88 
F. W. Oliver. 
away, whilst the sky-line is formed by the higher slopes of the bank. 
Beyond the sky-line is the crest, from which the bank drops abruptly 
to the sea. 
The mechanism by which these ravines are carved and shingle 
dislodged and shot out to form fans projecting from the terrace is 
percolation by the sea. At high tide and especially with heavy 
on-shore gales, sea-water runs through the bank like a sieve, gushing 
forth at many places at the foot of the landward slope. At the 
points where the water flows out, shingle is displaced, and this 
causes the shingle above to be dislodged and slide down its “ slope 
of repose.” In this way the ravines are usually cut deeper and 
deeper into the bank, the dislodged material being vomited forth 
by the gully on the floor as detrital fans. 
The efficiency of this mechanism will depend on the height of 
the breaking waves on the sea front above the water-level of the 
Fleet. The latter—which shows little oscillation—is about 4 feet 
below ordinary high-water-mark outside. As a consequence, 
percolation takes place in both directions, i.e., from the sea into 
the Fleet when the waves are breaking at a high level, and at low 
water a gentle percolation in the opposite direction—indicated by a 
reversed flow in the gullies or channels on the floor of the ravines. 
In the absence of systematic records it is impossible to make 
any quantitative estimate of the amount of shingle that is shifted 
by the operation of this process ; but that the amount is large and 
the result important in relation to the vegetation is evident from 
an inspection of the ground. 
It will be understood that the result of the operation just 
described is the transport of shingle from the higher parts of the 
bank on to the terrace below, and especially the projection of this 
detritus in the form of fan-like platforms beyond the general edge 
(the dotted line in Text-fig. 7, a section of the bank, illustrates the 
nature of this transfer). By this means the terrace advances 
landwards at the expense of material derived from the higher slopes. 
Though it is not possible to say for how long a time a given ravine 
remains in the active phase, it is manifest that this activity is 
intermittent and that ravines go out of action and remain quiescent, 
probably for years. 
Terrace formation by ravine-agency is of course only one part, 
and a minor part, of the big phenomenon of landward creep which 
is always in progress. That the latter is possible depends primarily 
on the constructive action of the waves, which pile up the shingle 
