214 Natural Terraces on the Eildon Hills. 
would be ground against each other by every tide, into 
smooth pebbles and shingle. Nor is it easy to conceive how 
terraces of 100 or 120 yards broad, as on the Eildons, formed 
of angular stones detached and precipitated from above, could 
have been made to assume by the action of water, even the 
irregular horizontality they do actually possess, when falling 
upon a slope having an angle of thirty or forty degrees; and 
this, without the stones showing any marks of attrition. On 
a gently inclined beach, where the tidal wave is ever and 
anon rolling such fragments over a considerable area, they 
would soon be converted into rounded pebbles; but on a 
steep rocky shore they would fall at once into deep water and 
assume the shape of a conical talus or “scree,” where the 
tide would have comparatively little effect upon them. Their 
rough angular surfaces would lock into each other, and pre- 
vent them from being scattered over so broad a space as we 
see them on the Middle Eildon. It must also be borne in 
mind, in reference to the terraces on the eastern hill, which 
appear to have slipped down from their original situation, that 
the probability of their having done so is much weakened by 
their being composed of angular stones. 
It struck me as singular, that all the terraces I examined, 
should be found on the sides of the respective hills most ex- 
posed to the strong currents that may be assumed to have 
been then in action; those on the north side of the Kildons, 
facing the great valley of the Tweed; those on William- 
law, overhanging the more circuitous one of Gala water. Of 
course I conclude they do exist on the retired sides of some 
of the hills. One should have supposed, @ priori, that the 
currents would have swept away the fragments of rock as 
they fell from above, and would have prevented them from 
accumulating into projecting shelves. Indeed, several of the 
best developed are widest precisely at the point where they 
project into the valley, and would come in contact with the 
current. I was also surprised to find no trace of terraces in 
other situations, apparently more favourable to their produc- 
tion. Immediately to the west of Williamlaw, and seen to 
advantage from its summit, is a wide and deep circular am- 
phitheatre, formed by the smooth grassy sides of several 
neighbouring hills which environ it with very uniform a 
except on the side that connects it with the valley of the Gala 
water. Ifthe sea ever occupied the latter, it must also have 
filled this hollow, and converted it into a spacious, though 
sheltered and tranquil bay, round whose encircling sides, well- 
developed terraces might be expected to be found, Their 
total absence, therefore, from so favourable a locality, leaves 
room to inquire whether those which occur in more equi- 
ESS ae cis ares 
