325 
Perhaps no cave-deposits that we know are quite so old as the 
oldest river terrace that has yielded traces of man, still all the 
earlier ones may be included in the same bracket, and referred 
to the oldest stone or palaeolithic times. 
From the caves we cannot get much evidence of the lapse 
of time. The circumstances that affect the mode and rate of 
their formation, or the growth of travertine, or the slow in- 
filling of the cave with mud, are far too variable, and depen- 
dent upon too many local causes to found on them a date. I 
have myself found modern bottles under as great a depth of 
stalagmite as elsewhere covers mammoth bones. 
But from the terraces we may derive some help to form an 
estimate of the great lapse of time, though we may not as yet 
assign a term of years. What, then, are these terraces, and 
how formed ? It might appear at first an explanation not quite 
consistent with known facts to state that all the valleys with 
which we are concerned in this inquiry were scooped out by 
the gradual action of the streams, and that the terraces but 
mark old margins, where the streams once ran at higher 
levels. Why, it is said, if so, do we not find at every inter- 
mediate step of this continuous gradual waste the marginal 
deposits ? Elsewhere* I have more fully dwelt upon this ques- 
tion, pointing out that every river only just hands on along the 
flat the mud and gravel it receives from higher lands, but at the 
rapids and the waterfalls it still cuts back its channel, lengthen- 
ing the lower reaches of the river at the expense of the upper. 
The terrace generally marks the vertical height of the higher 
above the lower reach. It is clear that synchronous deposits 
may be found at the two levels, but it is also clear that, if we see 
a terrace far above the level of the present stream and far down 
the valley from the waterfall or rapid that tumbles from the 
level of that terrace higher up the stream, then we may mea- 
sure the antiquity of that terrace by the time that it would 
take the waterfall or rapid to cut back from where it was when 
the terrace was being formed to where we find it now. Some 
circumstances we must take account of which would increase 
the rate of waste, and so reduce the time. If an upheaval 
take place near the sea where formerly the long low flats 
were added to, not cut through by the river, then the flood, 
tumbling over the now-raised soft deposits of mud or sand or 
gravel into the sea, would soon cut back its channel. Also 
movements in the hills might cause some changes; or again, 
a not unimportant thing in chalk districts, the gradual removal 
of a clay covering which caused the water to collect in runlets 
first, then streams, would let the water soak into the porous 
* Royal Institution, March 24th, 1876'. 
