Ash-F low Tuffs: Their Origin Geologic Relations and Identification {fit-9‘ "GEOLOGICAL SURVEY BROFESSIONAL PAPER 366 GEOLOGICAL SURVEY PROFESSIONAL PAPER 366 FRONTISPIECE Thin section showing oxidation and later partial reduction of the iron in a glassy welded tufi. lddings (1899, pl. 50, fig. B) shows the same sec- tion in black and white. His belief that this was all originally red and that bleaching took place after deposition is confirmed by the presence of very fine grained microlites of magnetite in the light-colored area. From Yellowstone National Park. The dull brown cores of the large shards near the middle of this photomicrograph of glassy welded tufi represent the original character of the glass, and the red areas are a result of oxidation. The horizontal alinement of shards indicates marked compaction. USNM specimen 38771 from Rio Blanco, 10 miles north-northwest of Guadalajara, Mexico. OXIDATION AND REDUCTION IN GLASSY WELDED TUFFS Ash-Flow Tufts: Their Origin Geologic Relations and Identification By CLARENCE S. ROSS and ROBERT L. SMITH GEOLOGICAL SURVEY PROFESSIONAL PAPER 366 A stna’y oft/2e emplacement, oyflowage, ofnot gas—emitting ‘volcanic aslz; its inclnration oy weta’ing anaI crystallization, anaI criteria for recognizing t/ze resulting rocé UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1961 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY Thomas B. Nolan, Director For sale by the Superintendent of Documents, US. Government Printing Office Washington 25, DC. Q/Ifr/s v. it!” CONTENTS EARTH SCIENCES LIBRARY Page Abstract 1 Recognition of ash—flow tuffs—Continued . Microscopic characteristics _________________________________________ t . In reduction 1 Pyroclastic character,” __ Acknowledgments 2 Pumice fragments .................................................. N . Welding, distortion, and stretching __________________ omenclature ~ 2 Phenocrysts and foreign materials ____________________ Glossary of terms 3 Devitrification Development of concepts 8 Physical chemistry _ . . Heat Histgrilc, eruptlons of ash-flow tufl’s __________________________________ 15 Viscosity . eee 15 Volatile components Valley of Ten Thousand Smokes ________________________________ 16 Welding Cotopax1 16 Devitrification Source vents for ash-flow tufl's __________________________________________ 16 . . Physmal propertles Recognition of ash-flow tufl‘s _______________________________________________ 18 Indices of refraction Field characteristics 18 Specific gravity Pyroclastic character ____________________________________________ 18 Porosity Degree of sorting ______________________ '_ ___________________________ 19 Thickness 20 Ash-flow materials of intermediate composition ____________ Layering 20 Andesitic ash flows of Costa Rica ,,,,,,,,,,,,,,,,,,,,,,,,,,,, t Areal ”l” 22 Distribution in time Gentle dips 23 M 0 . Welding and deformation of pumice __________________ 24 P els 201? Fused tufl‘s contrasted with welded tuffs _____ 26 Pa e020? . Devitrification and vapor-phase minerals ,,,,,,,, 26 recam rlan Jomting 28 Selected references Eroswn forms 29 Fossil fumaroles 30 Index III '60. Same area as illustrated in figure 59, showing a coarse-grained radlal aggregate 01 Leluspur 'm'cxlo- tobalite laths developed independently of the original shard structure ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, ; ,,,,,,,,,, 61. Photomicrograph of welded tuff from southeastern Idaho, showing devitrified shards and partial de- struction of pumice structure . . 62. Photomicrograph of welded tuff from the Jemez Springs quadrangle, Valles Mountains, N.Mex., showmg complete destruction of original structure by development of coarse-grained devitrification products. 63. Photomicrograph of Welded tuff from the Valles Mountains, N.Mex., showing a very large spherullte in which radial aggregates of feldspar and cristobalite are well represented ______________________________________ Page 31 32 33 33 3‘5 36 38 39 40 40 41 44 45 45 46 46 47 48 48 49 49 49 50 79 68 68 68 69 IV VI 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 7 5. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 92. 93. 94. 95. 96. 97. 98. CONTENTS ILLUSTRATIONS CONTENTS Photomicrograph of welded tuft“ from the Valles Mountains, N.Mex., showing the development of groups of spherulites Photomicrograph of welded tuff from the Lake Toba region, Sumatra, showing spherulitic areas ............ Photomicrograph of welded tufl" from the Valles Mountains, N.Mex., showing an open cavity in which platelike crystals of tridymite have developed Polished section through a so-called thunder egg, a type of hollow spherulite which commonly develops in glassy volcanic rocks Photomicrograph of the outer rim material of figure 67, showing section perpendicular to the plane of deposition Photomicrograph of the outer rim material of figure 67, showing section parallel to the plane of deposi- tion Photomicrograph of welded tufi‘ from southeastern Idaho, showing parallel structure produced by mold- ing of shards against each other... . Photomicrograph of tufl" from the Valles Mountains, N.Mex., showing welding and devitrification, but without distortion of the shard structure Photomicrograph of specimen shown in figure 71, under higher magnification, illustrating axiolitic struc- ture Photomicrograph of devitrified welded tuff from the Valles Mountains, N.Mex., showing molding of the shards around a feldspar phenocryst Photomicrograph of devitrified welded tuff from the Western United States showing uncommonly per- fect axiolitic structure Photomicrograph of welded tuif from Australia, showing partial loss of structure by devitrification ........ Photomicrograph of devitrified welded tufl" from Hungary, showing obscured shard structures ................ Photomicrograph of devitrified welded tuif from Korea, showing dim retention of shard structure..- Photomicrograph of welded tuff from Sweden, showing local preservation of shard structure ................... Photomicrograph of a welded tufl“ of intermediate composition from Russian Armenia, showing slight welding but distortion of pumice Photomicrograph of a welded tufi' of intermediate composition from Russian Armenia, showing moder- ate welding and warping of shards Photomicrograph of andesitic welded tufi‘ from Japan, showing lineation produced by thorough com- paction Photomicrograph of welded tufl" from Calaveras County, Calif., illustrating the type biotite-augite latite described by Ransome Photomicrograph of welded tufl‘ from Argentina described by Bonorino as andesite, showing section parallel to the plane of deposition Photomicrograph of same specimen shown in figure 83, showing section perpendicular to the plane of deposition Hand specimen of andesitic welded tuff from Costa Rica . Photomicrograph of section from specimen shown in figure 85, showing destruction of pumice structure by welding Photographic reproduction of a colored drawing by Iddings Photomicrograph of one of Iddings’ thin sections of rocks from Yellowstone National Park ........................... Copy of a photomicrograph of Iddings’ “welded pumice with axiolites” Photomicrograph of specimen from the Valles Mountains, N.Mex., studied by Iddings .............................. Photomicrograph of welded tufl? from Yellowstone National Park studied by Iddings, showing typical large inclusion of andesite Photomicrograph of welded tufl" from Yellowstone National Park, studied by Iddings, showing incipient devitrification and thorough welding Photomicrograph of glassy welded tufl" from Wadsworth, Nev., studied by Zirkel, showing section diag- onal to the plane of deposition Photomicrograph of welded tuff from West Humbolt, Nev., studied by Zirkel, showing extreme flatten- ing and stretching of shard structures Photographic reproduction of a drawing presented by Zirkelyshowing axiolitic structure .......................... Photomicrograph of ash-flow tufl" from the Valley of Ten Thousand Smokes, showing little induration--- Photomicrograph of welded tufl" from New Zealand, representing materials which were the basis for Marshall’s outstanding contribution to an understanding of ash-flow tuffs ............................................ Photomicrograph of welded tuff from Morocco, described by Bouladon and Jouravsky .......... L .................... Page 69 69 69 76 70 70 70 71 71 71 71 72 72 72 72 73 73 73 73 74 74 n 75 75 75 76 76 76 76 77 77 77 77 ASH-FLOW TUFF S: THEIR ORIGIN, GEOLOGIC RELATIONS, AND IDENTIFICATION By CLARENCE S. Ross and ROBERT L. SMITH ABSTRACT Pyroclastic materials, which are interpreted as having been deposited by flowage as a suspension of ash in volcanic gas, are becoming widely recognized as major geologic epi- sodes. These may be unconsolidated, indurated by partial Welding, or welded into a compact rock. Many students are working on these materials and the interest in them is so widespread that need for a coordinated treatise on them has developed. This report deals with the history of the concept of their origin; gives detailed descriptions of their character and mode of occurrence; gives criteria for their recognition; and considers their distribution and consolidation. The terminology to describe ash flows and the reasons for avoiding the invention of new specific names, or the re— defining of old ones, are stated. The terms used are given and defined, so far as possible, by quotations from authorita- tive sources. The use of descriptive phrases, where gener- ally accepted specific names are unavailable, is the usage adopted. The evolution of geologic concepts which led eventually to an understanding of the distribution and, in some de- posits, the welding of tufl’s is traced. Prior to about 1900, geologists argued as to whether these materials represented peculiar volcanic lava flows, or were pyroclastic in origin. The development of a better understanding is traced through the discussion of geologists who reported on the great eruptions of Pelée, Soufriere, and the Valley of Ten Thou- sand Smokes, and by other geologists who studied analogous ancient eruptions. Recognition of ash-flow materials involves both field ob- servations and laboratory studies. These are presented in some detail, and lead to the conclusion that the ash-flow materials may be recognized by either of these approaches, but for some occurrences both are required. The field rela- tions that are presented for evaluation are: extent, thickness, the relation of overload to welding, jointing, erosional forms, inclusions of alien materials or pumice fragments, devitrifi- cation, and the relations to source centers. Experience with distinctive features commonly allows identification of ash- flow deposits even at a distance of many miles. Laboratory studies that require description and evalua- tion include chemical analyses, the relation of composition to other features, porosity and its relation to overload and welding, and the temperature of welding. Microscopic studies reveal a large number of variables that demand con- sideration and discussion of their relative value in identifi- cation. These include a variety of primary glassy shard structures, the greatly varied modification of the shards by compression, welding, and devitrification. These factors may leave clearly significant evidence of origin, or this may be obscured in varying degree. Experience will allow identifi- cation even in the presence of severe modifications, but the development of uncommonly coarse devitrification products may destroy significant structures. The significant charac- teristics are discussed, and anomalous features which may lead to misinterpretation are discussed and evaluated by the use of photographs. The minerals in these tufi’s comprise the primary minerals ‘ (phenocrysts), those resulting from devitrification after em- placement, and those formed in the presence of a vapor phase. The bearing of these on the preeruption, emplace- ment, and subsequent mineral development is traced. The wide distribution of ash flows, and the immense vol— ume of the materials represented by some deposits pose many problems in volcanology. The mechanism of emplace- ment and physical chemistry of the deposits are considered. The source of heat required for welding has long posed a problem to geologists, but a summation of all the geologic and physical factors leads to the conclusion that no exo- thermic heat source is required. The inherent heat and its conservation by physical relations seems to provide ample heat for welding. The physical chemistry of the welding process is presented. Typical occurrences of ash-flow deposits, several repre- sentative hand specimens, and a wide range of welded and nonwelded tufi's as seen under the microscope, are illus- trated, and discussed Where clarification of their significance is required. INTRODUCTION The study of volcanism as an observable geologic process, and of volcanic rocks, readily available at the earth’s surface, provides a direct approach to the understanding of igneous processes and rocks. Consequently, a great deal of geologic work has been concerned with volcanism and its products, but more emphasis has been given to the lavas or flow rocks than to the fragmental rocks. One reason for this is that the pyroclastic rocks are a transitional group, composed of lava fragments, but laid down generally like sediments. These rocks have many of the characteristics of sedimentary deposits. Silicic pyroclastic rocks are, in many regions of the world, more widespread than other types, therefore an understanding of their origin is of major impor- tance. Among these, the group known variously as ash flows, welded tuffs, glowing avalanche deposits, incandescent tufl" flOWS, ignimbrites, and others, has 1 2 ASH—FLOW TUFFS been the subject of intensive study by many geolo- gists, and several excellent papers describing specific occurrences have been published. Throughout the literature, however, an ever—in- creasing diversity and duplication in terminology has been used to describe ash—flow materials, and to designate different origins, owing in part to the development of criteria for recognition, and in part to the evolution of ideas on their origin. For sev— eral years, the authors have been engaged in a de- tailed study of the Valles Mountains region of New Mexico where ash—flow deposits reaching a maxi— mum thickness of nearly 1,000 feet cover more than 350 square miles. The area is dissected by many canyons where pyroclastic rocks and sections are well exposed. The experience gained in this study, supplemented by observations in many other areas of the Western United States, especially the Basin and Range province, and in Mexico, New Zealand, and Iceland, has led the authors to believe that the time is long overdue for a consideration and evalua- tion of the nomenclature, characteristics, and origin of ash-flow tufi's. Many ash-flow tuffs have not been mapped as such in the past because they were not recognized, and many areas previously mapped aslava flows are now known to be ash-flow tuf‘fs. Other areas will doubtless be found after more de— tailed laboratory study and areal mapping. The data acquired in these studies will have more mean- ing if time is taken now to unravel the threads of the problem, define the terms, sort out the ideas, and reach a state, however temporary, from which fur— ther progress can be made. This report is concerned with the results of ob- servations of pyroclastic eruptions as reported in the literature by earlier writers. Geologic and petrologic studies of several great eruptions were made and correlated with physical and chemical data. The report also gives the nomenclature of pyroclastic materials, and describes the genetic his- tory of ash—flow tuffs, that is, eruption, trans- portation, consolidation (including welding), and devitrification. Because the varied and distin- guished characteristics can be more adequately and precisely described by illustrations than by words, the report contains definitive photographs of hand specimens and thin sections. ACKNOWLEDGMENTS / Geologists who have observed and studied volcan- ism and volcanic rocks in many parts of the world have contributed generously during the preparation of this report, and this help, either as criticism, suggestions, advice, or specimens supplied, is here appreciatively acknowledged. Some contributions should be mentioned particularly, and first among these is the collection of 800 thin sections assembled by F. Zirkel for his petrographic studies that ac- companied the geologic report on the exploration of the 40the parallel published in 1876. It is histori- cally valuable as one of the earliest collections of volcanic rocks from the Western United States. The fine collections at the US. National Museum were available through the kindness of the late W. F. Foshag, curator of geology. For many areas throughout the world that are classic localities for pyroclastic volcanic rocks, the authors have had the benefit of personal discussions with geologists familiar with the areas, or of speci- mens, and personal thanks are extended especially to Ivan Wilson, for ash-flow tuffs from Baja Cali- fornia; J. Hoover Mackin, for specimens from the Iron Springs area of southwestern Utah; W. H. Bryan, for a collection of tuifs from the Brisbane area, Queensland, Australia; the late C. N. Fenner, a pioneer in the study of ash flows, for specimens from Peru; F. G. Bonorino, for andesitic welded tufi's from Argentina; J. Bouladon, for ignimbrites from Morocco; S. Hjelmqvist, for ignimbrites from central Sweden; C. G. Johnson, H. L. Foster, H. Kuno, and T. Matumoto, for specimens from Japan; J. Westerveld, for discussions of Sumatran welded tufts, and the late C. P. A. Zeylmans van Em- michoven, for specimens from the Toba region of North Sumatra, which were collected in 1892 by N. Wing Easton, and for specimens from the West Pasoemah region, South Sumatra, collected in 1933 by Musper and Westerveld. NOMENCLATURE Early in the authors’ study of the Valles Moun- tains region, it was evident that the terminology of pyroclastic materials was inadequate. For some materials or processes, there was no term; for others, there were several. As the study and un— derstanding of pyroclastic rocks had progressed over the years, new terms had been suggested, old terms had not been completely discarded, terms in use for certain ideas or substances had been ex- tended in their meaning, and others were no longer applicable in their original definition. As a result, an extended examination of the terms reported in the literature has been made to face the problem of the confusion of terminology that is the natural result of work carried on in diiferent parts of the GLOSSARY OF TERMS 3 world on a subject that lends itself so easily to differences in description and interpretation. Two alternatives were open—either to bring to- gether the terms in the literature, with their defini- tions, show how they had been used originally and evolved, and modify them where necessary by de- scriptive words and phrases—0r to devise an en- tirely new systematic nomenclature which would require discarding some terms, redefining others, and coining new ones. No set of terms will encom- pass all the variables that are found in a highly detailed description of any group of phenomena. If specific names are proposed, there will be a need for even finer splitting, and the making of still more names. The latter course was therefore not desir- able particularly for the reason that Wentworth and Williams (1932, p. 44) have put so succinctly, However, no great acquaintance with geologic literature is needed to be aware that only a small fraction of the terms which are proposed are destined to come into general use. The majority are only likely to add to the already burden- some list. Experience has shown that complete, systematic, novel series of terms to cover a particular field are rarely adopted. The use of modifying words and phrases, usually requiring only 1 or 2 words to describe a variation from the norm, offers a flexibility not possible With specific terms. Thus it was decided to use the exist- ing. nomenclature and modify it by descriptive ad- jectives and phrases, a course that has Wide acceptance and that gives others the opportunity to adapt terms to their own problems and circum- stances. However, no system of nomenclature, no matter how well devised, can obviate the need for detailed description wherever that will give a more accurate picture. Among the contributors to the development of the nomenclature of pyroclastic rocks are Blyth (1940), MacGregor (1955), Pirsson (1915), and Wentworth and Williams (1932). The following glossary of terms gives the deriva- tion and current usage of the term together with the authors’ comments on interpretation or suggested usage. GLOSSARY OF TERMS Ash fall.