Having ascertained the lifting effect of wooden aeroplanes of various forms and at varying velocities of the wind, and, also, the resistance offered by various bodies driven through the air, I next turned my attention to the question of condensation. I wished to recover as much water as possible from my exhaust steam. I had already experimented with Mr. Horatio Philipps’ sustainers, and I found that their lifting effect was remarkable. A curious thing about these aeroplanes was that they gave an appreciable lift when the front edge was rather lower than the rear. I therefore determined to take advantage of this peculiar phenomenon, and to make my condenser tubes as far as possible in the shape of Mr. Philipps’ sustainers. Fig. 30 shows a section of one of these tubes, in which _a_, _a_ is the top surface, _b_ a soldered joint, and _c_ the steam space. These were mounted on a frame as shown at _a_ (Fig. 31). I had already found that bodies placed near to each other offered an increased resistance to the air, but by placing these sustainers in the manner shown this was avoided, as the air had sufficient space to pass through without being either driven forward or compressed. It was found by experiment that the arrangement of tubes or sustainers, shown at _d_, _d_ (Fig. 31), was very efficient as a condenser, but it gave a very heavy drift and no lifting effect at all; whereas, on the other hand, the arrangement shown at _a_ was equally efficient, and, at the same time, gave a decided lifting effect. When twelve of these tubes or sustainers were placed at an angle of 1 in 12, the lifting effect was 12·63 lbs. and the drift 2·06 lbs. It was found, however, that a good deal of the drift was due to the wind getting at the framework that was used for holding the sustainers in position. With a wind velocity of 40 miles an hour and a temperature of 62° F., 2·25 lbs. of water were condensed in five minutes, and, while running, the back edge of the sustainers was quite cool. At another trial of the same arrangement under the same conditions, the lift was 11 lbs. and the drift 2·63 lbs. It is quite possible on this occasion that the metal was so extremely thin that the angles were not always maintained; consequently, that no two readings would be alike. It was found at this point that the belt was slipping, and a larger pulley was put on the driving shaft of the screws; and under these conditions, with a wind of 49 miles per hour and an angle of 1 in 8, the lifting effect ran up to 14·87 lbs. with a drift of 2·44 lbs., and the condenser delivered 2·87 lbs. of water from dry steam in five minutes. The weight of metal in this condenser was extremely small, the thickness being only about 1/500 of an inch. This condenser delivered the weight of the sustainers in water every five minutes. They should, however, have been twice as heavy. Cylinder oil was now introduced with the steam in order to ascertain what effect this would have.