THE COLLAR-CELLS OF HETEROCGLA. 37 
membrane. Similar results were obtained on clearing the 
sections in benzol and in olive oil. 
I am inclined to consider the chief engine of distortion to 
be the passage from alcohol into benzol, chloroform, or turpen- 
tine, and vice versa. The cells of fig. 15 probably escaped, 
not only because the passage into benzol was effected by very 
gradual dialysis, but because they were first hardened in 
alcohol between 85 per cent. and absolute strength for 
some eighteen hours. It may be noted that the tendency 
of all the cells to assume a drop-like form proves that the 
force effecting their distortion is surface-tension. 
It does not seem unlikely that the reduction in volume is 
due to the abstraction of water and soluble matters by the 
alcohol. It is not due to shrinkage of the paraffin block, for 
from the standard tables contraction through 45° C. would be 
in wax to ‘96, and in paraffin not more than to ‘99 of the 
original linear dimensions. I have no reason to suppose that 
there was any appreciable compression in cutting the sections ; 
and since the nuclei remain spherical, and the collars are un- 
altered in length, this cannot be assumed. But it must be 
pointed out that the mean contraction-ratio is less certain than 
the amount of distortion, since it involves the measurement of 
the living cell-height, which can only be done accurately in 
fortunate instances. 
The collar rarely contracts in length; this may either be 
due to its thinness, or to the nature of the rods which com- 
pose it. Sollas’s membrane may be due to either a local con- 
striction of the collar or the forcible contraction of its base 
throwing out the free lip; it should be noticed, however, that 
in such a section as is drawn in fig. 18, the chamber has so far 
contracted as a whole, that where the free ends of the collars 
remain of their living diameters, they must be pushed into 
contact. 
By the definition of “ basal width” employed, it will be 
seen that this measurement expresses the linear contraction of 
the wall of the chamber as a whole. There is generally least 
contraction in this plane, the tendency of the cell to become 
