136 PROCEEDINGS OP THE AMERICAN ACADEMY 



To reduce these to the mean temperature of 0° to 40°, I have used 

 the rate of increase found by Bede for copper. They then become, 

 for the mean from 0° to 40°, — 



.0897 



.0897 



.0878 



.0893 



.0883 



.0906 

 Mean .0892 ± .00027 



As the capacity of the calorimeter is about four per cent of that of 

 the total capacity, including the water, this probable error is about 

 ¥«W °f ^ ne total capacity, and may thus be considered as satisfactory. 



I have also computed the mean specific heat as follows, from other 

 observers : — 



Copper between £0° and 100° nearly. 



.0949 Dulong. 

 .0935 



15) 



>2J 



qq 5? i Kejrnault. 



.0933 Bede. 

 .0030 Kopp. 



Hence we have the following for the ca 



.0940 



r ' by JLiede's I'ormi 



■alorimeter : * — 



The close agreement of this number with the experimental result 

 can only be accidental, as the reduction to the air thermometer would 

 decrease it somewhat, and so make it even lower than mine. IIow- 



* The cast brass was composed of 28 parts of copper, 2 of tin, 1 of zinc, and 

 1 of lead. The rolled brass was assumed to have the same composition. The 

 solder was assumed to be made of equal parts of tin and lead. 



