LEGUMIN whic: . A485 
cked on to them—in fact a bell pomibly alkaloidal in 
nature, not completely separated from the proteid in the 
preparation of the pure poison. 
The globulin is separated by extracting the crushed and- 
while the albumose remains in solution. The globulin is now 
moved by filtration, and washed with distilled water (previ- 
ously boiled to sterilise it) for two days, in order to remove 
any albumose or sodium chloride clinging to the precipitate, 
the absence of the albumose being tested by the washings’ 
_ giving no reactions for a soluble proteid, and the absence of 
the salt by a negative reaction with silver nitrate. The glo-~ 
bulin is then removed from the filter and dried over sulphuric 
acid, Prepared in this way, it is a whitish-yellow, amorphous 
rwder, soluble for the most part in 15 per cent. sodium chlo- 
nide, Sind giving the reactions. previously described. In the 
dried state, it may be kept for a long time without losing its 
Bpticlopical properties. Specimens prepared for more than © 
so months are as active now as when first dried. This fact 
is, indeed, only in accordance with the behaviour of other dried 
| Decca they can be kept an indefinite time in the dried state 
Without undergoing any chemical change. 
The albumose was- prepared by making a concentrated 
tery extract of the seeds, and filtering the clear infusion 
direct into an excess of absolute alcohol, thus throwing down 
both proteids as a white precipitate. After a few days the 
precipitate was removed, redissolved in water, and reprecipi- | 
tated by alcohol, this process being repeated at intervals of a 
months. The precipitate was allowed to stand under 
hol for about eight months or longer, at the end of which 
_ the globulin was completely coagulated, while the 
mose was still soluble in water. Dried over seacihee” acid, 
