CORRESPONDENCE OF WATT WITH SMEATON. — FAILURE OF CONDENSATION BY SURFACE. — IMPROVEMENTS IN CONSTRUCTION OF PISTON. — METHOD OF PACKING. — IMPROVEMENTS IN BORING THE CYLINDERS. — DISADVANTAGES OF THE NEW COMPARED WITH THE OLD ENGINES. — GREATLY INCREASED ECONOMY OF FUEL. — EXPEDIENTS TO FORCE THE NEW ENGINES INTO USE. — CORRESPONDENCE WITH SMEATON. — EFFICIENCY OF FUEL IN THE NEW ENGINES. — DISCOVERY OF THE EXPANSIVE ACTION OF STEAM. — WATT STATES IT IN A LETTER TO DR. SMALL. — ITS PRINCIPLE EXPLAINED. — MECHANICAL EFFECT RESULTING FROM IT. — COMPUTED EFFECT OF CUTTING OFF STEAM AT DIFFERENT PORTIONS OF THE STROKE. — PRODUCES A VARIABLE POWER. — EXPEDIENTS FOR EQUALISING THE POWER. — LIMITATION OF THE EXPANSIVE PRINCIPLE IN WATT'S ENGINES. — ITS MORE EXTENSIVE APPLICATION IN THE CORNISH ENGINES. (77.) In a letter addressed by Watt to Smeaton, dated April, 1766, Watt refers to some of these practical difficulties which he had to encounter. "I have been," says he, "tormented with exceedingly bad health, resulting from the operation of an anxious mind, the natural consequence of staking everything [Pg146] upon the cast of a die; for in that light I look upon every project which has not received the sanction of repeated success. "I have made considerable alterations in our engine lately, particularly in the condenser. That which I used at first was liable to be impaired, from incrustations from bad water; therefore we have substituted one which works by an injection. In pursuing this idea I have tried several kinds, and have at last come to one, which I am not inclined to alter. It consists of a jack-head pump, shut at bottom, with a common clack bucket, and a valve in the cover of the pump, to discharge the air and water. The eduction steam pipe, which comes from the cylinder, communicates with this pump both above and below the bucket, and has valves to prevent anything from going back from the pump to the eduction pipe. The bucket descends by its own weight, and is raised by the engine when the great piston descends, being hung to the outer end of the great lever: the injection is made both into the upper part of this pump and into the eduction pipe, and operates beyond my ideas in point of quickness and perfection." Besides the difficulty arising from incrustation, Watt found the tubulated condensers, and indeed all other expedients for condensing by cold surfaces, subject to a fatal objection. They did not condense instantaneously, and although they were capable of ultimately effecting the condensation, yet that process was not completed until a great part of the stroke of the piston was made. Thus during more or less of the stroke the uncondensed steam resisted the piston, and robbed the moving power of a part of its effect. This objection has ever attended condensation by surface. [Illustration: _Fig._ 25.] [Illustration: _Fig._ 26.]