—Deposition of volcanic ash directly from the air, generally, but not always, resulting in a stratified deposit showing crude to very complete sorting of its component parts. Unconsolidated deposits are called ash; consolidated deposits are called tufi". The term “ash fall” has been used by Griggs (1922, p. 25), Fenner (1923, p. 26—28), Kozu (1934, 552858-61—«72 p. 136), and MacGregor (1952, p. 70). Some writ- ers use the term “airborne” to describe their ma- terial. Ash flow—A turbulent mixture of gas and pyro- clastic materials of high temperature, ejected ex- plosively from a crater or fissure, that travels swiftly down the slopes of a volcano or along the ground surface. The solid material in an ash flow, although unsorted, is dominantly of particles of ash size (less than 4 mm in diameter) but generally contains different amounts of lapilli and blocks. Fisher (1954, p. 74), Taylor (1954, p. 86), and MacGregor (1955), among others, use the term “ash flow.” The flowage principle is used in industry for the transportation of solids suspended in a gas, and is known as fiuidization (Work and others, 1949). Reynolds (1954) has applied the concept to the dis— persal of ash-flow materials. Ash-flow taff.—The consolidated deposits of vol- canic ash resulting from an ash flow are called ash- fiow tufl". Ash flow is here used as an adjective to indicate the mechanism of dispersal, and tufi‘ indi— cates the state and size of the material. Ash-flow tuff is an inclusive, general term for consolidated ash-flow beds that may or may not be either com— pletely or partly welded. Ash, volcanic—Wentworth and Williams (1932, p. 45) define ash as “uncemented pyroclastic ma- terial consisting of fragments mostly under 4 mm in diameter.” Blyth (1940, p. 148) described ash (volcanic ash) as “unconsolidated pyroclastic ma- terial consisting mainly of fragments less than 4 mm in size.” He adds, “The use of the term ‘ash’ to connote a consolidated pyroclastic deposit is not to be recommended; * * * the word tuff should be used in such cases.” (See Pyroclastic materials.) Aso lava—Indurated pyroclastic materials sur- rounding the Aso caldera of Kyushu, Japan. Matu- moto (1943, p. 6) states, “The name Aso lava was originally given to the obsidian-like agglomeritic, eutaxitic, or welded mud lava * * *.” Aso lava was originally described by Iki (1899). The origin of this and many other deposits of simi— lar materials is now well known in Japan and most of these deposits are described under the terms “welded tuffs” and “pumice flows.” Avalanche, volcanic.—The sudden avalanching of large amounts of any volcanic material down the slopes of a volcano. The term has been used by many authors, among them Lacroix (1904, p. 350) who described the material in the avalanche at Mount Pelée as being composed of earth and blocks, 4 ASH-FLOW TUFFS and by Anderson and Flett (1903) who described an avalanche of dust, sand, stones, and burnt timber at the 1902 eruption of La Soufriere on St. Vincent. Some volcanic avalanches have been described as glowing avalanches (Williams and Meyer-Abich, 1955, p. 33) where the material in them is red hot and contains gas. Other avalanches, that occur where large quantities of water are involved, are known as lahars. Axiolite, axiolitic—Zirkel (1876, p. 173, 176) de- scribed “linear aggregates of axially grouped fibers” as axiolites, and illustrated them in plate 7, figure 4 of that report (reproduced in fig. 95 of this re- port). Iddings (1899, p. 419) has given a detailed description of “axiolitic structure” as follows: In certain kinds of rhyolite apparently composed of welded glass fragments, there is a microspherulitic growth which bears a definite relation to the form of the supposed glass fragments. The feldspar fibers are in groups which are ap- proximately normal to the outline of the fragments and radiate inward. Where the fragment had a rudely tri- angular shape the central part often attained a greater degree of crystallization than the margin, and sometimes consists of distinct crystals * Marshall (1935, p. 23) has described axiolitic struc- tures under the term “pectinate structure.” Blocks—Blyth (1940, p. 147) defines blocks as follows: “Fragments of cognate or accidental ma- terial, larger than lapilli and usually angular, which have been erupted in a solid or nearly solid state.” (See Pyroclastic materials.) Block and ash flow—See Ash flow and Pumice flow. Dust clouds—Airborne pyroclastic material of dust size (see Pyroclastic materials, also Dust, ’vol- canic) that characterize explosive volcanic erup- tions. 1 The dust clouds associated with ash flows are not basically different from glowing clouds, glutwolken, and nuées ardentes (see Nue’e ardente) although these terms emphasize the glow that is normally reflected from the underlying incandescent ash flow. Dust, volcanic—Wentworth and Williams (1932, p. 47) have defined volcanic dust as follows: “Pyro- elastic detritus consisting mostly of particles less than 14 mm in diameter, that is, fine volcanic ash.” Blyth (1940, p. 150) gives the dimension of dust as less than 0.05 mm. (See Pyroclastic materials.) Eutaxite, eutaxitic.——Fritsch and Reiss (1868, p. 414) proposed the name eutaxite for a volcanic rock composed of ejected fragments of different colors, and texture as follows: “The different frac- tions in general lie beside one another as streaks, n sed’ bands, and lenses in seemingly well ordered dis- tribution.” This type of material is illustrated in figures 1 and 11. Piperno seems to be a similar rock but may, however, differ slightly by having larger lenses of glass known as fiamme. Fiamme.—The Italian name used to describe black glassy inclusions in piperno and which have a cross section shaped like tongues of flame. These are often several centimeters in length, but may range from microscopic ’size to several feet in length. Zavaritsky (1947, p. 11) has proposed to include all such glass lenticles (collapsed pumice frag— ments) under the term “fiamme.” Matumoto (1943, pl. 24) illustrates “obsidian spindles” in the “Aso lava” from the Aso caldera of Kyushu, Japan. Fluidization.—See Ash flow. Wt filowing avalanche—See Avalanche, volcanic. Glowing clouds—See Dust cloud, also Nue’e ar— dente. Glutwolken.—See Dust cloud, also Nue’e ardente; FIGURE l.—A specimen from the Battleship Rock tufi (fig. 3) north- northeast of Jemez Springs, N.Mex. This represents the ”thoroughly welded part, and figures 42 and 43 represent specimens from the non- welded basal part of the same flow. In the nonwelded tufi the pumice contains about ’70 percent of pore space, while in this specimen there has been complete collapse and welding with the elimination of all pore space. The black lenses of collapsed pumice in a fine-grained ground- mass represent the structure known as piperno. Ignimbrite.——-Marshall (1935, p. 4—10) proposed the name “ignimbrite” (literally fiery rain cloud rock). He defined ignimbrites (1935, p. 1) as fol- lows: "‘ * * they are thought to have been deposited from immense clouds or showers of intensely heated, but generally minute fragments of volcanic magma. The temperature of these fragments is thought to have been so high that they were viscous and adhered together after they reached the ground. GLOSSARY OF TERMS 5 Elsewhere he states (1935, p. 38), Ignimbrite is used as a name for a tuffaceous rock of acid composition that has been formed from a ‘nuée ardente Katmaieene’ in the nomenclature suggested by A. Lacroix. Thus Marshall visualized ignimbrites as a rock of “acid composition” formed by the fall from a cloud of hot viscous material. The term “ignimbrite” has been used by Bonorino (1944), Bouladon and Jouravsky (1955, p. 25), Hjelmqvist (1956), and others. Incandescent tnfi flow—Fenner (1948a, p. 879) described the “tuff flows” of the Arequipa region of Peru as “incandescent tuff flows,” and stated, “The tuff deposits are the result of a series of fragmental outbursts of rhyolitic lava, similar to that which occurred in the Valley of Ten Thousand Smokes in Alaska.” He also stated (1948a, p. 882), “An es- pecially important feature * * * is the ability of such tuff flows to spread widely over level or gently inclined surfaces.” Jenks and Goldich (1956, p. 156) use the term “tuff flows” for the same deposits. “Tuif” is not a suitable term to describe uncon- solidated material (see Tufl; also Ash, volcanic) during fiowage. Only after emplacement and con- solidation should the material properly be called (‘tufi'.’, Lahar.—A term used in Indonesia to designate a volcanic mudflow. The term has been used by Cur— tis (1954, p. 458) “* * * for any volcanic breccia with a matrix of tuifaceous aspect which came to rest as a single unit and was originally mobilized by addition of water, gravity alone being the moti- vating force.” Lahars may originate in different ways such as mobilization of rain-soaked debris on volcanic slopes (cold) or by eruption through crater lakes (hot) (van Bemmelen, 1949, p. 191). Hot lahars may also be initiated by nuées ardentes en- tering streams. Lapilli.——Wentworth and Williams (1932, p. 33) state, “According to most writers they [the lapilli fragments] may consist either of juvenile lava, still liquid or plastic when ejected, or of broken rock of any sort from the walls of the vent or from the bed rock; that is, they may be essential, accessory or ac— cidental ejecta.” Blyth (1940, p. 147) describes lapilli as “Cognate or accidental ejecta ranging mainly from 32 to 4 mm in size.” (See Pyroclastic materials.) The lapilli of ash-flow tuffs are commonly pumice fragments, as are most included blocks, but acci- dental rock fragments may also be present. Mndflow, volcanic.——Volcanic mudflows or lahars are not basically different from nonvolcanic mud- flows except for the volcanic origin of their ma- terials. Nonvolcanic mudflows or “mud spates” have been described by Rickmers (1913, p. 195) as follows: It is not dry nor is there much water, but the whole mass appears like a rapid flush of mud, although frequently the rock waste is so rough as to suggest that it is properly mud. Enormous boulders will float in this thick porridge. The term “mudflow” was also applied to ash flows before they were fully understood. Thus Marsters (1912) called the ash flows of Peru “mudflows” and Griggs (1917, pl. 2) in his early studies of the vol- canic deposits of the Valley of Ten Thousand . Smokes called this material “mudflow.” This was later called “sand flow” by Griggs and “tuff” deposit by Fenner (1923). Mud lava—The term “mud lava” has been used to describe some materials in Japan which are now known to be ash-flow materials. Matumoto (1943, pl. 30) referred to welded mud lava. (See Aso lava.) Nonwelded tnfi‘.—The term “nonwelded tuff” will be applied to those ash flows or parts of ash flows that have not become welded. The mode of deposi- tion and consequent cooling of ash-fall materials precludes welding, in contrast with ash-flow ma- terials which may or may not become welded. Thus there should be no confusion, and it seems unneces- sary to use the term “nonwelded ash-flow” materials each time they are mentioned. That is, “nonwelded tuffs” are always to be‘ understood as the corollary of welded tuffs. ‘ An alternative would be to extend the term “sillar” (see Sillar) to include all nonwelded tuffs. Fenner (1948a, p. 883) proposed this name for the poorly indurated tufi‘s of Peru, but did not suggest its use as a general term for these materials. Also, some of the tuffs of Peru that were included as sil- lars seem to have undergone slight welding. How- ever, one of the principal characteristics of the Peruvian tuifs is induration by vapor-phase crystal- lization. This was recognized by Fenner and it may be that this term could be more appropriately used for tuffs indurated by this process than as a gen- eral term for nonwelded tuffs. The term “sillar” has been used by Barksdale (1951) and Williams (1952, p. 176) for nonwelded ash-flow tuffs. Nnée ardente—Lacroix (1903a, p. 442—443) pro- posed the term “nuée ardente” to describe the previously unrecognized type of volcanism that characterized the eruptions of Soufriere and Pelée in 1902. Lacroix (1904, p. 350) describes the erup- tion at Pelée as follows: 6 ASH-FLOW TUFFS A nuée ardente is made up of an emulsion of solid materials in a mixture of water vapor and gas at high temperature, about the nature of which I shall not refer. The shape and dimensions which it presents at the moment of its emission from the carapace of the dome, show that it has been sub- jected to an enormous pressure, for after a few seconds, it occupies a volume greater by many thousand times that which it had at its inception. He also says, At the base [of a nuée ardente] is found a zone at very high temperature, in which the solid materials predominate (blocks of all dimensions, very small fragments, fine cin- ders); each of these pieces, or the solid particles of which it is formed, radiate heat, and must be surrounded by an atmosphere of gas and vapors, extremely compressed at the beginning, but expanding rapidly; it is this atmosphere which prevents the solid particles from touching one another, maintaining the mass in a state of mobility which allows it to flow over the slope almost in the manner of a liquid. The term “nuée ardente” is well established as the name for a type of volcanic eruption, a usage with which we are not concerned (see Pelean eruptions). We are, however, concerned with an understanding of the mechanism of deposition of welded tuffs. The second part of the quotation from Lacroix shows that he clearly recognized the dual character of the nuée ardente—the overriding dust clouds, and the dense basal part. The full significance of the concept of LacroiX has not always been understood. Perret (1937, p. 5) pointed out the common fail- ure to recognize the dual character of nuées ar- dentes in the following statement: The term nuée ardente for want of something more pre- cisely descriptive, has been quite generally adopted. ‘Nuées ardentes’ as well as ‘Glutwolken’ and ‘Nuages denses’ all refer in too restricted a sense to the cloud aspect of the phenomenon, losing sight entirely of the source of energy within the avalanche, the inherently active mass of denser material, hidden, all but completely, by the magnificently spectacular convolutions of vapor and ash * ’1‘. Again (1937, p. 20) in describing the eruptions of January 5, 1930, he states, A strong west wind i“ moved the ash cloud above, while the heavy mass of the avalanche beneath was quite un- affected. The dust cloud ’3 * * was not only checked, but forced backward even beyond the point of issue. Figure 2 shows the dual character of a Pelean nuée ardente. Fenner (1948a, p. 881) made the following com- ment: The terms ‘nuées ardentes’ and ‘glowing clouds’ were ap- plied to these phenomena, but though these clouds are spec- tacular and terrifying, to call the whole group of actual manifestations ‘nuées ardentes’ tends to place the emphasis FIGURE 2.~A “small typical nuée ardente" of Feb. 8, 1930, at Pelée, illus- trated by I‘erret (1937, fig. 16). The arrows indicate the ash-flow part of the nuée ardente which has sped in advance of the overriding dust clouds. on the relatively superficial clouds that rise from the plung- ing mass rather than on the mass itself * i" *. Notwithstanding the well-established recognition (beginning with Lacroix) of the dense basal part of an ash flow as the effective medium of transporta- tion, the strictly airborne concept has found its way into some geologic interpretations, and beclouded the whole approach. The geologist must study and interpret the deposits formed by the dense under- lying part of the eruptions, and is in need of clearly distinguishing terms. This means that, as Mac- Gregor (1955, p. 10) has pointed out, “* * * the classification of nuée ardente eruptions should be quite distinct from the classification of the products produced by such eruptions [and including the mechanism that emplaces these products].” In summary, nuée ardente has come to have two distinct usages; one for a special type of volcanic eruption, and the other for the clouds that ordinarily accompany these eruptions—glowing cloud being the commonly used English equivalent. Usage has not always differentiated between the clouds themselves and the dense ash- or block— and ash-transporting basal part. If so used, this basal part would con- stitute the noncloud portion of a glowing cloud, which may not even be glowing. In general, it glows only by reflecting the incandescent underly- ing ash flow. Obsidian spindle—See Fiamme. Pectinate.—See Am’olite. Pelean eruption.—This term has been used to de- scribe certain types of eruptions resulting in ash- flow or avalanche deposits, and similar terms have ( GLOSSARY OF TERMS 7 been proposed for several other types of volcanic eruptions. The classification of types of eruption has been outlined by Lacroix (1930) , Fenner (1937, p. 236), Perret (1937, p. 3—5), and Williams (1941b). A detailed classification has been pre- sented by MacGregor (1952, tables 1 and 2). These terms are widely used in descriptions of the erup- tions of active volcanoes, but have also been applied to older ones where the character of the eruption could be verified. Piperno.—A rock first described from the Phle- grean Fields in Italy is locally known as piperno. It is characterized by conspicuous lenses of glass (fiamme). “Pipernoid” is a phase of the, same rock in which a similar structure is revealed under the microscope. . Dell’Erba (1892), and later Zambonini (1919, p. 72), concluded that piperno was a tuif, and was deposited at a high temperature. Rocks of this texture are illustrated in figure 1. Pumice.