332 
PROFESSOR O. REYNOLDS AND MR. W. H. MOOR BY 
the observed mean difference of temperature between the laboratory and the hraJce, 
it was easy to obtain an approximate constant, H, for radiation for each degree of 
difference of temperature, and so to introduce a correction, II (Tj — T^), in each 
trial for the radiation resulting from the observed mean difference of temperature of 
laboratory and brake, lb — T^. 
These corrections would serve two purposes—first, affording a better comparison 
of the results of the separate trials for future guidance, and secondly, by recording 
the mean differetjce of temperature, would show how far the mean differences of 
temperature in the large trials had differed from those in the small trials, and thus 
how far the radiation had been eliminated. 
Lagging the Brakes. 
33. In order to obtain still more definite assurance as to the elimination, it was 
arranged that after consistent results had been obtained in several groups of four 
pairs of trials, as above, with the brake naked, the brake should be covered with 
non-conducting material, in the best way practicable, so as greatly to reduce the 
radiation, at the same time leaving it definite, and then similar trials should be run. 
If the coefficient of radiation could in this way be reduced to one-fourth that of 
the naked brake, such error as there might be remaining in the mean results with the 
naked brake wnulcl be reduced to one-fourth with the lagged brake. 
In this, however, there was danger of introducing errors of other kinds. 
The non-conducting material would absorb heat slowly and take a long time to 
arrive at a state of equilibrium, and during the interval the rate of loss of heat from 
the brake would be irregular. The total error that could result from this cause 
would be the product of the specific heat of the material used multiplied by the 
weight, and again by the 75°, or the half of wdiatever was the difference in temperature 
of the brake and the air. This decided the choice of the material to include cotton 
v/ool. Two pounds of this would, if not too tightly pressed, cover the brake about 
Ig inches thick, and the total heat it would absorb would be less tlian 0‘4 lb. of 
water raised from 32° to 212° Fahr., and would then be only 0*0008 of the heat 
generated by 30 h.p. in an hour, while it would reduce the radiation to about y. 
But as the cotton wool would gradually collapse if subjected to any elastic pressure, 
it was decided only to use this to such thickness as it could be protected by light 
cotton strings extending in axial planes round the brake, and to prevent 
absorption of moisture by the cotton wool, to cover it with tliick anti-rheumatic 
flannel, about 1 inch to 1^ inches in thickness, as shown in Plate 5, which would 
raise the capacity for heat of the entire lagging to about that of the heat 
generated in the small trials, and as the brake was kept at steady temperature for 
about one hour or more before the trial commenced, the actual differences would not 
exceed some one ten-thousandth part. 
