HOW THE TESTS WERE CONDUCTED. 17 



glass for showing the height of the water inside, the same being 

 graduated in inches measured from the face of the orifice plate. 

 A valve in the 4-in. supply-pipe served to regulate the height of 

 water in the stand-pipe, and consequently the amount passing 

 through the orifice. During the progress of the test, the head 

 of water in the stand-pipe was maintained at such a point as to 

 supply the required quantity of water ; and a careful record 

 was kept of the height indicated in the gauge glass. Subse- 

 quently, when the pump was stopped, the orifice was calibrated 

 by observing the quantity of water which flowed into the tank 

 under conditions of the average head, the contents being pre- 

 viously known. 



Whatever method is pursued in determining the quantity of 

 water pumped into the boilers on a feed-water test, a determi- 

 nation should be made of the leakage of the boilers, stop valves, 

 safety valves, steam-pipe joints, blow-off cocks, etc., concerned 

 in the plant, so as to correct for such leakage, and charge the 

 engine with only that quantity of feed-water which actually 

 passes into it as steam through the throttle valve. To accom- 

 plish this object a leakage trial should be made immediately 

 after the engine is shut down at the close of the test, the 

 pressure being maintained in the boilers at a point nearly, if 

 not quite, as high as the working pressure, and no change made 

 in the stop valves, etc., concerned, or in the drips or other 

 avenues of escape. Observations should then be made of the 

 height of water in the gauge glasses, taking readings at inter- 

 vals of ten minutes for a period of one hour, or until successive 

 differences in the ten-minute periods show a uniform rate of 

 leakage. By calculating the weight of water corresponding to 

 the volume lost, as found by this test, which can be done know- 

 ing the dimensions of the boilers, the desired correction for 

 leakage is determined. To make this test reliable it is neces- 

 sary, of course, that the throttle valve at the engine should be 

 tight. The tightness of the throttle valve can readily be deter- 

 mined by observing whether steam blows from the open indi- 

 cator-cock of the cylinder when the steam valve is wide open, 

 this observation being made at the end of the cylinder which is 