—Webster’s New International Diction- ary (1933, p. 1735) defines pumice as follows: “A highly vesicular volcanic glass produced by the ex- travasation [more properly exsolution from the glass] of water vapor at a high temperature as ' lava comes to the surface: hardened volcanic glass froth.” Wentworth and Williams (1932, p. 39) quote Lacroix (1930, p. 437) as follows: The glass is not only puffed up, but spongy, drawn out, sometimes fibrous or filamentous; the cavities are extremely abundant and their form varies according to the composi- tion. The types with elongated, tubular parallel vesicles have a fibrous or silky aspect; these are characteristic of rhyolite pumice, whereas the cavities are more or less spheri- cal in trachytic and phonolitic pumice. The term “pumice” implies no specific size of ma- terials and a pumice structure is retained by some ash-size materials. In general, however, explosive eruption has disrupted the original vesiculated magma into individual segments of the cell walls, and so the distinctive pumice structure has been destroyed to form glass shards. The material which results is known as volcanic ash. (See Ash, vol- canic.) If the retention of pumice structure is conspicuous the material could be called a pumi- ceous ash. Pumice fall—Deposition of pumice directly from the air. (See Ash fall and Tephra.) Pumice flow—Flows with a conspicuous propor- tion of pumice fragments of lapilli size (4 to 32 mm) or block size (>32 mm) have been called pumice flows. (See Pyroclastic materials.) Tsuya (1930) applied the term “pumice flows” to the ma- terials formed by the 1929 eruption of the Komaga— take volcano in Hokkaido, Japan. Other Japanese geologists, including Kozu (1934, p. 136) and Kuno (1941, p. 148), have described pumice flows. AS used by the Japanese geologists “pumice flow” is synonymous with “ash flow.” Flows characterized by coarser material have been called lapilli flows, and block flows. Williams (1942, p. 79) referred to the Crater Lake “glowing avalanches” as “pumice flows” and “scoria flows.” Perret (1937, p. 5) applied the term “block and ash, flow” to some materials at Pelée. Pumiceous ash.—See Pumice. Pyroclastic.——Wentworth and Williams (1932, p. 24) state, “Pyroclastic is an adjective commonly applied to rocks produced by explosive or aerial ejection of material from a volcanic vent.” The noun form pyroclasts is defined by Holmes (1920 p. 193) as “A general term for fragmenta] deposits of volcanic ejectamenta, including volcanic con- glomerates, agglomerates, tufts, and ashes.” (See Tephra.) Pyroclastic materials—The following size classi- fication of volcanic fragmental materials is given by Wentworth and Williams (1932), and by Blyth (1940) : Size classification of volcanic fragmental materials, in millimeters Name Wentworth and Williams (1932) Blyth (1940) Block >32 >32 Lapilli Mostly 32 to 4 32 to 4 Coarse ash 4 to 1%; 4 to 0.5 Flne ash <14 0.5 to 0.05 Volcanic dust <0.05 Wentworth and Williams, and Blyth are virtually in agreement, except that Blyth preferred decimal fractions, using the figure approximating the value of the common fractions used by Wentworth and Williams. Sand flow.—Griggs (1922, p. 253—254) describes the “great hot sand flow” as follows: The bulk of the deposit is composed of fine fragments, many of them dust-like, but there are included numerous lumps of pumice which in places make up a considerable fraction of the whole. There is no trace of stratification of the ma- terials, except where they were obviously subject to sec- ondary readjustments after deposition * * *. At the foot of the valley we found clear—cut and positive evidence of the manner of the tufi' making it certain that the sand of which it was composed must have flowed down the valley like a viscous liquid. 8 ASH-FLOW TUFFS Griggs (1922, p. 261) also referred to the tufl’ as “the incandescent sand flow.” Scoria flow—See Pumice flow. Sillom—Fenner (1948a, p. 883) in describing the “incandescent tuff flows” of Peru writes, For those [tufl‘s] in which induration is primarily the result of recrystallization, and for those in which the fragments have little cohesion, another term is desirable [that is, other than welded tufl‘s]. The local term ‘sillar’ (pronounced seelyar), commonly used in the Arequipa region, has been applied in the present paper. Tephra.—Thorarinsson (1955, p. 12) proposed “the term tephra as a collective term for all elastic volcanic material transported from a crater through the air * * *.” Tnfi.—Wentworth and Williams (1932, p. 50) define tuff as “indurated pyroclastic rocks of grain generally finer than 4 mm, that is, the indurated equivalent of volcanic ash or dust.” (See Blyth, 1940, p. 153; also Ash, volcanic, and Incandescent tnfi“ flows.) Tufi‘ land—Abich (1882, p. 39) described the rocks of the Alaquez massif, Russian Armenia, as “tuflava,” and Zavaritsky (1947, p. 12) has used the same term and shown that the materials are characteristic welded tuffs. Tufi‘, crystal.—In describing crystal tuffs Pirsson (1915, p. 199) states, Tuffs composed entirely of crystals, must be very rare ’i‘ i“. On the other hand, crystals of minerals, the kinds depend— ing largely on the nature of the magma, either perfect in form or more or less fragmental, are found in nearly all tuffs; and when they become a dominating or striking fea- ture of them the rocks are referred to as crystal tufts. Crystals or crystal fragments characterize almost all ash—flow tuf‘fs and range from a trace to more than 50 percent of the rock in some tufl’s. L Tnff, lithic—In describing lithic tuffs Pirsson (1915, p. 201) states that “The essential feature of tuffs of this class is the presence in them of a strik- ing or dominating degree, of fragments of previ- ously formed rocks * * *. These fragments may be holocrystalline or partly glassy.” Rock fragments are present in nearly all ash-flow tuffs but rarely exceed a few percent of the rock. Tnfi, nitric—A name proposed by Pirsson (1915, p. 194) for “Tuffs produced by the sudden and violent explosion of a more or less viscous magma * * *. The explosion of gas and rupturing began in a liquid medium; the resulting product falls as a rigid glass.” In general, ash-flow tufl"s are dominantly vitric or derived from originally vitric material. As Pirs- son states, such material would fall from the air as a rigid glass (ash-fall tufts) . However, where flowage occurs, the glass may retain plasticity until after deposition. Tnfi‘, welded—Mansfield and Ross (1935, p. 308, 321) described welded volcanic tufts—that is, those in which individual fragments had remained plastic enough to become partly or wholly welded i‘ * In a few specimens the original forms are unmodified; in others there is flattening, but Without obliteration of characteristic ash structure; and in a few, extreme flattening and slight flowage has almost obscured the original structure. The term “welded tuff” is self explanatory, and has been Widely used by many geologists including Gilbert (1938, p. 1851), Westerveld (1942), Wil- liams (1942, p. 60), and Enlows (1955). Boden- hausen (1955, p. 44) used the French form of the same term (tufs cineritiques soudes). Welded pumice—Collapsed pumice has been called welded pumice by Iddings (1899). The materials called welded pumice by Iddings are discussed in detail on pages 11—12 of this report. The full relationships of these materials were not fully understood by Iddings, but the subject of welding interested him greatly, and he returned to it several times. Iddings is widely recognized as discovering the phenomenon of welding, and of hav- ing proposed the term “welded pumice,” although the term “welded” was used long before by Zirkel (1876, p. 267). DEVELOPMENT OF CONCEPTS Present-day concepts of the origin of ash-flow tuffs and of processes which lead to welding in these rocks have developed through a series of discon- nected observations and events over a long period. An attempt is made here to integrate a series of early interpretations, key studies, and important natural events into a chronology showing the gen- eral evolution of thought regarding these rocks. A review of the literature indicates that welded tufl’s have caused much speculation and difference of opinion among geologists at least as far back as the late 1860’s. Several workers have been con- fronted with rocks that showed most of the charac- teristics of pyroclastic materials and yet, in the light of existing knowledge, could only be explained as lava flows. There was at times actual contro- versy, one man claiming the rock to be a tufl’, and the other one calling it a peculiar type of lava. Many rocks have been described that show pecul- iar “flow” structures or “ash” structures, or what DEVELOPMENT OF CONCEPTS 9 has been considered normal “rhyolitic” structure. Many of these rocks, from their descriptions, illus— trations, or by reexamination, are now known to be welded tuffs. The writers believe that the at- tempts of some of the early workers to interpret these materials are an essential part of the historical background of welded tuffs. Thus the following review will include quotations from the older works as well as discussions of the papers dealing directly with welded and nonwelded ash-flow tuffs, and the relations of these rocks to volcanic eruptions of “Pelean” and “Katmaian” type. Special emphasis is given those older works where reexamination by the authors or by other workers has shown what is probably the true nature of the rocks described. Fritsch and Riess (1868), working on Tenerife, one of the Canary Islands, described rocks that they named eutaxites. These rocks, which to the au- thors’ knowledge have not been reexamined, seem almost certainly to have been welded tuflt's or at least welded clastic rocks. This seems to be one of the first recorded examples of an attempt to understand the origin of such rocks and is important because from this work comes the term “eutaxitic.” This term has since been used by many authors in different ways but has come to be applied most frequently to the structure seen in many welded tuifs (fig. 1). For further discussion of this term see page 4. During 1867—73, those engaged in the US. Geo- logical Exploration of the 40th Parallel, under the leadership of Clarence King, made extensive collec- tions of rocks in several of the Western States. These rocks were studied petrographically by Zir- kel (1876) and the field studies were discussed by King (1878). Of the more than 2,500 thin sections studied by Zirkel, the authors do not know how many were of rhyolitic rocks. However, more than 800 of these thin sections are now in the collections of the US. National Museum. Reexamination of these sections by the authors has revealed that nearly 200 are of welded rhyolitic tuffs, most of which came from the 40th parallel in Nevada (figs. 93 and 94). Many of Zirkel’s descriptions and illustrations (see reproduction in fig. 95) are clearly those of welded tuffs. It is doubted that Zirkel observed anything unusual about the origin of these rocks, but he did observe many textures that were new to him. Zirkel (1876, p. 168) says of the rhyolites in general, Of all the rocks, these rhyolites most excel in variety and diversity of microscopical structure; and since better facili- ties for investigation than had ever before been enjoyed, were furnished by the extraordinary number of occurrences at hand, it is highly probable that the following pages will be found to explain all, or nearly all, the most characteristic types of which the rhyolitic structure is capable. Particular attention has been paid to these interesting varieties, ex- amples of which will doubtless be found in studying the comparatively unknown rhyolites of other countries. Zirkel has given one of the most remarkable ac- counts of variations in texture in welded tuifs ever presented, but he was unaware of their true origin. From Zirkel’s work has also come the term “axio- litic” which he applied to a certain type of de- vitrification in glass shards. This structure is com- mon in certain welded tuifs and is discussed on pages 4 and 37, and illustrated in figures 71—74. King (1878) gives detailed megascopic descrip- tions and accounts of the field occurrences of the rocks that had been studied petrographically by Zirkel. Several welded tuffs are discussed under the classification of breccias and cemented breccias by King. The reader is profoundly impressed by these accounts and by the widespread distribution of these rocks in the Western United States, es- pecially in Nevada. Dell’Erba (1892), in a report on the origin of the much discussed piperno of Italy, concluded that this rock was a tuff rather than a lava as had been supposed by other workers. He also concluded that it had been hot at the time of its deposition. Piperno contains many lenticular fragments of black glass known locally as fiamme and Dell’Erba thought that it was these fragments that supplied the heat, the effects of which could be seen in certain peculiar characteristics of the tufi‘. Dell’Erba deserves great credit for recognizing that this rock had a pyroclastic origin. Several studies of the Italian piperno, particularly those of Zambonini (see p. 12 of this report) and of Zavaritsky (see p. 14), suggest that these lenticular glass fragments are collapsed and strongly welded pieces of pumice (fig. 1). Turner (1894), in discussing some of the rhyo- lites of the western slope of the Sierra Nevada in California, says, The rhyolite flows evidently followed the old river channels to a remarkable extent. The exact nature of these flows is not in all cases determined. They have been considered as tufi's or mud flows, but in thin section some specimens show trains of spherulites and other evidences of having been molten lava. Reexamination of similar and probably related rhyolites on the western slope of the Sierra Nevada 10 ASH-FLOW TUFFS by several workers indicates that these rocks can now be regarded as welded tuffs. Ransome’s (1898) study of a series of volcanic rocks of the western slope of the Sierra Nevada in California, provides the most outstanding example of careful consideration, indecision, and wrong con- clusion caused by a welded tuff, that has yet come to the attention of the authors. Among the vol— canic rocks Ransome discovered one that he named “biotite-augite latite.” He seems to have had difl‘i- culty classifying it from the very beginning, and says of his rock (1898, p. 27), * * ‘1‘ the biotite—augite latite has characteristics that render its distinction from a tuff not always easy. The fact that it has a wider distribution than the two undoubted massive flows, also suggests that it may have had a different origin. But the field evidence does not support the View that it is in any sense a water-deposited tuff. It is entirely devoid of the horizontal bedding characteristic of the known tuf‘fs of the region, and on the other hand it possesses the more or less uneven upper and lower surfaces, and the ability to lie upon perceptible slopes, which are characteristics of true lava streams. It also resembles lavas in being coarsely columnar in structure, although, as has been shown this structure may also occur in undoubted tuffs. The basal con- tact of the biotite-augite latite was nowhere seen sufficiently well exposed to give a decisive answer to the question whether it is tuffaceous or massive in character. The next quotation (Ransome, 1898, p. 42—44) is long, but is so pertinent to an understanding of the problems that faced earlier students of these rocks that the authors believe it should be included. This quotation not only indicates the views of some of the earlier European geologists, but shows the intensity of Ransome’s thinking on the controversial nature of some tuffs and lavas. The general microstructure of the biotite-augite latite ap- pears to approach very closely to the eutaxitic structure as the term was first employed by Fritsch and Reiss (1868), who applied it to certain facies of phonolites, andesites, and trachytes possessing some resemblances to clastic rocks and inclosing undoubted clastic fragments. They distinguished agglomerate lava, in which the structure was supposed to be due to partial refusion of clastic material, from piperno, in which portions of the magma in different stages of crystallization are brought into juxtaposition by the motion of the flow: Common usage, however, has somewhat altered the original application of the term eutaxitic, so that it fre- quently implies merely a fluidal banding brought about by alternating streaks of microfelsitic and clear glass (Zirkel 1893), or is even used as ‘a general name for banded vol- canic rocks’ (Kemp 1896). That such can hardly have been the earlier application of the word appears from the discus- sion that has centered about the original piperno of Pianura in the Phlegraean Fields, near Naples, a rock that Fritsch and Reiss took as the second type of eutaxitic structure, and which appears to resemble in some ways the biotite-augite latite of California. Luigi dell’Erba (1892) considers this rock a tuff, but the greater number of petrographers have regarded it as a lava flow. Zirkel (1893) refers to it as a sanidine trachyte with eutaxitic structure, remarking that Luigi Dell’Erba’s View is not even probable, while Rosen- busch (1896) cites it as an example of his Ponza type of the trachyte family. Such a discussion could scarcely have arisen were the structure concerned a simple banding such as is observed in many vitrophyres, and which no petrog- rapher could for a moment regard as indicating a tuffaceous origin. Kuch (1892), following the terminology of Reiss, describes both kinds of eutaxitic structure, the agglomerate lava and the piperno, as occurring in the dacites and ande- sites of Colombia. The latter form is particularly abundant, and Kuch remarks that it is at times impossible to dis- tinguish the two varieties, as the piperno often contains in- cluded fragments of other andesites. Wickmann (1897), in describing an augite-mica—andesite from the Indian Archi— pelago having a structure apparently nearly identical with that of the biotite-augite latite of the preceding pages, re- fers to it as having piperno structure; and there seems to be little doubt that this latter name, as used by the petrog- raphers cited, expresses accurately the structure of the second flow of the Sierra Nevada latites. When, on the other hand, a thin section of the last-named rock is compared with that of a lava having typical flow banding—as, for example, the beautiful vitrophyre of San Lugano, in the Tyrol, or the pitchstone-vitrophyre occurring between Lake Lugano and Lago Magiore, both of which, according to Rosenbusch (1896), often show eutaxitic facies —the chief difl’erence seems to lie in a greater irregularity and brecciation of the bands in the latite, and in the pres- ence of included rock fragments. In a later paragraph Ransome (1898, p. 44), says, As an historical illustration of the difficulty that sometimes exists in distinguishing a tufl" from a massive lava may be recited the case of the so-called ‘peperino’ near Viterbo, in Italy, which, according to