AND ON A PHiENOMENON OF ROTATING BODIES. ^1^ 



would have left the field of the current, had it not been arranged 

 that the centrifugal fan could so follow it that the current always 

 acted in the direction of the beam yz. For this purpose the fan 

 was placed on a board AB, capable of turning on a vertical axis 

 c, and carrying on its other extremity a counter weight G. 



When the cylinder rotated without the current of air being 

 directed against it, the foi'mer remained stationary. It was also 

 stationary when the current was directed against it during its 

 non-rotation. If the current, however, struck against the rota- 

 ting cylinder, the latter, together with the beam yz, moved to- 

 wards the side on which the air, by rotation, and that in the 

 current, had a like direction, as shown by the arrows in fig. 3, 

 Plate III.; a indicates the direction of the current of air ; b that 

 of the cylinder's rotation ; and c that of the lateral motion^ 



If, by turning the board AB, the centrifugal fan was made to 

 follow the cylinder, the latter moved laterally as long as its 

 rotation continued to be pretty strong, often describing a com* 

 plete circle. When made to rotate in an opposite direction, it 

 moved towards the opposite side. When the current was in- 

 terrupted, the motion of the beam yz, to which the cylinder 

 was fixed, owing to its rather high moment of inertia, did not 

 immediately cease. If, during the rotation of the cylinder and 

 its lateral motion, together with the beam yz, the latter de- 

 ceived a push towards the opposite side — the centrifugal fan, 

 however, made to follow it — the motion caused by the push 

 soon ceased, and the primitive one was re-established. It cannot, 

 therefore, be doubted that the lateral motion of the cylinder was 

 caused by its rotation during the action of the current upon it. 



It is true that the current of air exercised also a pressure against 

 the surface of the ring in which the cylinder rotated, and when 

 this surface was not perpendicular to the direction of the cur- 

 rent, cylinder and ring moved towards one side, even when the 

 former did not rotate. That the observed deviation of the ro- 

 tating cylinder, however, was not caused by such a position of 

 the ring, is evident from the fact, that even when the plane of 

 the same, during the rotation of the cyUnder, made an angle of 

 45 degrees with the current, a lateral deviation towards the right 

 or left always ensued, according as the rotation of the cylinder 

 was to the right or left ; hence the lateral pressure against the 



R 2 



