260 



Mr. J. Joly. 



[Nov. 18, 



The experiments are so close that any of the recorded values are 

 fairly in the proportion, and some evidently better than the means ; 

 thus if my own results and Neumann's be alone considered, it will be 

 found that the ratio is almost absolute. As it stands, on the mean 

 values, the products are as 638 to 640. 



If Neumann's result for marcasite, 0*133, be included, we find the 

 more extended proportion — 



Pyrites. Gralena. 

 126-5 : 130 : 133 :: 491 : 506 : 522. 



But in this the last number is a little excessive, which, as observed, 

 is probably due to impurity. 



Two distinct values, very different, are on record for sphalerite : — 



N 0-114 R 0-123 



J 0-115 K 0-120 



and finally the proportion with the extreme values of pyrites and 

 galena obtains. 



Sphalerite. Pyrites. Gralena. 



114 5 : 121-5 :: 126'5 : 133 :: 491 : 522. 



If this proportion be examined by products of consequent and 

 antecedent of first term with sum of antecedents and consequents, 

 the products obtained are as 88,938 to 88,909. One value only, it will 

 be seen, is omitted from the ratios, that of Neumann, for galena. It 

 appears to lie outside the proportion, and may either be erroneous or 

 simply indicative of corresponding terms not yet found for the other 

 sulphides. 



In conclusion, I would ask in reference to the foregoing remarks, 

 if in the present state of our knowledge there is anything a priori 

 improbable in different molecular arrangements or orientations ob- 

 taining under the same crystalline form and affecting to definite 

 extents the thermal freedom of the molecule. Such differences of 

 arrangement, hardly detectable perhaps by any other means of investi- 

 gation, might exist unnoticed except when, as possibly occurs in the 

 case of marcasite, prevailing to a degree competent to determine a 

 different symmetry for the aggregate. 



Regarding the experiments which follow, I have only now to 

 observe on the precautions observed in effecting them. 



The thermometers, certified at Kew, read directly to 0'1° C. ; the 

 place of hundredths being obtained by estimation. 



The balance used was trustworthy to 0-0005 gram. To secure t Y 

 accurately indicating the temperature of the substance, thermometer 

 and specimen were left together in the calorimeter — starting at air 

 temperature— never less than one and a half hour; large specimens, 