Washington (1896), is probably a tufi', although it possesses some apparent flow structures, and has been by earlier investigators frequently designated a massive rock. As a result of reading the writings of Ransome on the biotite-augite latite and a group of associated rhyolites, Smith reexamined these rocks in the field in 1950 and concluded that they were welded tuffs. The biotite-augite latite shown in figure 82 is from an exposure near McKays, Calaveras County, Calif. This rock is a fine example of a crystal-rich, glassy, welded tufl’ and after examining it in the . field and laboratory it is not difficult to understand Ransome’s uncertainty about it. He finally con- cluded it was a lava because at that time no vol- canic mechanism was known that might produce a rock intermediate between a lava and a tufl". Part of the group of rhyolites studied by Ran- some are located near Altaville and Vallecito, Cala- veras County, Calif, and are within an area recently mapped by Lorin Clark (oral communication, 1950) of the US Geological Survey. These rocks have DEVELOPMENT OF CONCEPTS 11 been recognized by Clark as welded tuffs, and are probably closely related to the rhyolites discussed by Turner (1894), and the welded tuffs mentioned by Curtis (1954, p. 453—454). The studies of Diller and Patton (1902) of the volcanic rocks of the Crater Lake region give us one more example of the uncertainties of nomen- clature attached to these rocks. In writing of the Wineglass dacite flow Diller says, This peculiar tufl’aceous dacite occurring along much of the crest of the rim all belongs to one flow, which spread as a uniformly thin sheet over that portion of the base of Mount Mazama. It is altogether unlike the other flows of dacite and appears to be intermediate between them and tuff. Reexamination of this same dacite was made by Wil- liams (1942) and shown by him to be a welded tuff. The first major contributions to the understand— ing of welded tuffs were the works of Iddings (1885—86, 1899, 1909), an outgrowth of his studies of the volcanic rocks of Yellowstone National Park (figs. 87—89, 91 and 92). He must be credited with developing the concept of welding of pyroclastic materials as well as being the first to use the term “welded pumice.” His first use of the term “welded” seems to have been in connection with a specimen from the rhyolite flow of Obsidian Cliff. The fol- lowing quotation is important only because it seems to have been Iddings’ first recorded use of the word “welded,” and his concepts of welding of hot clastic rocks apparently developed from that time. Iddings says (1885—86, p. 274) , in reference to colored bands in obsidian, In some the colored streaks are in broad, thin bands, either straight or twisted, according to the last movements of the viscous glaSS. In others, they are in the most delicate threads, alternating with streaks of black grains running continuously through the rock, though sometimes interrupted by streaked patches of other character or appearing as though the rock had been broken into fragments and welded together again (pl. 16, fig. 2). This figure is reproduced in this report as figures 87 and 88. This rock from the Obsidian Cliff rhyolite flow is not a welded tuff, but rather a welded breccia which is not uncommon on the margins of rhyolite domes and flOWs or in sheared zones in rhyolite flows. Such welded breccias may often superficially re- semble coarse-grained welded tuffs, especially if they are composed of pumiceous fragments. In a more detailed study of the rhyolitic rocks from Yellowstone National Park, Iddings (1899, p. 403— 404) says in relation to pumice, “In some instances it is evident, from the confusedly twisted and curved arrangement of the glass fibers and films, that the inflated glass mass settled back upon itself, or collapsed, after the escape of much of the gas.” To explain this interpretation of his observations he says, “When we remember the enormous extent of many of the streams of rhyolite in this region, we may easily imagine the formation of pumice over the surface of an intensely heated area of lava, thus permitting its subsequent welding.” These statements were made with reference to collapsed or deformed pumices, rather than to the typical shard structures of welded tuffs. However, Iddings states (1899, p. 405—406), In numerous cases a pumiceous character is entirely want- ing. The mass is a compact glass, but it consists of ir- regularly shaped streaks and patches of different color. These twist and curve about one another and appear like a perfectly welded mass of strips or ribbons and irregular fragments of variously colored glass. In some cases their shape closely resembles that of fragments of pumice pressed together and welded (pl. 50, figs. 1, 2, and 3). In others it appears as though such fragments had been drawn out and twisted by a movement of the mass (pl. 51, fig. 2). Un- doubtedly this has been the case, but it is doubtful whether all the streaked and variegated glasses have passed through the process of inflation, collapse, and welding with subse- quent flow. However, the distinctly outlined and strongly contrasted streaks and ribbons of variously colored glass, are with difficulty explained in any other manner. Although Iddings recognized the phenomenon of the Welding together of vitric fragments, he ap- parently was uncertain as to the causes of welding and the origin of the resulting rocks. This is un- derstandable since Iddings’ accounts show that he was dealing with three processes of welding. These processes are: (a) the brecciation of glass in shear zones, or at the base of rhyolite flows and the sub- sequent welding of the resulting breccia; (b) the formation of pumice on the surface of lava flows and the deformation or collapse of such pumice due to the heat from the flow and its subsequent movement. Although some of Iddings’ “welded pumice” was formed in this manner probably many of the rocks examined by Iddings and believed by him to have formed in such a way were true welded tuffs; (c) the explosion of material from a vent into the air, the fall of the material while still hot back into the vent, and its subsequent welding, followed by ex- trusion from the vent in the manner of a. lava. Con- cerning the third process, Iddings (1909, p. 331) says, When exploded fragments of molten magma, large or small, fall together in a still heated condition, as may readily happen within the crater of a volcano or in the mouth of a fissure, they may be plastic enough to weld together into a more or less compact, coherent mass. This may subse- 12 ASH—FLOW TUFFS quently flow like other lava, and is known as FLOW—BRECCIA. * * ’1‘ [1909, p. 333] The same operation may result in the welding of exploded pumice, or of collapsed pumice that was highly inflated. The distinction between this and a flow breccia is chiefly in the size of the welded fragments. The explanations given by Iddings are inadequate to explain the process of welding involved in the for- mation of welded ash-flow tuifs but they are im- portant historically. Twenty-two of the original thin sections studied by Iddings are still available and some of the best of these are shown in the frontispiece, figures 88, 91, and 92. Although more than 50 years old, they are very well preserved. It is very fortunate that the thin sections, which were the basis for Iddings’ recognition of the fact of welding together of frag- ments of glassy pumice and ash and for his proposal of‘the name “welded pumice,” are still available. The eruptions in May 1902, of Mount Pelée on the Island of Martinique and of La Soufriére on the Island of St. Vincent, marked a turning point in the understanding of the mode of formation of many of the world’s deposits of pyroclastic rocks. The observations made at Sufriere and Pelée by Ander- son and Flett (1903), and at Pelée by Lacroix (1904), and at a later date by Perret (1937), gave to the science of volcanology an understanding of a new type of volcanic eruption, and formed the basis for the present-day concepts of the origin of ash—flow materials, including welded tuifs. The studies of Anderson and Flett, Lacroix, and Perret are discussed on pages 15—16 in this report. The pyroclastic deposits formed by the eruptions of Pelée and Soufriere are characterized by their complete lack of sorting and at Pelée, at least in part, by the widely variable dimensions of the con- stituent blocks, fragments, and dust. Such deposits came to be known as Pelean tuffs and in the years following the eruptions of 1902 several writers de- scribed tuff deposits from other parts of the world and attributed their origin to eruptions of Pelean type. Dakyns and Greenly (1905) seem to have been the first to further develop the concept in a paper on the “felsitic slates” of Snowdon, Wales. In an earlier paper Dakyns (1900) had recognized that these rocks were elastic and concluded that they were “felsite tuffs.” Greenly, inspired by the account of Anderson and Flett on the eruptions of Pelée and Soufriere, sought to reconcile certain un- explained features of the Snowdon rocks with the observed features of the ejecta of Pelée and Sou- friére. In so doing, he concluded that the Snowdon felsites were due mostly to eruptions of Pelean type. Greenly says of these rocks (Dakyns and Greenly, 1905, p. 548), “* * * in the felsitic slates of Snow- don we have a Pelean deposit of the Ordovician period.” Lacroix (1906, p. 1020—1022) attributed a Pelean origin to a group of ancient volcanic rocks of the Auvergne region in France. Zambonini (1919) reviews the problem of the origin of the Italian piperno and concludes, as had Dell’Erba, that it is a tuff and that it was deposited in a hot condition as shown by the occurrence of several late-stage minerals, some of which contain fluorine. He also concludes that it probably origi- nated as a “Pelean cloud” deposit. Although the eruptions of Mount Pelée and Sou- friere and the observations by Anderson and Flett, Lacroix, and the later studies by Perret were of fundamental importance in establishing a new type of volcanic action, and especially a new mechanism for the transportation of pyroclastic materials, it was the formation in 1912 of the “sand flow” in the Valley of Ten Thousand Smokes, Alaska, ac- companying the eruption of Mount Katmai, that provided the necessary link for our present under- standing of the origin of ash— and pumice-flow ma- terials in general. The first account of the discovery of the Valley of Ten Thousand Smokes and its now famous de- posit of volcanic ejecta, by members of a National Geographic Society Expedition was by Griggs (1917, p. 12—68). In a series of papers Griggs (1917, 1918a, b, c, 1919, 1921, 1922) discussed other phenomena of the region and the deposit of volcanic tuff which he termed the “great hot mudfiow” (Griggs, 1918b). Fenner (1920) and his associates, especially E. G. Zies of the Carnegie Institution of Washington Geophysical Laboratory, noted the effects of heat as shown by the so-called mudflow, but were unable to reconcile the observed heat effects with hot water- bearing mud and concluded that the deposit, while rich in gas, was virtually dry and called it the “great sand flow” (fig. 96). Fenner (1923) finally concluded that the sand flow was deposited as a mixture of gas and ash and was similar in many respects to the deposits formed by the eruptions of Pelée and Soufriere. However, he noted a differ- ence in the materials deposited and therefore con- sidered them to be the result of a typekof, or modified, Pelean eruption. This conclusion has had a marked influence on nearly all the geologists who have subsequently worked with ash-flow pyroclastic materials because, in comparing most welded tufl’s DEVELOPMENT OF CONCEPTS 13 and their nonwelded equivalents with the deposits of Pelée and the Valley of Ten Thousand Smokes, a greater similarity is seen with the sand flow than with the Pelean deposits. As an example of this comparison Williams (1927) , in a detailed study of the rocks of Snowdon, Wales, discusses the origin of some rhyolites that Dakyns and Greenly had previously concluded were tuffs formed from an eruption of Pelean type. Wil- liams noted that the tuffs were composed largely of glass fragments and glass dust rather than crystals, which are more typical of Pelean tuffs, and con- cluded that while they were probably erupted in a manner somewhat analogous to the Pelean ejecta, they resembled more strongly the rhyolite tuffs of the Katmai area. Moore (1934) discusses the probability of a nuée ardente origin for the “older” pumice of Crater Lake, Ore., as opposed to an origin of direct fall from the air as shown by the “younger” pumice. His conclusions are based on several differences be- tween the two types, especially with regard to size and sorting. The younger pumice is more uni- form in size and is better sorted, whereas the older pumice is widely variable in size and shows no bedding or sorting. The work of Marshall (1932, 1935) on the vast deposits of rhyolitic tufi's on the North Island of New Zealand is a milestone in volcanic geology. Marshall proposed the name “ignimbrite” to include a large group of rhyolitic rocks that cover approxi- mately 10,000 square miles of the North Island and that he regarded as formed by an eruptive process similar to that proposed by Fenner for the sand flow of the Valley of Ten Thousand Smokes. One of Marshall’s greatest contributions was his correlation of field studies with petrographic studies. He noted welding in tuffs as observed by Iddings and he was the first to recognize the im- portance of devitrification in obscuring the frag- mental character of these glassy pyroclastic rocks (fig. 97). Richards and Bryan (1934) in discussing the origin of the Brisbane tuff (fig. 75) of Queensland, South Australia conciude that * ’1‘ * the combination of tuffaceous and non-tuifaceous char- acters presented by the Massive Tuff could be most readily explained as due to an enormous eruption of the .incandes- cent avalanche type * * and that The Massive Type of the Brisbane Tuff presents many fea- tures closely analogous With those of the Ignimbrites of New Zealand and has also much in common with the Hot Sand Flow of Alaska. As did Marshall, these writers related petrographic to field characteristics and recognized materials closely resembling the welded pumice of Iddings. At about the same time and without knowledge of the work being done in New Zealand and Aus- tralia, Mansfield and Ross (1935) described a group of rocks from southeastern Idaho for which they used the term “welded tuff” (figs. 9, 10, 17, 18, 49— 54, 61, 70). These tuffs were compared to the “welded pumice” of Iddings and are possibly re- lated to the vast fields of Yellowstone rhyolite, much of which is now known to be welded tufl". These writers concluded that the Idaho tufi's have many features in common with the Katmai sand flow and may have had a similar origin. The integration of field studies with petrographic studies, as in the works of Marshall, Richards and Bryan, and Mansfield and Ross, established a direct relation between welding in pyroclastic materials and an origin of some of these materials from a particular type of volcanic eruption. It provided a correlation between some of the welded pumice of Iddings and volcanic eruptions of a type exemplified by Pelée, Soufriére, and the Valley of Ten Thousand Smokes. This relation seems true although no weld- ing was observed in the deposits at Pelée, Soufriere, or at Katmai by the early workers there. The pres- ent authors have examined material from Katmai and have found evidence of slight plasticity and possibly very slight local welding of some of the shard fragments. Although in 1935 there was much yet to be learned about welded and nonwelded ash-flow tuffs, the outstanding features of these rocks had been presented in the above discussed papers on the New Zealand ignimbrites, the Brisbane tuff, and the welded tufi's of southeastern Idaho. There followed a series of papers by several authors describing new localities for welded tufi's and presenting detailed studies of tuffs from specific areas. Fenner (1937) proposed the term “Katmaian type of eruption” to apply to eruptions that formed deposits having the characteristics of the sand flow of the Valley of Ten Thousand Smokes. He sug- gested that several of the pyroclastic deposits of Yellowstone National Park had such an origin. Gilbert (1938) described a group of welded tuffs from eastern California. These tuf‘fs comprise the Bishop tuff that covers 400 square miles and has an average thickness of 500 feet. Gilbert’s paper is 14 ASH-FLOW TUFFS an excellent treatment of the subject of welded tuffs and probably is one of the best in its organization and presentation of the facts, both field and petro- graphic. Stearns, Crandall, and Steward (1938) mention welded tufl‘s in the Snake River Plains area, and the Mud Lake region, Idaho. Anderson and Russell (1939, p. 243—247), in describing the Tertiary for- mations of northern California, were the first to use a recognized ash-flow deposit as a stratigraphic marker. The “Nomlaki tuff,” they believe, has “covered an area of at least 2,000 square miles.” This study is valuable as it points out the tremen- dous potential use of ash-flow tulfs for stratigraphic correlation. Feitler (1940) briefly described glassy welded tuffs from Bare Mountain, Nev. Kuno (1941) discusses the different characteristics of de- posits formed by “pumice flows” and those formed by “ejected pumice” as observed in the deposits formed by the eruption of the volcano Komagatake in 1929. He also compares these pumice—flow de- posits with Pleistocene and Recent pumice deposits of some of the other Japanese volcanoes and finds that they have very similar features. Ross (1941) found that the well-known “thunder eggs” (chal- cedony-filled spherulites) from central Oregon formed in a welded tuff (figs. 67—69). Williams (1941a) in his excellent paper on cal- deras discusses welded and nonwelded ash-flow tuffs and their common association with collapse calderas of the Krakatau type. In his study of the geology of Crater Lake National Park, Williams (1942) describes both welded and nonwelded ash- flow tuffs and discusses their origin from “glowing avalanches.” Westerveld (1942) describes “welded rhyolitic tuffs” from South Sumatra and discusses their origin, chemistry, and similarity to similar deposits in other parts of the world (figs. 16 and 22). Matumoto (1943) describes in some detail the so-called Aso lava which forms vast deposits around the Aso caldera. He says, “* * * the name Aso lava was originally given to the obsidian—bearing, piperno—like, agglomeratic, eutaxitic, or welded mud lava found along the top of the somma.” In an elaborate classification, he subdivides the Aso lava into several types. The paper is well illustrated with photographs of polished specimens of the Aso lavas and photomicrographs of thin sections (fig. 81). Bonorino (1944) describes andesitic welded tuffs from Argentina (figs. 83 and 84). Zavaritsky (1946, 1947) adequately explains many hitherto anomalous features of the Armenian “tuff lavas” as compatible in origin with an eruption of Katmai type. Following Fenner’s usage, Zavarit- sky (1947 ) classifies as deposits of the Katmai type several of the better known ash-flow deposits. His list includes the New Zealand ignimbrites, Bishop tuff of Gilbert in California, Crater Lake welded tuffs and pumice flows, Japanese “Aso lavas” and the classical Italian piperno. Following a review of the main features of the above—named deposits he describes the principal physical features of the Armenian tuffs and “tufi' lavas” (figs. 79 and 80). Westerveld (1947 ) concludes that the vast de- posits of “acid” volcanic rocks around Lake Toba, . Sumatra are welded rhyolitic tuffs (figs. 26, 27, 39, and 65). These rocks cover 20,000 to 30,000 square kilometers, and have an estimated volume of 1,500 to 2,000 cubic kilometers. Fenner (1948a), in describing the “incandescent tuff flow” deposits in the Arequipa region, southern Peru, concludes that these materials have had an origin similar to the tuff deposit of the Valley of Ten Thousand Smokes, Alaska. The areal extent ofi - these deposits is not known, but Fenner believes that these deposits rank high among the world’s greatest known examples of “incandescent tufl’.” Fenner found that» much of this tuff is composed of incoherent constituents and his studies indicated that the indurated varieties were due to recrystal- lization of shards and pumice fragments and the ., growth of secondary vapor-phase minerals. thought that welding was not an important process of induration in these particular rocks (see p. 8, 28 of this report). He thus decided that the term “welded tuff” was not satisfactory as a name for these materials and instead used the locally applied name “silla1” (figs. 24 and 25). Boyd and Kennedy (1951) were the first workers to approach the problem of welding in tuffs from a purely experimental standpoint. They found that by heating pulverized volcanic glass in a bomb they could effect welding. From this they reached con- clusions as to the temperature of formation of ‘ welded tuffs (see p. 42 of this report). Other noteworthy occurrences of welded tufi's that have been described or briefly mentioned in the literature in recent years are as follows: Recent descriptions of welded tufls Location United States: Little Hatchet Mountains, N.Mex .................................. Lasky (1947). Reference He '2 HISTORIC ERUPTIONS OF ASH-FLOW TUFFS Recent descriptions of welded tufls—Continued Location Reference United States—Continued Iron Springs district, Mackin and Nelson (1950); Utah1 ___________________________________ Mackin (1952). Rattlesnake formation, Oregon2 ____________________________ _Wilkinson (1950). Lewis and Clark County, Mont1,._.._____.