Among his multifarious pursuits, Watt had experimented a little in the powers of steam; but it was not till the winter of 1763-4, when a model of Newcomen's engine was put into his hands for repair, that he took up the matter in earnest. Newcomen's engine was then about the most complete invention of its kind; but its only value was its power of producing a ready vacuum, by rapid condensation on the application of cold; and for practical purposes was neither cheaper nor quicker than animal power. Watt, having repaired the model, found, on setting it agoing, that it would not work satisfactorily. Had it been only a little less clumsy and imperfect, Watt might never have regarded it as more than the "fine plaything," for which he at first took it; but now the difficulties of the task roused him to further efforts. He consulted all the books he could get on the subject, to ascertain how the defects could be remedied; and that source of information exhausted, he commenced a series of experiments, and resolved to work out the problem for himself. Among other experiments, he constructed a boiler which showed by inspection the quantity of water evaporated in a given time, and thereby ascertained the quantity of steam used in every stroke of the engine. He found, to his astonishment, that a small quantity of water in the form of steam heated a large quantity of water injected into the cylinder for the purpose of cooling it; and upon further examination, he ascertained the steam heated six times its weight of well water up to the temperature of the steam itself (212 deg.). After various ineffectual schemes, Watt was forced to the conclusion that, to make a perfect steam engine, two apparently incompatible conditions must be fulfilled--the cylinder must always be as hot as the steam that came rushing into it, and yet, at each descent of the piston, the cylinder must become sufficiently cold to condense the steam. He was at his wit's end how to accomplish this task, when, as he was taking a walk one afternoon, the idea flashed across his mind that, as steam was an elastic vapour, it would expand and rush into a previously exhausted place; and that, therefore, all he had to do to meet the conditions he had laid down, was to produce a vacuum in a separate vessel, and open a communication between this vessel and the cylinder of the steam-engine at the moment when the piston was required to descend, and the steam would disseminate itself and become divided between the cylinder and the adjoining vessel. But as this vessel would be kept cold by an injection of water, the steam would be annihilated as fast as it entered, which would cause a fresh outflow of the remaining steam in the cylinder, till nearly the whole of it was condensed, without the cylinder itself being chilled in the operation. Here was the great key to the problem; and when once the idea of separate condensation was started, many other