1875.] Niagara. 149 
the old blocked-up channel, across which it had evidently 
been at one time continuous. 
This terrace must have been formed when the river was 
much wider than now, and flowing so slowly, that it had 
only power to cut through the unconsolidated sandy clay, 
without stones. The lower terrace marks a later stage, 
when, by the cutting back of the gorge, the river ran with a 
swifter current, at a lower level, and in a narrower channel, 
andcutthroughthe lower till to thesolid rock below. Itmarks 
the last stage in the present river’s course, before it occupied 
the gorge cut through the Silurian limestones. The ridge 
capping the bank of the old river-bed must have been heaped 
up during floods. It is well exposed, though in danger of 
total obliteration, as it is being carted away for gravel. The 
stones in it are all rounded, like true river gravel. Mixed 
through it, often filling the interstices between the stones, 
are multitudes of fresh-water shells belonging to the genera 
Melama, Limnea, Umio, and Cyclas. I was able to confirm 
the observation of Lyell, that the assemblage of species is the 
same as that now found in the river above the falls. 
Amongst the stones of the present beach above Goat Island, 
I found shells of the same species as I did in the old ridge 
above the whirlpool, and in not much better state of preser- 
vation; indeed, excepting that I had labelled the boxes con- 
taining the different sets, I could scarcely now tell which 
were the older of the two. This assemblage of dead shells 
in the gravel of the beach differs, both in the older and 
newer deposits, from that found living in the present river, 
in that many delicate shelled species are scarce, or not 
found at all, owing, no doubt, to only the more robust shells 
being preserved. Thus, thin-shelled species of Physa 
abound in the river, but I did not find any of their shells, 
either amongst the stones of the present or the old beach. 
A little above the whirlpool, the gorge widens out abruptly, 
as shown in plan, and the terraces and ridge are cut off by 
it, so that they cap the gorge, and are exhibited in section 
almost as clearly as I have depicted them in fig. 4. The 
river ridge, composed entirely of loose gravel and sand, is 
seen running to the edge of the cliff. The widening of the 
gorge extends for some distance beyond it, cutting off the 
upper terrace in the same manner. At this spot the upper 
layers of limestone project beyond the lower beds, just as 
they do now at the table rock at the falls, forming an over- 
hanging precipice, so that the widening of the gorge cannot 
have been caused by weathering. I cannot conceive how 
the present river could excavate the gorge beyond the loose 
