We have traced how succeeding engineers tried to prevent loss of heat. Trevithick took the first bold step, and aiming at the same object, made the boiler the steam-jacket for the cylinder, and in his patent of 1802 went still further and protected the boiler from external cold, and thus describes it:--"The steam which escapes in this engine is made to circulate in the case round the boiler, where it prevents the external atmosphere from affecting the temperature of the included water, and affords by its partial condensation a supply for the boiler itself."[162] So that a quarter of a century before the date of those Binner Downs experiments he had patented an engine having neither cylinder nor boiler exposed to the cooling atmosphere. The flues around the Binner Downs cylinder were difficult of control. Trevithick says the piston packing had not been injured, showing that observers thought it would be, and even the cylinder was endangered, for the writer, who stoked those heating flues, recollects the fires burning very brightly in them. The ready transmission of heat through thin metal, used by Trevithick in 1802 for heating feed-water, and in the cellular bottom of the iron ship of 1808, serving as a surface condenser,[163] and his experience in 1812, that "the cold sides of the condenser are sufficient to work an engine a great many strokes without any injection,"[164] still followed up in 1828 by condensing steam without the use of injection-water, led to what is since known as Hall's surface condenser. [Footnote 162: See patent specification, vol. i., p. 132.] [Footnote 163: See vol. i., p. 335.] [Footnote 164: See Trevithick's letter, 7th December, 1812, vol. ii., p. 18.] The following letter is in the handwriting of the present writer; it is the only one of Trevithick's numerous letters not written by himself:-- "MR. GILBERT, "HAYLE, _December 30th, 1828_. "Sir,--On the 28th inst. I received your printed report on steam, and have examined Farey's publication on sundry experiments made by Mr. Watt, which are very far from agreeing with the actual performance of the engines at Binner Downs. Mr. Watt says that steam at one atmosphere pressure expands 1700 times its own bulk as water at 212°, and that large engines ought to perform eighteen millions when loaded with 10 lbs. to the inch of actual work, the amount of condensing water being one-fortieth part of the content of the steam in the cylinder at one atmosphere strength, the cold condensing water at 50°, and when heated 100°. This would give for the Binner Downs engine, with a 70-inch cylinder, 10-inch stroke, 11 lbs. effective work on the inch (this load being one-tenth more than in Watt's table, by Farey, for an engine of this size and stroke), 57 gallons of injection-water for each stroke, and when working eight strokes per minute, to do eighteen millions would consume 11-1/4 bushels of coal per hour.