To comprehend it we must first consider two geological facts. The first is that no dawdling modern Merced cut this chasm, but a torrent considerably bigger; and that this roaring river swept at tremendous speed down a sharply tilted bed, which it gouged deeper and deeper by friction of the enormous masses of sand and granite fragments which it carried down from the High Sierra. The second geological fact is that the Merced and Tenaya torrents sand-papered the deepening beds of these canyons day and night for several million years; which, when we remember the mile-deep canyons which the Colorado River and its confluents cut through a thousand or more miles of Utah and Arizona, is not beyond human credence, if not conception.
But, objects the sceptical, the Merced couldn't keep always tilted; in time it would cut down to a level and slow up; then the sand and gravel it was carrying would settle, and the stream stop its digging. Again, if the stream-cut valley theory is correct, why isn't every Sierra canyon a Yosemite?
Let us look for the answer in the Sierra's history.
The present Sierra Nevada is not the first mountain chain upon its site. The granite which underlay the folds of the first Sierra are still disclosed in the walls of the Yosemite Valley. The granites which underlay the second and modern Sierra are seen in the towering heights of the crest.
Once these mountains overran a large part of our present far west. They formed a level and very broad and high plateau; or, more accurately, they tended to form such a plateau, but never quite succeeded, because its central section kept caving and sinking in some of its parts as fast as it lifted in others. Finally, in the course, perhaps, of some millions of years, the entire central section settled several thousand feet lower than its eastern and western edges; these edges it left standing steep and high. This sunken part is the Great Basin of to-day. The remaining eastern edge is the Wasatch Mountains; the remaining western edge is the Sierra. That is why the Sierra's eastern front rises so precipitously from the deserts of the Great Basin, while its western side slopes gradually toward the Pacific.
But other crust changes accompanied the sinking of the Great Basin. The principal one was the rise, in a series of upward movements, of the remaining crest of the Sierra. These movements may have corresponded with the sinkings of the Great Basin; both were due to tremendous internal readjustments. And of course, whenever the Sierra crest lifted, it tilted more sharply the whole granite block of which it was the eastern edge. These successive tiltings are what kept the Merced and Tenaya channels always so steeply inclined that, for millions of years, the streams remained torrents swift enough to keep on sandpapering their beds.
The first of these tiltings occurred in that far age which geologists call the Cretaceous. It was inconsiderable, but enough to hasten the speed of the streams and establish general outlines for all time. About the middle of the Tertiary Period volcanic eruptions changed all things. Nearly all the valleys except the Yosemite became filled with lava. Even the crest of the range was buried a thousand feet in one place. This was followed by a rise of the Sierra Crest a couple of thousand feet, and of course a much sharper tilting of the western slopes. The Merced and Tenaya Rivers must have rushed very fast indeed during the many thousand years that followed.
The most conservative estimate of the duration of the Tertiary Period is four or five million years, and until its close volcanic eruptions continued to fill valleys with lava, and the Great Basin kept settling, and the crest of the Sierra went on rising; and with each lifting of the crest, the tilt of the rivers sharpened and the speed of the torrents hastened. The canyon deepened during this time from seven hundred to a thousand feet. The Yosemite was then a mountain valley whose sloping sides were crossed by cascades.
Then, about the beginning of the Quaternary Period, came the biggest convulsion of all. The crest of the Sierra was hoisted, according to Matthes's calculations, as much as eight thousand feet higher in this one series of movements, and the whole Sierra block was again tilted, this time, of course, enormously.
For thousands of centuries following, the torrents from Lyell's and McClure's melting snows must have descended at a speed which tore boulders from their anchorages, ground rocks into sand, and savagely scraped and scooped the river beds. Armed with sharp hard-cutting tools ripped from the granite cirques of Sierra's crest, these mad rivers must have scratched and hewn deep and fast. And because certain valleys, including the Yosemite, were never filled with lava like the rest, these grew ever deeper with the centuries.