In regard to the curvature of narrow aeroplanes, I have found that if one only desires to lift a large load in proportion to the area, the planes may be made very hollow on the underneath side; but when one considers the lift in terms of the screw thrust, I find it advisable that the planes should be as thin as possible, and the underneath side nearly flat. I have also found that it is a great advantage to arrange the planes after the manner shown in Fig. 87. In this manner the sum of all the spaces between the planes is equal to the whole area occupied by the planes; consequently, the air neither has to be compressed, spun out, nor driven forward. I am, therefore, able by this arrangement to produce a large lifting effect per square foot, and, at the same time, to keep the screw thrust within reasonable limits. [Illustration: Fig. 87.--The position of narrow aeroplanes arranged in such a manner that the air has free passage between them, and this arrangement has been found superior to arranging one above the other after the manner of a Venetian blind.] A large number of experiments with very narrow aeroplanes have been conducted by Mr. Horatio Philipps at Harrow, in England. Fig. 88 shows a cross-section of one of Mr. Philipps’ planes. Mr. Philipps is of the opinion that the air, in striking the top side of the plane, is thrown upwards in the manner shown, and a partial vacuum is thereby formed over the central part of the plane, and that the lifting effect of planes made in this form is therefore very much greater than with ordinary narrow planes. I have experimented with these “sustainers” (as Mr. Philipps calls them) myself, and I find it is quite true that they lift in some cases as much as 8 lbs. per square foot,[13] but the lifting effect is not produced in the exact manner that Mr. Philipps seems to suppose. The air does not glance off in the manner shown. As the “sustainer” strikes the air two currents are formed, one following the exact contour of the top, and the other that of the bottom. These two currents join and are thrown downwards, as relates to the “sustainer,” at an angle which is the resultant of the angles at which the two currents meet. These “sustainers” may be made to lift when the front edge is lower than the rear edge, because they encounter still air, and leave it with a downward motion. [13] In my early experiments I lifted as much as 8 lbs. per square foot with aeroplanes which were only slightly curved, but very thin and sharp. [Illustration: Fig. 88.--The very narrow aeroplanes, or sustainers, employed by Mr. Philipps. It has been supposed that the air in striking at A was deflected in the manner shown, but such is not the case. The air in reality follows the surface, as shown in the dotted line in the second illustration.]