VOL. LXXIII.] PHILOSOPHICAL TRANSACTIONS. 3t>3 



a perpendicular shaft or roller, which turned freely in brass sockets, fixed into 

 the floor and bridge, on pivots of hardened steel -^ of an inch in diameter. On 

 each side of this roller was extended an arm of deal, feather-edged, and supported 

 by stays of the same material, feathered in the same manner, to oppose as little 

 surface as possible to the air when in motion. Round the upper part of this 

 roller was wound a string of cat-gut, which, passing over pullies properly dis- 

 posed, was fastened to a scale that descended into the well of an adjoining stair- 

 case. The extremity of these arms described a space of more than 40 feet in 

 every revolution, the weight descending in the same time only 6 inches. The 

 time in all the following experiments was the same; and, as each revolution was 

 performed in 4 seconds, the velocity of the end of the arm on which the sur- 

 face was fixed, was at the rate of about 7 miles an hour.* 



The first figure that I tried was a parallelogram of tin, 9 inches long, and 4 

 inches wide. Its longer side was placed parallel to the floor, at the extremity 

 of one of the arms. Its shorter sides were inclined to an angle of 45 degrees 

 from the perpendicular, and in this situation it was carried round with its surface 

 against the air. After -\- suffering it to revolve till its motion was become uni 

 form, I put as much weight into the scale as moved it with a velocity of 5 turns 

 in 20 seconds. I then changed the situation of the parallelogram, placing its 

 shorter sides parallel to the floor, and inclined to the same angle as before. I 

 now found, that more weight was required to produce the same velocity, though 

 the quantity of surface was the same as in the preceding experiment. The 

 weight necessary to put the machine alone in motion, with the velocity above- 

 mentioned, was 2± lb. When it carried the parallelogram with one of its shorter* 

 sides downwards, it required 4-L lb. additional weight; and when the parallel- 

 ogram was reversed, another half pound was barely sufficient to give it the same 

 velocity. The difference therefore, occasioned by placing the same parallel- 

 ogram with its longer or shorter sides inclined from the direction of its motion, 

 was equal to T V °f tne greatest resistance. 



It is to be observed, that in these two experiments the mean velocity of the 

 plane was not the same, as its extremity extended farther from the centre of the 

 machine in one than in the other. This is strictly true; but the size of the 

 parallelogram bore so small a proportion to the length of the radius to which it 

 was fastened, that the error arising from this circumstance is scarcely perceptible, 



* This contrivance of machinery is manifestly of nearly the same nature as that invented by Mr. 

 Robins for the same use, and described in the 1st volume of his works, p. 200, new edition, in the 



year 1805. 



f Instead of saying after suffering it to revolve he put the weight into the scale, this ingenious gen- 

 tleman must have meant to say, that after putting in the weight he suffered the machine to revolve. 



J Query, longer 5 



• 3 A 2 