___.-....;........___ Barksdale (1951) . Sierra Nevada, Calif ____________ Hudson (1951). Yellowstone National Park, Wyo __________________________ Boyd (1954). Lake Valley quadrangle, southwestern New Mexico ________________________________ _Jicha (1954). Blake Range, New Mexico_Kuellmer (1954). Chiricahua Mountains, Ariz ___________________________________ _Enlows (1955). Basement rocks of Texas and southeastern New Mexico __________________________ _Flawn (1956). Dragoon Mountains, Ariz....Gilluly (1956). Other countries : Near Parral, Chihuahua, Mexico __________________________________ Wilson and Rocha (1948). Baja California, Mexico ______ Wilson and Veytia (1949). Islands of the Gulf of California and neigh- boring land areas, Mexico ________________________________ _Anderson (1950). ;.Hunter-Karuah district, New South Wales ,,,,,,,,,,,, Osborne (1950). Sikhote-alin Mountains, eastern Siberia ________________ _Solovev (1950). Costa Rica 3 ,,,,,,,,,,,,, ._-Williams (1952). Eastern Iceland ‘ ________________ _ Dearnly (1954) ; Tryggvason and White (1955). English Lake district, northern England __________ Oliver (1954). Balsam Chain, Salvador ______ Weyl (1954). Precambrian of southern Morocco 5____________________Bou1adon and J ouravsky (1954, 1955). Asia Minor_______________-___~_____Westerveld (1955). Salvadorm.m.-.___________-_._______Williams and Meyer-Abich - (1955) . Northern J apan________.______- Ishikawa and Minato (1955) ; Yagi (1956). Corsica ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Bodenhausen. Oita Prefecture, Japan ______ Ishii and others (1956). Peru _________________________________ _Jenks and Goldich (1956). Japan ____________________________________ __ Matumoto and others (1956) . Sudbury, Canada __________________ Thompson and Williams ' (1956). Precambrian of Sweden ‘_-__Hjelmqvist (1956). 1Illustrated in figures 32, ‘33, 40, and 41 of this report. 2 Illustrated in figure 29 of this report. 3 Illustrated in figures 85 and 86 of this report. 4 Illustrated in figure 19 of this report. 5 Illustrated in figure 98 of this report. “Illustrated in figure 78 of this report. 15 HISTORIC ERUPTIONS 0F ASH-FLOW TUFFS A study of the mechanisms that produce ash-flow tuffs must begin with the available records of direct observations of volcanic eruptions. These eruptions are rare and minor events compared with those in some earlier geologic periods, however, such erup- tions assume great importance, and the following section Will present some of the conclusions of the few geologists who have had the good fortune to make direct observations. PELEE The great eruption at Pelée on May 8, 1902, on the Island of Martinique, is recognized as having initiated an understanding of nuées ardentes and the related dispersal of pyroclastic materials. The great initial eruption was not directly observed by volcanologists but a few of them promptly began investigations and gathered valuable information from residents, and the long continued observations on later eruptions built up an outstanding fund of information. Anderson and Flett, Lacroix, and Perret are those who contributed most to an under- standing of the eruptions at Pelée. In discussing the factors that give nuées ardentes the properties of flowing, Anderson and Flett (1903, p. 508—509) say, As this turbulent mixture of expanding gases and fine dust pours down the surface of the mountain, the small solid grains are unable at first to rest on the ground, even when they may have sunk to the base of the cloud, and they are swept up again and borne along till they reach some shel— tered hollow or the violence of the expansive force lessens and the turmoil diminishes. The fundamental question remains to be discussed—what is the source of the energy which drives the cloud along? To this we believe there is only one answer—the motive power is supplied by the weight of the mass. It is in a con- dition comparable to that of a heavy and mobile fluid which has been elevated by the volcanic forces and poised on the edge of the crater and proceeds to flow downward in obedi- ence to the law of gravitation. The foregoing interpretation of ash-flow mecha- nisms is evidently a synthesis of observations of the deposits at Soufriere, and, in particular, the later and lesser eruptions at Pelée. The nuées ardentes eruptions at Pelée were de- scribed in great detail by Lacroix (quoted on p. 6 of this repOrt). Lacroix (1904, p. 354) granted the effect of gravity, but emphasized the great im- portance of an initial eruption and the Violent pro- jection of the debris that resulted from this ex- plosion. The directed effect of the blast was the result of its coming from the base of the incipient 1 6 ASH-FLOW TUFFS spine, whose influence has been discussed by Mac- Gregor (1952, p. 54). Continuous observations of eruptions at Pelée were made by Perret from 1929 to 1932, and re- sulted in valuable contributions to the understand- ing of volcanism and, in particular, nuées ardentes of which he observed several hundred examples (fig. 2). Perret (1937, p. 22—23) says, First of all it should be realized that the horizontal move- ment, three kilometers in three minutes, is due to an ava— lanche of an exceedingly dense mass of hot, highly gas- charged and constantly gas emitting, fragmental lava, much of it finely divided, extraordinarily mobile, and practically frictionless, because each particle is separated from its neighbors by a cushion of compressed gas. For this reason too its onward rush is almost noiseless. To the end of its course the ‘auto-explosivity’ continues * ‘ Elsewhere Perret (1937, p. 87) says, At the moment of explosion liquid masses of lava, instead of being hurled high into the air to form bombs are con- verted more or less completely into ash by the rapid dis- charge of gas * * *. VALLEY OF TEN THOUSAND SMOKES An important advance in our understanding of the eruptive processes that give rise to ash flows, was a result of studies in the Katmai region of Alaska, where the eruptions more nearly resembled the great ones of the past, than did those at Pelée. However, this eruption occurred in June 1912, and the area was first visited by Robert Griggs in 1915; thus no one directly observed the “sand flow” of that eruption. However, significant relations were observable and some of the later phenomena of the eruption were still active. A series of papers resulting from several succes- sive expeditions to the Katmai region of Alaska, under the auspices of the National Geographic Society, by Griggs, and by Fenner and Zies of the Carnegie Institution of Washington Geophysical Laboratory, constitute outstanding contributions to the understanding of ash flows. In discussing the relationship Griggs (1922, p. 253—254) says, Surrounded as it is by high and rugged mountains, the most striking feature of the conformation of the Valley of Ten Thousand Smokes is the flatness of its floor. One could ride a bicycle for miles along its smooth surface * * *. Alto- gether it occupies an area of 53 square miles (137 sq km) * * *. The bulk of the deposit is composed of fine fragments, many of them dust-like, but there are included numerous lumps of pumice, which in places make up a con- siderable fraction of the whole. There is no trace of strati- fication of the materials, except Where they were obviously subject to secondary readjustments after deposition. In discussing the “sand flow” Fenner (1923, p. 72) says, I would, therefore, stress the remarkable character imparted to the dust and gas mixture by a continuous evolution of gas, which must not only have eliminated almost completely the contact friction of the particles, but also have tended to drive them apart somewhat forcibly and caused the mass to spread almost as freely as a true liquid. Naturally, gravity would act upon a mixture that was undergoing the assumed reactions just as it would act upon a liquid and would direct its course down the slope of the Valley, but the important fact to be emphasized is the man— ner in which internal friction of the mass was eliminated, and it seems to me that only in some way as that suggested would it be possible for this to be accomplished. COTOI’AXI Fenner (1923, p. 65) pointed out that although the mechanism of an ash flow was not understood at that time, Wolf has given an excellent description of such a flow in his discussion of an eruption of Cotopaxi, Ecuador. In the discussion Wolf (1878, p. 131—132) says, About 10 o’clock in the morning, while strong subterranean detonations were heard * ‘i‘ i“ the crater of Cotopaxi boiled over [iibersprudelte] with fluid, glowing lava, and this pre- cipitated itself with furious velocity down the declivities of the cone. I chose the words ‘boiled over’ intentionally, be- cause they indicated best the manner and form in which the effusion of lava occurred in this extraordinary outbreak i“ * the mountain as it suddenly came to effervescence [ebullicion] and threw out a ‘black mass’ smoking and ‘ steaming over all parts of the crater edge simultane- ously * *. It is indeed one of the singular features of this eruption that the lava poured out of the crater not in one or several streams, but symmetrically in all directions, over its lowest edge as over its highest points. On that ac- count the floods were so general around the mountain * * *. Very many phenomena indicate that the new lava must have possessed a very high temperature at its exit from the crater and must have been almost as fluid as water. Its expulsion occurred suddenly with a fearful welling-up of the fluid, glowing masses; for only thus is it explicable that in a quarter, or at most, a half an hour, such a fabulous amount of lava was delivered * * * and that it flowed out over the highest crater edges like the foam from a boiling over rice pot. One of the significant observations by Wolf is the “boiling over like foam.” Not knowing the identity or fluidity of an ash flow, Wolf believed that ex- cessively high temperature was necessary. SOURCE VENTS FOR ASH-FLOW TUFFS No eruption of ash flows, that has produced welded tufi's, has ever been directly observed. Thus the nature of volcanic vents that have erupted ash flows in which welding has occurred has long been a subject of speculation. Lacking direct observa- SOURCE VENTS FOR ASH—FLOW TUFFS 17 tion, analogy must be drawn among the observed eruptions of nonwelded ash flows, pumice flows, and other types of deposits formed by certain historic outbursts of volcanoes such as Pelée, Soufriere, Komagatake, Merapi, and more recently Lamington and Hibok-Hibok. These are the nuée ardente, Pelean, ladu, glowing avalanche, and other deposits of various writers. The vents that gave rise to these historic erup- tions are immediately divisible into two types, those characterized by open craters, and those superim- posed on volcanic domes. Eruptions from domes seem to follow more than one pattern. The specific characteristics and origins of these eruptions have been a subject for controversy, and they have been . discussed and classified by several authors: Lacroix (1930), Escher (1933a, b), Kemmerling (1932), Grandjean (1931), MacGregor (1946, 1952), van Bemmelen (1949), and Weyl (1954b). A lengthy discussion of these differences is not pertinent to the main topic of this report. Briefly, they concern the directed-blast hypothesis of La- croix, the simple, as well as explosive, avalanching of the sides of the volcanic dome, and eruptions from fissures and subsidiary craters in the dome. These processes all seem to produce the so-called nuée ardente phenomenon, but the pyroclastic prod- ucts vary greatly in size range and composition. The products of greatest significance to students of ash-flow tuffs are those that consist primarily of new magma. This condition is met in certain erup— tions from the domes and, to a greater degree, in eruptions from open craters. Deposits formed from dome eruptions seem to be restricted in volume and distribution, the latter being dependent on the loca- tion of the vent or vents. Eruptions from open craters, being less restricted, tend to produce de- posits with a more symmetrical distribution around the crater, with probably greater volume and greater areal extent. As mentioned above, eruptions of both types have been observed and others have taken place in his- toric time, but in no known instance has either type produced a welded ash-flow tuff. Therefore vents that produced deposits of welded ash-flow tuf‘f re- main to be discussed. Small deposits of welded tuff (covering a few square miles) are probably not uncommon and they may have been derived from vents such as were dis- cussed above. A small deposit of welded rhyolite tuff (less than 10 square miles) in the Valles Moun- tains, N.Mex., probably came from a small crater that was later filled and covered by a large rhyolite flow. However, a dominant proportion of welded tuffs are found in ash-flow sheets of great areal extent. The great extent and tremendous volume of some of these deposits are incompatible with erup- tions from single domes or single craters. A third type of vent, the fissure type, offers an adequate explanation. Fissure vents have been assumed by van Bemmelen (1949), Westerveld (1952), Mac- Gregor (1952), and others. The field evidence for fissures as a source for ash-flow tuffs is meager and indirect evidence is most suggestive. Williams (1952, p. 155) states, “There is no way of telling whether the voluminous Costa Rican ava— lanches issued from the summit vents of Poas and Barba or from fissures on the flanks.” Wilbur Burbank, of the US. Geological Survey, believes the rhyolite dikes he studied in the In- dependence Pass region of Colorado were feeders for the extensive welded tuffs closely associated with them. A suite of these rocks made available by Burbank to the authors was, studied, and a thin section cut from one of the dikes is illustrated in figure 37. This shows intricate flow structures sweeping around the abundant phenocrysts. There is no sign of vesiculation or of the collapse of bubble fractures. Other evidence of the existence of fissure feeders is less direct, but still convincing. This indirect evidence is found in the large collapse structures so commonly associated with extensive pyroclastic deposits, especially ash—flow tuffs. The large size of some of these deposits precludes their origin from a single vent; they were probably erupted from a series of vents or from fissures. The linear charac- teristics of some of these collapse structures clearly indicate control by regional faults or rift zones and strongly suggest that the ash flows were erupted along these linear zones of weakness. The relation between voluminous pyroclastic erup- tions and certain collapse calderas has been dis- cussed by Williams (1941a) and van Bemmelen (1939, 1949). Some welded-tuff deposits, especially the more andesitic ones, are traceable to definite volcanic centers, but here their specific point of emission is generally obscured by a collapse caldera. Where the caldera is smaller the eruptions probably came from a single crater, Whereas in a larger more com- plex caldera the eruptions came from a group of craters or fissures, possibly ring fractures. In the 18 ASH—FLOW TUFFS Valles Mountains, N.Mex., the welded tuffs came from an area now occupied by a large caldera 14 miles in diameter. Within this caldera there is a ring of rhyolitic centers about 8 miles in diameter that formed after the caldera. This arrangement of centers within the area of caldera collapse surely represents a ring fracture and the eruptions that produced the welded-tuff deposits may have been localized along the same fracture. A similar origin from a ring dike was postulated for the rhyolitic pyroclastic rocks on the Island of Mull (Richey and Thomas, 1930, p. 370). In the Western United States and Mexico, es- pecially in the Basin and Range province, there are vast deposits of welded tuffs that have not been correlated with any known centers. These deposits occupy very large areas on the ranges but are cov- ered with immense thicknesses of fill in the inter- vening trenches, and offer many challenging prob- lems. RECOGNITION OF ASH-FLOW TUFFS The recognition of ash-flow tuffs has long pre- sented a difficult problem because of a lack of knowl- edge concerning their mode of deposition and the reliability and limitations of field and laboratory criteria. As has been discussed on page 13, a real understanding of the geologic relations of these rocks did not come until the early and middle 1930’s, resulting from the correlation of field and laboratory