64 TENOPYR: CONSTANCY OF CELL SHAPE 
The transitional leaf is more than four times as long as broad, yet 
its cells are of almost the same shape as those of the basal leaf, 
being slightly longer than wide. In the linear leaves, which are 
about 14.15 times as long as they are broad, the cells are only .07 
longer than wide. That is, the difference between the length and 
the width is 287 times greater in the leaf than in the cell. More- 
over this difference in cell dimensions is least in the basal leaves 
and greatest in the transitional leaves. There seems to be no rela- 
tion here between the shape of the cell and that of the leaf. 
The slight difference between the ratios of length to breadth 
of cells obtained in the different leaves is almost equaled by the 
variability of the individual cells in any one leaf. In some leaves 
the cell proportions, as compared with the leaf proportions, may 
be found reversed in certain areas. Thus fifteen measurements of 
cells in a basal leaf E, which was 10 mm. long by 12.5 mm. broad, 
showed that the cells averaged 0.042 mm. long and 0.037 mm. wide 
(a ratio of 1.13 : 1) while similar measurements of cells of a linear 
leaf, 50 mm. by 3 mm., showed that they had an average length of 
0.032 mm. and an average width of 0.033 mm. (a ratio of .97 : 1). 
The forms of the three types of leaves found in Campanula rotundi- 
folia are not therefore directly correlated with the shape of their 
cells. 
The cells of the stem leaves show a smaller average size than 
those of the basal leaves, the cells of the linear leaves being most 
reduced. In Taste II are given the relative length and relative 
TABLE II 
CAMPANULA ROTUNDIFOLIA 
Relative size of the lower epidermal cells of the three types of leaves. 
Length of Width of 
cells cells 
Transitional leaf : ei leaf . 04:1 86:1 
Linear leaf : basal leaf... .......; 86:1 Ae Bede 
width of the cells of the three types of leaves. The difference in 
the dimensions of the three types of cells is very slight, and might 
seem at first to fall within the limits of individual cell variability. 
I therefore plotted frequency curves (Fic. 1) showing the number 
of times that I found a given number of cells to a line of 0.232 mm., 
in the long and in the broad axis of the basal and the linear leaves. 
A greater number of cells to a given line indicates smaller cell size. 
