stage of evolution. The experimenter does not 
exchange one structural unit for another, e.g. horns for ears, but a 
structural unit at one stage of evolution for a homologous unit at 
another stage, e.g. long staple and short staple in cotton, presence 
or absence of horns in cattle. 
Once more, it may be allowable to emphasize the fact that it is 
1 H. F. Osborn, American Naturalist, 1917, p. 449. 