studies with information gathered from earlier ob- servations of significant volcanic eruptions. Al- though more and more geologists are becoming familiar with these rocks, and the actual existence of such rocks is recognized by most geologists, many workers are still confronted with problems of recog— nition. The discussions in the following pages to— gether with the illustrations may help clarify some, if not most, of these problems. Nonwelded ash-flow tuffs are often confused with tuffs of other origins and welded tuffs are often con- fused with lava flows. This is especially true of the devitrified tuffs. The ash flows of a more hetero- geneous nature may be confused with lahars and transitions may exist between one and the other. Normal vitric tuffs are sometimes fused in contact with lava flows or shallow intrusions and these may be mistaken for welded ash-flow tuffs. The following section is divided into two main parts. The first part pertains to field character- istics or those features of ash—flow tuffs readily discernible in the field. The second part is a treat— ment of microscopic characteristics. The sections on field characteristics, microscopic characteristics, and the one on physical chemistry must all necessarily involve a certain amount of recapitulation, but an effort has been made to keep this at a minimum. FIELD CHARACTERISTICS PYROCLASTIC CHARACTER The pyroclastic nature of ash-flow tuffs is gen- erally discernible in the field. In unaltered non- welded tuffs there is no problem because the presence of unconsolidated pumice fragments, ash, volcanic dust, and the ubiquitous foreign rock frag- ments would indicate pyroclastic origin for any tuff. These features are discernible through-milder types of induration and alteration, and become 0b- scure only when advanced stages of welding and devitrification have been reached. However, even in advanced and extreme examples of welding and devitrification a careful search will generally reveal rock fragments which, even if not diagnostic in themselves, should cast doubt on the rock being a lava, and together with other features to be dis- cussed, should give the correct answer. The most important single criterion for recogni- tion of the pyroclastic nature of ash—flow tuffs in the field seems to be the presence of pumice frag— ments. These pumice fragments occur in nearly all ash—flow tuffs and commonly persist in some form through extreme conditions of welding and vapor-phase mineralization, and are generally 0b- scured only when extreme devitrification follows extreme welding. In very fine grained tuffs the pumice fragments are too small or too few to be of much aid in field studies and other field or micro— scopic criteria must be used. A few crystal-rich tuffs also show few pumice fragments. Success in the use of pumice fragments as a criterion for pyro— clastic character of a welded tuff depends on a familiarity with the extremely diverse appearance of pumice fragments under different conditions of welding, devitrification, and vapor-phase crystalliza- tion. This knowledge comes with the study of many diverse types of welded tuff. In some localities all the above mentioned condi- tions may be observed in a single rock outcrop be- cause of varying degrees of compaction, welding, and devitrification in different parts of a single ash RECOGNITION OF flow. Where this occurs the relation between seem- ingly different structures becomes obvious, and their dissimilarities become understandable. Ash flows of the type exemplified by Battleship Rock in the Valles Mountains, N.Mex., present an opportunity for studying pumice relations under varying conditions of welding and compaction. Here a single sheet is nonwelded at the base and top, whereas a central part shows marked welding. The character of the tufl" scarp is shown in figure 3; the highly pumiceous character of a hand speci— men from near the base of the unit is shown in fig- ure 42; and a thin section of the same material is shown in figure 43. A hand specimen of a thor— oughly welded part of the same unit is shown in figure 1. Here the pumice fragments have com- pletely collapsed, and welding has eliminated much of their internal structure. Were it not for a complete transition from typical pumice to a material resembling blebs of glass, the origins of the black lenses shown in figure 1 would not be recognizable. The characteristics of pumice that has undergone several modifications are dis- cussed in some detail on page 33. However, nearly all lenticular elements in a welded tufi‘ (un- metamorphosed) that are observable in the hand specimen or the outcrop, probably are, or were, pumice fragments. The one exception to this is in those rare welded tuffs that have developed litho- physae. Most lithophysae are nearly spherical, but FIGURE 3.—Scarp locally known as Battleship Rock in Canon de San Diego, 5 miles north of Jemez Springs, Valles Mountains, N.Mex., showing columnar structure characteristic of many ash-flow tufi's. A single sheet is represented here. The incoherent nonwelded top has been removed by erosion and the nonwelded base is concealed. Near the lowest visible part of the scarp (about one-third above the base) is a zone of maximum welding. The erratic jointing near the base was probably due to uncommon cooling conditions as this material descended a canyon, whose wall was not far to the left and so presented a lateral cooling surface. Material from the nonwelded base of this unit is illus- trated in figure 42, and the thoroughly welded central part in figure 1. ASH-FLOVV TUFFS 19 a few may be flattened and thus appear lenticular in cross section. DEGREE OF SORTING A principal characteristic of ash-flow tuffs is their common occurrence in thick units (tens of feet) of typically nonsorted or nonbedded materials. This characteristic is in direct contrast to ash-fall tuff de- posits of comparable thickness, in which pronounced bedding is nearly always present, as shown in fig— ure 4. Ash-flow tuffs commonly show a wide range in size and relative amounts of constituent materials. However, the dominant material is generally ash or fine ash size (<4 mm) although some types are composed predominantly of pumice lapilli or blocks of different sizes, and for these the term “pumice flow” may be more suitable than ash flow. Dust or fine ash-size material is nearly always present. All gradations exist between those deposits that consist primarily of ash and dust and those that are predominantly pumice lapilli or blocks, although the ash-size types seem to be the most common. In- cluded accidental rock fragments, that commonly add to the heterogeneous appearance of the ash-flow tufi's, may range from microscopic size to large boulders, but are most often 1 inch or less in diam— eter. They are generally present in amounts less than 5 percent of the whole, but the authors have observed as much as 20 percent in some ash-flow tufi's and this figure is probably much less than the maximum amount. FIGURE 4,—Tuff section on the eastern wall of Colle Canon, near its junc- tion with Peralta Canyon, Valles Mountains, N.Mex. The lower part of the section is composed of material with the distinct bedding character- istic of ash-fall tufi‘s. In the ash-flow materials above this bedding is absent. The tendency for poorly welded or nonwelded tufis to form conical erosion forms is shown in the light-colored ash flows. The dark- colored material above is also part of an ash flow. 20 ASH-FLOW TUFFS THICKNESS Some ash flows are only a few feet thick, but in general they are many feet thick. Thus Macdonald and Alcaraz (1956, p. 174—175), in discussing the 1951 eruptions of Hibok-Hibok, Philippine Islands, state that the avalanche materials were 100 to 150 feet thick on the higher slopes of the mountain, but thinned toward the coast, and at Baylao had an aver— age thickness of “60 centimeters.” In the Valles Mountains, N.Mex., single flows range from a few feet to 300 feet or more in thickness. Ash flows shown in figures 5 and 6 range from 50 to 200 feet in thickness. FIGURE 5.—Welded-tuff scarp in Canon Media Dia, Valles Mountains, N.Mex., showing columnar jointing in thoroughly welded tufts. The visible part of the scarp represents the densely welded part of an ash- flow unit about 300 feet thick. Unexposed parts, less thoroughly welded at the base, would add nearly 150 feet. Note level top. FIGURE 6.—Ash-flow tuflf scarp below the Puye Pueblo ruins, Valles Moun- tains, N.Mex. The lower bench represents the thoroughly welded part of an ash flow; the debris-covered slope below represents the poorly welded or nonwelded part of the same flow. The upper cliff is the welded part of a later ash flow, and below it another poorly welded part of that flow. Columnar jointing near the top and middle of the section. Gilbert (1938, p. 1849) states, “In this section, the Bishop tuff is composed of several members. Some are less than 100 feet in thickness, while others reach nearly 200 feet in thickness.” In dis- cussing the thickness of the tuff deposit in the Valley of Ten Thousand Smokes, Fenner (1923, p. 33) says, “* * * it seems indeed quite probable that in the upper valley the tuff may attain a depth of several hundred feet in places.” The base is nowhere exposed in this area. Marshall (1935, p. 6) finds that the New Zealand ignimbrites range from about 60 to about 500 feet in thickness, but states, “* * * in general it appears that formations of the ignimbrites are not often thicker than 100 feet.” Of those welded-tuff deposits visited by the authors, single units 200 to 300 feet thick are common and units 500 feet thick are probably not rare. In Yel- lowstone National Park, and the San Juan Moun- tains of Colorado, greater thicknesses can be found. The thickness of an ash flow depends on the volume of material erupted and the type of topography over which it is emplaced; over gentle surfaces it will tend to spread laterally and form thinner units. If confined to canyons they Will flow farther from the source than those flowing over a plain or plateau. If confined to a topographic basin the ash flows will no doubt be thicker. The foregoing statements assume that the initial volume is the same in each situation. There is no reason to doubt that units more than 1,000 feet thick may be found. LAYERING Different zones of single flow units have com‘ monly undergone various degrees of consolidation, some parts remaining unconsolidated, others slightly or thoroughly welded. Columnar structures have formed in some zones and very often these zones show different colors or different shades of the same color, with shades of brown and gray occur- ring most frequently. These features tend to give a layered appearance to many ash—flow tuif units, which is commonly mistaken for the bedding of several flows. This effect of “layering” is further accentuated by weathering in fresh rocks and, no doubt, by metamorphism in older rocks. Many tuffs show a case—hardening effect on the surface, due no doubt to the release of silica from the ash and its redeposition as opal or chalcedony by evaporation at the surface. This thin siliceous coating, having erratic distribution, causes inequalities in the erosion patterns, especially in the nonwelded parts of the ash flows. These are vulnerable to wind erosion that pockmarks their surfaces, commonly RECOGNITION OF ASH-FLOW TUFFS 21 with deep holes. These pitted nonwelded tuffs con- trast sharply with the more firmly welded parts that react differently to wind and other erosive agents. The net effect is to increase the illusion of layering. In many ash flows, especially those in which the welded zone is in the central part of the flow, the transition from soft rock to hard rock commonly results in the formation of benches or ledges along this zone. These benches are often covered with rubble that accumulates from falling blocks. The harder zone commonly forms vertical cliffs because it has developed a more systematic pattern of pris- matic joints than the softer underlying material, as shown in figure 6. Also, these softer zones may support vegetation; thus it is easily seen why these contrasting types of material have been mistaken for separate flow units and even separate rock types. They commonly have a different mineralogy, color, texture, joint pattern, and erosion pattern; in short, they look different and are so interpreted. The transition zone may be surprisingly narrow and only a careful examination reveals that it really is a transition and not a contact. In the immediate vicinity of volcanic vents some ash- and pumice-fall tufi's and breccias may tend to form thick beds of a heterogeneous or nonsorted nature, and these may be confused with tuffs of ash-flow origin. However, if it is possible to trace these beds laterally from their vents, their true nature will become manifest with the gradual ap- pearance of graded bedding. The distance from the vent at which bedding will appear will depend on many factors and so it is not possible to give an arbitrary figure. However, within the limits of the authors’ experience, bedding will be clearly shown within a few miles from a vent and more often much closer. Ash-flow tuffs have rarely been traced directly to a specific vent so that only under uncommon circumstances should vent tuffs be con- fused with them. Volcanic mudfiows (lahars) and ash flows have not always been distinguished. However, lahars are typically nonsorted, but tend to be more heterogene- ous in composition. They commonly contain a high percentage of rock fragments of boulder and cobble size and, although they may be emplaced in a hot condition, they rarely show any lasting effects of heat. On the contrary, ash-flow tuffs commonly show some indication of having been hot, such as devitrification, vapor-phase minerals, or welding and compaction of pumice shards. In general, ash-flow tuffs occur typically without bedding within a single flow unit. However, as in all other geological phenomena there are exceptions to the rule. Many ash-flow units are underlain by ash— or pumice-fall beds. Normally there is a clear contact between the fall tuffs and the flow tufi‘s, but in others the fall and the flow are transitional (fig. 4). The bedding in the ash or pumice becomes more and more obscure upward until it merges im- perceptibly with the nonsorted flow tuff above. Probably the best explanation for this gradual transition is that the nonsorted tuff immediately above and transitional with the bedded tuff is also ash-fall material that was deposited in such large volume by eruptions of increasing intensity that gravity sorting was inhibited. The ash fall was then followed by ash flows of the same composition and the contacts between the nonsorted fall tuff and the flow tuff are not discernible. Where unconsoli- dated material is overridden by an ash flow, there may be an incorporation of ash-fall material in the ash flow. Some ash flows incorporate rubble of different sorts from the terrain over which they travel. This material, especially when large boulders are present, may be concentrated in lenses or linear zones within the ash flow. These are commonly near the base, but may occur well up in the flow and give an im- pression of bedding. A real sorting can take place near the distal ends of ash flows and this commonly results in the partial separation of pumice fragments and blocks. The coarser pumice seems to concentrate on the tops, bottoms, and sides of flow units. In the Valles Mountains, alternating layers, in single flow units, of thin pumice zones and typical nonsorted ash-flow tuff have been attributed to the overlapping of lobes or surges of an ash flow near its distal end. The entire flow near the end of its course was probably separating pumice and perhaps crystals. Thus lobes of the flow, having top and bottom concentrations of pumice and overriding one another, would leave a final product showing a distinct bedding. This bed- ding is traceable in some localities for several miles back from the ends of the flows and gradually merges with typical nonsorted ash flows. Probably a great many ash flows would show these features. However, as the ends of the flows are in general thinner and have lost more heat during emplacement than have the thicker parts of the ash sheet, they show less intense welding and vapor-phase crystallization or none at all and are 22 ASH-FLOW TUFFS more readily eroded, and thus are less often pre- served in older deposits. A significant paper on sorting during flowage is by Kuno (1941) who .contrasts the distinctive fea— tures of “pumice flow” and “ejected pumice” de- posits, especially those formed by the 1929 eruptions of Komagatake, Hokkaido, Japan. These “pumice flow” deposits show a definite lenticular banding caused by accumulations of larger pumice frag- ments into longitudinal ridges at the tops of, and within the flows, and formed parallel to the direc- tion of flow. Kuno (1941, p. 147) attributes these to “* * * differential movements of a number of streams within each pumice flow, overlapping one another or merging the one into the other.” He further states, “Obviously, the differential move- ments are the result of differences in the grain-size of the components that formed the pumice flows.” The pumice flows of Komagatake described by Kuno are small compared to many prehistoric ash flows. The Osidasi-zawa pumice flow ends 4 kilome- ters from the crater and reaches a maximum thick- ness of only 4 meters. However, these pumice flows illustrate a process that could well be expected in any pyroclastic deposit of similar origin. Any differences would be of degree rather than kind and would depend on the expectable variables of size range, thickness of flow, terrain, distance from source, speed of flow, and probably others. Another form of overlapping of parts of the same ash flow may take place in terrains characterized by canyons. Ash erupted from a high mountain source may flow dOWn adjoining canyons of different lengths. The flow traveling down the shorter can- yon reaches its destination first and is overridden minutes later by the flow in the longer canyon. This process can give a compound effect in the final product and may make proper interpretation of the welding and vapor—phase crystallization processes difficult. Such a process can be demonstrated in the Valles Mountains and is probably not uncommon where ash flows have been erupted in mountainous regions where the valleys or canyons open on a common plain or plateau. In many ash flows probably some lateral sorting occurs, although there is generally no apparent vari- ation except thinning over long distances. In some deposits the size of shard and pumice fragments decreases away from the source area, but this is less conspicuous than might be expected. The less- ening momentum of an ash flow near its distal end may result in a greater accumulation of foreign materials near the base. Similarly crystals may concentrate near the base of units; this process has been suggested by some writers to explain crystal accumulations at the base of some ash—flow tuffs. An alternate explanation for some basal accumula- tion of crystals is well illustrated by an occurrence of tuffs in Media Dia Canyon in the southwestern