16 Starkweather — RegnauWs Calorie and Our 



over Henrichsen makes the ratio of the mean specific heat 

 from 0° to 100° to that from 0° to 25° equal to 1*025, a num- 

 ber which is markedly larger than has been obtained by any 

 other investigator. 



So far we see that little or no dependence can be placed on 

 the experiments. Rowland's work should be more reliable. 

 His experiments make <? 28 _ 100 equal to 1*0024 c 18 _ 28 , or reduced 

 to <? 1B as a unit, <? 28 _ 100 is 1*004:9 c 1B , which makes tf _ 100 equal 

 to 1*0034 c lb . He mentions two other determinations and 

 gives as the mean of all three that <? _ 100 is equal to 1*0041 <? 16 . 



The very recent work of Ludin was performed with great 

 care. He gives for c _ 100 the value 1*00543 c 1% . 



Joly's determination of <? _ 100 was obtained by the use of his 

 steam-calorimeter, and gives for its value 9957 c 16 . Concern- 

 ing the accuracy of this instrument we know nothing, and no 

 details are given. He himself says he obtained much larger 

 values in earlier experiments, but he considers this to be more 

 accurate. 



We have now left the determinations of Dieterici and 

 Reynolds. These were both by obtaining the mechanical 

 equivalent. Dieterici found by electrical methods (and using 

 the ice calorimeter) that c _ 100 was equal to 42*436 megalergs. 

 Griffiths* has pointed out that he made the error of using the 

 " legal" instead of the "true" ohm, which reduces his value 

 to 42*33. There is also a constant errorf in experiments made 

 by electrical methods of about one-quarter of one per cent due 

 to the determination of the electrical standard of electromotive 

 force, which reduces this to 42*225. This makes the mean 

 specific heat between 0° and 100° equal to 1*00799 c 1B . The 

 experiments of Dieterici were performed with great care, were 

 sixteen in number, and deviated from the mean at most 0*37 

 per cent, the probable error being 0*05 per cent. But they 

 have one fault, that there is lacking a conclusive determination 

 of the quantity of quicksilver displaced in the ice-calorimeter 

 by a quantity of heat equal to the mean specific heat from 0° 

 to 100°. Bunsen has found it to be 15*41 milligrams, Schueller 

 and Wartha 15*442, and Yelten 15*47 ; Dieterici takes the 

 second value. Velten's values being uncertain, as has been 

 shown, it seems as if less weight should be given to his deter- 

 mination, which would still further reduce Dieterici's value for 

 the mean specific heat from 0° to 100°. 



In total opposition to Dieterici stand the recent results of 

 Reynolds. According to him the mechanical equivalent of the 

 mean specific heat from 0° to 100° is 41*83 megalergs, which is 



* Phil. Mag., xl, p. 446, 1895. 



fPhil. Mag., xl, p. 449, 1895; Proc. Roy. Soc, lxi, p. 479, 1897, and Johns 

 Hopkins Univ. Circular, No. 135, June, 1898, p. 54. 



