VOL. XXXI.] PHILOSOPHICAL TRANSACTIONS. 433 



If there be a funnel, as abc, fig. 1, pi. 12, full of water, of which the 

 wide end stands in a vessel of water bc, and the top of the funnel a ends in a 

 capillary tube open at a; the whole water will be sustained: viz. the column 

 Aa by the attraction of the circle of glass within the tube immediately above 

 it ; and all the rest of the columns of water, as rf, od, ec, Gg, &c. in some 

 measure by the attraction of the parts of the glass above them, as f, d, e, g: 

 and that the small columns or threads of water, Dd and ec, do not slide down 

 to pf and Gg, and so go quite down, seems to be owing to their cohesion with 

 the column Aa, which is sustained by the capillary tube a : for if you break 

 off the said tube at db, the whole water will presently sink down. 



As this solution is very different from what I had before given, and the re- 

 putation of that gentleman, whose great knowledge in experimental philosophy 

 is generally known, was sufficient to give weight to any of his opinions, I 

 thought myself under an obligation to examine his account of the experiment, 

 in order either to demonstrate its insufficiency, or to retract my own solution. 

 Accordingly at the next meeting of the society, I produced the following ex- 

 periment. 



The funnel, afgbc, fig. 2, whose lower part bcfg, was cylindrical to a con- 

 siderable height, and the top drawn out. into a fine tube at a, being filled with 

 water to the height bp, so that the surface of the water pg did not reach to 

 the arched part of the funnel, I touched the end a with a wetted finger, by 

 which a small quantity of water being insinuated into the capillary tube at a, 

 the water contained in the funnel was suspended above the level of the water 

 in the cistern de, as in the former experiment. 



In this experiment it is manifest, that the little columns, into which we 

 may suppose the cylinder of water, fgbc, to be divided, are no way sustained 

 by the attraction of the arched part of the glass above them, since they have 

 no contact with it. Nor is there any such middle column of water, which, by 

 its contact with the tube at top, is both sustained itself, and helps to support 

 the columns about it ; on the supposition of which two particulars that gen- 

 tleman's solution was founded. 



This experiment may be thus accounted for. The cylinder of water fgbc, 

 by its weight balances a part of the pressure of the atmosphere, which is in- 

 cumbent on the water in the cistern, and endeavours to force that cylinder 

 upwards. The rest of that pressure is balanced by the spring of the air, afg, 

 which is included between the cylinder of water fgbc, and the little column 

 of water in the capillary a. But, as this air by its spring presses equally every 

 way, it must balance as much of the pressure of the atmosphere on the small 

 column of water at a, as it does of that on the water in the cistern. The 

 VOL. VI. 3 K 



