268 PROCEEDINGS OF THE AMERICAN ACADEMY, 



The auramin pyrophosphate of sodium, on boiling with baric chlo- 

 ride, gives a yellow crystalline salt; with mercuric chloride, a fine pale 

 yellow crystalline salt which becomes very distinct on boiling. 



When the auramin salt is*boiled with mercurous nitrate, some mer- 

 cury is reduced to metal, and at the same time very characteristic 

 white prismatic crystals are formed. Chlorhydric acid does not sensi- 

 bly dissolve the pyrophosphate, but changes it to a pale yellow flocky 

 crystalline body. Chloride of luteo-cobalt gives on boiling an orange 

 crystalline salt. 



When the crystalline yellow or dull orange baric salt is well washed 

 with hot water and then filtered off, the colorless filtrate on evaporation 

 gives with sodic hydrate a very distinct reaction for ammonia. Part 

 of the ammonia in the salt must therefore have been in the form 

 of ammonia, unless we admit that the auramin is decomposed under 

 the circumstances. When treated with a cold solution of argentic ni- 

 trate reaction sets in at once, and a fine yellow flocky crystalline salt is 

 formed ; but no trace of ammonia is obtained from the filtrate after 

 washing the salt with cold water and separating the excess of silver. 

 Both the barium and silver salts were partially analyzed. 



The barium salt was analyzed by Mr. G. W. Patterson : 



0.6472 gram gave 0.1015 gram BaS04 = 10.30% BaO. 

 0.6472 gram gave 0.1426 gram PoOyMg., = 14.09% PA • 

 0.6472 gram gave 0.3900 gram gold = 67.63%. 



Here the ratios are 9 AU2O3 : 6 P2O5 : 4 BaO. 

 In the silver salt (Patterson) : 



0.7718 gram 0.0915 gram AgCi = 9.58% AgA 

 0.7718 gram 0.4470 gram gold = 64.98%. 



0.7718 gram 0.1584 gram PaO^Mga = 13.13% PA- 



The ratios are approximately 16 AuAs '■ 10 P2O5 '■ 5 AgoO. 



The analyses of the barium and silver salts are at least sufficient to 

 show that no simple double decomposition takes place in either case. 

 The formula which I have given for the gold salt is to be regarded 

 as purely empirical, and does not explain the explosive character 

 of the salt. No compounds falling under the general expression 

 Au2(NH3)n03 are at present known. Dumas and Raschig give to 

 fulminating gold the formula NIIo • Au'" . NH, which may be written 

 N2II3AU'". This is equivalent to 2NHr; the corresponding ammo- 

 nium must be NgHsAu^'Ha, and the oxide of this (N2H5Au'")0. If 



