Rowland's Value of the Mechanical Equivalent of Heat. 169 



be partly due to the nature of the glass used in these ther- 

 mometers, which seem more dependent upon their previous 

 use than the hard glass thermometers. The depression of the 

 zero in these thermometers is large. Thus in one experiment 

 with 6163 the zero at the beginning, after the thermometer 

 had been at about 20° G. for several weeks, was 60*11 millim. ; 

 after heating to 40° G. the zero was 5977, a depression 

 corresponding to '0'68° G. 



A similar experiment with Baudin (No. 6165), which was 

 exactly similar to these two thermometers, gave a depression 

 of -021° C. after being raised to 30° G. 



It will be observed that from 14° to 25° the range on the 

 two air scales does not differ by more than a few thousandths 

 of a degree. We must therefore conclude that the difference 

 between the mechanical and electrical determinations of the 

 mechanical equivalent of heat cannot be accounted for by 

 differences in standards of thermometry, but must be sought 

 in the energy determinations. 



XXA^. ^L Recalculation of Rowland's Value of the Mechanical 

 Equivalent of Heat, in terms of the Paris Hydrogen-Ther- 

 mometer. By W. S. Day*. 



THE measurement of the mechanical equivalent of heat 

 made by Rowland in 1877-79 (Proc. Am. Acad. xv. 

 p. 75, 1879) is probably the best one in which the heat was 

 produced by the expenditure of mechanical energy. Later 

 determinations made with great care, in which the heat was 

 produced by the expenditure of electrical energy, give results 

 higher by about one part in four hundred. Rowland's mea- 

 surement of temperature was based on comparisons made 

 between an air-thermometer and three Baudin mercurial 

 thermometers, by which he reduced his measurements to the 

 absolute thermodynamic scale. It was the object of the 

 present investigation to compare his thermometers with the 

 hydrogen scale of the International Bureau of Weights and 

 Measures, at Sevres, near Paris, and make a recalculation of 

 his value of the mechanical equivalent accordingly. 



For this purpose, three Tonnelot thermometers which had 

 been carefully studied at the International Bureau, and com- 

 pared with their standards at several points of the scale, were 

 obtained and compared with the three principal thermometers 

 used by Rowland in his experiment. These comparisons 

 were made in a horizontal comparison tank, designed and 



* From the Johns Hopkins University Circulais, June 1897. 

 Phil. Mag. 8. 5. Vol. 44. No. 267'. August 1897. N 



