14 Prof. S. U. Pickering. Determinations to test [Dec. 11, 



large. Special sources of error and uncertainty might be pointed oat 

 in each particular case (except that of water, perhaps), but there is one 

 fatal objection to all Person's results, namely, that the heat capacity 

 of both solids and liquids varies considerably with the temperature, 

 and that he mnde his determinations at any temperature which hap- 

 pened to be most convenient, that, for instance, at which the heat 

 capacity of the solid was determined varying between 11 and 280 

 below its melting point. 



One case will be sufficient to illustrate the effect of this. With 

 solid phosphorus determinations have been made by Kopp ('Liebig's 

 Annalen,' Snppl. 3, p. 1) and Regnault ('Ann. Chim. Phys.,' vol. 73, 

 j>. 56 ; and (3) vol. 26, p. 269), and the latter, at any rate, must com- 

 mand as much confidence as Person's (indeed Person adopts some of 

 Regnault's determinations with this substance) ; these results are 



c = 0-2020 at ( + 36 to + 13 = ) + 24- 5 C. . . Kopp, 



r = 0-1895,, ( + 7'15,, + 30-21 = ) + 18-68 .. Regnault, 



c = 0-1783,, (-21 + 7 =)- 7 .. Person, 



c = 0-1699 (-77-75,, +10 =)-33-88 .. Regnault, 



and, taken in their order, they give 1969, 292, 152, and 

 102 for the temperature of no crystallisation,* which results clearly 

 show that no value can be attached to Person's figure, 152. 

 Person's and Regnault's results lie in a fairly straight line which 

 gives c = 0*1979 at the fusing point, and 719 as the temperature 

 of no crystallisation, but it is impossible to accept even this value, 

 as there are not sufficient data for calculating the heat capacity of 

 the liquid at the fusing point. 



C, c, and I should evidently be determined at the same temperature, 

 and this temperature must necessarily be that of fusion (f), but, inas- 

 much as the heat capacities near this temperature may be abnormally 

 high owing to the fusion or solidi6cation being sometimes a gradual 

 process, the determinations should not be made too near the fusing 

 point,t and the only means of ascertaining their true value at this 

 point is to determine them at several different temperatures, and 

 from the rate of change thus obtained to calculate their value at the 

 fusing point itself. This method has been adopted in the present 

 work. 



C = O2045 (98 to 48), I = 5'034, and t = 44'2. 



t The values obtained by Person for beeswax afford a striking instance in point ; 

 from 9 to + 50 the heat capacity of this solid increases from 0'43 to 1'72, exceed- 

 ing at this latter temperature that of the liquid (0'50) by a very large amount. Ice 

 chows a similar increase, but in a much smaller degree (see Person, 'Ann. Chim. 

 Phys.' (3), Tol. 30, p. 80). 



