THE main cone of Mount Mazama had perhaps reached its full height when the centers of principal activity moved from the summit region to the northern flank. When this took place, the glaciers were thin and small and none extended far beyond what is now the caldera rim. Possibly this northward shift of the vents was brought about by enlargement of the magma chamber in the same direction, either as a result of internal assimilation of the cone or, more likely, as a consequence of ring-fracture stoping. Whatever the cause, an arcuate line of vents opened along what later became the north wall of the caldera, approximately 5000 feet below the original summit of Mount Mazama.

It is surely no mere coincidence that the rim of Crater Lake cuts across this Northern Arc of Vents. On the contrary, the position of the rim seems to have been controlled thereby, for the center of Crater Lake does not lie immediately below the former summit of Mount Mazama, but approximately a mile to the north. The eccentricity of the caldera with respect to the main cone appears to have been determined by a predisposing plane of weakness, possibly a continuous ring fracture, the surface expression of which was the Northern Arc of Vents. This eccentric relation accounts in large part for the fact that the north wall of Crater Lake is so much lower than the south wall.

At least six of the Northern Arc of Vents are visible on the caldera walls. In a clockwise direction, these are as follows: the vents of the Watchman andesite, of the andesitic cone forming Hillman Peak, of the Llao Rock, Cleetwood, and Redcloud dacite flows, and lastly of the Sentinel Rock andesite. Possibly another vent lay close to the Wineglass. Possibly, also, such thick flows as those of the Palisades and Roundtop were erupted from vents adjacent to the main arc.

The evidence concerning the exact order of eruption of these vents is incomplete, though the andesitic vents were active before those which erupted dacite. Some of the vents, notably those of Redcloud and Llao Rock, erupted pumice before extruding lava; at other vents, the only activity was the effusion of viscous lava; at Hillman Peak, explosions and effusions alternated many times.

The age of the Northern Arc of Vents relative to that of the other parasitic eruptions on the sides of Mount Mazama is not easy to determine. No doubt the long dacite flows on the south side of the cone, described in an earlier section, are much older, for they were erupted before the period of maximum glaciation. Perhaps the eastern cluster of dacite domes was active about the same time as the Llao, Cleetwood, and Redcloud vents or a little earlier. Most of the basaltic cinder cones must also be approximately coeval, though some may have developed later.

We may now continue to a discussion of the principal vents of the Northern Arc, beginning with those which erupted andesite.