with the sheet of tin as unity, and gave to every other material experimented with, a coefficient of the unity thus established. Upon testing a frame covered with tightly-drawn white silk, a considerable amount of air passed through, and with an angle of 1 in 14, the lift was only about double the drift. A piece of very open fabric, a species of buckram, was next tried, and with this the lift and drift were about equal. With closely-woven, shiny satin the coefficient was about ·80; with a piece of ordinary sheeting the coefficient was ·90; with closely-woven, rough tweeds, ·70; and with glass-paper about ·75. With a piece of tracing linen very tightly drawn, results were obtained identical with those of a sheet of tin, and with Spencer’s balloon fabric the coefficient was about ·99. I, therefore, decided to cover my aeroplanes with this material. It will be observed that the apparatus is so arranged that both the lift and the drift can be easily measured. [Illustration: Fig. 22.--Apparatus for testing the lifting effect of aeroplanes and condensers in an air blast. _k_, _k_ show two aeroplanes in position for being tested.] [Illustration: Fig. 23.--Apparatus for testing aeroplanes, condensers, etc., in an air blast. The opening is 3 feet square. Thin brass sustainers are shown in position for testing.] [Illustration: Fig. 24.--Cross-sections of bars of wood employed for ascertaining the coefficient of different forms.] [Illustration: Fig. 25.--Transverse sections of bars of wood experimented with for the purpose of ascertaining their coefficients as relates to a normal plane.] [Illustration: Fig. 26.--A flat aeroplane placed at different angles.]