The high-pressure steam required a less amount of injection-water to condense it than the low-pressure steam, in proportion to the work done, showing the Watt rule and the Watt experience to be inapplicable to high-pressure engines; for instead of 57 gallons of injection-water the Binner Downs engine with steam of 45 lbs. to the inch required but 15-1/2 gallons of injection-water, and this amount was further reduced to 13 gallons by superheating the steam; this roughly agrees with the coal consumed, or in other words, with the amount of heat to be carried off by injection-water: the Watt rule giving 11-1/4 bushels as the fair allowance for low-pressure steam vacuum engines, while the high-pressure steam vacuum engine burnt but 4-1/4 bushels. This was further reduced to 3 bushels by superheating. Those facts led to the idea that if the steam pressure was sufficiently increased, condensation might be carried out without any injection-water, by the transmission of the heat in the steam through the metal sides of the condenser. An experiment was at once made by removing the Watt condenser and injection-water, as he had done seventeen years before,[169] using in their stead a thin copper surface-condenser immersed in cold water, producing, within 1/2 lb. on the inch, as good a vacuum as when injection-water was used, leading to the conclusion,-- [Footnote 169: See Query 3rd, vol. ii., p. 19.] "It is my opinion that high steam will expand and contract with a much less degree of heat or cold, in proportion to its effect, than what steam of atmosphere strong will do. I intend to try steam of five or six atmospheres strong, and partially condense it down to nearly one atmosphere strong, and then by an air-pump of more content than is usual to return the steam, air, and water back into the boiler again, and by a great number of small tubes, with greatly heated surface sides, to reheat the returned steam." This, in practical words, is the surface condenser by which the used steam is returned to the boiler in the form of water. The more general use of high-pressure steam of 70 or 90 lbs. to the inch, increasing its expansive force on one side of the piston by superheating it on its passage through numbers of small tubes, and decreasing its expansive force on the other side of the piston by cooling it in passage through similar tubes exposed to cold, is partly effected in steamboats, but has not yet been attempted in engines on the road. After a month's further consideration he wrote:--