Denny was remarkable, in so far as the firm had never produced, or had to do with, any craft faster than 15 or 16 knots. The attainment not only of the speed guaranteed, but of the better part of a knot in excess of that speed, was triumphant testimony to the skill and care brought to bear upon the undertaking. In this case, at least, the result was not one due to a previous course of "trial and error" with actual ships, but was distinctly due to superior practical skill, backed and enhanced by knowledge and use of specialized branches in the science of marine architecture. Messrs. Denny are the only firm of private shipbuilders possessing an experimental tank for recording the speed and resistance of ships by means of miniature reproductions of the actual vessels, and to this fact may safely be ascribed their confidence in guaranteeing, and their success in obtaining, a speed so remarkable in itself and so much in excess of anything they had previously had to do with. Confirmatory evidence of their success with the Belgian steamers is afforded by the fact that they have recently been instructed to build for service between Stranraer and Larne a paddle steamer guaranteed to steam 19 knots, and have had inquiries as to other high speed vessels. In estimating the power required for vessels of unusual types or of abnormal speed, where empirical formulae do not apply, and where data for previous ships are not available, the system of experimenting with models is the only trustworthy expedient. In the case of the Czar's extraordinary yacht, the Livadia, already referred to, it may be remembered that previous to the work of construction being proceeded with, experiments were made with a small model of the vessel by the late Dr. Tideman, at the government tank at Amsterdam. On the strength of the data so obtained, coupled with the results of trials made with a miniature of the actual vessel on Loch Lomond, those responsible for her stipulated speed were satisfied that it could be attained. The actual results amply justified the reliance placed upon such experiments. The design of many of her Majesty's ships has been altered after trials with their models. This was notably the case in connection with the design of the Medway class of river gunboats. The Admiralty constructors at first determined to make them 110 ft. long, by only 26 ft. in breadth. A doubt arising in their minds, the matter was referred to the late Mr. Froude, who had models made of various breadths, with which he experimented. The results satisfied the Admiralty officers that a substantial gain, rather than a loss, would follow from giving them much greater beam than had been proposed, and this was amply verified in the actual ships.