Conglomerates, and Marls of Devonshire. 
49 
they were derived. Moreover, as both the conglomerates and the 
sandstones of Torbay belong to quite the base of the formation as 
developed in Devonshire — the former being, so to speak, scarcely a 
step above the latter,— we are under the necessity of believing 
either that the process of formation was so extremely slow that a 
bed of sand was converted into indurated sandstone before the 
conglomerate, almost immediately over it, was formed ; or that the 
sandstone was formerly overlain by a thick series of beds of which 
every trace was torn down before the deposition of the conglomerate. 
There does not appear to be any other alternative ; but, whichever 
view we adopt it cannot but enhance our notion of the time-value 
of the formation. 
Nor are these nodules the only evidence of intercalated periods 
of unbuilding. Some of the cliff sections are found to consist of 
two series, one of which is composed of course materials, the other 
of fine. That, however, which resolves them into two distinct sets 
of Triassic deposits, is the fact that the upper surface of the lower 
group has been cut into a series of steps, which, to use the ter- 
minology of architects, are irregular in their rise and in their tread. 
They are formed of the truncated ends of a series of partially and 
unequally denuded beds, and into these inequalities the lower beds 
of the upper series are dovetailed. A fine example of this kind 
exists in a low headland immediately south of Paignton harbour, 
and necessitates the belief that during a pause in sedimentation, 
the upper beds of the lower series were partially destroyed, after 
which the overlying beds were deposited under somewhat different 
conditions, as is evidenced by the dissimilar characters of the two 
groups. It is worthy of remark that the dip of the upper beds is 
precisely the same, in amount and in direction, as that of the 
lower ; hence, had the denuded strata been completely, instead of 
partially, destroyed there would have been the most exact confor- 
mability between the two series, and not a trace of evidence that 
there had been any denudation at all ; a fact which suggests the 
possibility that the destruction is only partially represented by the 
