Page:Popular Science Monthly Volume 12.djvu/147

Rh of musical instruments, making himself familiar with the sciences, and devising improvements in the construction of organs.

His reading was still very desultory; but the introduction of the Newcomen engine in the neighborhood of Glasgow, and the presence of a model in the college collections, which model was placed in his hands in 1763 for repairs, led him to study the history of the steam-engine, and to conduct for himself an experimental research into the properties of steam, using a set of improvised apparatus.

31. The Newcomen model, as it happened, had a boiler, which, although made to a scale from engines in actual use, was quite incapable of furnishing steam enough to work the engine.

It was about nine inches in diameter, and the steam-cylinder was two inches in diameter, and of six inches stroke of piston, arranged as in Fig. 13. This is a picture of the most carefully-preserved treasure in the collections of the University of Glasgow. Watt at once noticed the defect referred to, and immediately sought first the cause and then the remedy.

32. He soon concluded that the sources of loss of heat in the Newcomen engine—which loss would be greatly exaggerated in a small model—were: first, the dissipation of heat by the cylinder itself, which was of brass, and was both a good conductor and a good radiator; secondly, the loss of heat consequent upon the necessity of cooling down the cylinder at every stroke in producing the vacuum; and, finally, a loss of power was due to the existence of vapor beneath the piston, the presence of which vapor was a consequence of the imperfect method of condensation which characterizes the Newcomen engine.

He first made a cylinder of non-conducting material—wood soaked in oil and then baked—and found a decided advantage in the economy of steam thus secured.

He then conducted a series of experiments upon the temperature and pressure of steam at such points in the scale as he could readily reach, and, constructing a curve with his results, the abscissas representing temperatures, and the pressures being represented by the ordinates, he ran the curve backward until he had obtained