VOL. XVI. (3) THE LOWER SEVERN 245 
principal stream has been settled. The channel which 
‘becomes the river or principal stream in the system will 
be the one best adapted to meet the requirements both of 
this river or principal stream and, also, of the lateral or 
tributary streams flowing into it. 
Let me now call attention to a large map on the 
half-inch scale, extending from the Bristol Channel to 
the Dee and Mersey, and from the Welsh Hills to a line 
from Nottingham to Windsor.’ This is coloured so as 
to mark out heights in different shades of green and 
brown, and it is clear that, if the area were submerged, 
at every level we should have a number of islands forming 
an archipelago. If in imagination we go back to the time 
when the area was first uplifted above the sea, we may be 
sure that the surface was not an even one, nor can we 
believe that it came up suddenly, “per saltum,” so to 
speak, and we may suppose that the Welsh Mountains, as 
we now call them, were apparent while other parts of the 
area were still covered. But whatever the nature of the 
surface when it was first clear of the surrounding sea, we 
may be quite sure that it was not a level area. As the 
rain-fall, coming upon an irregular surface as yet un- 
covered by vegetation of any kind, required, more than it 
now does, channels to carry it off, the streams, even if 
their general tendency had been in one direction, would 
not all have gone straight; they would have been diverted 
some to the right and some to the left, and so become 
united and, in uniting, would form, for a time, a real net- 
work, such as we still see, for instance, on the upper 
Thames at Oxford and by Newbridge. 
Now, it is of the very essence of a net-work of streams 
that water at a given point has an opportunity of more 
than one route to the same destination: for a time these 
routes may afford equal facility, but, sooner or later, one 
will afford an easier passage than the other: the relative 
1 Bartholomew’s Sheets. 
T2 
