^ PROFESSOR O. REYNOLDS AND MR. W. H. MOORBY 
which was brought down to three-quarters of an inch diameter and carried two round 
pins, a sixteenth of an inch diameter, standing out radially, the engagement being 
effected by pushing the counter forward till the wire crank engaged on one of the 
pins. (Owing to the wire being pointed and the pins rounded, the chance of the wire 
striking plumb on to the pin and so preventing engagement was reduced to a minimum,) 
This engagement was the result of a great deal of experience, and answered perfectly, 
but it involved the mean chance of a quarter of a revolution of the engine shaft after 
the wire had passed the j)in before the actual engagement was effected, whereas on 
coming off the disengagement was instantaneous, the counter stopping by the friction 
of the worm before the momentum had carried it through any appreciable angle. 
This would leave a mean error of the work done during one-fourth of a revolution 
on each trial, whence, the number of revolutions during the trial being 18,000, the 
relative mean correction would be one seventy-two thousandth part, or 0‘000013. 
As, however, when the two operations were executed by different observers on a 
signal, the peisonal equations might amount to more than this, although it involved 
a difficult piece of linkage, an automatic connection was effected, as shown in Plate 5, 
the pushing of the counter into engagement shifting the switch, so that in making the 
trials no error was introduced. 
The Leakage of Water. 
35. As the loss of any of the water, which had entered the brake before it was weighed, 
would conscitute a corresponding error in the results, the perfect tightness, not only 
of all the fixed joints, but of the casting and the pipes, was a matter of first considera¬ 
tion and of continual care. This was one of the reasons why the lagging was delayed 
till after consistent results had been obtained; for, as long as the brake and pipes 
weie naked, such leakage could not fail to be observed on close inspection, and before 
lagging it was arranged to test the brake and pipes to an excess of pressure, so as to 
insuie perfect soundness. Besides the fixed joints there were only two working 
joints, in addition to the openings into the switch and again into the tank. 
(1) The working joints were : The stuffing-box on the main shaft and the stuffing- 
box on the automatic cock on the outlet from the brake. 
Any leakage from these was open to observation both before and after lao’O'ino- as 
tney were m no way covered; and arrangements were made so that such leakage 
could be sejiarately conducted by pipes and caught in bottles. With care such 
leakage could be reduced to insignificant quantities. 
The absolute loss of heat resulting from a leakage of lbs. of water from the 
stuffing-box on the shaft was equal to the product of the difference of temperature of 
the stuffing-box Ts.b°, and inlet (T^®) multiplied bj^ 
- T,°), 
