218 MAGNUS ON THE DEVIATION OF PROJECTILES, 



ficiently established, as I think, by the described experiments, 

 is, it is true, contrary to the hitherto accepted notions regarding 

 the resistance of the air, but is not on that account the less 

 correct. In order to free it from every objection, however, it has 

 yet to be proved that the difference of atmospheric pressures 

 against the opposite hemispheres is sufficiently great to produce 

 a deviation of the projectile itself. As we cannot yet measure 

 the magnitude of this difference, I have endeavoured to produce 

 in miniature a lateral motion in order thus to demonstrate that 

 the difference of atmospheric pressures is sufficiently great to 

 deflect the projectile from its course. 



To produce such a deflection, a light, hollow brass cylinder, 

 3 in. high and 2 in. in diameter, ab, fig. 2, Plate II., was used ; 

 very easily moveable around its axis, which turned on two 

 pivots, fixed in a metal ring. Cylinder and ring were fixed to 

 the end of a light wooden beam, yz, 4c. feet long, suspended in 

 its centre from a fine wire vw, 8 feet long, and thus forming 

 a kind of torsion balance. To keep the beam horizontal, it 

 carried at its other extremity a moveable counter weight p, 

 and it was fastened to the vertical wire by two strings, rt and 

 ru, from the points t and m, equidistant from the centre v. The 

 object of this connection was to avoid, as much as possible, 

 all oscillatory motion. For the same purpose a silk thread vc 

 descended from the centre v of the beam, and was fastened ex- 

 actly vertically under the wire w, so that the position of the 

 suspension-point v of the beam could not vary more than the 

 elasticities of the wire wv and silk thread vc permitted. 



To set the cylinder in rotation, a small pulley, e, was fixed on 

 the production of its axis, and a fine silk thread was wrapped 

 round the same. A sharp pull at this string gave the cylinder a 

 rotation, which continued for two or three minutes. During its 

 rotation a current of air was directed against it. To generate the 

 latter, the small centrifugal fan EF, described above (page 213), 

 was used, and so placed that the air moved in the direction of 

 the beam yz, from the point of suspension v towards the extre- 

 mity z, to which the cylinder was fixed ; the latter, in its ring, 

 whose plane was perpendicular to the current of air, was placed 

 vertically, at a distance of 2 inches, before the orifice mn of the 

 centrifugal fan. As soon as the cylindei; moved to one side, it 



