2 Art. 2.— T. Suzuki : The Change of 



residue treated with a little water, filtered from some calcium 

 sulphate, aud left in the desiccator again to crystallise. The salt 

 is too changeable to be safely recrystallised. It forms small very 

 dark red crystals, ve]y soluble in water, and remarkable for being 

 alkaline to litmus. 



To anal^'se the salt, the cobalt Avas precipitated as cobaltous 

 ammonium phosphate and weighed as cobaltous pyrophosphate 

 (DiRVELL and Claeke's process as given in Dittmar's Quantilaiice 

 Analysis, p. 309). The little cobalt left in solution was obtained 

 by adding a drop of ammonium sulphide, and burning the cobalt 

 sulphide to the black oxide, C03O4. The cobaltic cobalt in the salt 

 was separately determined by digesting the salt with a con- 

 centrated warm solution of sodium carbonate and treating the 

 residue twice Avith fresh sodium carbonate, the last time bringing 

 the whole to boiling. The washed residue was dissolved in dilute 

 hydrochloric acid holding potassium iodide in solution and the 

 cobalticum estimated iodometrically. The nitrite of the salt, 

 which had been converted by the sodium carbonate into sodium 

 nitrite, was estimated by the permanganate method. 



The results of analysis are expressed by the formula, 



Co'i.Co' ; (N0.,)3,(0H)i„ 34 H,0, 



in which Co:N = 2:3 and the ratio of the valencies of the cobalt to 

 those of the nitrogen atoms are as slightly more than 3:2. The 

 tabulated percentage numbers are — total cobalt,* f. 34-13, c. 34-26; 

 cobaltic cobalt, f. 10-42, c. 10-28; nitrogen, f. 12-31, c. 12-21. 

 The formula calculated for the salt is resolvable into 



2 Co", Co'"., (NO,)k(OH\+ Co",Co"'.,(NO,)io(OH)s + 34H,0, 



* The letters, "f," and "c." stand for "found '"and " calculated." 



