200 PROCEEDINGS OP THE AMERICAN ACADEMY 



improved in two ways ; first, by the use of a motive power more 

 regular in its action ; and, second, by a more exact determination of 

 the loss due to radiation. The effect of the irregularity of the engine 

 has been calculated as about 1 in 4,000, and I suppose that the error 

 due to it cannot be as much as that after applying the correction. 

 The error due to radiation is nearly neutralized, at least between 0° 

 and 30°, by using the jacket at different temperatures. There may 

 be an error of a small amount at that point (30°) in the direction of 

 making the mechanical equivalent too great, and the specific heat may 

 keep on decreasing to even 40°. 



Between the limits of 15° and 25° I feel almost certain that no 

 subsequent experiments will change my values of the equivalent so 

 much as two parts in one thousand, and even outside those limits, say 

 between 10° and 30°, I doubt whether the figures will ever be changed 

 much more than that amount. 



It is my intention to continue the experiments, as well as work up 

 the remainder of the old ones. I shall also use some liquids in the 

 calorimeter other than water, and so have the equivalent in terms of 

 more than one fluid. 



Baltimore, 1878-79. Finished May 27, 1879. 



