32 PROCEEDINGS OP THE AMERICAN ACADEMY 



Tliese were first analyzed and described in my laboratory by Sadtler, 

 who found for the luteocobalt salts the formula, — 



Co,(NH3),,(NO,)„+Co_,(NO,),+OH,. 



In resuming the study of this subject, I found the method of prepara- 

 tion at first adopted somewhat uncertain, because the excess of the alka- 

 line nitrite acts readily upon the cobaltamine, forming other products 

 not always easj' to sejjarate. The following method gives better re- 

 sults. One molecule of any soluble salt of the cobaltamine — the 

 nitrates are to be preferred — is to be dissolved with two molecules of 

 cobaltic chloride or nitrate, and a little acetic acid. A solution con- 

 taining as nearly as possible twelve molecules of sodic nitrite is then to 

 be added. The luteocobalt salt is precipitated almost immediately ; 

 the corresponding salt of roseocobalt after a short time. 



The salt of luteocobalt obtained in this way is a yellow crystalline 

 body very slightly soluble in cold water, and easily purified by wash- 

 ing. Boiling water dissolves it in very small quantity, giving a j^ale 

 yellow solution. I have usually obtained it in rather larger deep 

 orange granular crystals by adding a solution of the corresponding 

 much more soluble roseocobaltic salt, Co2(NIJ„)j^,(NO.,)g-|-Co2(N02)g, 

 to a hot neutral solution of nitrate of luteocobalt. It may also be 

 formed by adding a solution of Co,,(N02)i;-|-6NaN02 in excess of sodic 

 nitrite to a solution of luteocobalt, there being in this case less danger 

 of the formation of other products than with the other cobaltamines. 

 I find the formula of this salt to be 



lCo,(NH3),,nCo,(NO,),,| or Co,(NH3),,(NO,),+Co,(NO,)„ 



as the following analyses, made with three different preparations, 

 clearly show : — 



0-3312 gr. gave 0-2006 gr. SO, Co = 23-789^ cobalt. 

 0-3776 gr. gave 0-236'J gr. SO,Co = 23-89% cobalt. 

 0-1785 gr. gave 0-1115 gr. SO,Co = 23-78% cobalt. 



The formula requires 23-79% cobalt for the anh^-drous salt. In 

 one preparation, however, I obtained a salt in which 



0-4291 lost 0-0087 gr. at 125° C. = 2-16. 



This would corrcs])ond to Sadtler's formula, which requires 1-78%, if 

 we consider a part of the water as hygroscopic, or that, as is more proba- 

 ble, there was a slight decomposition. In the dried salt, — 



0-4204 gr. gave 0-2643 gr. SO.Co = 23-93% cobalt. 



