100 
BULLETIN OF THE NATURAL HISTORY SOCIETY. 
few miles down it makes a great bend to the eastward, and here 
its course is quieter, in part at least a gravel-bottomed quick- 
water. Then it turns again southward and cuts into a rocky 
bed in a country of increasing height. For a few miles below 
I have not seen it, but it is said here to be very rough, and to 
include some falls, one of which is 18 feet in height. Below, 
where the portage road is near, it is a quieter, smooth, gravel- 
bottomed stream, canoeable and pleasing. Farther down it be- 
comes again rough, and for the lower six miles, as I know in 
part by observation and in part from report, it is everywhere 
a rough, rocky strong-running brown stream in a wide valley 
winding among great ridges. 
Such is the arrangement of those waters at present. There 
is no doubt, I believe, that the uppermost waters, and even those 
down to the easterly bend, originally and in part pre-glacially, 
were tributary to the Dungarvon, and this is no doubt true of 
the easterly bend itself, as shown by its direction and the 
presence of the re-entrant little stream at the turn. Moreover, 
the directions of the branches below strongly suggest that parts' 
at least, e, g. the Tahoe-Hurd valley, were also tributary to the 
Dungarvon. The lowest part, however, appears to have had the 
same evolution as the Sisters, Clearwater and Burnt Hill, viz., 
a stream let down between great intrusive ridges. But even 
here is a problem, for the parallelism and alignment of the 
streams of the region would suggest for Rocky Brook a former 
outlet through the present course of Fall Brook. 
One of the most interesting physiographic features of this 
stream, however, is the nature of the watershed between it and 
Dungarvon. For this watershed, which continues south and 
determines the direction of the Miramichi between Fall Brook 
and Boiestown, is extremely narrow and parallel with Rocky 
Brook, while the Dungarvon branches start off close bv and 
nearly at right angles. The explanation, however, is not difficult 
on a supposition which I have advanced in a preceding note 
(No. hi) and which I repeat a page or two later. 
