Trevithick's first idea of steam navigation appears to have arisen in 1804, though his specification was not dated till 1808;[152] and in this, as in almost every other important step in his life, he relied very much on the sound judgment and sympathetic advice of his friend Davies Gilbert. Paddle-wheels, however (the original mode of propulsion), were found cumbrous to ships, especially in heavy weather; and this led Trevithick to the invention of the screw-propeller, a design for which he laid before the Navy Board in 1812, but without effect. The patent was not dated till the 6th of June, 1815; nor was success assured until after the busy engineer had left his native county on a voyage to Peru, which seemed to hold out promises of proving an El Dorado for him. His son and biographer, taking into consideration the numerous marine inventions and appliances of his progenitor, claims for his father, and not without much show of justice, that he may be regarded as the originator of our present iron steam fleet. But, in fact, the man's versatility in his profession seems to have been unbounded. In 1813 he was the life and soul of the arrangements for constructing the Plymouth Breakwater; then he turns his attention to the manufacture of agricultural engines, and constructs the first steam thrashing-machine, which was until recently at work at Trewithan, in Probus, but now occupies a place of honour in the South Kensington Museum of Patents, together with a boiler of curious construction by him.[153] This invention, like some others of Trevithick's, seems to have been almost still-born; yet it was destined, as we now know, to re-appear as a powerful factor in the development of agriculture. His next great stride was the new 'pole-puffer-engine' of 1816, in connection with which squabbles arose between the engineer and his relations, Henry Harvey and Andrew Vivian, both of whom were said to have been moved to jealousy by Trevithick's having arranged to get the castings for the first of these engines (viz., that for Wheal Herland, near Gwinear) made at Bridgenorth, instead of at Hayle. The first trial of the new engine, when at length set up, seems to have been somewhat of a failure, owing to the inaccurate way in which it was made. Here is an eye-witness's amusing account of its starting: