580 report— 1884. 



our data in this direction. The writer, studying- these results, found 

 that the cylinder-condensation varied sensibly as the square I'oot of the 

 ratio of expansion, and this is apparently true for other forms and pi-o- 

 portions of engine. The amount of such condensation usually lies between 

 one-tenth and one-fifth the square root of that ratio, if estimated as a 

 fraction of the quantity of steam demanded by a similar engine having a 

 non-conducting cylinder. 



The state of the prevalent opinion on this subject, at the time of this 

 work of Clark and of Ishervvood, is well expressed by the distingnished 

 German engineer, Dr. Albans, who, writing about 1840, says of the 

 choice of best ratio of expansion : ' Practical considerations form the 

 best guide, and these are often left entirely out of view by mathematicians. 

 Many theoretical calculations have been made to determine the point, 

 but they appear contradictory and unsatisfactory.' Rcnwick, in 1848, 

 makes the ratio of initial divided by back pressure the proper ratio of 

 expansion, but correctly describes the effect of the steam-jacket, and 

 suggests that it may have peculiar value in expansive working, and that 

 the steam may receive heat from a cylinder thus kept at the temperature 

 of the ' prime ' steam. John Bourne, the earliest of now acknowledged 

 authorities on the management and construction of the steam-engine, 

 pointed out, at a very early date, the fact of a restricted economic expan- 

 sion. Rankine recognised no such restriction as is here under considera- 

 tion, considered the ratio of expansion at maximum efficiency to be the 

 same as that stated by Carnot and by other early writers, and only perceived 

 its limitation by commercial considerations, a method of limitation of 

 great importance, but often of less practical effect than is the waste 

 by condensation. In his ' Life of Elder ' (1871), however, he indicates the 

 existence of a limit in practice, and places the figure at that previously 

 given by Isherwood, for unjacketed engines. By this latter date, the 

 subject had become so familiar to engineers that a writer in ' London 

 Engineering,' in 1874, contemns writers who had neglected to observe 

 this limitation of efficiency as indulging in • mediaeval twaddle.' 



A few writers on thermodynamics finally came to understand the 

 fact that such a limitation of applied theory existed. M. G. A. 

 Hirn, who, better than probably any authority of his time or earlier, 

 combined a knowledge of the scientihe principles involved with practical 

 experience and experimental knowledge, in his treatise on thermodynamics 

 (1876), concludes: l Qu'il est absolument impossible d'edijier a priori tme 

 theorie de la machine a vapeur d'eau douce d'un charactere scientifique et 

 exact,' in consequence of the operation of the causes here detailed. While 

 working up his experiment upon the performance of engines, comparing 

 the volume of steam used with that of the cylinder, he had always found a 

 great excess, and had, at first, attributed it to the leakage of steam past 

 the piston ; but a suggestion of M. Leloutre set him upon the right 

 track, and he came to the same conclusion as had Watt so many years 

 before. He explains that errors of 30, or even up to 70 per cent, may 

 arise from the neglect of the considei'ation of this loss. Combes had 

 perceived the importance of this matter, and De Freminville suggested 

 the now familiar expedient of compression, on the return stroke as nearly 

 as possible to boiler pressure, as a good way to correct the evil. The 

 matter is now well understood by contemporary writers, and it has 

 become fully agreed, among theoretical writers as well as among prac- 

 titioners, that the benefit of extended expansion in real engines can only 



