Ganges and Burrampooter Rivers . ior 
within the compafs of his obfervation *. Some of thefe iflands, 
four or five miles in extent, are formed at the angular turnings 
of the river, and were originally large fand banks thrown up 
round the points (in the manner before defcribed) but after- 
wards infulated by breaches of the river. Others are formed 
in the ftraight parts of the river, and in the middle of the 
ftream ; and owe their origin to fome obftruftion lurking at the 
bottom. Whether this be the fragments of the river bank ; a 
large tree fwept down from it ; or a funken boat ; it is fufficient 
for a foundation : and a heap of fand is quickly colledled below* 
it. This accumulates amazingly faft : in the courfe of a few 
years it peeps above water, and having now ufurped a confi- 
derable portion of the channel, the river borrows on each fide 
to fupply the deficiency in its bed ; and in fuch parts of the 
river we always find fteep banks on both fides f. Each periodi- 
cal flood brings an addition of matter to this growing ifland ; 
increafing it in height as well as extenfion, until its top is per- 
feftly on a level with the banks that include it : and at that 
period of its growth it has mould enough on it for the pur- 
pofes of cultivation, which is owing to the mud left on it when' 
the waters fubfide, and is indeed a part of the ceconomy which 
nature obferves in fertilizing the lands in general. 
Whilfl: the river is forming new iflands in one part, it is 
fweeping away old ones in other parts. In the progrefs of this 
deftruftive operation, we have opportunities of obferving, by 
means of the feftions of the falling bank, the regular diftri- 
* Accordingly, the laws refpefting alluvion are afcertained with great pre- 
fcifion. 
f This evidently points out the means for preventing encroachments on a 
river bank in the ftraight parts of its cpurfe, viz , to remove the ihallows in the 
middle of its channel* 
bution 
