ON THE EARLY STAGES OF INFLAMMATION 213 
plates of glass, when they may be observed sticking for a longer or shorter time 
to one of the surfaces of the glass, each one dragging behind it a short tail-like 
process ; and as the movement of the blood diminishes so as to permit the 
formation of rouleaux, the latter may be not unfrequently seen ad- => 
hering in the same way by one of their terminal corpuscles, as a> 
represented in the accompanying diagram. aS 
That the cylindrical character of the aggregated masses is an accidental 
result of the shape of the blood-discs, is evident from 
the fact, that in the frog, although the same ten- 
dency to agglutination of the corpuscles exists as 
in mammalia, yet, as their biconvex form renders 
it mechanically impossible for them to be applied 
to one another throughout their entire circumference, 
they become arranged in groups of an irregular 
form, as is shown in the annexed sketch of blood 
contained in a small vein of the frog’s web. 
Again, different specimens of mammalian blood differ very much in the 
amount of adhesiveness of their corpuscles; and when this property exists 
beyond a certain degree, the discs stick together by any parts that happen to 
come first in contact, and retain that position more or less, so that the result 
is the formation, not of rouleaux, but of irregular confused masses. The most 
striking example which I have seen of this was presented by the blood of a bat, 
which had lived some days after having been severely wounded. In that case, 
chains of red discs might be seen adhering firmly by their edges, notwithstanding 
considerable force of traction operating upon them, and before they at last gave 
way tail-like processes of considerable length were drawn out between every 
pair of corpuscles, indicating that they were very adhesive. These facts seem 
sufficient proof of the correctness of the view above expressed regarding the 
cause of the rouleaux. 
The adhesiveness of the red corpuscles does not appear to be a vital pro- 
perty. When the fibrine has been removed from a drop of blood during the 
progress of coagulation, the rouleaux will form again, after being broken up, 
as many times as the experiment is repeated, until the blood becomes thick 
from dryness ; and if evaporation be prevented by Canada balsam placed round 
the plate of thin glass, with suitable precaution against the approximation of 
the two plates, the rouleaux will remain perfect for several days (e.g. fourteen 
in one experiment of the kind), after which the very slow chemical action of 
the balsam upon the blood gradually renders it confusedly red and opaque. 
Gum mixed with blood seems to preserve it, like a pickle, from decomposition 
