324 Oti certain Stony and Metalline Subjlanccs. 



iirtance of potafli, and by evaporating to drynefe. The oxide 

 of nickel was precipitated by hydrogenized fulphuret of am- 

 monia. 



Under all circum (lances, I am induced to (late the pro-tf 

 portions of condiment parts thus : * 



Silica - • - - - £d 



Magnefia - - i^ 



Oxide of iron 34 



Oxide of nickel » - - %\. 



io\\ 



The excefsof weight, inftead of the ufual lofs, is owing to 

 the difference of oxidizement of the iron, in the (tone and in 

 the refult of the analyfis ; which will be found to be the cafe 

 in all analyfes of thefe fubitances ; indeed it is always necef- 

 fary to reduce the oxide to the red ftate, as being the only 

 one to be depended upon. To avoid future repetition, I (hall 

 alfo obferve, firft, that by preliminary experiments I could 

 not detect any other fubflance than thofe mentioned. Se- 

 condly, that the earth obtained as alumina appeared to me 

 to be moftly, if not entirely, filiceous; becaufe, after it had 

 been ignited, and again treated with potafh and muriatic 

 acid, 1 found it was very nearly all precipitated by evapora- 

 tion. Thirdly, I examined, and judged of, the filica col- 

 lected from the oxide of iron, in the fame way. Fourthly, 

 the weight of the magnefia is given, not immediately, as ob- 

 tained by evaporation, but after a fubfequent folution in an 

 acid, and precipitation by potafli. And, fifthly, the propor- 

 tions are taken from the mean of two analyfes. 



Nothing remains to be examined, of the ftone from Be- 

 nares, except the earthy matter, forming a cement or ma- 

 trix for the fubftances already examined. 100 grains of this 

 matter were, by mechanical means, feparated as perfectly as 

 pomble, from the pyrites, iron, and globular bodies, and ana- 

 Jyfed as above. The mean refult of two analyfes gave, 



Silica ----- 48 

 Magnefia - - - -18 



Oxide of iron 34 



Oxide of nickel - - - %i 



XXXVII. Me* 



