Should the poison consist of an alkaloidal body existing 

 as a salt in the acid solution, the possibility of extracting 

 the base with ether or chloroform, after the solution had 

 been made alkaline with ammonia, is apparent. This 

 was tried with negative results. To a water solution of 

 colon poison, acid in reaction, ammonia was added, drop 

 by drop, to a slightly alkaline reaction, the mixture shaken 

 with ether, the ether separated and evaporated. The residue 

 remaining was non-toxic. The ammoniacal water solution 

 was next shaken with chloroform, the slightly colored 

 chloroform drawn off and evaporated at low temperature, 

 leaving a small amount of a dark, thick, semiliquid, which 

 was not poisonous either as it was or after faintly acidify- 

 ing with hydrochloric acid. The water solution remaining 

 being still poisonous, it is evident that the toxic part is not 

 an alkaloidal body capable of being extracted directly. 



Potassium bismuth iodide in acid solution of the crude 

 soluble poison produces an abundant precipitate, apparently 

 more or less soluble in excess, and soluble in ammoniacal 

 water. 



Kowalewsky has shown that uranyl acetate will com- 

 pletely remove from various albuminous fluids every trace 

 of protein giving a biuret reaction, while Jacoby and others 

 have used this reagent for the removal of proteins from 

 faintly alkaline solutions. Abel and Ford used it to remove 

 protein from an extract of poisonous fungi. In a slightly 

 alkaline solution of albumin poison, uranium acetate gave 

 an abundant precipitate, but not a complete separation, 

 as both precipitate and filtrate still gave the Millon and 

 biuret tests, and the filtrate, after removal of excess of 

 uranium with a solution of di-sodium hydrogen phosphate, 

 filtration, evaporation, solution in alcohol, and reevapora- 

 tion, was still poisonous. In acid solution, the precipitation 

 was complete, the filtrate no longer giving the protein 

 reactions. 



Freshly prepared metaphosphoric acid also produced 

 an abundant precipitate, but not a complete separation, 

 the filtrate showing both Millon and biuret reactions. 