slopes of the Valles Mountains, N.Mex. Here in an SOD-foot section of ash flows, one flow has a non- welded base and overlies another flow from which a nonwelded top has been stripped (locally) by erosion. Several feet of tuff, showing a heavy con- centration of crystals in a matrix of ash and pumice, occurs at the base of the upper flow. This mixture of crystals, ash and pumice, contains pumice typical of the upper flow and crystals derived from the under- lying flow as is clearly shown by the presence of large quantities of chatoyant sanidine crystals which do not occur in the upper flow but are characteristic of the lower flow. Weathering of the tuffs in the Valles Mountains commonly produces a pavement of ‘ crystal fragments on the tops of the exposed sur- faces. These accumulations have locally been in- corporated in subsequent ash flows. The contact between 2 ash-flow tuff units is not always easily discernible, especially if the materials of the 2 flows are similar. In the specific occurrence described above the disconformity was marked by several angular blocks of dense welded tuff that were clearly derived from the underlying unit and, no doubt, littered the surface of that unit prior to the deposi— tion of the upper ash flow. Without these blocks to mark it, the contact could easily have been missed even with good exposures. This would have been especially true had the base of the upper unit been welded. Comparable examples must be very com- mon in many occurrences of ash-flow tuff. AREAL EXTENT The areal extent of an ash-flow unit depends pri- marily on the volume of ash erupted and type of terrain over which it is emplaced. Many workers engaged in the study of welded tuffs have described their great lateral extent. In some of these areas the lack of detailed examination has failed to show whether similarity of character represents wide extent of groups of flows or of single flows. Some ash flows, however, have distinctive characteristics which make it possible to identify single flows that have spread for long distances. In the Valles Moun- tains a single flow characterized by uncommonly complete welding and abundant chatoyant sanidine feldspar phenocrysts is traceable for about 15 miles west from the eruptive center. In the same region RECOGNITION OF other flows are known to have traveled at least 20 miles from the center of eruption. Griggs (1922, p. 253), and Fenner (1937, p. 236) described the “sand flow” of the Valley of Ten Thousand Smokes as extending for 14 miles with a slope of 2 to 3 percent. According to Williams (1942) pumice flows from Mount Mazama reached a maximum dis- tance of 35 miles from their source. Matumoto (1943, p. 4) states that the so-called Aso lava ex— tends as far as 100 kilometers (62 miles) from the center of Aso caldera. Anderson and Russell (1939) have traced the Nomlaki tuff member of the Tuscan formation for a distance of about 50 miles. The Nomlaki tufl’ member is partly welded over a large part of this distance, which has allowed correlation of the Ter- tiary sedimentary section across the northern part of the Sacramento Valley in California. This study is very significant in pointing out the usefulness of ash—flow tuffs in the correlation of geologic sections. . These distances are probably not excessive for ash and pumice flows, and may be exceeded when some of the world’s great ash-flow sheets are studied in detail. Less is known about the maximum area] distri- bution of single ash-flow units, but the quaquaversal distribution around a crater or along a fissure, of material from a single eruption could conceivably cover several thousand square miles. The large ash-flow sheet deposits, such as those of New Zealand, Sumatra, Asia Minor, and others, cdver thousands of square miles and are made up of many flows from different centers and different eruptions. The uniformity of these tuff sheets over large areas and distances is an important criterion for their recognition. This uniformity in both welded and nonwelded types is not to be found in either ash-fall 'tufi‘s or flow rocks of silicic composition and are rarely found in flow rocks of intermediate types. GENTLE' DIPS Ash flows tend to have very even upper surfaces and very low angles of dip except where deforma- tion has modified their original attitude. When erupted upon uneven topography the ash flows show evidence of having flowed around obstacles and down drainage channels. They may therefore have a very uneven base and nearly level top in contrast to the blanketing of topography as with ash-fall tuifs (fig. 7). Neither do normal ash-flow tuffs ASH—FLOW TUFFS 23 show steep primary dips such as exhibited by many lava flows. The very even surface of the “sand flow” of the Valley of Ten Thousand Smokes has been described and illustrated by Griggs (1922, p. 257). He re- marks that a bicycle could have been ridden over much of the surface. In discussing the New Zea- land ignimbrites Marshall (1935, p. 4) states, “Physiographically it may be said that within this district the outcrops of the ignimbrite rocks can generally be recognized in the field, even before they are closely approached. Usually they have a surface that is approximately level; * * *.” The even upper surface of ash flows has also been described by Gil- bert (1938, p. 1837) for the Bishop tuff of Cali— fornia, and by Zavaritsky (1947, p. 11) for ash flows of Russian Armenia. They are well shown by the Crater Lake “pumice and scoria” flows and by the ash-flow tuff deposits of the Valles Moun- tains, N.Mex. (figs. 6 and 8). The nearly level tops have also been observed by the authors in other parts of New Mexico, Arizona, Utah, Nevada, Colo- rado, Oregon, and Wyoming. FIGURE 7.—Ash-flow tuff scarp of Capulin Canyon, Valles Mountains, N.Mex. Three ash flow units are shown in this figure, the upper two are welded wits, and the lower light-colored one is a nonwelded tufi'. The columnar structure commonly developed in welded tufi‘s is well shown in the two upper beds. The level surface is characteristic of the top of an ash flow. In deformed rocks an evenness of surface may be reflected in the continuity of outcrop over long dis- tances or persistent recurrence of the same bed in fault blocks. ‘ This even upper surface is comple- mentary to the lack of a scoriaceous surface that Marshall (1935, p. 7) cites as an important differ- 24 ASH-FLOW TUFFS FIGURE 8.7Ash-flow tuff scarp, Valles Mountains, N.Mex.; illustrates the level surface of an ash flow and even bedding in a series of ash flows. The dark upper part is thoroughly welded. The light-colored part is The peak in the background is a quartz latite volcano, formed before deposition of the tuff. nonwelded. ence between the New Zealand ignimbrites and true rhyolitic lava flows. WELDING AND DEFORMATION 0F PUMICE Many ash flows contain a zone of welding which is easily distinguished in the field. With others it may be very difficult or impossible to detect welding from field examination alone. The boundary be- tween welded and nonwelded material is always a transitional zone of indeterminate character and the point at which incipient welding begins cannot be located with any high degree of accuracy, even though the transition may be only a few feet thick. The welded zone may occur at the base of the flow or occupy a zone within the flow unit and show transitions above and below into unconsolidated nonwelded tuflfs. The position of the zone of weld- ing depends on several factors that are discussed on page 47. All ash flows that completed their cooling history without burial by other ash flows, or other materials, probably had nonwelded tops which may or may not have been removed by erosion. How- ever, in some areas the emplacement of flows in rapid succession results in a complex of flows welded in their entirety, especially near the source areas. Compaction and flattening of the glassy com- ponents normally accompanies welding, although specimens obtained in the transition zones from welded to nonwelded material may at times show a high degree of welding without obvious deformation of the glassy parts (fig. 28). The most striking field characteristic that marks the transition from nonwelded to welded material is the change in appearance of the pumice frag- ments. Welding and compaction together, without devitrification and vapor-phase crystallization, cause the pumice to change in shape and color. Fresh glassy pumice normally ranges from white or gray to some shade of brown, depending on its chemical composition. However, during flattening and weld— ing the pumice darkens until in the strongly welded zones it becomes black and obsidianlike. These black glass lenses have heretofore been commonly mistaken for, and referred to as “clots” or “shreds” or “pasty glass fragments.” They are the “fiamme” of the Italian piperno (Zambonini, 1919, p. 66—68; Zavaritsky, 1947, p. 12), and the “obsidian spindles” 0f the “Aso lavas” (Matumoto, 1943, p. 7). It is this flattening of pumice fragments and other con- stituents by load compaction that imparts a foliate (eutaxitic) structure to most welded tuff's. Viewed normal to the plane of foliation the collapsed pumice fragments appear as discs or irregularly shaped flattened plates; viewed parallel to the plane of foliation they are lenticular. The difference in the two planes is illustrated in figures 83 and 84. Extreme conditions of welding may produce a zone of dense obsidianlike glass. This may be homogeneous or it may range into a “porphyritic” glass with a high percentage of crystals or rock fragments, depending on the makeup of the magma at the time of its eruption. Examination with a hand lens of glasses formed by extreme welding will usually reveal their fragmental nature. Even in those rocks with the outward appearance of a normal, nearly black obsidian, the collapsed pumice is commonly a more intense black than the material comprising the groundmass shards. This contrast is heightened by wetting. The welded tqu" from southeastern Idaho, illustrated in figure 9, would undoubtedly be mistaken for a normal obsidian. However, where water or oil is applied to a ground or otherwise smooth surface, examination under a low-power lens reveals a fine-grained ghostlike structure that represents dim, but definitely recog- nizable, slightly flattened lenses a fraction of a millimeter in length. They show a foliation in one plane, but no semblance of a flow structure is re- vealed. That is, the rock dimly retains a very fine grained fragmental structure that is more clearly revealed by microscopic study (fig. 10). The welded zones commonly differ in appearance from the nonwelded parts because of differences in color, density, porosity, jointing, and the secondary effects of weathering and erosion. These differences give the appearance of gross bedding within a single RECOGNITION flow, and frequently lead to a misinterpretation of the number of flows involved. This is especially true in areas where exposures are poor and the rocks have been deformed. In most of the glassy welded tuffs studied by the authors, welding to a compact glass has not taken place. Instead, the FIGURE 9.—Photograph of a hand specimen of welded tuflf with conchoidal fracture, and superficially resembling a typical obsidian. However, a polished or smooth wet surface shows a very fine grained, but recog- nizable, pyroclastic structure when examined under a low-power micro- scope. Collected by Ross and Smith from the Ammon quadrangle, southeastern Idaho. X 1. FIGURE 10.——Photomicrograph of thin section of glassy welded tuft illus- trated in figure 9. Very complete welding has eliminated all pore space and the different shards are molded one against another with moderate compression. The shard structure is well retained. Compare the dis— tortion with that shown in figure 36. OF ASH-FLOW TUFFS 25 rock has a blotched or streaky appearance due to the contrast in color between the darker collapsed glassy pumice fragments and the lighter matrix of glass shards and crystals (fig. 1). This type of glassy welded tufl’ is the most common and easiest to recognize. The size of the pumice fragments varies greatly from locality to locality as well as within the same rock specimen. In most welded tuffs, load compaction and flatten- ing of the shard structure occurs without complete elimination of all pore space, but in some tuffs all pore space has been eliminated, and distortion may stop here, or the pumice and more rarely the shards, may be further squeezed and distorted. When this happens the pumice and fragments will show as greatly flattened and enlarged discs in the plane of foliation and as greatly elongated lenses in the other plane. The result is a disc or plate with a vertical dimension many times smaller than the other twO dimensions. Most pumice fragments have primary shapes that range from those that are virtually equidimensional to those that may have their short dimension 2 or 3 times smaller than their long dimension. During emplacement of the ash flow there will be a tendency for the pumice lapilli or blocks to orient with their short dimensions in the vertical plane. These pumice fragments may be compressed and pore space eliminated during welding by a reduction of vertical dimensions of as much as an order of 2 to 6 times. Deformation during welding added to an original flat shape may give an end product that will have a vertical dimension perhaps 3 to 20 times smaller than the dimensions in the horizontal plane. Collapsed pumice fragments having a discordance in dimension of this order should be considered normal and are probably the result of simple elimi- nation of pore space. However, some welded tuffs are characterized by pumice fragments that have vertical dimensions ranging from 20 to 60 times smaller than their other dimensions. These pumice fragments have been flattened beyond the point of simple loss of pore space, and there has been a stretching of the structure, which may be mistaken for the flow lines of a lava flow (figs. 40, 41). The deformation or flattening of pumice will re- sult in a higher concentration of crystals per unit volume in welded zones than in nonwelded zones. In tuffs with a low initial concentration of crystals the difference in apparent content of crystals be- tween nonwelded and welded zones will not be obvious, but in those tufl’s whose initial crystal con- centration was high the difference may be striking. 26 ASH-FLOW TUFFS The ratio of pumice fragments to shards will have a significant influence on the amount of crystal en- richment per unit volume of tuff in welded zones, because of the difference in porosity between the pumice fragments and packed shards. Thus other factors being equal, the ash flow with the highest volume percentage of pumice will have the highest potential for crystal concentration per unit volume during welding. Some densely welded tuffs have been found to occur in very thin units of a few feet to a few tens of feet in thickness. No detailed studies of such rocks have been published. Probably some very thin tuffs are products of fusion and are more prop- erly classified with the fused tuffs discussed on page 26. These thin tuffs are sometimes found at the base of a normal welded—tufi unit that has a strongly welded bottom where there was enough heat to cause fusion of a few inches or a few feet of underlying glassy ash or pumice. This material may be either of ash-fall or ash—flow origin. The prime requisite is, of course, fragmented glass. A 2—foot thick bed of this origin has been briefly described by Enlows (1955). This process of fusion of underlying ash could lead, in some instances, to misinterpretation of the structure of an ash-flow tuif deposit. As already discussed some ash-flow units have a truly nonwelded top and bottom with the zone of maxi- mum welding somewhere above the base. Fusion of the top of a nonwelded unit by the overlying welded base of another flow of the same general appearance could produce a zone of dual origin that would appear to be a single flow with a nonwelded top and bottom. This process, together with the added effects of devitrification and vapor-phase ac- tivity, could produce some real complications in a sequence of ash-flow tuffs. A second type of thin welded-tuff unit is of true ash-flow origin; the only reasonable explanation for flows of this type seems to be that they were em- placed at temperatures high enough to induce com— plete welding without load being an important fac- tor in the elimination of pore space. Some examples of this type of welded tuff form dense obsidianlike glass (fig. 10) without showing the high degree of flattening of shards as is commonly shown in the basal glass of thick welded-ash flows. Other ex- amples develop lithophysae and spherulites in the glass and still others develop gas pockets in pore spaces where welding has not been complete. Ex- amples of the latter type are usually devitrified. The complete lack, or rarity of phenocrysts is a further suggestion of a high temperature of em- placement. FUSED TUFFS CONTRASTED WITH WELDED TUFFS Rocks identical with or closely resembling welded tuffs and of the same composition may be formed by fusion under conditions quite different from the process which is known as welding. This fusion may occur: (a) at the contacts between vitric tuff and rhyolitic domes or dikes; (b) in the granulated breccias at the base of rhyolitic flows; and (c) in the brecciated zones in the glassy parts of flows and domes. The first process is common in areas of intense rhyolitic or dacitic volcanism. The second and third are comparatively unimportant and their mode of derivation obvious. DEVITRIFICATTON AND VAPOR-PIIASE MINERALS The postdeposition mineral changes that com- monly affect ash-flow tuffs are devitrification and the development of vapor-phase minerals, which are best shown by microscopic studies (p. 36—38). Devitrification and vapor-phase mineralization are distinct episodes, even though they are related processes in the cooling history of many ash-flow tufl’s. The distinction made by the authors is that in devitrification the formation of crystals takes place within the boundaries of the glass shards or glass mass. In vapor—phase crystallization the for- mation of crystals takes place in open spaces under the influence of a vapor phase. Devitrification has obliterated the original glassy character in most welded tuffs, but in a few the pumice fragments have been devitrified, while the groundmass shards have remained glassy. Devitri- fication imposed on zones of intense welding and flattening of the structure results in a rock that closely resembles a flow lava. However, silicic flow rocks nearly always show flow banding. This means that if linear elements (exclusive of horizontal joint- ing) in unmetamorphosed rock extend over several feet, the rock almost surely is not a welded tuff. An unusual exception is a pumice-block flow or tuff con- taining large pumice blocks that have undergone extreme flattening and stretching. Even in these exceptional occurrences the linear elements are lens- like without great continuity. Devitrification as previously defined simply turns pumice fragments into crystalline material, and the pumice structure may be preserved in its entirety. However, the texture of recrystallized pumice frag- ments is greatly influenced by the degree of welding, or the extent of flattening, as there is a tendency RECOGNITION OF ASH—FLOW TUFFS 27 for the pumice vesicle walls to coalesce during weld- ing. The vesicles of the pumice become smaller and fewer and this may result in the complete oblitera- tion of pumice structure. The formation of de- vitrification or vapor-phase minerals will tend to further obliterate pumice structure and only ghosts of this structure remain. Vapor-phase crystalliza- tion occurs in open spaces, and although it may happen in any of the pumice fragments, except where there has been complete collapse, it is most characteristically shown where there has been no collapse (fig. 11). Where vapor-phase crystals form, the pumice structure may be partly, or in some tuffs, wholly destroyed by the growth of dis- crete crystals and crystal aggregates. These show a coarser texture than the products of devitrifica- tion. Vapor-phase crystallization commonly results in cavities that are lined with or contain mesh works of crystals. The cavities may lack direct evidence of derivation from pumice areas, and may be mis- taken for the lithophysal cavities and the vapor- phase crystallization common to flow rocks. The presence of vapor-phase minerals in pyroclastic rocks is commonly a good criterion for ash-flow origin. Some ash flows have retained volatiles dur- ing welding, and these may build up vapor pressure so that cavities are developed, and in these typical lithophysal minerals will be found. Spherulites and lithophysae may form in some welded tuffs but they are less common than might be expected, perhaps because flow rocks tend to FIGURE 11.