^o New Comlinalion difcovend in Zaffer, 



reduced to one-half it became turbid : the e? aporation being 

 continued till there remained only one- third of the liquid. 

 It was taken from the fire, and it then depofited a white 

 matter, which was colle<Sled on the filter. The liquor which 

 pafl'ed through had a yellow colour and a tafte fenfibly acid, 

 and exhibited the fame phaenomena in every thing as the 

 cobaltic acid obtained by the procefs before indicated. 



The following are fome of, its properties, which Brugnatelli 

 confiders as chara^leriftic : 



ill. It precipitates a folution of filver. 



2d, It precipitates lime water in a white coagulum, info- 

 luble in water and in exccfs of acid. 



3d, It can be feparated by alcohol from its aqueous folution. 



4th, It precipitates acetite and muriate of barytes. 



I fliall now dcfcrlbe the experiments I made on this fub- 

 jeil, and I fubmit the confequences I have deduced from 

 them to the examination of the Inftitute. 



Exp. I. I formed ammoniure of cobalt by fuffering am- 

 monia to macerate, in the fun, over zafi'er. I took care to 

 ftir the mixture, which was contained in a matrafs : it foon 

 aflumed a red colour, and, forty-eight hours after, a bright 

 red colour. I remarked in this liquid a pretty abundant 

 cryftallization under the form of white and brilliant needles. 

 This cryftallization was permanent until the fun had again 

 heated the liquid. It then difappeared, and diifolved in the 

 liquor. In this ftate the ammoniure of cobalt was filtered^ 

 and put into a retort : as it was heated it aftumed a violet red 

 cclour, which. became darker till it aflumed a beautiful red 

 wine colour. When the greater part of the ammonia was 

 evaporated the mixture acquired a greenifh colour, and, by 

 reft, there was precipitated a matter of the fame colour as 

 the folution. The liquid being then filtered while warm, 

 and evaporated to drynefs, there were obtained the rudiments 

 of cryftals, the form of which could not be determined. 

 White and brilliant particles were remarked. The reft of 

 the matter had a yellowifli colour. I poured water over this 

 rcfiduum, ftirring it with a platina fpatula. The part leafl: 

 coloured dilfolvcd entirely, and communicated to the water 

 a ftraw colour. This liquor was acid, and poflefled fome of 

 the properties announced by Brugnatelli. I fhall defcribe 

 hereafter the experiments to which I fubjefted it, and the 

 refults of which, caretuUy examined, prove that it was not 

 an acid formed by cobalt. The two rcfiduums obtained in 

 this experiment, one of a green and the other of a yellowifh 

 rolour, were found nyt to be pure oxide of cobalt, asBrugna- 

 fXW favs, but a combination of that oxide with arfenic acid. 



