whereas in the largest paddle boats a dip of wheel of six feet is generally sufficient. Hence it is that nearly all fast steamers plying on rivers or shallow estuaries, and channel steamers running to ports where there is little water when the tide is low, are of necessity paddle-wheel. By employing two screws (one on each side instead of one amidships) the draft of water can be reduced by at least thirty per cent. Likewise, by increasing the number of revolutions smaller screws will do, and the draft of water may be still less, so that some thirty years ago, on the introduction of twin-screws, there were soon many ships built for services that had hitherto been monopolized by paddle boats;[15] and to-day, when there is a demand for higher speed and more power, and where paddle-wheels are not admissible, three screws are being employed. Ships have also been employed with four screws, viz., two at the bow and two at the stern, and, for the purpose for which they were required, answered very well indeed; but the worst possible place for a propeller is obviously at the bow, and therefore in these ships the bow screws were not very efficient, but they undoubtedly added somewhat to the power of the ship. In the same way some tug-boats have been fitted with a screw at each end. All attempts at propulsion with internal propellers—that is, by turbine wheels, pulsometers, ejectors, or by pumps—have failed in consequence of the great friction set up by the water in its rapid passage through the pipes from and to the sea; the motion must be rapid owing to the size of the pipes being necessarily restricted. The best experiment with this kind of propeller was made on a costly scale by the British Admiralty in 1866, when they fitted the iron-clad gun-boat Waterwitch, of 1,200 tons displacement, with a Ruthven’s hydraulic propeller, consisting of a horizontal turbine wheel drawing its water through the bottom of the ship and discharging it fore-and-aft-ways at each side, and driven by an engine of 160 nominal horse-power; and although this vessel was only 162 feet long, 32 feet broad, and drew 11 feet 4 inches of water, her speed was only a little over 9 knots, with an indicated horse-power of 801. The speed co-efficients whereby her performances could be compared with that of other ships were most disappointing. [Illustration: The Twin Screws of the City of New York]