40 PROCEEDINGS OF THE AMERICAN ACADEMY 



1-4510 gr. gave 59-5 c.c. oxygen at 22*^ C. and 758 """ = 5-25% 

 oxygen. 



This would correspond to one atom of oxygen for the formula given 

 above. Fremy also determined the amount of oxygen given otF by the 

 action of dilute sulphuric acid, and obtained as a mean of two analyses 

 r)-20%. So far as can be judged from the analyses above, the formula 

 which I have given deserves the preference. On the other hand, 

 Fremy's formula, which I double for the sake of comparison, 



Co,(NH3),„0,(N03),+20H„ 



is perhaps somewhat simpler, and explains all the known reactions at 

 least equally well. The corresponding structural formulas may be 

 written, — 



Co, 



NH3— NFI3— NO3 



N H3— NII3— NHg— NO3 



O, C02 



N H,— NH3— NH3— NO3 



NH3— ^H3— NO3 



NH,— NH3— NO, 



NII3— NH3— Nri3— N ), 



O., 



NH,— N H, — N 1 1 .,— NOg 



NIL— N a, —no! 



I adopt provisionally the formula which agrees best with the analy- 

 ses, fully recognizing the possibility that the other may prove correct. 

 In this connection I may mention that when iodine is added to a solu- 

 tion containing cobaltic nitrate, ammonic nitrate, and ammonia, an 

 olive-green crystalline precipitate is thrown down which contains 

 iodine, and which may prove to be either the iodide corresponding to 

 the nitrate above discussed, or an iodo-nitrate corresponding to the 

 oxy-nitrate. The formation of this nitrate may be expressed by the 

 equation, — 



4Co(N03),+24NIl3+20H,+80 = 2.|Co,(NH3),,.0,.(N03) J 

 +20(N1I,),. 



It is remarkable that the action of iodine upon an animoniacal solu- 

 tion of cobaltic nitrate is not analogous to its action upon an amnioiii- 

 acal solution of the sulphate. When the nitrate is heated with clilor- 

 hydric acid, a small quantity of cliloride of purpureocobalt is, as I 

 find, always formed. We may have the reaction expressed by the 

 equation : — 



Co,(NIL)„0,(N03),+12IICl = Co,(NIl3),„Cl,+4N03lI+ 

 40II,+6C1 ; 



but the greater part of the nitrate is decomposed^ The formation of 

 the chloride of puipureocobalt under the circumstances is, I think, a 



