76 The Right Hon. Lord Rayleigh [March 20, 



altogether from that prescribed by the classical hydro-dynamics, 

 according to which the character of the motion should be the same 

 behind as in front. The liquid refuses to close in behind, and a 

 region of more or less "dead water" is developed, entailing a greatly 

 increased resistance. To meet this Helmholtz, Kirchhoff, and their 

 followers have given calculations in which the fluid behind is sup-- 

 posed to move strictly with the advancing solid, and to be separated 

 from the remainder of the mass by a surface at which a finite slip 

 takes place. Although some difficulties remain, there can be no 

 doubt that this theory constitutes a great advance. But the surface 

 of separation is unstable, and in consequence of fluid friction it soon 

 loses its sharpness, breaking up into more or less periodic eddies, 

 described in some detail by Mallock (Fig. 8). It is these eddies which 

 cause the whistling of the wind in trees and the more musical notes 

 of the a^ohan harj). 



The obstacle to the closing-in of the lines of flow behind the 

 disk is doubtless, as before, the layer of liquid in close proximity to 



M5/fo 



Fig. 8. 



the disk, which at the edge has insufficient velocity for what is 

 required of it. It would be an interesting experiment to try what 

 would be the efi^ect of allowing a small "si3ill." For this purpose 

 the disk or blade would be made double, with a suction applied to 

 the narrow interspace. Relieved of the slowly moving layer, the 

 Uquid might then be able to close in behind, and success would be 

 witnessed by a greatly diminished resistance. 



When a tolerably fair-shaped body moves through fluid, the 

 relative velocity is greatest at the maximum section of the solid 

 which is the minimum section for the fluid, and consequently the 

 pressure is there least. Thus the water-level is depressed at and 

 near the midship section of an advancing steamer, as is very evident 

 in travelling along a canal. On the same principle may be explained 

 the stability of a ball sustained on a vertical jet as in a well-known 

 toy (shown). If the ball deviate to one side, the jet in bending 

 round the surface develops a suction pulling the ball back. As Mr. 

 Lanchester has remarked, the effect is aided bv the rotation of the 



