388 
PROFESSOR 0. REYNOLDS AND MR. W. H. MOORBT 
be safely assumed that in the first series of trials this error did not occur to any 
sensible extent. 
I think that, especially with the above result in view, the loadino- of the brake 
may be taken as absolutely accurate. 
As to the limit of accuracy of the weighings in the 600 ft.-lb. trials, the weight 
of water dealt with was approximately 470 lbs. On this quantity the maximum 
probable error was 002 lb. in any trial. This gives greater accuracy than 1 part in 
20 , 000 . 
The Adjustments of the B^^ahe. 
(1.) Length of the Lever .—(Part I., par. 45.) 
28. This length was required between the centre line of the eno-ine shaft traversino- 
the brake and the V-groove carried by the lever. 
It bad been previously observed that both the shaft and the brake shifted a little 
horizontally when the engine was started, from the positions occupied with the engine 
stationary. It was therefore necessary to make the comparison between the length 
of the lever and our standard 4-feet with the engine running. Also, since the length 
of the lever varied with the temperature of the brake, this temperature was main¬ 
tained, as in all the trials, at 212° Fahr. 
Between the brake and the adjacent bearing the shaft is 4 inches diameter within 
XoVo of inch. 
At a distance of 3 feet 10 inches from one of its scjuare ends a fine line was scribed 
on a steel straight edge. Ihis straight edge was then held with the square end 
aforesaid butting against the shaft, the length being horizontal and perpendicular to 
the line of shafting, and the distance between the straight edge and the lever being 
10 inches. At a distance of 11 feet from the other side of the lever a theodolite was 
set up and adjusted so that the vertical plane of collimation of the instrument was 
parallel with the shaft and contained the line scribed on the face of the straight edge. 
A steel scale, graduated to g-Q of an inch, was fixed firmly on to the lever, and a 
reading of this scale was taken through the telescope without altering the adjust¬ 
ments merntioned. This reading, of course, referred to the point on the scale just 
4 feet distant from the centre line of the shaft. By a slight rotation about the 
vertical axis the line of collimation was then made to cut the centre line of the 
groove, and then a vertical rotation enabled a second reading of the scale to be taken. 
A number of these observations were made while the brake was subjected to 
moments of 1200, 600, and 400 ft.-lbs., and they all indicated that the length of the 
lever in the trials made was 4' -j- 0‘02". 
A correction to the value of the equivalent derived directly from the trials is 
therefore necessary on this account. It amounts to f 0'0417 per cent. 
With this correction added, I think that tbe length of the lever can be assumed 
accurate to inch, or 1 part in 10,000 nearly. 
