2 Messrs. C. W. Waidner and F. Mallory on 



of Joule's thermometers with a thermometer standarized at 

 the Bureau International, an indirect connexion was obtained 

 between Rowland's Air-scale and the Nitrogen-scale of the 

 Bureau International. This comparison pointed to differences 

 in the two scales as great as 0°05 C. Inasmuch as the details 

 of the method employed by Joule in his comparison are not 

 known, Professor Ames suffsested that another more direct 

 comparison of Professor Rowland's thermometers was 

 desirable. 



Method of Comparison. 



As we wished to use the results of this experiment to 

 reduce Rowland's values of the equivalent to the scales used 

 by experimenters employing the electrical methods, and thus 

 render the results comparable, it was deemed best to make the 

 comparisons of these thermometers under conditions as nearly 

 as possible the same as those under which they were employed 

 by him in his experiments on the mechanical equivalent. For 

 this reason, a Callendar-Griffiths' platinum thermometer was 

 selected for the comparison. This instrument offers many 

 advantages for the standardization of calorimetric thermo- 

 meters. The platinum thermometer was subsequently compared 

 with a Tonnelot thermometer standardized at the Bureau 

 International. A connexion was thus obtained between 

 Rowland's air-scale and the Callendar-Griffiths' air-scale, as 

 well as with the nitrogen-scale of the Bureau International. 



The zero readings of the mercury thermometers were taken 

 in the same way that Rowland used them, before each 

 comparison. The thermometers were placed in a metallic 

 vessel well wrapped with thick felt and filled with finely 

 cracked ice of great purity steeped with distilled water. After 

 they had remained in ice about one and one-half hours, the 

 zero readings were taken by means of a micrometer- telescope. 

 The mercury thermometer was then hastily transferred to a 

 well stirred calorimeter with its bulb near that of the platinum 

 thermometer. The temperature of the water in the calorimeter 

 at the start was generally about 2° or o°G. A series of five 

 or six readings on the mercury thermometer was then taken 

 by one observer with the aid of a micrometer telescope, while 

 the other observer took simultaneous readings of the resistance 

 of the platinum thermometer. 



The temperature of the water in the calorimeter was then 

 raised at a rate of about 10° to 15° per hour to the next 

 temperature at which a comparison was to be made ; this was 

 accomplished by sending a suitable current through a coil 

 wrapped on the outside of the calorimeter. When the desired 

 temperature was reached, the current was cut off or so 

 regulated that the temperature was very slowty rising, when 



