In the second place, it should be noted that in the basic inclusions hornblende is rare and largely replaced by granular pyroxene and ore. Resorbed crystals of oxyhornblende are also typical of the basic inclusions in the dacites of the Lassen region. By far the commonest type of inclusion in the Mazama lavas consists of a crisscross felt of hypersthene and labradorite crystals with accessory augite, a little glass, and either tridymite or cristobalite. Hornblendic inclusions are unknown in hornblende-free lavas. Similarly, the hornblende-rich dacites which form the main dome of Lassen Peak are crowded with basic inclusions characterized by abundant resorbed hornblende, whereas the 1915 lava of Lassen Peak, which presumably rose from very shallow depth, is almost devoid of hornblende and such as there is shows extensive resorption. Further, the basic inclusions in the 1915 lava, which came from still shallower depths and presumably represent fragments of the semiconsolidated roof of the chamber and walls of the conduit, are quite devoid of hornblende. These occurrences lead us to agree with MacGregor’s view that close to the surface hornblende breaks down unless the magma is drastically and suddenly chilled. The occurrence of plentiful tridymite or cristobalite, or both, in the basic inclusions is also in accord with the idea of a shallow origin, for siliceous vapors would normally be concentrated at the roof of the magma chamber and in the overlying conduit.
A further analogy may be drawn with the magmas of Montserrat, for at Crater Lake also the explosive magma of the final eruptions was characterized by paucity of any but fresh, green hornblende.
When the climactic eruptions began, the magma in the upper part of the chamber was of dacitic composition and the ratio of crystals to melt was approximately 3 to 7. What little hornblende floated in the magma was still in a stable condition. Below this acid fraction lay more basic magma, of the composition of basaltic andesite, which had crystallized to an even greater extent. At these depths, hornblende was by far the dominant ferromagnesian constituent. Between these two types of magma there was little or no material of intermediate composition. Intermediate magma, hypersthene andesite, like that which had built the main cone of Mazama, was either absent altogether or present in such small amount that it has not been recognized among the ejecta. What agencies brought about such a clear differentiation is not understood. Yet they must have operated earlier, for the basic inclusions in the dacite flows erupted from the Northern Arc of Vents have approximately the same composition as the hornblende-rich scoria. This implies that the semicrystallized roof of the dacite magma chambers had virtually the same composition as the hornblende-rich scoria. Olivine-bearing basaltic andesite, erupted from parasitic cinder cones, basic inclusions in the andesite and dacite flows, and the hornblenderich scoria of the culminating eruptions are therefore merely heteromorphs.
After the summit of Mount Mazama disappeared, new eruptions took place on the floor of the caldera. Unfortunately, only the products of the youngest of these are visible for inspection. Nothing can therefore be said of the trend of post-caldera differentiation. The visible products on Wizard Island consist of hypersthene andesite. Thus, as its final act, the volcano reverted to eruption of an intermediate type of magma essentially identical with that which had characterized its growth from the beginning until the time it reached maturity.
***previous*** — ***next***
