Trevethick in 1804 constructed a locomotive for the Merthyr and Tydvil Rail in South Wales, which succeeded in drawing ten tons at five miles an hour. The boiler was of cast-iron, with a one-cylinder engine, spur gear and a fly-wheel on one side. He sent the waste steam into the chimney, and by this means was very nearly arriving at the blast-pipe, afterwards the great and important discovery of George Stephenson. The jumping motion on the bad roads, however, caused it constantly to be dismounted, and it was given up as a practical failure, being sent to work a large pump at a mine. Trevethick was satisfied with a few experiments, and then gave it up for what he thought more profitable speculations, and no further advances were made in locomotives for some years. An imaginary difficulty seems to have been among the obstacles to its progress. This was the supposition that if a heavy weight were to be drawn, the grip or bite of the wheels would not be sufficient, but that they would turn round and leave the engines stationary, hence Trevethick made his wheels with cogs, which of course tended to cause great jolts, as well as being destructive to the cast-iron rails. A Mr. Blenkinsop of Leeds patented in 1811 a locomotive with a racked or toothed rail. It was supported on four wheels, but they did not drive the engine; its two cylinders were connected to one wheel behind, which was toothed and worked in the cog-rail, and so drove the engine. It began running on Middleton Coal Rail to Leeds, three and a quarter miles, on the 12th August 1812, and continued a great curiosity to strangers for some years. In 1816 the Grand Duke Nicholas of Russia saw this engine working with great interest and expressions of no slight admiration. An engine then took thirty coal-waggons at three and a quarter miles in an hour.