366 BOTANICAL GAZETTE [NOVEMBER 
development of a new race within the range of the old but centering 
about a different mean. It is also believed that skewness may result 
from physiological causes having no direct bearing upon the origin 
or modification of species. While in no specific case may the 
suggested interpretation be the correct one, these different views may 
at least be accepted as evidence that skewness may result from various 
causes, and that it is therefore not self-explanatory. 
If the 1903 curves are compared with those for 1900 in figs. 11-13, 
it will be seen that in every case the positive sides of the curves are 
approximately coincident, but on the negative side there is a very 
material disagreement. According to the recent discussion of skew 
variation by Lutz (1904), we have here a case of skewness produced 
by the addition of variates, and this addition of such magnitude as 
already to overtop the 1900 population, thus giving a fine example of 
“historic” skewness; but no one can be convinced that this is here 
due to the “starting of a new race about a mean within the range 
of the old race.” A 
It is evident that the skewness is here the result of direct physio- 
logical reaction to the changed environment. Not all individuals 
are alike sensitive to changed conditions, some being more, some less 
affected by a given amount of change; so that while many individuals 
respond to the less favorable conditions by the production of heads 
with smaller numbers of parts, there is still a considerable number 
of conservative individuals which are little or not at all affected. The 
positive skewness of these curves is due to the fact that only 4 — 
proportion of the population is conservative. If the great mass : 
variates had been comparatively conservative and only a small ae 
centage sensitive to the changed conditions, it is plain that the ge 
tion of the principal modes would have been little affected, while a 
mean would have been lowered and negative skewness would ee 
been the result. This would then have been a case of onl a e 
“prophetic” skewness. We may say then that in cases of age 
physiological variation, prophetic skewness indicates slight mee 
ness, and historic skewness great sensitiveness,* to the chang' pene 
ditions, provided always, of course, that “under ordinary co 
the distribution of the variates affected is normal. pe 
5 As measured by the number of sensitive individuals, not by the degree 
sitiveness of each individual. 
