214 MAGNUS ON THE DEVIATION OF PROJECTILES, 



greater atmospheric pressure existed than when the cylinder 

 was at rest. 



These differences of atmospheric pressure were most obser- 

 vable at the points where the direction of the current was tan- 

 gential to the cylinder. 



When the velocity of the air in the current was very great 

 in proportion to that generated by rotation, the vanes were 

 but very little deflected from the positions which the current 

 caused them to assume before the cylinder rotated. But when 

 the motion of the air produced by rotation was not much less 

 than that in the current, the vane on the side where both motions 

 had a like direction was driven strongly towards the cylinder ; 

 and then also the flag on the opposite side was strongly repelled. 



In passing to an explanation of these phaenomena, it may be 

 remarked, that they are intimately connected with those described 

 and explained by me in a memoir * On the Motion of Fluids*.' 



If a fluid issues through an orifice with a certain velocity 

 into a similar fluid mass, the pressure which exists, perpendi- 

 cular to the direction in which the mass moves, is smaller than 

 that which would be present at the same place in a condition 

 of rest. 



We can most easily convince ourselves of this decrease of 

 pressure by blowing air through a tube, and bringing the flame 

 of a candle near the end from which the air issues. As soon as 

 the velocity of the air is sufficiently great, the flame is seen to 

 move towards the current, and at last, by increasing the velo- 

 city, can be made to place itself quite at right angles to the 

 direction of the same. A similar decrease of pressure shows 

 itself when a current of air is directed against a fixed wall ; for 

 a flame held before a wall near the current does not move from 

 the wall, but towards it. 



Exactly the same occurs on the surface of a fixed cylinder, if 

 a current of air be directed against it. If the direction of this 

 current be perpendicular to the cylinder's axis, and at equal 

 distances from its surface, at opposite sides, two flames or two 

 vanes be placed, both move towards the cylinder, at least 



* Abhandlungen der Akademie der Wissenschaften zu Berlin fur 1848, 

 p. 135. Pogg. Ann. Ixxx. p. 1 ; and Phil. Mag. for January 1851. 



