in the boiler, the feed-water entering through a spring valve, the spring valve being adjusted in such a manner that the pressure on the water is always 30 lbs. per square inch in excess of the boiler pressure. This fall of 30 lbs. in pressure acts upon the surrounding hot water which has already passed through the tubes, and drives it down through a vertical outside tube, thus ensuring a positive and rapid circulation through all the tubes. This apparatus is found to work extremely well. A little glass tube at the top provided with a moving button, indicates exactly how many pounds of water per hour are passing into the boiler. By this means, the engineer is not only enabled to ascertain at a glance whether or not the pumps are working, but also to what degree they are working. Water may be considered as 2,400 times as efficient as air, volume for volume, in condensing steam. When a condenser is made for the purpose of using water as a cooling agent, a large number of small tubes may be grouped together in a box, and the water may be pumped in at one end of the box and discharged at the other end through relatively small openings; but when air is employed, the tubes or condensing surface must be widely distributed, so that a very large amount of air is encountered, and the air which has struck one tube and become heated must never strike a second tube. In order to accomplish this, I make my condenser something in the form of a Venetian blind, the tubes being made of very thin copper and each tube in the form of a small aeroplane. These were driven edgewise through the air, so that the actual volume of air passing between them is several thousand times greater than the volume of water passing through a marine condenser. I find that with such a condenser I can recover the full weight of the copper tubes in water every five minutes, and if I use aluminium, in half that time. Moreover, experiments have shown that a condenser may be made to sustain considerably more than its own weight and the weight of its contents in the air, and that all the steam may be condensed into water sufficiently cool to be pumped with certainty. I find that the most advantageous position for the condenser is immediately after the screw propellers. In this case, if the machine is moving through the air at the rate of 50 miles an hour, and the slip of the screws is 15 miles an hour, it follows that the air will be passing through the condenser at the rate of 65 miles an hour. At this velocity, the lifting effect on the narrow aeroplanes forming the condenser is very great, and at the same time the steam is very rapidly condensed. The tubes are placed at such an angle as to keep them completely drained and prevent the accumulation of oil, the steam entering the higher end and the water being discharged at the lower end. .       .       .       .       .       .       .       .       . EXPERIMENTS WITH SMALL MACHINES ATTACHED TO A ROTATING ARM.