in condensing the steam and carrying its heat back again to the fire. [Footnote 175: See Trevithick's letters, July 8th, 1815, vol. ii., p. 80, and 7th, vol. i., p. 364; and 16th May, vol. i., p. 370; and patent of 1815, vol. i., p. 375.] The object or principle of this engine was to avoid the loss of heat, and the necessity for either condensing water or feed-water, as described in the letter and drawing of August 19th, 1830, but the detail was changed, mainly to facilitate construction. As in practice it might be impossible to fully attain those objects, preparation was made to get rid of the salt from such water as might be required as feed-water to make good the loss from leakage or other defects in the working of marine steam-engines. The specification states: "For the purpose of supplying the boiler with distilled water, in case there should be a deficiency in it, a small vessel made of two upright tubes, one within the other, is placed on the cap-piece. The inner tube is of the same diameter as the flue, and forms a continuation of it. The water being heated by the flue is converted into steam, which, passing into the condenser, is there reduced to water again, leaving the sediment or salt in the supplying vessel." Where water condensation was preferred the surface-air condenser could be converted into a surface-water condenser by a current of cold water in place of the air; in which case the air from the blowing cylinder was taken direct in to the fire-place or other means used for giving the necessary draught. Steam of about 135 lbs. to the inch was to be so expansively worked as at the finish of the stroke, on its escape to the condenser, to be no more than atmospheric pressure, or 15 lbs. to the inch--just the strength with which Watt preferred to commence his work in the cylinder.