the conditions were those of to-day intensified.  In summer, then, the amount of water seeking outlet by these drainage channels to the sea was enormously multiplied, and the corrasive power was correspondingly augmented.  When the ice caps finally began to permanently diminish, the summer floods were doubtless terrific.  The waters of the Colorado now rise in the Grand Canyon, on the melting of the snows in the distant mountains, from forty to one hundred feet; the rise must then have amounted to from one hundred to four hundred or more.  The Kanab heads in two very high regions—­the Pink Cliffs and the Kaibab.  Though probably not high enough to be heavily glaciated they were high enough to receive an increased snowfall and to hold it, or a portion of it, over from one year to another.  Thus the canyons having their origin on these high regions would be given perennial streams, with torrential floods each summer, compared with which anything that now comes down the Kanab would be a mere rivulet.  The summit of the Kaibab is covered with peculiar pocket-like basins having no apparent outlets.  These were possibly glacial sinks, conducting away some of the surplus water from the melting snow and ice by subterranean channels.  It seems probable, therefore, that glacial flood-waters were an important factor in the formation of the canyons of the Colorado.  If this supposition is correct it would account, at least in a measure, for that distinct impression of arrested activity one receives from the present conditions obtaining there.*

* Some canyon floors, where there is no permanent large stream, appear to have altogether ceased descending.  Dutton says of those which drain the Terrace Plateaus:  “Many of them are actually filling up, the floods being unable to carry away all the sand and clay which the infrequent rains wash into them.”—­Tertiary History, p. 50.  See also pp. 196 and 228 Ib.

The drainage at the edges of most canyons is back and away from the gorge itself.  The reason is that the rains cannot flow evenly over a canyon brink, owing to irregularities of surface, and once an irregular drainage is established, the water seeks the easiest road.  A side canyon is formed, draining a certain area.  Another is formed elsewhere, and another, and so on till all drainage is through these tributaries and away from the brink, by more or less circuitous channels to the main stream.  This backward drainage leaves the immediate brink, or “rim,” till the last, in its work of erosion and corrasion, and the rim consequently is left higher than the region away from it.  This effect of a backward drainage is very plain on both sides of the Grand Canyon, though it is somewhat assisted, on the north at least, by the backward dip of the strata.  It may be modified by other conditions, so that it would not always be the case.