142 SOME AERONAUTICAL EXPERIMENTS. 



ers, and after less than one minute's actual practice had made a glid( 

 of more than 300 feet, at an angle of descent of 10-, and with { 

 machine nearly twice as large as had previously been considered safe 

 The trouble with its control, which has been mentioned, we believec 

 could be corrected when we should have located its cause. Severa 

 possible explanations occurred to us, l)ut we tinally concluded that th( 

 trouble was due to a reversal of the direction of th(^ travel of the centei 

 pressure at small angles. In deeply curved surfaces the center o: 

 pressure at 90° is near the center of the surface, but moves forvvart 

 as the angle becomes less, till a certain point is reached, varying witl 

 the depth of curvature. After this point is passed, the center o: 

 pressure, instead of continuing to move forward, with the decreasing 

 angle, turns and moves rapidly toward the rear. The phenomena ar( 

 due to the fact that at small angles the wind strikes the forward pari 

 of the surface on the upper side instead of the lower, and thus thi.- 

 part altogether ceases to lift, instead of being the most effective 

 part of all, as in the case of the plane. Lilienthal had callec 

 attention to the danger of using surfaces witli a curvature as greai 

 as 1 in 8, on account of this action on the upper side; but h( 

 seems never to have investigated the curvature and angle at whicl 

 the phenomena entirely cease. My brother and I had never madf 

 any original investigation of the matter, but assimied that a curvatu ( 

 of one in twelve would be safe, as this was the curvature on whicL 

 Lilienthal l)ased his tables. However, to he on the safe side, insteac 

 of using the arc of a circle, we had made the curve of our machine 

 very abrupt at the front, so as to expose the least possible area to thi.- 

 downward pressure. While the machine was Iniilding Messrs. Huff- 

 aker and Spratt had suggested that we would lind this reversal of the 

 center of pressure, but we believed it sufficiently guarded against. 

 Accordingh^ we were not at first disposed to believe that this reversa, 

 actually existed in our machine, althougli it offered a perfect exphma- 

 tion of the action we had noticed in gliding. Our peculiar plan oi 

 control by forward surfaces, instead of tails, was leased on the assump- 

 tion tiiat tlic center of pressure would continue to move farthei' and 

 farther forward as the angle of incidence became less, and it Avill be 

 readily perceived that it would make quite a difference if the front 

 surface instead of counteracting this assumed forward traxel should 

 in reality be expediting an actual 1)ackward movement. For seNcral 

 days we were in a state of indecision, but wtn'e tinally convinced by 

 o])serving the following phenomena (tig. 1): We had removed the 

 upper surface from the machine and were flying it in a wind to see at 

 what angles it would be supportcnl in winds of different strengths. 

 We noticed that in light winds it tlew in the upper position shown in 

 the figure, with a strong upward pull on the cord r. As the wind 

 became stronger the angle of incidence became less, and the surface 

 tlew^ in the position shown in the middle of the figure, with a slight 



