436 Mr. E. H. Griffiths on 



evidence does not justify the use of the formula given hy 

 them. This adverse opinion has been strengthened by some 

 charts, brought under my notice by Mr. Spencer Pickering, 

 F.R.S., which show with great clearness the discrepancy 

 between the results obtained by those observers when using 

 different methods of experiment. It is, however, noticeable 

 that in addition to using thermometers standardized by the 

 Bureau International, they also used a thermometer standar- 

 dized by Rowland, and no systematic difference is observable 

 in the conclusions drawn from the employment of the different 

 standards. It would therefore appear as if the discrepancy 

 between their conclusions and those of Rowland could not be 

 explained, as is usually the case in such investigations, by a 

 difference in the thermometric standards employed, but that 

 it must arise from experimental errors. 



A further indirect comparison of Rowland's air-thermometer 

 with the nitrogen standard of the Bureau International is 

 given in Professor Schuster's paper on the scale value of 

 Dr. Joule's thermometers*. The result indicates a difference 

 of about 0°*05 in elevation over the range 10° to 30° C, 

 although near the bottom and top of the range the difference 

 alters but slightly. 



1 st Series. 2nd Series. 



At 10° it is about . . '025 — 



„ 15° „ . . -045 -045 



25° „ . . -038 



„ 20° „ . . -042 -051 



» 



These results, however, are dependent upon previous obser- 

 vations made by Joule when comparing his thermometer with 

 Rowland's. As we have no details concerning this comparison, 

 Professor Schuster points out that " it would be necessary to 

 have further information before any definite conclusions can 

 be drawn" [ibid. p. 499). Assuming the validity of the above 

 corrections, they would indicate that Rowland's rate of decrease 

 in the capacity for heat of water over the above range is 

 slightly too large. 



The most recent determination with which I am acquainted 

 is my own completed in 1892. The method adopted was 

 an electrical one, and singularly suitable for an investi- 

 gation into the changes in the thermal capacity of water. 

 Errors in the assumptions as to the magnitude of the 

 various constants which would affect the "value of J would 

 leave unaltered the relative thermal capacities of water at 

 different temperatures. 



* Phil. Mag., June 1895. 



