ON THE MECHANICAL EQUIVALENT OF HEAT. 
331 
The Method of Coyiducting the Trials—Elimination of Radiation. 
32. The entire system of working was designed to secure the most perfect elimi¬ 
nation of radiation possible. Thus, it was arranged in the first place that the trials 
oe made in pairs, one heavy trial and one light trial, made under circumstances as 
nearly similar as possible, except in respect of load and water. The loads in the 
first instance being 1200 and 600 foot pounds, and the quantities of water such that 
the final temperature should be as nearly as possible 212° Fahr., and, after the pre¬ 
liminary trials, 300 revolutions per minute was adopted as the speed for all the trials, 
60 minutes as the time of running. The inlet and outlet thermometers to be read 
after the first minute, and every two minutes; also the temperature of the laboratory 
as shown by a thermometer in a carefully-chosen place. This temperature to be 
maintained as nearly constant as possible. The setting of the regulators during 
each trial to be recorded; also the pressure of the artificial atmosphere, and that 
in the supply pipe after jDassing the coil; and, subsequently, the reading of the 
thermometers in the stuffing-box and bearings taken every five minutes, and the speed 
gauge every two minutes. The observations and incidents being recorded by the 
rules in surveying, in ink, in a book, and distinct from any reductions. The initial 
and final reading on the scales and counter being included, as were also the initial 
and final readings of the inlet and outlet thermometers and speed gauge for the 
purpose of determining the terminal differences of the heat in the brakes. 
As it was impossible to make trials simultaneously, and so secure similar conditions 
in the laboratory, it was at first arranged that the trials should be made in groups, 
including four pairs of trials. 
The regular work in the laboratory monopolised the engines and brakes on all days 
in term time, except Mondays and Thursdays, so that the trials were confined to two 
days in the week. There was a certain likelihood of the state of temperature of the 
walls and objects in the laboratory being systematically different on the Mondays, 
after the laboratory had been without steam all Sunday, from what it would be on 
the Thursday, after the steam had been on for three days. And besides this, there 
would be a systematic difference in the temperature of all the objects during the first 
trial in the day, although the brake had been running for an hour before, from that 
which would hold in the following trials. In the first instance, therefore, it was 
arranged that a heavy and a light trial should be made on the same day, and a light 
and a heavy trial on the next available day, under as nearly similar circumstances as 
possible, except for the inversion of the order. Then again, a light and a heavy trial 
on the next day, followed by a heavy and a light on the following, so as to break the 
order and secure the same arrangement, in days of the week as well as in hours of 
the day, for the four light trials as for the four heavy trials. 
As the results of any group of four pairs of trials would furnish a tolerably close 
approximation to the loss of heat by radiation, assuming this to be proportional to 
2 u 2 
