ON THE MECHANICAL EQUIVALENT OF HEAT. 
397 
Series II. 
37. With the exception of the addition of the lagging, no alteration was made in 
either apparatus or method between trials 11 and 12, 
SufSicient experience and confidence in the apparatus had now been gained to 
enable me to make three trials per day, as a rule two being made in the morning 
and one in the afternoon, a stop of about one hour being made after the second trial. 
The brake was not allowed to cool down during this interval; the hot water 
contained on finishing the morning’s run being shut in. 
In Table A, the value 787*4 is given as the result of the combination of trials 12 
and 14. There was evidently something amiss with this result, and as the com¬ 
bination of trials Nos. 13 and 14 gave the result 779*4, which agrees fairly closely 
with those given in Series L, the explanation which at once suggested itself was 
that the new lagging was damp when the day’s running began and had dried before 
the commencement of trial 13. On this account, trial No. 12 has been expunged 
from the final Table B, and takes no further jjart in the investigation. 
Series III. 
38. As it had by this time been found possible to run three satisfactory trials per 
day, the most obvious way of combining them was to make three trials, all carrying 
the same load, on the first day; while the trials recjuired to complete the three 
determinations were run on the next convenient day. 
This method was pursued during the whole of the subsequent course of the 
investigation. 
From this series onward T made an attempt to keep the temperature gradient 
along the shaft, between the brake and the adjacent bearing, the same in each pair 
of trials. In trial No. 21, I took observations for the first time of the temperature of 
the lower brass in the main bearing. In these trials also, the possible importance 
of the small leakage of water occurring along the spindle of the lower regulating 
cock, for the first time became apparent. The weight of water actually leaking 
away had not, I think, any appreciable effect, but owing to its high temperature it 
was nearly all evaporated, and, consequently, may have had a sensible effect in the 
lowering of the temperature of the water discharged from the brake. No successful 
means were yet devised for catching this water. So, in this series, it still remains 
as a possible source of error. 
Series IV. 
39. For use in the regular engine trials the brake is provided with a rider 
weighing 48 lbs,, which can be traversed along a graduated scale on the lever by 
means of a leading- screw. In order to maintain the balance of the brake, it carries 
at the back a second fixed load of 74*6 lbs. 
