CELLULAR STRUCTURE OF THE RED BLOOD-CORPUSCLES. 47 
seems impossible to account for this stroma assuming a glo¬ 
bular shape when acted upon by water, since the full diameter 
of the sphere, when formed, is occupied by the least amount 
of distended matter, -while the dimension, in which lay the 
greatest bulk of stroma previous to the addition of water, 
is actually diminished; on the other hand, any hypothesis 
that the porous substance coalesces, on the removal of its 
haemato-crystallin, into a jelly-like drop, is negatived by the 
fact that occasionally, as is also described by Professor J. C. 
Dalton, one side of a corpuscle, rendered colourless by water, 
fails to assume a convex form, being apparently sucked in by 
the other side, which becomes exaggeratedly convex, until 
the whole corpuscle resembles a bell, or more accurately a 
liberty cap, in shape, without any tendency to present the 
outline of a sphere. Second, the appearance of the speci¬ 
mens I have examined strongly indicates the existence of a 
dense membrane, thrown into folds around the extremities 
of projecting crystals, just as a loosened tent cloth would be 
around the point of a cane thrust against it from the inside; 
and further, the movement of the crystals, when partly dis¬ 
solved, around the nucleus, but confined within the corpuscle, 
as described in the early part of this paper, both tend to 
show that the cavity of the corpuscle between the nucleus 
and the membranous envelope is quite unoccupied by solid 
matter. Third, the perfect freedom with which one side of 
the cell w r all of a red blood-globule from the Menobranchus 
when acted upon by water may float in until it touches the 
nucleus, and out again to its old place, will, I think, furnish 
conclusive evidence to any one who sees it as I have done, 
against the existence of a porous substance which maintains 
the shape of the blood disk. 
From these researches I therefore conclude that the older 
theory, which asserts that the red blood-corpuscles of the 
vertebrata generally are vesicles, each composed of a delicate, 
colourless, inelastic, porous, and perfectly flexible cell wall, 
enclosing a coloured fluid, sometimes crystallizable, cell con¬ 
tents, which are freely soluble in water in all proportions, 
explains the physical phenomena presented by red blood- 
globules far more satisfactorily than any other hypothesis 
which has hitherto been advanced; and, moreover, that the 
usual bi-concave discoid form of the corpuscles in most 
mammals, as well as the change of shape which they undergo 
in fluids of greater or less specific gravity than the liquor 
sanguinis, becoming crenated in denser, and globular in rarer 
liquids, are such as to be perfectly explicable by the light of 
our present knowledge in regard to the laws of the exosmosis 
