16 THE ROYAL SOCIETY OF CANADA 



1. 2. 



CaO 10.00 10.14 



FcA ^^•'^•72 36.81 



AsA 34.06 33.83 



A more recent analysis made of another specimen since supplied by 

 Mr. T}Trell yielded results which are essentially similar,thoughthe As^Og. 

 was higher and the Fe-^O., lower than in the above. The differences 

 observed are no doubt due to small amounts of foreign admixtures in 

 the samples taken — either of the yellow ochreous material in which the 

 mineral is embedded, or of thin films of the green mineral referred to 

 below (symplesite) . 



The total water was determined by Penfield's method of direct 

 weighing, and also by a])S()rption. Some difficulty was at first ex- 

 perienced owing to the fact that the mineral, even when finely powdered ^ 

 "spits" and flies about a lot on warming, and it further decomposes 

 on heating more strongly, losing, in addition to water, a part or the whole 

 of its arsenic which deposits on the cooler parts of the apparatus as 

 transparent octahedra of ASjO,. 



Total water, by Penfield's method 17 .57% 



" " by absorption method 20 .28% 



The latter is regarded as the more correct; in this determination 

 a long tube was introduced between that containing the heated mineral 

 and the absorption tube, and the water was drawn off from this by 

 suction at a low temperature; the arsenious oxide which volatilised 

 during the ignition was in this way deposited and left behind. The 

 lower result by the Penfield method may be due to a slight loss, by 

 evaporation, of the expelled water before weighing, and in any case 

 the two determinations are not strictly comparable having been made 

 on two different samples. 



The amount of water lost at 100° C and at 110° C was obtained by 

 noting the loss experienced by weighed amounts kept at these tem- 

 peratures for two or three hours; and in one experiment by heating 

 in a vacuum over phosphorus pentoxide on a water bath for 2^ hours 

 till constant. 



Loss after 5 hours on water l)ath in vacuo over P^Og . . 11. 96% 



" heating for 2 hours at 100°C 10 . 52% 



" 2 " " 110°C 10.91% 



" 4 " " 100°C 11.00% 



The fact that only al)0ut one half of the total water is driven off 

 at 100°C, and that there is comparatively little further loss even after 

 prolonged heating at 160°C, indicates that a part only of the water is 



