24 Reynolds and Moorby, Equivalent of Heat. 



of the error that might arise from each of the circumstances 

 affecting the relative accuracy were provided, thus assuring 

 the accuracy of the results obtained within these limits, in 

 terms of the standard of measurements — length or tem- 

 perature — to which they are referred. These limits are 

 given in a table at the end of this Part. 



The Standards of Measurements. 



During the inception of the research, the appliances 

 for standardising the results had been in course of prepa- 

 ration, in accordance with the system already mentioned 

 as designed by the author. 



Since the absolute quantity of mass does not enter 

 into the results, it was only necessary to use a common 

 standard in comparing the masses of water with the 

 masses which formed the load on the brake. And for this 

 purpose 12 of the 251b. cast-iron weights used for the 

 load on the brake were adjusted by balance, with a limit 

 of accuracy of 00 lib. of each other; then the scales, after 

 being reset by Mr. Foster, were surveyed with the adjusted 

 weight and the corrections fully determined. As the 

 weights on the scales, as well as the adjusted weights 

 exclusively used on the brake, were of the same material, 

 it was not necessary to apply any correction for the weight 

 of the atmosphere to these. And, since the water is 

 balanced by cast-iron and the load is balanced by cast-iron, 

 a correction for the weight of the atmosphere displaced by 

 the iron would leave the relation between the load and 

 the water the same ; so that the only correction was for 

 the weight of the atmosphere displaced by the water. 



The standard capacity for heat being that of distilled 

 water, this would have been employed in the experiments 

 had it been practicable. But since, letting alone the extra 

 appliances, the cooling the 100 tons of water from a tem- 

 perature, say, of 72° to 32^, would have required more than 



