OX THE THEORY OF THE STEAM-ENGINE. 579 



steam by expansive working, and it affords a sufficient explanation 

 of the fact, in engineering practice, that expansive working has been 

 fonnd to be expensive working, and that, in many cases, an absolutely 

 o-reater quantity of fuel has been consumed in extended expansion work- 

 ino-, while less power has been developed.' He states that high speed 

 reduces the effect of this cause of loss, and indicates other methods of 

 checking it. He states that ' the less the period of admission relative to 

 the whole stroke, the greater the quantity of free water existing in the 

 cylinder.' His experiments, revealing these facts, were, in some cases, 

 made prior to 1852. But the men handling the engines had observed 

 this effect even before Clark ; he states that they rarely voluntarily 

 adopted ' a suppression of above 30 per cent.,' as they found the loss by 

 condensation greater than the gain by expansion. Describing the method 

 of this loss, this author goes on to say that ' to prevent entirely the 

 condensation of steam worked expansively, the cylinder must not only 

 be simply protected by the non-conductor — it must be maintained by 

 independent external means, at the initial temperature of the steam.' 

 He thus reiterates the principle expressed by Watt three-quarters of a 

 century before, and applies it to the newly-stated case. 



The same author, writing in 1877, says : ' The only obstacle to the 

 working of steam advantageously to a high degree of expansion in one 

 cylinder, in general practice, is the condensation to which it is subjected, 

 when it is admitted into the cylinder at the beginning of the stroke, by 

 the less hot surfaces of the cylinder and piston, the proportion of which 

 is increased so that the economy of steam by expansive working ceases to 

 increase when the period of admission is reduced down to a certain 

 fraction of the stroke, and that, on the contrary, the efficiency of the 

 steam is diminished as the period of admission is reduced below that 

 fraction.' The magnitude of this influence may be understood from the 

 fact that the distinguished engineer Loftus Perkins, using steam of 300 

 pounds' pressure, and attaining the highest economy known up to his 

 time, found his engine to consume 1'62 pounds of fuel per hour per 

 horse-power ; while this figure is now reached by engines using steam at 

 one-third that pressure and expanding about the same amount, and 

 sometimes less. 



Mr. Humphry s, Avriting a little later than Clark, shows the con- 

 sumption of fuel to increase seriously as the ratio of expansion is increased 

 beyond the very low figure which constituted the limit in marine engines 

 of his time. 



Mr. B. F. Isherwood, a chief engineer in the United States Navy, 

 and later chief of the Bureau of Steam Engineering:, seems to have been 

 the first to have attempted to determine, by systematic experiment, the 

 law of variation of the amount of cylinder-condensation with variation 

 of the ratio of expansion in unjacketed cylinders. Experimenting on 

 board the U.S.S. Michigan, he found that the consumption of fuel 

 and of steam was greater when the expansion was carried beyond about 

 one-half stroke than when restricted to lower ratios. He determined the 

 quantity of steam used, and the amount condensed, at expansions ranging 

 from full stroke to a cut-off at one-tenth. His results permit the deter- 

 mination of the method of variation, with practically satisfactory accu- 

 racy, for the engine upon which the investigation was made, and for 

 others of its class. It was the first of a number of such investigations 

 made by the same hand, and these to-day constitute the principal part of 



pp2 



