250 PROCEEDINGS OF THE AMERICAN ACADEMY. 



glass sometimes' used for increasing the range of the instrument, and 

 the pyrometer was sighted on the carbon. The current was then adj usted 

 till the two fields in the pyrometer had the same appearance. The 

 small furnace was kept heated during each run, and before measuring 

 the temperature of the vacuum furnace a comparison was always made 

 with the thermo-electric couple. It was found by this means that the 

 current in the lamp as adjusted from day to day was much more con- 

 stant than when the amyl acetate lamp was used as a standard, but it 

 will be seen from the tables below that the current was not absolutely 

 constant. This must have been due to a variation in the electric lamp 

 itself The following table shows the effect of a change of the current 

 in the lamp on the apparent temperature as measured by the pyrom- 

 eter. The temperature of the body measured was held constant in the 

 vacuum furnace. The current is given in arbitrary units equal ap- 

 proximately to tenths of an ampere. 



The current could not be set better than to 0. 1 of a unit, corresponding 

 on the average to 10 degrees. 



In order to see how the temperature varied in the auxiliary furnace, 

 the temperature was taken with the junction by placing it at different 

 distances from the carbon block, which was placed near the center of 

 the tube. The results are given in Figure 3, which shows that no 

 appreciable error would be introduced if the junction had not been 

 exactly in the same plane as the surface of the carbon plate. 



The calcium carbide used in the following experiments was all made 

 at one time from Merck's lime and turnings from Acheson graphite 

 electrodes. It was analyzed by the loss in weight method and found 

 7^1 per cent pure. It was ground up and passed through a sieve with 

 36 meshes to the linear inch before the analyses were made. 



The method of carrying out the experiments was as follows : the 

 crucible was filled nearly to the top with the carbide, was covered with 



