20 THE ROYAL SOCIETY OF CANADA 



analysis, on 0-3392 grnis., was made, and the water was estimated by 



difference. 



Found. Calculated. 



FeO 21-55 21-55 



NiO 8-87 8-94 



CoO 4-57 4-49 



AsA 41-09 41-31 



H^O (Bydiff.) 23-92 23-71 



The figures obtained do not lead to any very simple formula, but 

 agree in a general way with those calculated for R"3(As04)2.SH20; 

 this is the general formula for erythrite (R" = Co, Ni, Fe) and also for 

 forbesite (R" = Ni, Co, Hj), the former being a normal arsenate, and 

 the latter an acid salt, which is described as a substance of greyish- 

 white colour and dull to silky or resinous lustre (Dana, 6th Ed. p. 834). 

 The figures given in the second column above have been calculated for 

 an acid arsenate of general composition R"3(As04)2.7H20, where R" 

 is represented by (Hj, 5Fe, 2Ni, Co) in the proportions here given. 

 This as nearly as possible represents the result of the analysis of the 

 material, though there is no special reason for believing it to be an acid 

 salt, related to forbesite, rather than a normal hydrous salt. The 

 substance is mainly of interest as an unusual oxidation product of smal- 

 tite, and also from its physical resemblance to yukonite. 



The associated scorodite is in the form of almost microscopic crys- 

 tals not more than 0-2 m.m. in diameter; they were identified by 

 measurement on the goniometer and by a determination of the specific 

 gravity. This mineral has not been hitherto recorded from Canada. 



The crystals are pale yellowish, transparent, with a very high 

 lustre; in habit they are prismatic, terminated by pyramid faces, the 

 only forms present being d -j 120 \ and p ^ 111 ^ a common combina- 

 tion with scorodite. The angles as measured are not identical with 

 those usually given for scorodite, as the following axial ratios will show: 



a h c Sp. G. 



Scorodite 0-8658 :1 : 0-9541 3-1—3-3 



Crystals from Cobalt 0-8785 :1 : 1-0555 3-235 



Of these, the ratio a : 6 is probably about correct for the crystals, since 

 the prism angle can be measured with fair accuracy and was observed 

 on four crystals. The length of the vertical axis, c, is more doul)tful, 

 owing to the extremely small size of the pyramid faces. 



It is possil)le that in these crystals, the iron may be in part replaced 

 by cobalt, and this might account for the fact that the interfacial 



