Since Diller wrote, similar deposits have been found around other volcanoes, and the clue to their mode of origin has been given by Fenner’s study of the tuff erupted in 1912 in the Valley of Ten Thousand Smokes, Alaska. It is now apparent that magma may be erupted and flow over the surface in all degrees of comminution, from the coarse, blocky state seen in many lavas to the form of ultra-pulverized dust seen in certain glowing avalanches (nuées ardentes). The expansion of gases liberated from incandescent particles of magma gives to these glowing avalanches an amazing mobility by reducing internal friction, so that they flow more freely than the most liquid lava. Long after the avalanches come to rest, the constituent fragments continue to give off gas, gradually adhere to one another, and become firmly welded, while &e larger lumps of viscous glass are flattened by the weight of material above. The final products are thus hardly to be distinguished from banded lava.
Diller regarded the Wineglass tuff as the latest lava on the walls of Crater Lake, for he saw that on Rugged Crest it partly overlies the Cleetwood dacite. How he reconciled this observation with his theory concerning the origin of the celebrated “backflow” in Cleetwood Cove is not clear. The present survey shows, however, that the tuff was erupted long before the summit of Mount Mazama disappeared and that in places it is capped by glacial moraines. Unfortunately, its age relative to the dacites of Grouse Hill, Redcloud, and Llao Rock remains in doubt. Presumably it is slightly younger. If so, then the Wineglass tuff was the product of the last eruption preceding the great pumice explosions which led to the destruction of Mount Mazama.
The type section at the Wineglass. One of the best and most easily accessible sections of the tuffaceous dacite flows and of the underlying pumice forms the brim of the Wineglass. To reach the type exposure from the Rim Road, it is necessary to descend over 50 feet of loose pumice and scoria, products of the culminating eruptions of Mazama, and then over an irregular layer of bouldery glacial debris up to 20 feet in thickness. Beneath these deposits, the flat-topped sheet of tuffaceous dacite forms a cliff between 20, and 25 feet high (plate 15, figure I).
Most of the dacite consists of compact and finely streaked glass in shades of pink, red, orange, brown, and gray. Here and there are short, thin lenses made up of angular lithic blocks of andesite up to 2 feet across. Black strings of obsidian accentuate the banding of the varicolored matrix. Some of. these measure a foot long though less than half an inch in thickness. Others are paper thin and have frayed ends. They do not seem to be any more flattened near the base of the cliff than at the top. Within 2 to 5 feet from the base, they disappear. If no more of the section were seen than that just described, one would not hesitate to classify the rock as dacitic lava. But in the lower-most part of the cliff, the massive, streaked glass passes within a foot or two, into pinkish, pulverulent containing many chips of andesite. Seeing only this part of the section, one would say, with no more hesitation, that the material was tuff, for it is crushed and flattened granular pumice. Yet the part which seems to be tuff grades imperceptibly into that which looks like lava. Suspicion that the main body is not really lava is increased when the streaked glass is traced laterally, for on the same horizon it is possible to follow all gradations between dense, varicolored obsidian and almost incoherent lump pumice. The conclusion is then inescapable that one is dealing with a pyroclastic flow compacted and welded to various degrees owing to differences in temperature, gas content, and thickness.
This conclusion gains weight when the Wineglass deposit is compared with welded tuffs recently examined in Japan, New Zealand, and California. Around the giant calderas of Kyushu, notably around &at of Aso, there are tremendous sheets of glassy ejecta almost identical with the Wineglass tuff. Mention may also be made of the welded tuffs on the “Rhyolite Plateau” of the North Island of New Zealand. Here, in Pliocene times, great volumes of fine tuff were poured from swarms of fissures and flooded an area of no less than 10,000 square miles. So closely do the tuff sheets resemble rhyolitic lavas that their origin has only lately been recognized. Marshall,1 who discussed them in detail under the name ignimbrites, has shown that, characteristically each sheet has a basal layer composed of quickly chilled and comminuted particles of glass. Above this, the tuff becomes increasingly welded, the glass shards are flattened, and the pumice lumps have collapsed. Near the top, because of the smaller load and quicker chilling, the welding gradually diminishes and the topmost layers may be as incoherent as those at the bottom. Gilbert2 has given a splendid account of comparable welded tuffs in the Owens Valley, California.
