74 THE SCIENTIFIC PAPERS OP 



densing or boiling point, being in the case of atmospheric steam 

 an average of 5 times the rate of air up to 230 and 3 times up to 

 260 ; diminishing in fact in a ratio that could be represented by 

 a hyperbolic curve, approaching gradually at higher temperatures 

 to the same uniform rate of expansion as air which would be 

 represented by the asymptote to the curve. On this assumption 

 it followed that there would be a certain advantage gained in 

 applying heat direct to expand the bulk of steam, instead of 

 employing the same heat to generate an additional quantity of 

 ordinary steam ; but the extent of gain that could be attained was 

 not large, since the rise of temperature was practically limited to a 

 moderate range, and the specific heat of steam near the point of 

 saturation was proportionately great. 



The most important source of economy in superheating steam 

 was no doubt to be found in preventing condensation in the 

 cylinder ; for the consequence of condensation of any portion of 

 the steam on entering the cylinder was the total loss of the power 

 of that steam during the remainder of the stroke. But in order 

 to effect this object completely he considered that it was not 

 enough to supply extra heat sufficient to heat the metal of the 

 cylinder up to the temperature of saturated steam of the maximum 

 pressure, whether by superheating the steam or by a steam-jacket 

 round the cylinder ; for although this were done, there must still 

 be condensation in the cylinder during the expansion of the steam, 

 on account of the loss of heat accompanying the development of 

 moving power from the pressure of the steam on the piston ; since 

 there would be as much heat lost from the steam (or so to speak 

 rendered latent in the form of power) as was equivalent to the 

 total work done by the steam, according to the accepted theory of 

 the mutual convertibility of heat and power, by which according 

 to Joule's results one unit of heat (the quantity of heat required 

 to raise the temperature of I Ib. of water 1 Fahr.) was equivalent 

 to or became converted into a power of 770 ft.-lbs. (770 Ibs. lifted 

 1 foot). This amount of heat lost during the expansion of the 

 steam must consequently be supplied by superheating the steam 

 originally admitted to the cylinder in order entirely to prevent 

 condensation ; and this appeared about the point to which super- 

 heating could be carried advantageously, and generally speaking 

 required an addition of about 100 to the natural temperature of 



