76 PROCEEDINGS OF THE AMERICAN ACADEMY 



the history of the subject carefully, we find that the only experimenter 

 who has made the determination with anything like the accuracy 

 demanded by modern science, and by a method capable of giving 

 good results, is Joule, whose determination of thirty years ago, con- 

 firmed by some recent results, to-day stands almost, if not quite, alone 

 among accurate results on the subject. 



But Joule experimented on water of one temperature only, and 

 did not reduce his results to the air thermometer ; so that we are still 

 left in doubt, even to the extent of one per cent, as to the value of the 

 equivalent on the air thermometer. 



The reduction of the mercurial to the air thermometer, and thence 

 to the absolute scale, has generally been neglected between 0° and 

 100° by most physicists, though it is known that they differ several 

 tenths of a degree at the 45° point. In calorimetric researches this 

 may produce an error of over one, and even approaching two per cent, 

 especially when a Geissler thermometer is used, which is the worst in 

 this respect of any that I have experimented on ; and small intervals 

 on the mercurial thermometers differ among themselves more than 

 one per cent from the difference of the glass used in them. 



Again, as water is necessarily the liquid used in calorimeters, its 

 variation of specific heat with the temperature is a very important 

 factor in the determination of the equivalent. Strange as it may 

 appear, we may be said to know almost nothing about the variation 

 of the specific heat of water with the temperature between 0° and 

 100° C. 



Regnault experimented only above 100° C. The experiments of 

 Hirn, and of Jamin and Amaury, are absurd, from the amount of 

 variation which they give. Pfaundler and Plattner confined them- 

 selves to points between 0° and 13°. Munchausen seems to have 

 made the best experiments, but they must be rejected because he did 

 not reduce to the air thermometer. 



In the present series of researches, I have sought, firstly, a method 

 of measuring temperatures on the perfect gas thermometer with an 

 accuracy scarcely hitherto attempted, and to this end have made an 

 extended study of the deviation of ordinary thermometers from the 

 air thermometer ; and, secondly, I have sought a method of determin- 

 ing the mechanical equivalent of heat so accurate, and of so extended 

 a range, that the variation of the specific heat of water should follow 

 from the experiments alone. 



As to whether or not these have been accomplished, the following 

 pages will show. The curious result that the specific heat of water 



