Recent claims as to the distribution of geologically old welded tuffs (Anderson, 1970) are apparently based largely on the use of “snowflake texture” as a criterion diagnostic of such deposits. Late Precambrian vesicular felsite lavas in Minnesota commonly show snowflake texture, and it should not be used as a criterion for recognition of devitrified welded tuffs.
Anderson (1970) has recently published a note that implies, if not claims, that very large volumes of altered Cambrian and late Precambrian felsic volcanic rocks in Oklahoma and Missouri are ash-flow tuffs. Before this claim becomes generally accepted, it is desirable to examine the available evidence, and I would like to suggest that the criterion for identification of ash-flow tuffs apparently relied upon by Anderson for many of the rocks is invalid (Anderson, 1969).
Although I am not acquainted with the particular rock groups described by Anderson, for several years I have been carrying out detailed mapping of the late Precambrian (Keweenawan) North Shore Volcanic Group in Minnesota (Goldich and others, 1961), with the support of the Minnesota Geological Survey and the National Science Foundation GP-5056 (Green, 1968; in prep.). These rocks are gently dipping and unmetamorphosed, except for local hydrothermal or pneumatolytic effects, or both, and contain many flows of felsic (quartz-latite) composition. They are well exposed along the shore of Lake Superior. All are devitrified, and a few are spherulitic. The rocks display grain sizes which are dense-aphanitic to very fine-grained, and most are porphyritic. The texture referred to as “snowflake” by Snyder (1962) and Anderson (1970) is common, with poikilitic patches of quartz, up to 2 or 3 mm across, enclosing small, blocky or tabular, alkali-feldspar euhedra in a pseudo-ophitic relationship. In contrast o t the suggestion of Anderson (1970, p. 287) that the quartz crystallized last, it would appear more compatible with crystallization from either a magma or a glass if the feldspar and quartz were contemporaneous during most of the crystallization, with fewer nuclei for the quartz. The feldspar may or may not, of course, have started to crystallize first.
I find nowhere in Ross and Smith (1961) or Anderson's other references, however, mention of snowflake texture as a criterion for ash-flow tuff origin, whether welded or not. Although a few flows with flattened lapilli and shards typical of welded tuffs have been reported from the Keweenawan (Foster, 1962; W. S. White, 1969, written commun.), the great bulk of felsic flows that I have examined shows no such fragmental textures either in outcrop or thin section (I have examined Paleozoic devitrified welded tuffs in England, Wales, and Maine, and vitric ones from New Mexico and in Japan). On the other hand, many of the Keweenawan flows are very massive and uniform, but develop a frothy, vesicular zone in the top few feet (Fig. 1A), clearly implying the presence of fluid lava and not a pyroclastic origin. These flows commonly show snowflake texture in their interiors (Fig. 1B). Lofgren's (1968) experimental work and conclusions, quoted by Anderson (1969, p. 2077–2078) show no essential correlation of snowflake texture with ash-flow origin. It appears that such a texture has little bearing on the eruptive characteristics of a flow, but perhaps is simply a common result of crystallization of any felsic glass.