342 Kirkwoop AND GIES: CHEMICAL STUDIES 
further purified by re-solution and re-precipitation. For quantita- 
tive analysis some of the final product was washed in water, alco- 
hol and ether, and dried at 100°—105° C. to constant weight. 
Sometimes the globulin prepared in this way was both crystal- 
line and amorphous. At other times it was entirely crystalline. 
Triangular, hexagonal and rhombohedral forms were frequently 
seen, although octahedra predominated.* The crystals so closely 
ee . 4 
Ss 
Fic. 4. Crystals of cocoa edestin. 
resemble those we have repeatedly made from hempseed and lin- 
seed by the same method, and are so like those given by Osborne 
for edestin,+ that we felt satisfied from the beginning our globulin 
would prove to be of the edestin type. Careful study of the t 
actions of the substance convinced us of this fact, for it gives all of 
those attributed to edestin by Osborne. 
* The large proportion of gum extracted by the saline solution made it difficult ree 
only to prepare the proteid in pure form but to obtain it quantitatively. Beste An 
edestin passed in part into an insoluble modification during the manipulations. — 25 
appreciable loss resulted, therefore, in each preparation. We obtained as much as 
grams of the purified product from the kernels of twelve nuts : on 
ft Osborne: Journal of the American Chemical Society. See also his paper 
crystalline vegetable proteids in the American Chemical Journal, 14: 28. 1893: 
