W. Gibbs on the Hexatomic compounds of Cobalt. 189 
This apparatus is especially useful where it is desired to ob- 
tain reliable results in a comparatively short time. As before 
mentioned, it was designed more especially for the analysis of 
illuminating gas. It will be noticed that the explosion pipette 
of this apparatus is nearly the same as the Cavendish eudiom- 
eter; the chief difference being in the manner of introducing 
and expelling the gas. 
Office of Gas Inspection, Boston, May 14, 1874. 
Arr. XVIII. — Researches on the Hexatomic compounds of Cobalt ; 
by Woxcorr Gisss, M.D. 
[Continued from vol. vi, p. 116, August, 187 3.] 
5. THE salts described by Fremy* under the names of chlo- 
ride, nitrate and sulphate of fuscocobalt contain also eight atoms 
of ammonia, and may be regarded as belonging to the octamin 
series. These salts have, according to.Fremy, respectively the 
formulas : ‘ 
Co,(NH,),-0-Cl,+30H, 
Co,(NH,),-0.(NO,),+30H, 
: Co,(NH;), .0.(50,4)2+40H, 
in modern notation. They are brown resinous masses, are dif- 
ficult to obtain in a state of purity, and have as yet been but 
little studied. If we admit that the formulas are accurate, we 
may write them in accordance with the theoretic views which 
I have adopted, as follows :+ 
NH,—C (NH,—NO NH, 
SG a cn| NE =itvo ool MENS 
—NH po i 
Co, NH,—NH.>° Co, ¢ NH.—NH,-? Co NH.—NH,-? 
NH,—Cl NH,;—NO, NHws0, 
NH,—-Cl | NH,—NO, | NH; 
Jérgensent suggests that these salts may contain hydroxyl in 
place of ae : There is at present no method of deciding 
structural formula: 
Co, NH,—NH, 
* Ann. de Chimie et de Physique, [3,] tome xxxv, p. 257. ; 
+ Blomstrand has given the sam Man with trifling variations. Chemie der 
