1886.] 



On the Specific Heats of Minerals. 



261 



two to three hours ; and the largest were reserved for leaving in all 

 night. 



Specimens presenting at all a porous or loose surface were carefully 

 dried, often for many days, in an oven, and subsequently preserved 

 over calcium chloride. Porous or loose substances, not of special 

 value, were, for greater surety in drying, broken into small bits — 

 as in the case of the few rock specimens investigated. To preserve 

 these dry in the calorimeter, while acquiring air temperature, 

 calcium chloride was in some cases placed in the little chamber pro- 

 jecting from the lower part of the sphere. If counterpoised imme- 

 diately on being placed in the calorimeter absorption of moisture may 

 evidently be detected, weighed, and its thermal capacity allowed for. 

 This plan was on a couple of occasions resorted to. In any cases 

 observed, however, entire neglect of this absorption would have pro- 

 duced but small error. 



Supposing errors of the same sign to accumulate through an 

 experiment, I do not think they can have in any case falsified the 

 value of the sp. h. above 0*4 per cent. This seems probable, too, from 

 the agreement observed on repetition. 



Allowance has in all cases been made for the variation of X with 

 £ 2 , according to Appendix of last paper. The effect of the low 

 specific gravity of steam on the apparent weight of condensation has 

 W 



been allowed for, as — = X 0'62 gram. With these deductions and 



Sp. gr. 



allowance for the precipitation on the carrier and bucket according 

 to range of temperature obtaining, the value of w is deduced as 

 recorded in the tables. 



The specific gravities recorded have been reduced to water at 4° C. 



