330 
PSOPESSOR O. REYNOLDS AND MR. W. H. MOORBY 
which will regulate an engine working against a resistance which is independent of 
the speed, without fluctuations, I arranged a hand-brake to be applied to the rope 
pulley, 3 feet in diameter, on the brake shaft, by one of the assistants in the 
laboratory during the trial. The amount of power to be absorbed by this being less 
than 2 h.p. at the most, a f-inch cotton rope, with oue end fast, passed round in one 
of the grooves of the pulley, the other end being attached to a spring balance, the 
position of which could be regulated with a screw, would answer the purpose (shown 
in Plate 5). 
In this way, as the natural variations of speed of the engines are very slow, 
Mr. Mathews was able, after a little exj)eiience, to keep the speed to within some¬ 
thing like one revolution, or 0*3 per cent. 
Tlie Corrections for the Terminal Heat of the Brahe. 
31. As the temperature of the effluent water could be continually regulated by 
regulating the supply of water to the brake, whatever might be the speed, the chief 
importance of keeping the speed regular arose from the errors (l) caused by small 
differences of temperature in the brake together with the water it contained at the 
commencement and end of the trial, and (2) by small differences in the weight of 
water in the brake at the commencement and end of the trial. 
Such errors belong to the class of casual errors to be eliminated in the mean of a 
number of trials. Still, it seemed desirable to have some assurance that such 
elimination wms effected, and, in order to obtain this, I proposed that the actual 
quantity of water in the brake for each of the loads used in the experiments should 
be determined experimentally at several speeds covering the range of variations 
likely to occur, and so to obtain a curve for each load, showing the water at each 
particular speed ; this to be done by running the brake as in the trials, steadily, at 
a particular speed, the water passing as in the trial. Then, suddenly, by forcing 
down the lever, to close the automatic outlet valve, and, at the same time shutting 
the inlet valve and stopping the engines, and thus trapping the working charge of 
w^ater in the brake. The water could then be drawn out and weighed. 
Putting B for the capacity for heat of the metal of the brake, id for the weight 
of water, and T for the temperature observed on the effluent thermometer, the total 
heat in the brake is expressed by 
(B + w) r, 
and, if Wi, T ° refer to the W(‘ight of water and temperature at starting, and Wf, '1/° 
to the corresponding quantities at the end of the trial, the correction which has to 
be subtracted from the heat observed is expressed by 
(B + IV,) - (B + ivf Tf 
