358 Gates, 
for example, differs from O.rubrinervis not only in leaf characters, 
but in amount of branching, thickness of stem, length of fruits, size 
of flowers and number of seeds produced. Several of the last-mentioned 
differences are due to the lessened vitality of O. oblonga. Whether 
the distinction between O. rubricalyx as a ”’quantitative‘“ mutant and 
O. oblonga as a qualitative one, is fundamental in character, involving 
a greater change in the germ plasm in the case of O. oblonga than in 
O. rubricalyx, depends upon the nature of the organic correlations 
involved. O. rubricalyx however, differs from O. rubrinervis quanti- 
tatively, in what may be considered a single character, namely, capacity 
for pigment production. 
III. Ontogeny of ©. rubricalyx, 
In order to understand the real nature of the difference between 
O. rubrinervis and its mutant, O. rubricalyx, this form was subjected 
to the closest scrutiny throughout its whole development. It was 
found that, when the seedlings of the mutant are only a few weeks 
old, a conspicuous amount of red appears on the under surface of 
the leaf petioles, a region which is always green in O. rubrinervis. 
As the rosettes develope, they usually show the maximum of red in 
this region when nearly full grown (four or five months after planting 
seeds), but there is always much variation in the amount of red 
appearing on the under surface of the rosette leaves in O. rubricalyx, 
some rosettes showing it only in certain leaves or in small quantities 
while others show a large and conspicuous development of it in nearly 
all the leaves. Red may also occur on the upper surface of the petioles 
and midribs of the rosette leaves, as is the case is O. rubrinervis. But 
in O. rubricalyx there is no correlation between the amount appearing on 
the lower surface and that found on the upper surface, the two varying 
quite independently of each other. In both O. rubrinervis and O. rubri- 
calyx as grown in my cultures, the amount of red on the upper sur- 
face was comparatively small. The wide variation in the development 
of red on the under surface of the leaves in O. rubricalyx made the 
determination of some of the rosettes doubtful, but since there is no 
red in this position in O. rubrinervis, the presence of a minimum 
amount sufficed to place a rosette in O. rubricalyx. When the 
characters of the mature plants were recorded, the provisional deter- 
mination made in the rosette stage was confirmed in practically every 
case. Thus by turning up the rosette leaves and examining their 
