Ne-w Comhinaiion dlfcovered in gaffer. 51 



JE.r/). II. Having prepared a new quantity of ammoniure 

 oF cobalt, as in the firft experiment, I fubjcfted it ta fpon- 

 taneous evaporation in a pldce where the fun accelerated the 

 volaiiiization of the ammonia. In proportion as the licjuid 

 evaporated, there was precipitated a flaky matter of a whitifli 

 rofe colour, wliich fenliblv increafed till the ammonia was 

 evaporated. T!ie hquor emitted no odour, but it had re- 

 tained a fuffioiently beau'iful rofe colour. The Hquor was 

 again evaporated to drynef?, and the remaining matter was 

 diflbh'ed in diliitlcd water. This foiution, of a light rofe co- 

 lour, was acid, and poflefTed the properties of the acid obtained 

 in the (irit experiment. 



The reiiduum left on the filter was carefully examined^ and 

 found to be arfeniate of cobalt. 



Fixp. III. I took, as the author of the memoir did himfelf, 

 a kilogramme of zafler, which was reduced to an impalpable 

 powder. I boiled it, for half. an hour, in three litres of di- 

 Itilled water. The liquor, which I filtered while warm, had 

 a light colour and a fenfible favour; and, being evaporated in 

 a porcelain capfnle, the liquid, towards the end of the opera- 

 tion, became turbid. I then continued till there remained only 

 about a heftogramme of the liquor. By cooling i obtained 

 a cryftallizatlon in needles. Having filtered the mixtures, 

 (lirring them, the cryitals remained on the filter, and the fu- 

 pernataut liquor pafl'ed through, very tranfparent, of a bright 

 yellow colour. Brugn^atelli, in making this experiment, did 

 not remark the cryllallization ; he only obfervcd a white 

 matter, which aflumed a rofe colour by coming into conta(^ 

 ■with the air, and which ho found to be oxide of cobalt. The 

 needle-formed cryflallization which I obtained produced, as 

 foon as it was heated, a confiderable dlfcngagement of arfeniq- 

 When carefully examined it was found to be arfeniate of 

 cobalt, and not pure oxide, as had been announced. The 

 fad is of little importance, no doubt, in regard to the dif- 

 covery of Brugnatelii ; but it may ferve to fupport my opi- 

 nion, if the refults of my experiments were not more than 

 fufficient to prove that the cobaltic acid does not cxift. 



The liquors of the firft, fecond, and third experiments, 

 found to be acid, and of the lame nature, were fubjeded to 

 the following proofs, and comparatively with the arfenic acid: 



I ft, This lifjuor was precipitated of a yellow colour, like 

 that of orjiiment or fulphuret of arlenic, l)y lulpliuratcd hy- 

 drogen and alkaline hydrofulphurets. This precipitate was 

 indeed fiilphurfjt of arlenic, and not fulphur precipitated, as 

 believed by Brugnatclli. 



2d, It prccipituies ammoniure of copper of a blueifh green. 

 D % This 



