262 W. Gibbs and F. A, Genth 
and second prism, and are larger than those which separate from 
pure solutions. From these measurements, it a aoe pears that the { 
chlorid of Purpureocobalt is isomorphous with the sulphate of | 
Roseocobalt. ‘This isomorphism is the more remarkable, imas- 
much as a precisely similar case occurs with the chlorid and sul- ’ 
hate of Luteocobalt, between which there is a similar difference 
of constitution. Thus we have 
5NHs.Co2Clzs=5NHs.Co20s, 8S03:+5HO 
6NHs.CozCls=6NHs3. Co20s3, 8503;+5HO. 
From this it appears.that in both cases we have the erystallo- ‘ 
graphic equallity re 
38C1=Os, 3803+5HO. : 
The density of the crystals of chlorid of Purpureocobalt, as | 
taken in alcohol, is 1802 at 28°-C.; the atomic volume of ‘the | 
chlorid is consequently 139° 0. | 
The chlorid of Purpureocobalt has the formula 
NHs.Co2Cls 
as appears from the following scien : | 
0°4962 gers. gave 0°3073 grs. | 
0°8514 grs. gave 0°5279 grs, . on meas: Hy . 
1-4116 grs. gave 08782 grs. “ = 2355 “ « 
0°7550 grs. gave 0°4105 ors. wae = 6°04 per cent hydrogen. 
06116 grs. gave 0°3412 grs. : = 619 “ “ 
0°6124 grs. gave 0°3865 grs, - = 610 “ 
14184 grs. gave 0°7800 grs. . om $41:.% 
16636 grs, gave 2°8540 grs. _chlorid of silver = 49-40 per cent chlorine. 
0°4754 grs. gave 0°8200 grs, ® “- {mit 4Sae..-.* : Sl 
0°1966 grs. gave 0°3365 grs. ” = rent ~ x 
04972 grs. gave 0°8553 or . 
05965 gre. gave 149 «.«. nitrogen at 18° C. and ramet (at 18° 0) =11819 
c.c. at 0° and "60m = 28°05 per ce 
0°9042 gre zave ¢ 124 c.o.nt nitrogen at 19° C. pa T66mm-56 (at 19°-4 0.) = 12896 
93 per cent nitr 
06036 rs re 16816 cc. “utrogen aT 752m™-60 and 15° - — += 
= 135-08 ¢.c, at 0° and 760™™ — 28-11 per ce 
0°5755 gre. ae 16039 ¢. c. nitrogen at 771mm- 8 and 15° O, 2 = 119mm, t= 
° C, = 12886 ¢. c. at 0° and 7 760"™ == 812 per cent nitrogen. 
The nitrogen in these as in all our Sas was moist. 
Hence we have 
Eqs. Theory. Mean. Found. 
Cobalt, 2 59 2855 2856 2358 2357 2855 2955 
Chlorine, 8 1065 4250 4243 4249 4231 4952 42-40 
ha is “ts 598 611 604 - 619 610 ° 61l 
N 5 0 2797 2801 2805 2793 9812 2811 
2505 10000 100-11 
The agreement of these analyses leaves no reasonable doubt 
that the true formula of ee chlorid of P tis 5NHs. 
Co2Cls, as first correc y determined by gojski, and subse- 
