252 



Mr. J. Joly. 



[Nov. 18, 



vue geographique et chronologiqne qui presentent parfois l'identite la 

 plus complete, de telle sorte qu'il est impossible d'en distinguer les 

 echantillons respectifs."* He adds a list of the few such occurrences 

 known. 



It is an important feature that the value of the thermal constant 

 admits of being approximately, sometimes closely, calculated on an 

 assumed chemical composition, and thus can be used in identification 

 where comparative data do not exist or are not at hand. This fact 

 rests on the experimental basis that the specific heat of the atom is 

 preserved nearly constant through various atomic groupings. This 

 applies indeed, only so far as the specific heats of the elements have 

 been determined in the solid state ; when the generalisation is 

 applied to molecules containing elements whose specific heats as 

 solids have never been directly determined it ceases to be accurate, 

 and it appears at present as if in each kind of grouping a particular 

 thermal capacity must be assigned to the atom. Approximate 

 values of these thermal capacities have, however, been calculated by 

 Kopp and others, and thus it is possible in all cases to pronounce for 

 or against an hypothesis as to the nature of a substance, and this not 

 alone in the case of simple salts, but when dealing with the silicates 

 where many different kinds of atoms are present. 



In the case of the bisilicate beryl, for example, the most probable 

 formula is 5BeO,2Al 2 3 ,H 2 0,llSi0 2 .t Of these molecules the specific 

 heat of BeO has alone not been directly determined. We take 

 for the specific heat of Be 0*430, J and taking the atomic heat of 

 solid oxygen as 4'5 — the value it affords approximately in com- 

 pounds having the constitution RO§ — we assume its sp. h. in 

 the molecule to be - 281. On these assumptions the sp. h. of BeO 

 would be 0"335. The atomic weight of Be is taken as 9*1. Taking 

 advantage, then, of direct determinations as far as possible, the 

 data are, sp. h. of BeO 0'335, of Al 2 O s 0-198, || of H 2 0'501,t of 

 Si0 2 , 0*188.** From these, by an equation similar to Woestyn's — ft 



WS = n^v^ -f- %w 2 s 2 -4- .... , 



where $ v s 2 , . . . are the specific heats of the several kinds of mole- 

 cules as above ; n v n 2 , w%, . . . the numbers and atomic weight of 



# ' G-eologie Experimentale, Constitution des Meteorites,' p. 507. 



f From Penfield's result that the composition is Be 5 ,Al 4 ,Si 11 34 , 'Nature,' vol. 

 30, 1884, p. 378. The ratio is not quite bisilicate. Mr. Penfield thinks, however, 

 that water should be included in the formula. 



X 11° to 100°, Humpidge, 'Eoy. Soc. Proc.,' vol. 39, 1886, p. 8. 



§ Kopp. 



|| From my own results on corundum, agreeing with Regnault's. 



% Person's result on ice, —20° to 0°. Landolt and Bornstein's 'Tabellen.' 



** My own result on quartz, agreeing with Neumann's. 



ft ' Annales de Chimie,' vol. 23, 1848, p. 295. 



