OM T^E TBEaBY OF C API WARY ATTRACTION. J3J 



may give it the convex form ^ s y; and if c d was the sur- 

 face in the branch A B, when that in the other branch was 

 Sey; let this lajjt become convex, as ^ y y, and the first 

 shall be « -B-; the difference of masses, in the latter case, 

 being to that in the former in a ratio approaching the more 

 nearly to that of 2 to I, the higher we can make the convex 

 surface J ^ y, and the nearer its sides at y and 5 are to per- 

 pendicularity with^ s y. 



Here then is the f rue and simple explanation ofAbat's 

 experiment, which in no respect depends on the figure of 

 the surface, in the fen fe that Mr, la Place means** 



Cor. 3rd. Let A B jt? r, fig. 6, represent a tube of such Corollary S> 

 intensity of attraction, that, if it be immersed in a fluid the 

 horizontal surface of which is y* c </, this surface shall 

 undergo no alteration. 



Suppose this tube cut off close to the surface; as, by 

 this, the part ofc (a qu;idrant to the radius of the sphere 

 of attraction) to which half its effect is owing (by the pro- 

 position) is taken away, the intensity of attraction of the 

 lower part /"op must be doubled to preserve the equilibrium : 

 and it plainly follows, that the intensity of attraction of the 

 fluid for itself is twice that which the tube before it was cut 

 off had for the fluid. 



May I not say, that so peculiar a demonstration, of a 

 theorem easily proved in other ways, is of itself sufficient to 

 establish the truth of this theory.^ 



Cor. 4th. If, in fig. 4, we suppose the branch E F to Corollary,4, 

 be cut off close to the surface (which I suppose horizontal) 

 and then to be of the same intensity of attraction with the 



♦ After the same manner is explained another experiment mentioned 

 by Mr. la Place, 1st. supplement, p. 60. Plunge a capillary tube into 

 water, and, having closed the lower, ocifice with the finger, draw it out 

 of the water. If we now remove the finger, the fluid will fall in the 

 tube, and form a convex drop at the lower orifice. But, when it has 

 ceased to descend, the height of the column always remains greater than 

 the height of the water in the tube, above the level, when it was plunged 

 in the fluid. " This excess (says Mr. la Place) is owing to the action of 

 the drop of water on the column.'''' 



The true explanation is the same as that I have pvet^^b.ove ofAbat's 

 ^periment, 



K 2 fluid 



