844 ■ Transactions of the American Institute. 



refuse. Suppose a liorse-power in a certain foreign steamsliip costs 

 2,8 pounds of bituminous coal per hour, and in an American vessel 

 it costs three pounds of coal, using anthracite, are we to say our 

 engines are inferior? Let us see. We iirst deduct the refuse from 

 the anthracite, for instance, twenty per cent, whicli leaves 2.4 pounds 

 of combustible. Tliis, then, is nine-tenths of the weight of coal 

 having ten per cent of refuse; so multiply 2.4 by V% gives 2.6T 

 pounds as the true cost of the power in the American engine, to 

 compare with 2.8 pounds used by the foreigner, when both are com- 

 pared by the same standard. 



We have been thus explicit, because the fuel is so generally used 

 in the comparison of the performance of steam engines. The coal 

 bills, of course, show the absolute cost of the power in any particular 

 case, no matter what quality of coal was used ; but under such cir- 

 cumstances, the weight of coal consumed, even when corrected as 

 above pointed out, is, as must be seen, but an imperfect comparative 

 measure. To make comparisons sufficiently correct to answer the 

 demands of science, we must measure the steam used in each case ; 

 in other words, compare engines by the number of pounds of steam 

 used per horse-power per hour. 



The calculations are usually made from the pressure shown at the 

 termination of the stroke; the assumption being that the engine 

 uses, at every stroke, one full cylinder of steam at that pressure. 

 In other cases, however, the initial pressure, and the portion of the 

 cylinder filled at the point of cut-off, are used in the calculation. 

 These methods of determination pre-suppose that dry or saturated 

 steam enters the cylinder, which may be true, and that the steam 

 continues in this state, through at least part of the stroke, without 

 condensation, which is never the case. Steam is necessarily condensed 

 to set free the heat transmuted into the work done ; and the tempera- 

 ture of the metal of tlie cylinder is a mean of the temperatures to 

 which it is subjected, and therefore the surfaces form a condenser with 

 respect to the initial steam. The consequence is, that there is always 

 more steam taken from the boiler than is shown by the indicator ; the 

 discrepancy increasing with the degree of expansion and amount of 

 external refrigeration. Clarke, in his work on the locomotive, points 

 out ffreat differences between the amount of steam calculated from 

 the initial and terminal pressures shown by the indicator; and yet 

 uses the first in all his calculations. Later experiments, where the 

 steam has been actually measured, show that in small engines twenty 



