Steerage is of course the main difficulty in these pairs. Three different sorts of apparatus have been used in them. Two of these are much of the same sort. One, generally in use to this day, consists of two bars projecting from the stretcher, and working horizontally in slits cut in the board. The foot presses against one bar or other to direct the rudder, Another process is to fix a shoe to the stretcher, in which the oarsman places his foot. This shoe works laterally. The third is one tried by the writer in 1868. Every inventor thinks his goose a swan, and possibly the writer is over-sanguine as to the merits of his own hobby. It consists of two bars laid on the stretcher, like a very widely opened letter V, the arms of the V pointing in the direction of the sitter. Each arm is hinged at the apex of the V. The stretcher is grooved, so that either arm can be pressed into the groove, flush with the surface of the stretcher. Behind each bar is a spring. The bars cross the stretcher just about the ball of the foot. The hinge is sunk deep in the wood, so that the arms of the levers do not begin to project above the wood till some 5 inches on either side of the centre of the stretcher. The feet are placed in ordinary rowing pose, in the middle of the V, where the levers lie below the flush surface of the stretcher. The strap, though tight, has a _wide_ loop, to admit of slight lateral movement of the feet. To put on rudder either foot is slipped half an inch or so outward. This brings it on to the lever of that side, and the pressure of the foot drives the lever flush. This pressure and movement of the lever, by means of another small lever and swivel outside the gunwale, in connection with it, works the rudder line. When steerage enough has been obtained, a half-inch return of the foot to its normal pose releases the lever, and the spring behind it at once brings it to _status quo ante_. Now in the other two mechanisms above cited, the same foot has to steer _both_ ways. Hence, for one of the two directions, the toe must turn in like a pigeon's. This must, for the moment, cripple leg-work, especially on slides. Again, with lateral movement in first and second machines, it is difficult for the steerer to know to exactness when his rudder is 'off.' He may, in returning it after steerage, leave it a trifle on, or carry it the other way too far. If so, he has to counter-steer a stroke or two later, till he feels that his rudder is free and trailing. The writer claims for his own invention that it never removes the feet from the proper outward-turned pose against the stretcher, and that the springs under the lever ensure the rudder swinging back and 'trailing' so soon as a lever is released. Whatever apparatus is used, _wires_, not strings, should lead the rudder, and should not be too tight; they will pull enough, though slightly loose.