AND ON A PHiENOMENON OF ROTATING BODIES. 213 



that the deviation occurs towards the side on which the direc- 

 tion of rotation is opposite to that of the projectile's progressive 

 motion. 



On this account it became necessary to investigate, experi- 

 mentally, the existing pressure. To effect this, a current of air 

 was directed against a sphere made to rotate around a fixed axis. 

 Afterwards, however, in place of a sphere, a cylinder was adopted ; 

 inasmuch as with the latter a greater certainty of observation 

 was attainable. By means of a simple contrivance, the cylinder 

 could be so fixed that its axis coincided as near as possible 

 with the axis of rotation ; or it could be slightly displaced, so 

 that at will a concentric or excentric rotation could be im- 

 parted. 



The current of air, directed perpendicularly against the axis of 

 the cylinder, was of such a breadth that during its most excen- 

 tric rotation the cylinder remained always within the current. 

 It was further arranged, that throughout its whole width the 

 motion of the air in the current was uniform. To generate 

 the latter, a small centrifugal fan was usually employed, a view 

 of which is given in fig. 1, Plate II. In the drum F of the 

 same (of 6 in. diameter), a shaft, on which six paddle-shaped 

 fans were fixed, revolved; and thus throughout the whole 

 breadth (5 inches) of the drum the air was uniformly impelled, 

 and forced out from the orifice m n, of equal breadth. 



Small vanes, which were very moveable, served to detect 

 the changes of pressure which took place in the current of air 

 during the rotation of the cylinder. Two such, a and b. fig. 1, 

 Plate II., were so placed on each side of the cylinder, that the 

 pivots on which they turned were equally distant from the orifice 

 m w, and at equal distances also from the plane passing through 

 the cylinder's axis of rotation and the middle of the current. 

 When the cylinder did not rotate, the two vanes assumed the 

 direction of the current. As soon as the cylinder, however, 

 began to rotate, the vane placed on that side of it which moved 

 in the same direction as the current, approached ; whereas the 

 vane on the opposite side, where the motions of cylinder and 

 current were opposed to each other, receded from the cylinder. 

 Consequently, on the former side a less, and on the latter a 



