ON THE MECHANICAL EQUIVALENT OF HEAT. 
329 
two tubes were attached, one to the top and one to the bottom brass of the bearing, 
holes being bored into the brass and the tubes screw'ed in. These tubes are shown 
in fig. 9. 
In this way, with a thermometer in one of the tubes on the stufiing-box and one in 
each of the tubes on the bearing, although the thermometers might not give the 
actual temperatures of anything in particular, still the steadiness of the conditions of 
the brake warranted the conclusion that the differences in the readings of the thermo¬ 
meters would serve to identify similar conditions as to slope of temperature, and 
this turned out to be the case. 
These thermometers threw a flood of light on to conditions which had before been 
hardly perceptible. Thus, after reading the thermometer during three large trials 
and three small trials, with the cocks set as before without having been displaced, 
and with the same pressures, it was found that the mean of the three large triads 
indicated 13° Fahr. greater slope from the stuffing-box to the brass than that 
indicated by the mean of the three small trials. 
The Constants and Limits of Error of Conduction. 
29. It thence became possible in the subsequent trials, by adjusting the cocks, 
to bring about a mean condition in which the mean slope in the large trials was the 
same as that in the small, and by comparing the mean results of those trials in which 
the difference of slope had been in one direction with the mean of those in which it 
had been in the opposite, to obtain a constant expressing the quantity of heat lost 
for each degree of the recorded slope. 
These thermometers, read to 1° Fahr. 7 times during the trial of each sort, would 
give a limit of error of the \ of a degree which, taking 12 thermal units per hour as 
the loss per degree, would give as limit of relative error on 100,000 thermal units of, 
on one trial, 
0 - 00002 , 
and these being casual, when taken over 40 trials would be less than a millionth. 
The ITonid-Brake for Regulatiraj the S2')eed of the Engines. 
30. Although it had been found possible to maintain the speed of the engine 
constant within 2 or 3 per cent, when the engines were working wfith a considerable 
margin of pressure in the boiler, by maintaining the pressure in the boiler constant, 
the care and attention on the part of Mr. J. FIall, who had charge of the engine, 
became excessive when the engines were indicating over 80 h.p., particularly as he 
could not be attending to the fire and lubrication, and at the same time watching the 
speed indicator. To meet this difficulty, as there is no known automatic governor 
VOL. CXC.—A. 2 U 
