Manchester Memoirs, Vol. xlii. (1898), No. 0. 13 



first instance were divided to fifths of a degree Fahrenheit ; 

 they were both immersed in the flowing water to within a 

 few degrees of the top of the mercury. They were com- 

 pared at equal temperatures, but otherwise subjected to no 

 tests for accuracy of scale. 



During the experiments, the link connecting the inlet 

 valve to the automatic gear was removed, the valve 

 being set open and the supply regulated by the hand-cock 

 below. The pressure in the brake being that of the atmo- 

 sphere, and the head of water on the inlet constant at 7ft., 

 when the hand-cock was set, the flow was steady. The 

 quantity of water in the brake then depended on the 

 automatic adjustment of the outlet valve, which, with the 

 exception of a little trouble at starting and stopping the 

 engines, soon overcome, kept the brake lever steady. 



To admit of catching the water after leaving the out- 

 flow thermometer, the extension-pipe turned downwards 

 over the side of the lever into a small basin, with its lip 

 above the mouth of the pipe and from the basin the water 

 flowed into a short trough, from which it was caught in 

 buckets and carefully weighed. This was a primitive 

 arrangement, and required several assistants, but was found 

 capable of considerable accuracy up to about 4olb. a minute. 



The engines were kept running at a constant speed 

 by keeping constant pressure in the boiler, the speed being 

 indicated on the speed gauge as well as recorded on the 

 counter. 



The water coming from the town's main was at nearly 

 constant temperature, between 40*^ and 50^ Fahrenheit, 

 according to the time of the year, and varying less than a 

 degree throughout several trials. 



The rise of temperature was adjusted by the quantity 

 of water admitted, according to the work, so that the final 

 temperatures, as well as the initial, were as nearly as possible 

 the same in the different trials. 



