396 
PROFESSOR O. REPROLDS AND MR. W. H. MOORBY 
In order to avoid any bias which might be given to the experiments by always 
combining a trial of one type with one of another type, trials of both of which types 
were always made at the same relative part of any day, the relative order of runnino- 
was changed as indicated by the dates and times given in Table B. (Part I., par. 32)° 
This method of combining the trials was adopted because at this time it was not as 
a rule possible to make more than two trials a day successfull}^ for breakdowns of a 
more or less serious nature were of fiequent occurrence. 
Referring now to the preliminary reduction shown in Table A, Series I. : 
lire values of K, Nos. I., III., IV., and V. are seen to be in close agreement, not- 
withstarrdino’ the comparatively rough method of reduction used. 
Determination No. II., however, stands out as very distinctly higher than the 
others, and the cause of this was fortunately eviderrt. 
In order to prevent the attempted rotation of the small handle shown in the illus¬ 
trations at the end of the brake lever, one revolution of which altered the load on the 
biake by 1 ft.-lb., one of my assistants had tied it to the hanger carrying the load, 
lire stiing making the connection was very tight, and the load was pulled perceptibly 
out of the perpendicular plane passing through the groove on the lever. 
riiis fault was sufficient to condemn the two trials Nos. 3 and 4, and they do not 
appear in the final table on that account. 
A wooden clip was subsequently added to prevent the rotation of the handle and 
its attached screw. 
Lagging. (Pa,rt I., par. 33.) 
36. ihe results given by the four accepted determinations of Series I. wmre so 
consistent that it was decided to proceed at once wdth the lagging of the brake, 
which, up to the present clme, had been deferred on account of want of confidence 
in the apparatus generally. 
Ihe lagging consisted of a layer of about 1|- inches of loose cotton wadding with 
V Inch the w^hole of the exterior of the body of the brake was covered, together 
with the discharge pipe betv'een the brake and the thermometer chamber. The 
cotton was all tied firmly in position, and the wdiole was enclosed in a covering of 
thick flannel. 
As will be seen later, this lagging reduced the radiation by nearly 75 per cent. 
Its weight, about 2 lbs., was inappreciable, and, being evenly distributed, could not 
affect the balancing of the brake to any extent which it would be possible to detect. 
1 he lagging was, I believe, of use, more especially in that it protected the bare 
metal from the strong draughts which often occurred in the engine-room. It 
lequired very careful attention, however, to protect it against dampness, and on 
this account 1 am not certain that better results would not have been obtained 
without it. 
