264 PROCEEDINGS OF THE AMERICAN ACADEMY. 



3.73 



4314 100.00 



In the analyses the salt was digested for some time at a boiling 

 heat with mercurous nitrate and mercuric oxide. The mercurous salt 

 formed contained mangauous oxide. The precipitate was fused with 

 a mixture of potassic and sodic carbonates to separate manganous 

 oxide. The filtrate from this last gave the phosphoric pentoxide. 

 Molybdic teroxide was determined by difference. As the salt could 

 not be recrystallized, the defects in tbe analysis are doubtless chiefly 

 due to traces of impurity. The formula of the salt should be written, 

 according to my view : 



20 M0O3 . 2 P^TMn . 2 (NH^)^© . 8 MnO . 3 (NH4)„0 + 12 aq. 



We may have here a double salt, as, for instance, 



10 M0O3 . P,07Mn(NH4)2 . 4 MnO . (NH4)20 . H^O 



+ 10 M0O3 . P20^Mn(NH4)2 . 4 MnO . 2 (NH4)20 + 12 aq., 



but of course other arrangements are possible. 



The manganese in the molecule containing P0O7 may be called, 

 for convenience, the internal or fixed manganese, to distinguish it from 

 the external or basic manganese. To determine if possible the ratio 

 between the external and internal manganese, I digested a weighed 

 portion of the salt in the cold for twelve hours with mercurous nitrate, 

 and then boiled, adding a little mercuric oxide in the usual manner. 

 In the filtrate, after separating the mercury, the manganese was 

 determined as PoOtMuo. In this manner, 



1.2306 grams gave 0.3520 gram PAMn = 11.08 %. 



From this it appears that about {h of the manganese was precipi- 

 tated by mercurous nitrate, in place of 1%. If, therefore, we consider 

 ^(f of the manganese to be present in the salt as PgOyMn, we must 

 suppose that this molecule is broken up, to a certain extent at least, by 

 boiling with mercurous nitrate, and it is very doubtful whether the 

 manganous oxide exists in any other form than as a base. Certainly 



