He first made a cylinder of non-conducting material--wood soaked in oil and then baked--and obtained a decided advantage in economy of steam. He then conducted a series of very accurate experiments upon the temperature and pressure of steam at such points on the scale as he could readily reach, and, constructing a curve with his results, the abscesses representing temperatures and the pressures being represented by the ordinates, he ran the curve backward until he had obtained closely-approximate measures of temperatures less than 212°, and pressures less than atmospheric. He thus found that, with the amount of injection-water used in the Newcomen engine, bringing the temperature of the interior, as he found, down to from 140° to 175° Fahr., a very considerable back-pressure would be met with. Continuing his examination still further, he measured the amount of steam used at each stroke, and, comparing it with the quantity that would just fill the cylinder, he found that at least _three-fourths was wasted_. The quantity of cold water necessary to produce the condensation of a given weight of steam was next determined; and he found that one pound of steam contained enough heat to raise about six pounds of cold water, as used for condensation, from the temperature of 52° to the boiling-point; and, going still further, he found that he was compelled to use, at each stroke of the Newcomen engine, _four times as much injection-water as should suffice to condense a cylinder full of steam_. This confirmed his previous conclusion that three-fourths of the heat supplied to the engine was wasted. Watt had now, therefore, determined by his own researches, as he himself enumerates them,[36] the following facts: [36] Robison's "Mechanical Philosophy," edited by Brewster. "1. The capacities for heat of iron, copper, and of some sorts of wood, as compared with water. "2. The bulk of steam compared with that of water. "3. The quantity of water evaporated in a certain boiler by a pound of coal. "4. The elasticities of steam at various temperatures greater than that of boiling water, and an approximation to the law which it follows at other temperatures. "5. How much water in the form of steam was required every stroke by a small Newcomen engine, with a wooden cylinder 6 inches in diameter and 12 inches stroke. "6. The quantity of cold water required in every stroke to condense the steam in that cylinder, so as to give it a working-power of about 7 pounds on the square inch."