60 TENOPYR: CONSTANCY OF CELL SHAPE 
any cell, the two fragments were recorded together as the measure- 
ment of the cell in that region. Frequently the end of the line of 
measurement did not reach the opposite wall of a cell. In such 
a case I estimated as accurately as possible what fractional part - 
of the length or of the breadth of that cell fell upon the line of 
measurement, and this fraction was recorded. Thus, if the line 
passed through nine cells and reached only one fifth of the way 
across the tenth cell, the cell count was recorded as nine and one 
fifth cells. This method cannot of course be regarded as giving 
absolute length and breadth of the epidermal cells especially, but 
is probably the best available. It seems on the whole preferable 
to record the fractional parts of cells lying on the scale unit. The 
_ alternative would be to make all the measurements on single cells, 
“and wh this method the problem of selection and the errors in deter- 
mining the cell axes would probably involve still greater inaccu- 
racies. 
The maximum length and width of each leaf examined was 
recorded, and an outline tracing was made. In most cases one 
hundred counts were made on each type of leaf as the basis for 
comparison, with the exception of Linum, in which case I made 
nineteen counts on the palisade cells and forty-five counts on the 
epidermis, and of the palisade cells of Plantago, on which thirty 
counts were made. 
Except in the comparison of leaf shapes in lobes and constric- 
tions, the cell counts were made in the middle region of the leaf, 
and about 2 mm. away from the midrib. As the comparison was 
between leaves differing from each other in length and breadth, 
only these two dimensions of their cells were measured. I have 
not concerned myself with the third dimension, as the difference 
in thickness of the leaves is too small for easy study. 
All the measurements were made from the surface of the leaf. 
The epidermal cells of the under surface of the leaf were measured 
from below. The palisade cells were measured from above. 
Thus in any one portion of a given tissue I could make measure- 
ments of the length and of the width of the cells at the same time. 
The length of the cell I considered to be that dimension of the cell 
which was parallel to the midrib of the leaf and the long axis of 
the entire leaf, the width of the cell as taken was the dimension 
to the width of the leaf. 
