46 Reynolds and Moorby, Equivalent of Heat. 



which were found to exist between the weights. The 

 extreme variation from 2 5 lbs. was 0'04lbs. 



{d) The counting of the revolutions. 



This was done by means of a counter, which was 

 pushed into gear with the shaft at the commencement of 

 a trial, and disengaged at its close. The engagement 

 was of such a character that the mean chance was that 

 y^ revolution was missed at the commencement of the 

 trials, while at the end the counter stopped as soon as it 

 was withdrawn from the shaft. The work obtained there- 

 fore needed a correction of 



+ —„ — = + = +o'ooooi nearly, 



18000 72000 ^ 



This correction was added to the mean values of K 

 derived from the trials. 



II. 



The measurement of heat. 



This involved the calibration of the scale of the 

 weighing machine, and the frequent determination of the 

 index errors of the two thermometers used in the deter- 

 mination of the rise in temperature of the stream of water 

 flowing through the brake. 



{a) The weighing inachiiie. 



It was of vital importance that the same unit of mass 

 should be used for the weighing machine scale and for the 

 weights used on the brake. The scale of the machine was 

 consequently checked against the 251b. weights and vice 

 versd. By this means a series of corrections to xJolt). was 

 obtained for all readings, and, with the exception of casual 

 errors occurring in any individual weighing, the deter- 

 mination of the mass of water may be taken as accurate 



