406 Particles of Fluids have no Cohefim. [Book VII. 



will defcend through the aperture, fliding between the 

 two columns a and 4, which are fupported by the 

 paiis c,i the bottom of the veffel b and r, all the 

 moveable parts of which become (if I may ufe the 

 expreliio") imall rollers, which retard the fall only 

 ir a v.Ty iiight degree. This effect is the refult of 

 tu,- fmail degree of cohefion between the parts of the 

 fluid. If the columns i and 2 on the one part, -and 

 4 and j ~n the other, were compofed of parts adhering 

 together, they v/culd retard each other in their defcent 

 during their whole length, in the fame manner as a 

 wax candle would do; and by the fall of the column 3, 

 a void would be made between them. But as all the 

 particles are exiremely minute, moving eafily upon 

 each ether, they defcend when the fummit of the co- 

 lurr.i: 3 begins to defcend, having no longer any fup- 

 port fi om that fide ; and the fuperficies of the whole 

 jiwfs a -icends in the fame manner, though only one 

 of the columns caufed the flow from 'its fall. When 

 the parts have a degree ofvifcofity, as thofe of oily 

 fluids, or when the mafs of the flowing liquor has 

 much more of breadth than of height, the void which 

 the defcending column leaves above it is eafily per- 

 ceived, for then the furface, inftead of being plane and 

 even, is noilow in the middle, and afliimes a funnel- 

 iike form, becaufe the adjacent parts do not arrive 

 with iufficient fwiftnefs to replace thofe which defcend 

 through the aperture ; befides the preffure of the 

 air above the aperture is ftronger than its refinance 

 below. 



From what has been now dated, it is eafy to per- 

 ceive how fluids differ from folids in the phenomena 

 of gravitation. If the verTel AB (plate V. fig. I.) 

 being full of water, and the tube E F being removed, 

 it was required to raife up the pifton G ; all that is 



neceffary 



