standard. 



Air thermometer. 



5-003° 



5-004° 



5-038° 



■ 5-036° 



216 W. G. Mixter — Heat of Combustion of Hydrogen, 



mometer graduated to tenths. This was also calibrated and its 

 0° and 100° points determined and the resnlts agreed well with 

 the certificate of the Reichsanstalt. The correction for No. 1 

 is +0-003° and for No. 172,683+0-038° for an interval of 5° 

 on the scale of each. 



By way of control comparisons were made with an air ther- 

 mometer consisting of the silver bomb connected with a con- 

 stant volume pressure gauge. As the air thermometer was in 

 fact the calorimeter the differences in temperature observed 

 with the mercurial thermometers were subject to the same 

 source of error as the temperature observations of the calori- 

 metric work. The following are the results : 



Interval. 

 No. 1 5° 



" 172,683 5° 



The error in the value of 1'8°, the interval in the calori- 

 metric work, probably does not exceed 0-001°. Thermometer 

 No. 3, used for finding the temperature of the gas measured, is 

 graduated to tenths. It was calibrated, and compared with 

 the standard, and the 0° point was from time to time taken in 

 ice. 



The temperature interval of the calorimetric work was deter- 

 mined as follows : The temperature observed at the instant of 

 an explosion was taken as the initial temperature. It was 

 assumed during the first minute after an explosion that the 

 gain and loss of heat from external causes were equal. To the 

 maximum temperature observed was added the average fall 

 noted subsequently for the number of minutes less one inter- 

 vening between the explosion and the highest temperature. 

 The errors due to the time required for the thermometers to 

 come to an equilibrium with the water are small, since No. 1 

 changed the' last 0-01° in 20 seconds and No. 172,683 in 30. 



The barometer used was one made by G-reen. The scale is 

 Q.^mm iggg ^Ijj^j^ 760"^"^ on the standard meter of the Physical 

 Laboratory, and the correction for the meniscus is 0-72™™. 

 Three sets of weights used in the investigation were tested and 

 found to be sufficiently in accord for the purpose. The bal- 

 ance is an old 5^ one in an ordinary glass case. It indicates a 

 difference of one milligram with a load of 2^ on each arm. 

 During the third series of experiments it was enclosed in a large 

 hood. The weighing of gas required much time, and proved 

 to be sufficiently accurate for the purpose. The weight of the 

 water used in the calorimeter was reduced to weight in vacuo. 



The mass of one liter of hydrogen in the latitude of New 

 Haven, 41-3°, is taken as 0-089844 gram and is derived from 

 Morley's figure 0-089873 gram for latitude 45°. The mass of 



