ON THE MECHANICAL EQUIVALENT OF HEAT. 
411 
The mean chance was therefore that at every engagement of the counter with the 
shaft one-fourth of a revolution would be lost by the instrument, while on dis¬ 
engaging the counter stopped the instant it was withdrawn. 
The work in every trial should therefore be increased to compensate for this loss. 
The number of revolutions was approximately 18,000. 
The correction factor is therefore 
1 + 
72,000 
= (1 -f 0-00001). 
65. A summary of these corrections is appended. 
Magnitude and sign. 
Cause of correction. 
+ 
— 
I. Lengtli of lever. 
II. Dissolved salts. 
0-00042 
0-00003 
III. Dissolved air. 
, • 
0-00021 
iV. Weight of atmosphere. 
V. Varying specific heat of water. 
Neglected. 
0-00120 
VI. {a) Effect of pressure on thermometers. 
(fe) Loss of head in the water. 
VII. Engagement of revolution counter. 
0-00007 
0-00001 
0-00037 
Totals. 
0-00053 
0-00178 
Therefore the final correction factor is 
(1 - 0-00125). 
6G. Applying this correction factor to the value obtained from the experiments, 
we get for the value of the mean specific heat of water between freezing and boiling- 
points, expressed in mechanical units, at Manchester, 
777*91 (1 — 0-00125), 
776 - 94 . 
3 G 2 
