Manchester Memoirs, Vol. xlii. (1898), No. 0. 39 



mined, so as to make the value of the equivalent given by 

 the heavy trials alone equal to that given by the light 

 trials alone. 



This quantity varied from 36*86, in trials made with 

 the brass of the brake exposed directly to the air, to 7*98 

 in trials made with the brake covered with loose cotton 

 lagging. 



(2) Loss by conduction along the engine shaft. 



This was assumed to vary as the difference in the 

 readings of two thermometers, placed, the one on the 

 stuffing-box making the joint between the brake and the 

 shaft, and the other on the lower brass of one of the*main 

 shaft bearings distant some 2j^ inches from the cover of 

 the stuffing-box. 



From the first 42 accepted trials this loss was calcu- 

 lated to be 12 thermal units per unit difference of 

 temperature between stuffing-box and bearing per trial. 

 The conical brass bushes forming the bearings of the 

 brake itself were lubricated by forcing a stream of ice-cold 

 water from the supply pipe, through each into the brake. 

 Consequently, by regulating the amount of this water 

 supplied to the stuffing-box, I had a very delicate control 

 over the temperature gradient along the shaft. In the 

 later trials I endeavoured to make this gradient zero, and 

 thus cut out altogether any loss of heat by conduction. 



(3) Loss by leakage of water from the envelope formed 

 by the supply and discharge pipes y the brake and the tank. 



The only leakage of consequence was that which 

 might occur after the water entered the brake. 

 The brake itself had three working joints. 

 1st. — The stuffing-box on the engine shaft. 

 2nd. — The smaller stuffing-box through which passed 



a pin, making the connection between the shaft 



and the revolution counter. 



