Much, he saw, yet remained to be done to obtain economical expenditure of steam. In particular the open-topped cylinder, whose walls were chilled at every descent of the piston by contact with atmospheric air, was an obvious source of inefficiency. He therefore determined not to expose the walls to the atmosphere at all, but to enclose all the space above the piston; and, thinking thus, he conceived the idea of replacing the air above the piston by steam, an equally powerful agent. The cylinder he proposed to maintain at a constant high temperature by means of a layer of hot steam with which he encased it, which he called a steam jacket. And so the atmospheric engine as left by Newcomen evolved into the _single-acting steam engine_ of Watt;--an engine in which steam was still used below the piston, only to displace air and provide a vacuized space for the downward motion of the piston; but in which steam now acted positively above the piston, in lieu of atmospheric air, to drive it down. It was still a sufficiently primitive form of prime mover. The piston was still lifted by the counterweight at the other end of the timber cross-beam; the engine had not yet developed the organs necessary for producing a satisfactory rotary motion. This step was shortly to follow. In 1769 Watt obtained his patent for the “double impulse,” as it was called; and by this step, by the transition from a single- to a double-acting engine, the possibilities of such machines for every variety of application first came into general view. This stage of the development showed to the full the ingenuity of Watt’s mechanical mind. By the invention of the slide-valve he distributed steam to the top and to the bottom of the cylinder, and in appropriate phase with these actions opened the two ends to the condenser; so that the piston was actuated positively and by an equal force on both up and down strokes. The chain by which the piston had been suspended was no longer adequate; it was replaced by a rod. A straight-line motion was required for the top end of the rod; so he formed a rack, to gear with the circular end or horse-head of the beam. But this noisy mechanism was soon superseded by another contrivance, the beautifully simple “parallel motion,” in which two circular motions are combined to produce one which is rectilinear. This was patented in ’84.