1896.] Scientific Instruments used by the late Dr. Joule. 



351 



foot-pounds at Greenwich at a temperature of 18'6°, but this assumes 

 the B.A. ohm (104*8 centimeters of mercury) to be correct. If we 

 take the present unit (106'3 centimeters of mercury), this reduces 

 to 771*4, which agrees very fairly with Joule's last value, 772'55 at 

 16*5°, obtained from the friction of water. If, however, Joule's 

 current meter measured wrongly, as now seems probable, by about 

 2*5 parts in a thousand, it would seem that the equivalent of heat 

 derived by Joule from the electrical method must be taken as 7753, 

 and Joule's results therefore seem to agree with those of more recent 

 experimenters in pointing to a systematic discrepancy between the 

 value obtained for J by mechanical methods on the one hand and 

 electrical methods on the other. The thermometer used in the experi- 

 ments has been examined by Professor Schuster, and his result (' Phil. 

 Mag.,' June, 1895) is that if Tj represent the reading on the Joule 

 thermometer, and T N the reading on the nitrogen thermometer of the 

 Bureau International, then Ty = T N (1 + 0*0024). Joule's value for the 

 equivalent then from the electrical method is 777*2 foot-pounds at 

 Greenwich on the scale of the "Paris" nitrogen thermometer at a 

 temperature of 18*6°. Professor Schuster and Mr. Gannon, at the 

 close of their paper on " The Determination of the Specific Heat of 

 Water in Terms of the International Electric Units " (' Phil. Trans.,' 

 vol. 186, 1895, A, pp. 415 — 467), compare the values obtained for J 

 by different experimenters, and give a table of equivalents in foot- 

 pounds at Greenwich at 19* 1°, referred to the " Paris " nitrogen 

 thermometer. 



Joule. Rowland. Griffiths. Schuster and Gannon. 



774 776-1 779-1 778*5 



Introducing the value obtained by Micelescu, which is given by 

 Schuster and Gannon as 776*6 at 15°, and which, if we take as the 

 temperature correction the mean of the values given by Rowland and 

 Mr. E. H. Griffiths (' Phil. Trans.,' vol. 184, 1893, A, p. 361), reduces 

 to 775*1 at 19*1°, the table becomes : — 



Mechanical method. 



r 



Joule. 



774 



Rowland. 



776*1 



— : ^ 



Miculescu. 



775-1 



Electrical method. 



Griffiths. Schuster & Gannon. Joule. 



779*1 778-5 777*1 



The first three numbers were obtained by mechanical methods, the 

 last three by electrical methods. 



Professor Schuster and Mr. Gannon remark that Rowland's num- 

 ber referred to the Paris nitrogen thermometer would probably be 

 smaller by one unit ; this being the case, the higher values given by 

 the electrical as compared with the mechanical methods are very 

 noticeable. 



