176 PROCEEDINGS OP THE AMERICAN ACADEMY 



As these temperatures usually contained fractions, the amount of 

 work necessary to raise one kilogramme of the water to the even 

 degrees could then be found from this table by interpolation. Joule's 

 equivalent at any point would then be merely the difference of any 

 two succeeding numbers ; or, better, one tenth the difference of two 

 numbers situated 10° apart, or, in general, the difference of the num- 

 bers divided by the difference of the temperatures. 



It would be a perfectly simple matter to make the record of the 

 torsion circle entirely automatic, and I think I shall modify the 

 apparatus in that manner in the future. 



It would take too much space to give the details of each experiment ; 

 but, to show the process of calculation, I will give the experiment of 

 Dec. 17, 1878 as a specimen. The chronograph sheet, of course, I 

 cannot give. The computation is at first in gravitation measure, but 

 afterwards reduced to absolute measure. 



The calorimeter before the experiment weighed 12.2733 kil. 



" " after " " " 1 2.271 G " 



Mean 12.2720 " 



Weight of calorimeter alone 3.8721 " 



.'.* Water alone weighed 8.3999 " 



.347 " 



Total capacity 8.7469 " 



The correction for weighing in air was .835 \ = .00106. 

 The total term containing the correction is therefore .99878. 



log 86.324 =1.9361316 



log .99878 =1.9994698 



1.9356014 

 log 8.7469 = .9418542 



log const, factor = .9937472 = log 9.85706. 



Hence the work per kilogramme is 9.85706 %Wn in gravitation 

 measure, the term 2 Wn being used to denote the sum of products 

 similar to Wn as obtained by simultaneous readings of torsion circle 

 and records on chronograph sheet. 



Zero of torsion wheel, 79. 3 mm \ 



Value of l mm - on torsion wheel .0118 kn -. 



The following were' the records of time on the chronograph sheet: — 



