MECHANICAL EQUIVALENT OF HEAT. 131 



correction detracts greatly from the independent authority 

 of the determinations by these two physicists. If it could 

 be shown that an error existed in the thermometry of 

 Rowland, the resulting correction would have to be applied 

 not only to his values but to those of Joule. The methods 

 employed by Rowland were otherwise so perfect that if 

 sufficient evidence could be produced to finally establish his 

 temperature measurements I, for one, should be inclined 

 to accept without reserve his conclusions as to the value 

 of the mechanical equivalent. 



Happily there is a good prospect of more light on this 

 point. Professor Schuster has for some time been engaged 

 in a separate determination of the mechanical equivalent, 

 and I anticipate with much interest the publication of his 

 results. At the same time he is (I believe) re-standardising 

 the Joule thermometers, and in my opinion the importance 

 of this portion of his investigation may prove as great as 

 the achievement of a revaluation of J. If, as I trust may 

 be the case, he is led to the conclusion that the temperature 

 range given by the previous standardisation is faulty, we 

 shall have the means of applying to the results of Rowland 

 and Joule an approximate correction which may bring them 

 into greater harmony with the latest investigation. 1 I indi- 

 cate later my reasons for believing that the discrepancy 

 between the results obtained by Rowland and myself has 

 its origin in thermometric errors. I have applied to my 

 own thermometry every test which appears to me to be 

 possible ; it remains to be seen whether Rowland's tempera- 

 ture range is capable of such a scrutiny, and I am glad to 

 think that this important investigation is in the able hands 

 of Professor Schuster. 



With regard to the temperature work of M. Miculescu 

 he adopted a thermo-electric couple of platinum and iron, 



1 No doubt changes have taken place in the Joule thermometers with 

 lapse of time. Although such time changes may have caused consider- 

 able alteration in the position of the zero points, it is improbable that they 

 will affect determinations of differences of temperature when made by the 

 same thermometer. 



