186 _ Mr. J. A. Craw. On the Physical [Mar. 22, 
being reversible, the velocity of dissociation will probably increase rapidly 
with rise of temperature. A temperature of 37° C. was employed as more 
likely to show reversibility if such exists. 
The methods employed were: (@) filtration through gelatine; (0d) diffusion 
through gelatine. 
(a) Filtration through Gelatine—The practically constant concentration 
of free lysin in the residues of Experiments Nos. 5 and 6, Table I, might 
well be due to reversibility. To test this the residue of No. 4 (index 31:25) 
was diluted with saline to the volume of the original mixture, and heated 
for 24 hours at 37°C. On again filtering through gelatine until the new 
residue had the same volume as the first residue, the hemolytic index 
was found to be (27°8).* On repeating this treatment the hemolytic index 
was found within the experimental error unchanged. The gelatine obtained 
from the filters used was strongly hemolytic (index 100). If the free 
lysin of the first residue had alone been available, viz., 33 x 31:25 = 1031 
“hemolytic units,” then as filtration removes 300 or more “ units,’ the 
1031—300 
32°5 
= (22°5) and the third residue (13). It would appear then that a portion 
of the apparently combined lysin has become free on heating at 37° C. 
I conclude, therefore, that in a neutral mixture the reaction is at 37° C. 
at least partially reversible. 
(b) Lxperinents on the Diffusion through Gelatine of Lysin and Antilysin.— 
Dr. G. Dean kindly suggested to me a suspension of red blood corpuscles 
in gelatine as a convenient indication of the diffusion of lysin. In applying 
this method I found it was necessary to allow the lysin to diffuse from a 
solid gelatine layer into the solid gelatine suspension, for when normal 
saline alone is placed above a 2°5-per-cent. suspension of corpuscles in 
gelatine (9 per cent.), hemolysis occurs. This proceeds from the surface 
downwards at a rate which is not altered when a strong fluid lysin is used 
instead of the saline. To avoid this effect, which is probably an imbibition 
phenomenon, the precipitated lysin was dissolved in 9-per-cent. gelatine 
at about 37°, and poured on to the strongly cooled columns of 9-per-cent. 
gelatine containing corpuscles. A rate of haemolysis was then obtained 
which decreased with decreasing concentrations of lysin, and was entirely 
absent when saline in gelatine was used. The course of the diffusion is 
evidenced by a zone of transparent gelatine coloured with the freed hemo- 
globin and a second zone in which partial hemolysis has taken place, which 
* Cf. Table I, No. 4a, 
second residue could not have had a higher index than 
