319 



I, however, hope I have effected some considerable im- 

 provement in this respect, by a method of working the valves 

 of air pumps, by gearing (which I will afterwards explain) 

 so as to avoid concussion, to give large air ways to both inlet 

 and outlet valves, and work the valves by the direct power of 

 the engine. I anticipate that this will save at least half the 

 power otherwise lost in opening the valves, or about five per 

 cent, of the power of the engine. 



Another and somewhat unexpected source of loss is that 

 the air, which by its expansion under exhaustion becomes 

 cooler, rapidly takes up heat from the traction tube, and 

 when again condensed in the air pump previously to being 

 expelled into the atmosphere, gives out its heat so as to 

 become very warm, and therefore requires much more power 

 to expel it into the atmosphere. The heat of the air thus 

 expelled sometimes exceeds that of boiling water. 



Mr. Robert Stephenson appears to have investigated the 

 loss by the opening of the pump valves, and of the heating 

 of the air together, by means of the indicator ; such results 

 vary considerably at the different degrees of exhaustion. 

 The maximum appears to be about 17 per cent., but I con- 

 ceive a mean of 15 per cent, to be rather too high. The 

 most important loss, however, arises from the leakage of air 

 into the tube; but this appears, after the most careful experi- 

 ments, and on a fair mean of workings at Dalkey, to be very 

 much less than could be anticipated, being at the rate of 

 about four cubic feet of air at the pressure of the atmosphere 

 per second, including leakage of air pump, valves, close con- 

 necting main, and traction tube — the leakage of the pump 

 valves and of the connecting main being greater than that of 

 the traction tube. The greatest amount of leakage observed 

 at Dalkey, when the apparatus was out of order, was only 

 seven cubic feet per second. But I am informed that at 

 Croydon the leakage is now extremely small, not proportion- 



A A 



