94 Afr. rennell’s Account of the 
experiment of my own on record, in which my boat was car- 
ried 56 miles in eight hours ; and that againft fo ftrong a wind, 
that the boat had evidently no progreffive motion through the 
water. 
When we confder, that the velocity of the ftream is three 
miles in one feafon, and five or more in the other, on the fame 
defcent of four inches per mile ; and, that the motion of the 
inundation is only half a mile per hour, on a much greater 
defcent ; no further proof is required how fmall the proportion 
of velocity is, that the defcent communicates. It is then, to 
the impetus originating at the fpring head, or at the place 
where adventitious waters are poured in, and fucceflively com- 
municated to every part of the ftream, that we are principally 
to attribute the velocity, which is greater or leffer, according to 
the quantity of water poured in. 
In common, there is found on one fide of the river an almoft 
perpendicular bank, more or lefs elevated above the ftream, 
according to the feafon, and with deep water near it : and on 
the oppofite fide a bank, Ihelving away fo gradually as to occa- 
lion fhallow water at fome diftance from the margin. This is 
more particularly the cafe in the moft winding parts of the 
river, becaufe the very operation of winding produces the 
fteep and Ihelving banks * : for the current is always ftrongeft 
on the external fide of the curve formed by the ferpentine 
courfe of the river ; and its continual action on the banks 
* Hence it is, that the fe&ion of a river, that winds through a loofe foil, 
approaches nearly to an obtufe angled-triangle, one of whofe fides is exceedingly 
Ihort and difproportioned to the other two » But when a river perfeveres in 
a ftraight courfe, the fedtion becomes nearly the half of an ediplis divided 
1 jngitudinally V J . 
either 
