10 PROCEEDINGS OF THE AMERICAN ACADEMY 



salt is almost completely decomposed. It is remarkably insoluble in 

 cold water. For analysis the salt was dried on bibulous paper, and 

 afterward in pleno over sulphuric acid. The analysis was effected by 

 boiling the salt with zinc as in the case of the jjlatinum salt, and 

 determining the chlorine in the filtrate by silver, but the gold precipi- 

 tated was found to contain much metallic cobalt. Another analysis 

 was made by simply heating with sulphuric acid, precisely as in the 

 process which Genth and myself introduced for the determiuatioji of 

 cobalt in these salts, and then washing and weighing the gold. This 

 method was found to give excellent results. In this manner, 



0-4791 gr. gave 0*3697 gr. silver= 25*36 per cent chlorine. 



0*3942 gr. gave 0*2483 gr. Au-|- CoSO^ and 0*1391 gr. gold, and by 



difference 0*1092 gr. CoSO^= 10*54 per cent cobalt and 35*30 per 



cent gold. , 



Calculated. Found. 



Cobalt 2 10*57 10*54 



Gold 2 35*30 35*30 



Chlorine S 25*44 25-36 



The formula of the salt is therefore : — 



Co.(NH3)8(NO,),Cl2 + 2AUCI3. 



Erdmami's Salt. — O. L. Erdmann,* in 1866, described a remark- 

 able salt which is formed when a solution of potassic nitrite is added 

 to a solution of cobaltic chloride containing an excess of ammonic 

 chloride. The liquid quickly assumes a dark orange color, becomes 

 strongly acid and evolves red vapors. Alter a time very beautiful 

 oblique rhombic crystals are deposited, which, according to Erdmann, 

 have the formula (old style) 



C02O32NH3, 3NO3 -f KCNOg. 



In modern notation the formula of Erdmann's salt may be written : — 



Co,(NH3),(NO,)3K,. 



Erdmann states that the potassium in this salt may be replaced by 

 other metals, and describes the corresponding ammonium and silver 

 salts. These compounds are especially interesting because they hold 

 an. intermediate position between the two series represented respec- 

 tively by the terms 



Co2(NH3)i2C]s and Co^CNOJiJ^e- 

 * Journal fur prakt. Cliemie, xcvii. 385. 



