286 BACTERIAL POISONS. 



silver salts is suspended in water, nitric acid added, decora- 

 posed with hydrogen sulphide (ammonium sulphide, or, 

 better, hydrochloric acid, may be used), and the clear filtrate 

 is concentrated on the water-bath to a small volume. It is 

 then saturated with ammonium hydrate and digested on the 

 water-bath for some time, whereby adenine and hypoxan- 

 thine go into solution, while the guanine remains undis- 

 solved (see p. 287). From the ammoniacal solution on 

 partial concentration and subsequent cooling, the adenine 

 crystallizes out first, whereas the more soluble hypoxanthine 

 remains in solution. If the adenine is still colored it can 

 be purified by dissolving in water and boiling with animal 

 charcoal. The hot aqueous solution is then rendered very 

 slightly alkaline with ammonium hydrate and allowed to 

 cool ; adenine crystallizes out, and can be still further puri- 

 fied by recry stall ization from water. 



Ammonium sulphide has been employed by SCHINDLER, 

 in place of hydrogen sulphide, in decomposing the silver 

 compounds of the above bases. BRUHNS recommends in- 

 stead warming with very dilute hydrochloric acid, espe- 

 cially if guanine is present. The solution can then be 

 neutralized with NaliCO 3 , using methyl-orange as indi- 

 cator, and the adenine separated from hypoxanthine by the 

 picric acid method described below. 



Another method for the separation of adenine from 

 hypoxanthine is based upon the behavior of the nitrates 

 of these bases in aqueous solution. From concentrated 

 aqueous solutions of the nitrates, free hypoxanthine crystal- 

 lizes out first, because the nitrate is decomposed ; whereas, 

 adenine, which is a stronger base, remains in combination 

 with the acid, in solution. 



SCHINDLER determines adenine and hypoxauthine indi- 

 rectly. The ammoniacal solution which is filtered from the 

 insoluble guanine is evaporated to dryness on a weighed 

 platinum dish, dried at 110, and weighed. A nitrogen 

 determination is now made of the mixed bases and from 

 these data the proportion of each is calculated. 



By far the best method for the quantitative separation of 

 adenine and hypoxanthine is the picrate method of BRUHNS. 



