The principle of the apparatus is not altogether new. In 1872 there was tried on the Ohio River an arrangement termed the Brooks motor. It was composed of two drums, placed horizontally and parallel to each other. Round these there passed endless chains at equal spaces apart on the length of the drums, and to these chains were fixed wooden blades or arms of a curved form, and so jointed to the frames that they opened when moving in one direction, and closed down on the chain when moving in the other. In this machine the weight of the chains was a serious obstacle to obtaining any large amount of power. The whole apparatus was mounted on a heavy wooden scaffold, which proved an impediment to the flow of the river. Again, the resistance due to the surface of the returning blades and to their stiffness was found to be far from insignificant. In the present system Mr. Jagn has found, after many experiments, that the best effect was obtained when the parachutes were spaced apart at twice their diameter, and when the rope made an angle of 8 degrees to 10 degrees with the current. It is found that when open and in motion the parachutes never touch the bottom. This was the case with a rope containing 180 parachutes of 4 feet diameter, and working in a depth of only 6 feet. This is easily explained by the fact that the velocity of a current always diminishes as it approaches the bottom. Hence the pressure on the lower part of the parachute will be less than that on the upper part; but the former pressure tends to draw the parachute downward, while the latter tends to raise it to the top of the water. Thus, the latter being the larger, the parachute will always have a tendency to rise. In fact, it is necessary to sink the return pulley sufficiently deep to make sure that the parachutes will not emerge from the surface. For the same reason no intermediate supports are needed over the driving span; if any are needed it is for the return span, on which the parachutes are closed. Of course, if metal were used instead of hemp, the case would be entirely different, and intermediate supports would have to be used for anything but very moderate lengths.