320 



ately half as much as at Dalkey. This is corroborated by the 

 fact, that at Dalkey a higher vacuum than 25| could seldom 

 be obtained, whilst at Croydon I have seen 27 J inches easily 

 obtained. 



It is also remarkable that the absolute quantity of air 

 which enters the tube is nearly constant at all pressures 

 above six inches of the gauge, evidently resulting from the 

 fact, that the higher pressures close the valve more perfectly 

 than the lower ones; but the higher the gauges the more 

 important the loss by leakage becomes; thus at 15 inches of 

 the gauge or half vacuum, the four cubic feet would expand 

 to eight, and require the action of eight feet of pump space 

 to remove or overcome such leakage ; at twenty inches, or 

 2.3rd vacuum, the four become twelve feet, and require 

 twelve feet of pump space; and at twenty-five or 5-6ths 

 vacuum, become twenty feet of rarefied air, and require the 

 exertion of twenty cubic feet of pump space. 



Therefore, if the leakage of three miles of pipe be six 

 cubic feet per second, the pumps capable of removing 120 

 cubic feet per second, at a vacuum of twenty inches, the six 

 feet become eighteen ; the loss is, therefore, fifteen per cent, 

 of the action of the pumps. But the loss by leakage is very 

 small during part of the prior exhaustion, and is proportion- 

 ally diminished as the train proceeds and shortens the length 

 exposed to leakage. 



The friction of the travelling piston amounts to about 35 

 lbs. out of 1,760 lbs. pressure, or about two per cent, of the 

 power, as carefully ascertained by the experiments of M. 

 Mallet. 



There is another source of loss of power which has been 

 the subject of much and very ingenious calculation and dis- 

 cussion ; I allude to the friction of the air in the pipe. From 

 the data given by the transmission of gas and of water through 

 tubes, it was found that the friction increased according to 



