ON THE MECHANICAL EQUIVALENT OF HEAT. 
same manner that I have described (paras. 17, 18, and 19), Some of the particulars 
mentioned were, however, omitted in the earlier trials, and were only recorded subse¬ 
quently after their importance had come to be recognised. 
In all, 80 trials were made on which any reliance has been placed, and these will 
be dealt with in dilferent series^ between any consecutive two of which some slight 
alteration had been made in tlie apparatus, the method of taking the observations, or 
of reducing the same; all these alterations leading up to the finally adopted 
methods which have been described. 
33. I must first mention two sets of trials which do not appear in the tables. They 
were commenced in December, 1895, and were made mainly with the object of gaining 
experience in the behaviour of the apparatus, and of determining the most favourable 
conditions under which the experiments could le conducted. 
The moments carried by the heavy and light trials in each set were 1200 and 
600 ft.-lbs. respectively. 
The speed was in the first set 230 revolutions per minute, and in the second set 
180 revolutions per minute. 
With the following exceptions the apparatus and methods were the same as 
described. 
I. Omissions and faults in apparatus. 
(1.) There were no thermometers on either the stutiing-box cover or on the 
main bearing, and consequently no effectual attempt could be made to 
keep these parts of the shaft at the same temperature in a pair of 
trials. 
(2.) There was no means of catching the leakage from the stuffing-box, or 
from the bottom regulating cock, 
(3.) The rising pipe at this time only maintained a head of about 5 feet of 
water over the thermometer in the discharge pipe. 
(4.) The hand brake had not been fitted to the shaft. 
1J. Omissions and faults in the methods employed. 
(1.) No corrections were added to the heat as given by the formula —^i) 
X (T^ - Tj). 
(2.) The heavy trials were of only half-an-hour’s duration, in order that the 
second reading taken of the weight of the tank should lie on the same 
part of the scale of the weighing machine, which had not up to this 
time been corrected, in both heavy and light trials. 
The results obtained were not very consistent, but, perhaps largely on that account, 
the trials admirably fulfilled the purpose for which they were made. 
The importance of the terminal corrections was clearly indicated when the results 
were considered, and consequently means were at orice taken to apply these correc- 
YOL. CXC.-A, 3 E 
