ON THE TRANSMISSION OF POWER BY COMPRESSED AIR. 455 



when it has fallen in the mains to the temperature which it had initially 

 at C. G H being again an adiabatic curve, the area A G H F (which is 

 61 per cent, of A B C F, and 52 per cent, of A B C D E) repi'esents the 

 maximum work which can be obtained from the air in a motor, the quan- 

 tity calculated in the last paragraph. 



If it bo assumed that the temperature in the mains is constant, then 

 any loss of pressure due to friction must be accompanied by an exactly 

 equivalent increase of volume. Thus if (Fig. 3) the pressure falls from 

 A to K, the volume increases from A G to K L, the point L l^ang on an 

 isothermal through G. The loss of possible work due to such a reduction 

 of pressure is represented by the difference between the areas A G H F 

 and K L M F, in both of which the expansion curves are adiabatic. The 

 change of temperature in passing through a reducing valve is so smaJl 

 that it may be assumed without sensible error that the loss due to such a 

 process may be calculated in the same fashion. 



I determined the loss of head in the mains by a series of observations 

 made simultaneously at known points in Paris and at St. Fargeau. Tho 

 pressure gauges used having been cai'efully compared, and all the necessary 

 corrections made, I found the loss of pressure to vary from 0'35 to 

 0'25 atmosphere, according to the distance from St. Fargeau and the 

 amount of air passing through the pipes. The average loss may be taken 

 at O'S of an atmosphere at three miles from St. Fargeau, when the indi- 

 cated horse-power there was about 1,250, and the maximum velocity of 

 the air in the mains about 1,550 ft. per minute. What proportion of 

 this loss of head may have been due to leakage, and what the amount of 

 leakage (if any) may have been, I had no means of determining. The 

 duplicate main to which T have referred is not yet completed, and it was 

 impossible to isolate any portion of the mains, even temporarily, to test for 

 leakage. From the figures I have given, as well as from the nature of 

 the pipe joint, I cannot doubt that the leakage must, under any circum- 

 stances, have been extremely small. 



In the Table I have given approximate values of the loss due to fall of 

 pressure in the mains and through the reducing valves with various values 

 of the total reduction of pressure. With a total reduction of half an atmo- 

 sphere the combined efficiency of mains and valves is 096, reducing the 

 maximum possible work at the motor to 0'5 indicated horse-power per 

 indicated horse-power at central station. Under these conditions the 

 minimum possible consumption of air per indicated horse-power at the 

 motor would be twice .348, or G96 cubic feet per hour. 



The motor on which I made most of my experiments was an ordinary 

 horizontal Davey-Paxman engine with a single cylinder 8 j in. in diameter 

 and 12 in. stroke, fitted with automatic cut-off gear. For convenience 

 sake I tested it at St. Fargeau and not in Paris, but I used a pressure 

 only of 4| atmospheres, which pressure I found to be exceeded on branch 

 mains 3), miles from St. Fargeau, where I made later experiments. The 

 position of the motor did not, therefore, put it under any conditions 

 difierent from those existing in the centime of Paris. I made a large 

 number of experiments on this motor, under various conditions, individual 

 experiments lasting from four hours down to half an hour. I shall here 

 give figures representing only the four most important of my experiments, 

 two with cold and two with heated air, averaging tho two experiments 

 in each case, so as not to burden my paper with unnecessary figures. 



The motor, when indicating 99 horse-power, and making about 125 



