136 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



Hydhogen. 



perature aDd pressure and errors of manipulation arising during the per- 

 formance of the experiment have a "probable error" of only one in a 

 hundred thousand. It seems plain, therefore, that farther advance in 

 accuracy can be attained only by improvements in the methods of meas- 

 uring volumes and the expansion of the various solid and liquid substances 

 connected therewith, and in the measurement of temperature. These dif- 

 ficulties must enter equally into any other investigation with the same 

 object. When these are improved, the method will give yet better 

 results. 



The difference between the final result, 0.0036G09, and that found in 

 the preliminary work, 0.003659, is no greater than the known experi- 

 mental error of the preliminary work, which made no pretensions to abso- 

 lute accuracy, but was intended only to outline the capabilities of a new 

 arrangement of apparatus. 



The comparison of this result with the work of others is much compli- 

 cated by the fact that no one else has worked under similar conditions. 

 Nearly every one has chosen a wider range of temperature, namely, from 

 0" to 100° ; and, moreover, very few have worked upon the volume- 

 increase, but have chosen rather the pressure-increase in constant volume. 



