REVIEW OF FEED -WATER TESTS. 257 



distinguish a difference, although in one case the exhaust 

 valve leaked no less than 10%. Practically the same effect, or 

 rather want of effect, has been noticed in one other case where 

 a broken packing-ring in a piston caused a leakage which 

 amounted to a still larger quantity. A close examination of 

 the expansion line of the diagrams, before and after, failed to 

 reveal any clearly defined difference. This is Engine No. 78. 

 In a case like that of the compound Engine No. 38, it cannot 

 be inferred from this that the influence of excessive leakage 

 could not be revealed by a study of the indicator diagrams. 

 Here it did produce a marked effect in the distribution of the 

 load between the cylinders, cutting down the power developed 

 in the L. P. cylinder, and increasing it to a corresponding ex- 

 tent in the H. P. cylinder. At the same time, it caused a 

 noticeable " drop " in pressure at the high-pressure release. 



In studying the effect of cylinder condensation and leakage, 

 and the extent of the loss which it produces, the quantity 

 shown at the cut-off point of the diagram is selected in prefer- 

 ence to that shown at the release, in the belief that at the cut- 

 off point the full extent of the loss is the more truthfully 

 indicated. If the steam accounted for at both points is identi- 

 cal, the loss is the same at one point as at the other, and it does 

 not matter which point is selected. If the quantity accounted 

 for is larger at the release than it is at the cut-off, which owing 

 to re-evaporation during expansion frequently occurs, the appar- 

 ent loss due to condensation and leakage is less at the release 

 than it is at the cut-off. Sometimes there is as much as 10 per 

 cent less apparent condensation and leakage at the release 

 than at the cut-off. In cases like this the release percentage 

 does not show the full extent of loss, because the work recov- 

 ered on account of re-evaporation is in no sense proportional to 

 the increase in the steam accounted for at that point. Neither 

 does the loss at cut-off in such cases represent the true loss, 

 and the reason is the same ; but the loss at cut-off furnishes a 

 closer indication of the true loss than the loss at release, and 

 a better basis for the study of the question of cylinder conden- 

 sation. 



