Stephenson continued to study and experiment, with a view to the improvement of his locomotive and the railroad. He introduced better methods of track-laying and of jointing the rails, adopting a half-lap, or peculiar scarf-joint, in place of the then usual square-butt joint. He patented, with these modifications of the permanent way, several of his improvements of the engine. He had substituted forged for the rude cast wheels previously used,[53] and had made many minor changes of detail. The engines built at this time (1816) continued in use many years. Two years later, with a dynamometer which he designed for the purpose, he made experimental determinations of the resistance of trains, and showed that it was made up of several kinds, as the sliding friction of the axle-journals in their bearings, the rolling friction of the wheels on the rails, the resistance due to gravity on gradients, and that due to the resistance of the air. [53] The American chilled wheel of cast-iron, a better wheel than that above described, has never been generally and successfully introduced in Europe. These experiments seemed to him conclusive against the possibility of the competition of engines on the common highway with locomotives hauling trains on the rail. Finding that the resistance, with his rolling-stock, and at all the speeds at which he made his experiments, was approximately invariable, and equivalent to about 10 pounds per ton, and estimating that a gradient rising but 1 foot in 100 would decrease the hauling power of the engine 50 per cent., he saw at once the necessity of making all railroads as nearly absolutely level as possible, and, consequently, the radically distinctive character of this branch of civil engineering work. He persistently condemned the "folly" of attempting the general introduction of steam on the common road, where great changes of level and an impressible road-bed were certain to prove fatal to success, and was most strenuous in his advocacy of the policy of securing level tracks, even at very great expense. Taking part in the contest, which now became a serious one, between the advocates of steam on the common road and those urging the introduction of locomotives and their trains on an iron track, he calculated that a road-engine capable of carrying 20 or 30 passengers at 10 miles per hour, could, on the rail, carry ten times as many people at three or four times that speed. The railway-engine finally superseded its predecessor--the engine of the common road--almost completely.