Mr. Garrett then interested Mr. Nordenfelt, the inventor of the celebrated Nordenfelt gun, in his boat. Mr. Nordenfelt improved upon Garrett's boat and built vessels for Greece, Turkey, and Russia. His first boat was sixty-four feet in length by nine feet beam, with a displacement of about sixty tons. The method of submerged control, which he devised, consisted of the use of two downhaul screws located in sponsons on either side of the vessel. These screws were operated by bevel gears and were run at sufficient speed to overcome the reserve of buoyancy. The vessel was intended to be always operated with a reserve of buoyancy. To submerge, therefore, it was necessary to run the propellers at a speed sufficient to exert a thrust to overcome this buoyancy and pull her bodily under water. After reaching the desired depth, forward motion was then to be given by the usual screw propeller, and she was expected to make progress on a level keel and in a horizontal plane. The level keel was to have been maintained by the use of a horizontal rudder placed in the bow. This method of submerged control for submarine vessels of moderate speed seems to me to be an excellent one in principle. I have been surprised that further development has not been made along these lines. I think the final abandonment of the Nordenfelt type of vessel was due to failure in carrying out the details of design rather than to faulty basic principles. A former chief engineer of Mr. Nordenfelt informed me that the heat from the large amount of hot water stored up in the reservoirs--for submerged power--made the interior of the vessels almost unbearable for the crew when the hatches were shut down, and that he did not believe the submarines ever made any submerged runs after being delivered. I also judge, from his description of his experiences with the vessels, that they lacked longitudinal stability and were difficult to hold in the horizontal position, which Mr. Nordenfelt claimed was a _sine qua non_ for a submarine boat. I concur in this claim. In an article on his boats, Mr. Nordenfelt stated that they were very sensitive, and that he had purposely made them so in order that the horizontal rudder might easily maintain the boat in a horizontal position. My experience has led me to prefer great statical stability rather than sensitiveness. Mr. Nordenfelt's boats had means for discharging the smoke from the fires under the water. This was done so as not to betray the submarine's position to surface vessels. He also seems to have been the first to incorporate torpedo tubes within his hull for the discharge of the Whitehead torpedo. The Spanish Lieut. Isaac Peral built, in 1887, a vessel in which the motive power was supplied from electric accumulators. It was operated by the usual vertical and horizontal rudders. Its submerged control was bad, but its electric propulsive system worked well.