7. Gray quartzose sandstone.

8. Red shaly sandstone and marl.

Before reaching the Whirlpool the mass becomes, practically, resolved into numbers three, four, and five, the limestone, as a general rule, growing thicker and harder, and the shale also, as we follow up the stream.

The reason why retrocession of the Falls is possible is found in the occurrence of the shale noted above as underlying the rock. It is a species of indurated clay, harder or softer according to the pressure to which it may have been subjected. When protected from the action of the elements it retains its hardness, but when exposed to them it gradually softens and crumbles away. After a time the superstratum of rock, which is full of cracks and seams, is undermined and precipitated into the chasm below. If the stratum of shale lies at or near the bottom of the channel below the Falls, it will be measurably protected from the action of the elements. In this case retrocession will necessarily be very gradual. If above the Falls the shale projects upward from the channel below, then in proportion to the elevation and thickness of its stratum will be the ease and rapidity of disintegration and retrocession. The shale furnishes, therefore, a good standard by which to determine the comparative rapidity with which the retrocession has been accomplished at different points.

From the base of the escarpment at Lewiston up the narrow bend in the channel above Devil's Hole, a distance of four and a quarter miles, the shale varies in thickness above the water, from one hundred and thirty feet at the commencement of the gorge, to one hundred and ten feet at the upper extremity of the bend. Here, although there is very little upward curve in the limestone, there is yet a decided curve upward in the Medina group, noticed above, composed mainly of a hard, red sandstone. It projects across the chasm, and also extends upward to near the neck of the Whirlpool, where it dips suddenly downward. The two strata of shale, becoming apparently united, follow its dip and also extend upward until they reach their maximum elevation near the middle of the Whirlpool. Thence the shale gradually dips again to the Railway Suspension Bridge, three-quarters of a mile above. For the remaining one and a half miles from this bridge to the present site of the Falls the dip is downward. We may then divide this reach of the Niagara River into three sections:

First. From Lewiston to the upper end of the Bend above Devil's Hole.

Second. Thence to the head of the rapid above the Railway Suspension Bridge.

Third. Thence to the present site of the Falls.

We are now prepared to consider these sections with reference to the retrocession of the fall of water. Through the first section the shale, as before noted, lying much above the water surface, and the superposed limestone being rather soft and thinner than at any point above, the retreat was probably quite uniform and comparatively rapid, about the same progress being made in each of the many centuries required to accomplish its whole length. Professor James Hall, in his able and interesting Report on the Geology of the Fourth District of the State of New York, suggests the probability of there having been three distinct Falls, one below the other, for some distance up-stream, when the retrocession first began. The average width of this section between the banks is one thousand feet. About one mile below its upper extremity is "Devil's Hole," a side-chasm cut out of the American bank of the river by a small stream called "Bloody Run," which, in heavy rains, forms a torrent. The "Hole" has been made by the detrition and washing out of the shale and the fall of the overlying rock. A short distance above, on the Canadian side, lies Foster's Glen, a singular and extensive lateral excavation left dry by the receding flood. The cliff at its upper end is bare and water-worn, showing that the arc or curve of the Falls must have been greater here than at any point below.

Near the upper end of this section there is a rocky cape, which juts out from the Canadian bank, and reaches nearly two-thirds of the distance across the chasm. At this point the great Fall met with a more obstinate and longer continued resistance than at any other, for the reason that the fine, firm sandstone belonging to the Medina group, as has been stated, here projects across the channel of the river, and, forming a part of its bed, rises upward several feet above the surface of the water. And here this hard, compact rock held the cataract for many centuries. The crooked channel which incessant friction and hammering finally cut through that rock is the narrowest in the river, being only two hundred and ninety-two feet wide, and the fierce rush of the water through the narrow, rock-ribbed gorge is almost appalling to the beholder. The average width between the banks of this section is about nine hundred feet.