34 HER OES OF INVENTION AND DISCO VER Y. 



quantity of steam for the creation of the vacuum. In order to 

 remedy this defect he was obliged, in repairing the model, to 

 diminish the column of water to be raised ; in other words, to 

 give the piston less to do, in compensation for its having to 

 descend, not through a perfect vacuum, but in opposition to a 

 considerable residue of undisplaced air. He also soon dis- 

 covered the reason why in this instance the steam sent up from 

 the boiler was not sufficient to fill the cylinder. In the first 

 place, this containing vessel, being made, not of cast-iron, as in 

 the larger engines, but of brass, abstracted more of the heat 

 from the steam, and so weakened its* expansion ; and secondly, 

 it exposed a much larger surface to the steam, in proportion to 

 its capacity, than the cylinders of the larger engines did, and 

 this operated still more strongly to produce the same effect. 

 Led by the former of these considerations, he made some ex- 

 periments in the first instance with the view of discovering some 

 other material whereof to form the cylinder of the engine which 

 should be less objectionable than either brass or cast-iron ; and 

 he proposed to substitute wood, soaked in oil, and baked dry. 

 But his speculations soon took a much wider scope; and, 

 struck with the radical imperfections of the atmospheric engine, 

 he began to turn in his mind the possibiHty of employing steam 

 in mechanics, in some new manner which should enable it to 

 operate with much more powerful effect. This idea having got 

 possession of him, he engaged in an extensive course of experi- 

 ments, for the purpose of ascertaining as many facts as possible 

 with regard to the properties of steam ; and the pains he took 

 in this investigation were rewarded with several valuable dis- 

 coveries. The rapidity with which water evaporates, he found, 

 for instance, depended simply upon the quantity of heat which 

 was made to enter it ; and this again on the extent of the surface 