—~Photograph of ash-flow tufi‘ from 4 miles north of Thumb Ranger Station, Yellowstone National Park, collected by Ross and Smith. The dark areas represent pumice fragments that have devitrified without collapse and in which vapor-phase minerals have formed. Tridymite formed in the cell walls giving them a very fragile texture that is readily broken down by weathering; this gives the tut-f a deeply pitted surface. The groundmass is composed of typical shards. 552858a61—73 retain a larger proportion of volatiles than do ash- flow materials. True lithophysae do develop in some localities in great numbers (figs. 49 and 52) where they may so dominate the outcrop that all or nearly all the diagnostic features of the welded tufi‘ as observed in the field may be obliterated. Typical round lithophysae are easily recognized and there should be no question as to origin. Hollow and expanded Spherulites appear much the same as they do in normal lava flows and call for no addi- tional discussion. However, in some rare occur- rences of welded tuifs, lithophysae become flattened and thus more or less lenticular in cross section and resemble flattened miarolitic pumice fragments. The origin of these cavities may be difficult to determine, especially in older rocks or rocks that have under- gone alteration or deformation. In many welded tuffs where vapor-phase minerals have formed in moderately to strongly flattened pumice fragments, pumice structure is not readily detected in the lenticular cross sections of the flat— tened pumice fragments. However, if the rock is split in the plane of flattening (foliation) the pumice fragments appear as discs or irregularly shaped plates, and ghost pumice structures are nearly always discernible. The vapor-phase minerals and minerals formed by devitrification are discussed in detail on pages 36—38. However, some of these minerals may be recognized in hand specimen and outcrop with the aid of a hand lens and so may be briefly mentioned here. The products of devitrification (as used in this report) are preponderantly cristobalite and feldspar and are too fine grained to be recognizable in the field. The main products of vapor—phase crystallization are tridymite, alkalic feldspar, and / cristobalite. Cristobalite that occurs as tiny whiteU balls or crystal rosettes and alkalic feldspar pre- dominate in lithophysae, while tridymite and alkalic feldspar predominate in the crystallized pumice and other porous areas of partly welded and nonwelded ash-flow tuffs. When these minerals are found in pyroclastic rocks it is an excellent indication that the rocks were of ash-flow origin. Here again exceptions may be found, but only rarely in such rocks as the fused tuffs that were discussed on page 26 and in other pyroclastic rocks that have undergone vapor— phase alteration originating from sources outside the body of rock in question. Such occurrences are not common and should be readily resolved in the field. Some ash flows come to rest and remain in an entirely unconsolidated or nonwelded state. Others 28 ASH-FLOW TUFFS may, depending on a number of factors, be partly or wholly welded and still remain in a glassy state. Other ash flows may be Wholly or partly devitrified and may develop vapor-phase minerals within their porous parts or develop spherulites or lithophysae within their glass zones. Devitrification during cooling is usually accompanied by the formation of vapor-phase minerals in pore spaces in certain zones within the tuff unit. Sections of ash-flow tufl’ units that are welded to the base commonly have a basal glass zone. This may grade upward through a densely welded and devitrified zone to a less welded or nonwelded zone that is devitrified and contains vapor-phase min- erals. Sections of units that have an unconsolidated base grading upward into a densely welded zone, that in turn grades upward into a nonwelded top, may have 2 zones of vapor-phaSe minerals, 1 above and 1 below the densely welded zone which may be either completely or partly devitrified. In thick flow units of this type the pumice and shards at the top and bottom of the unit (above and below the upper and lower zones of vapor-phase minerals re- spectively) may be unaffected by devitrification and vapor-phase activity. In those ash-flow tuffs that have developed two vapor-phase zones the lower zone is normally thinner than the upper. This fact is the result of the vapors having been squeezed out during compaction and welding and which tended to move upward, but were trapped below the densely welded zone and caused alterations similar or identi- cal to those above but within a more restricted area. Ash flows that have a high volatile content, but without the load or retained heat adequate to cause welding, may still devitrify and develop vapor-phase minerals. Such ash flows produce deposits as de- scribed by Fenner (1948a, p. 884) in the Arequipa region, Peru, and which he termed “sillar.” Thus we find the sillar-type ash flow forming complete units, or parts of units, that may also contain glassy or devitrified welded zones or both, as well as fresh unaltered pumice ‘and ash. Devitrified densely welded zones in many ash flows will be found to grade laterally into the sillar vapor-phase mineral zone. Examples of this type of transition are very common in the Valles Mountains, N.Mex. As shown in the foregoing discussion there may be a great deal of variety in individual ash-flow tuff deposits. When weathering, metamorphism, and deformation are superimposed on these rocks their complete history may be exceedingly difficult or im- possible to unravel. JOINTING Columnar jointing is a common feature of many welded tuffs. It has been described by Marshall (1935, p. 4) in the New Zealand tuffs; by Richards and Bryan (1934, p. 53) in the Brisbane tuff of Queensland, Australia; by Gilbert (1938, p. 1836) as being characteristic of the Bishop tuff of Cali- fornia; by Westerveld (1947, p. 19) in the Lake Toba region of North Sumatra; by Zavaritsky (1947, p. 9) in Russian Armenia, and by several other writers in more recent years. Fenner (1948a, pl. 2) presents an excellent illustration of jointing in the Arequipa region, Peru. Columnar structures are common and well de- veloped in the welded tufi's and in those parts of ash flows (sillar) indurated by vapor—phase minerals in the Valles Mountains, N.Mex. (figs. 5 and 7). Nor- mally they do not occur in the noncrystalline non- welded parts of the ash-flow units. No detailed studies of columnar structures in welded tuffs have been published, and there is little or no data on the details of spacing and attitude of joints or size of individual columns. Observations of the authors in the Valles Mountains indicate that joint spacings may range from a few inches to many feet. The more closely spaced joints are usually found in the zones of most intense welding. In any given locality the joints seem in general to be uniformly spaced, although the spacing may vary greatly over several miles of outcrop. Spacing is controlled by several factors such as rate of cooling, thickness, degree of welding, and others. Vertical jointing represents the conspicuous type, but departures from the vertical are not rare. Some welded tufl’s have developed fan jointing, while others have distorted vertical joints that give rise to bent or warped columns. These features are not common and are probably related to local deviations in the cooling surface. Fan jointing has been illus- trated by Richards and Bryan (1934, pl. 6, fig. 2) in the Brisbane tufl" of Australia. Fan jointing is well shown in the tuif scarp known as Battleship Rock in the Valles Mountains, N .Mex. (fig. 3). The local relations indicate that this tuff was emplaced against canyon walls, and the abnormal plane of cooling resulted in the fanlike group of curved columns. Bent columns are known from several localities in the Valles Mountains, N.Mex., and in some of these localities their relation to buried topographic highs seems well established, but in other localities their origin has not been definitely determined. Boden- RECOGNITION OF ASH—FLOW TUFFS 29 hausen (1955, p. 47, fig. 10) shows an excellent oc- currence of nearly horizontal columnar structure on the shore near the Island of Scandola, Corsica. An interesting type of jointing is represented by an obsidianlike welded tufl" from near Taxco, State of Guerrero, Mexico, which has been studied by Carl Fries, Jr. (written communication, 1958), of the US Geological Survey (figs. 12 and 13). This glassy layer is 10 to 20 meters thick and extends at least 10 kilometers. The jointing has produced a large proportion of plates about 1 inch in thick- ness, and some are as much as 2 feet square. FIGURE 12.—Photograph of a specimen collected by Carl Fries, Jr., US. Geological Survey, from a 10- to 20-meter-thick layer of welded tuif, at km 153 of Taxco Highway, State of Guerrero, Mexico. The photograph shows the obsidianlike character of this glassy welded tuif. At this locality jointing has resulted in a large proportion of platelike slabs that are commonly about 1 inch thick and cover as much as 2 square feet. For this reason the slabs are quarried and widely used in place of tile as flooring material in this part of Mexico. The joint surface is represented in the central part of the surface of this figure; the opposite joint surface is shown in figure 13. Many welded tuifs show a horizontal platy joint- ing in or near the zone of maximum compaction. This platy structure is accentuated by weathering and should not be confused with the platy structure that is commonly developed by weathering along the planes of foliation in zones of partial to com- plete welding and compaction, and where devitrifica- tion has furthered the inequalities of hardness in- herent in the eutaxitic structure of these rocks. This type of horizontal platy jointing has not been studied in detail but it seems to the authors that it marks the zone of maximum flattening; however, in a few rocks it is best developed a little below that i , 3cm? ., FIGURE 13.iThe reverse side of the specimen shown in figure 12 where there has been weathering along a joint crack that has brought out the eutaxitic structure. The lighter colored groundmass shows a fine- grained pryoclastic structure. The darker areas represent completely collapsed pumice fragments. The pumice character is shown by the fraylike ends of many of the pumice areas. zone. In some welded tuffs the vertical joints do not extend below this zone of horizontal joints. In others they continue downward through the horizon- tal joints to die out at the top of a nonwelded tuff that forms the base of the unit. Unlike the joint pattern seen in lava flows that often consists of the well known 5- or 6- (ideally) sided columns, many examples of welded tuffs have columns that are roughly to symmetrically rectangu- lar, and some are square. In general these are tensional cooling joints, but the reason for the wide occurrence of this pattern appearing so commonly in welded tuffs is not clear. It has been observed in many localities in the Western United States, and many illustrations in the literature of welded tuffs from other parts of the world indicate that the joint pattern exists elsewhere. However, Boden- hausen (1955, p. 47, fig. 11) has illustrated colum- nar jointing in welded tufi’s of Corsica that show closely spaced columns, a large proportion of which have 5 or 6 sides. EROSION FORMS The wide variations in texture and physical com- position, and different degrees of hardness of ash- flow tuffs will obviously produce a diversity of erosion forms. A complete discussion of these would fill many pages and does not seem justified here because of their questionable value as an aid 30 ASH-FLOW TUFFS in the recognition of welded tufi's. Locally a specific type of erosion form may aid in the correlation of flows, or may be very diagnostic in distinguishing ash-flow tuffs from rocks of other origins, but each locality must be evaluated by a study of its own peculiarities. A few specific examples of locally diagnostic forms should be cited. In the Valles Mountains, N.Mex., certain erosion forms are very charac- teristic and allow recognition of ash-flow tuffs at great distance. Foremost among these are conical shaped pinnacles (figs. 4, 14) that are locally known as tent rocks because of their resemblance to conical tents or tepees. There is a tendency for the formation of these in nearly all the nonwelded parts of the ash-flow tuif deposits but they are par- ticularly outstanding in two members of the Ban- delier formation. The Bandelier formation occupies an area of about 400 square miles and at least one of the “tent rock” members is present over most of this area. Some of the “tent rocks” are cones only a few feet high while others are 100 or more feet high (fig. 14). There are all gradations in size and of varying degrees of perfection in symmetry. Some cones are capped by large boulders that were rafted from the ground surface traversed by the ash flow during its emplacement. Some of these boulders weighing as much as several hundred pounds tend to retard erosion and produce boulder- capped pinnacles. Not rarely this results in whole colonies of pinnacles capped by boulders of varying s1ze. Another diagnostic feature in the Valles Moun- tains ash-flow tuifs is a magnified “swiss cheese” effect produced by wind erosion. This is common in zones of porous poorly weld‘éd tuff (fig. 14) and nonwelded tuif that has been subjected to vapor- phase activity, and to a lesser extent in the un- altered nonwelded zones. In the porous poorly welded zones where vapor-phase minerals have formed these minerals tend to be weakly coherent and thus they are very easily broken and vulnerable to the agents of erosion. This is especially true where these minerals occupy former pumice frag- ments. Here erosion forms pits that are further enlarged until some reach cavelike proportions. Many outcrops of ash-flow tuffs in the Valles Moun- tains have casehardened surfaces. Erosion may start at inequalities in these surfaces especially where pumice fragments occur. This gives wind and rain access to the poorly consolidated interior, and conspicuous cavities develop. Some of these cavities were enlarged and used as living quarters FIGURE 14.—Scarp in Wildcat Canyon, 11/; miles southwest of La Cueva, Valles Mountains, N.Mex. This represents the slightly welded to non- welded part of a single ash flow, about 400 feet thick. The tufi' at the top of the scarp is poorly welded, and an unknown amount of non- welded tufE has been eroded from the surface. In nonwelded and poorly welded tui‘fs, erosion localized along vertical joints tends to produce conical pinnacles. by the early Pueblo Indian inhabitants of the region. In a recent paper on the welded tuffs of Chiri- cahua National Monument, Ariz., Enlows (1955, p. 1229) says, “The unusual and spectacular rock for- mations for which the Monument is famous * * * re- sult from weathering guided by well-developed, ver- tical joints and the coarse eutaxitic structure.” This type of weathering, which is controlled by the foliate or eutaxitic structures of welded tuifs, is well illustrated in Enlows’ paper and is mentioned here because it is a common feature of many welded tuffs. Weathered eutaxitic foliation may appear in some welded tuffs like bedding in sedimentary or ash-fall tufi‘s or like flow banding in some lavas, and may make field determinations of some welded tuffs open to question. A few hard compact devitrified crystal-rich welded tuifs may superficially resemble intrusive rocks. These welded tuifs are usually dacitic or quartz latitic in composition. One occurrence seen by the authors in southern Nevada contained a large amount of biotite and resembled a granitic rock, even to weathering by exfoliation into rounded knobs typical of exfoliated granite. FOSSIL FUMAROLES Williams (1942, p. 86) in his study of the deposits of pumice and scoria flows surrounding Crater Lake, describes features which he concludes are “fossil fumaroles.” He says, MICROSCOPIC CHARACTERISTICS 31 Brown, pink, and white streaks cut the gray scoria where the gases rose to the surface. Some of the spires are hollow inside and have irregular openings at the top. The largest of these tubular spires is 8 feet across * * *. On Sand Creek, as many as 150 ‘fossil fumaroles’ may be counted in a distance of 1% miles along the canyon walls. Cementation of the materials surrounding these fumaroles takes place through the deposition of “iron oxides, kaolin, and opal.” Erosion of the deposits may leave the indurated fumaroles stand- ing as “spectacular columns and spires.” Similar features have not been reported from any other ancient ash-flow tuif deposit which seems odd. The nature of the ash-flow concept as it is under- stood at the present time implies that a deposit of this material would give off gases for a long time after its emplacement. Just how long may be seen from Kozu’s (1934) study of the pumice flows from the 1929 eruption of Komagatake and from the studies of the sand flow of the Valley of Ten Thou- sand Smokes. These deposits were still giving off vapors years after they were formed (see p. 41). These gases Should leave some trace of their pas- sage through the unconsolidated upper parts of an ash-flow deposit. In the Valles Mountains, N.Mex., the tops of ash-flow deposits are preserved in many localities, but examination of these by the authors has failed to reveal “fossil fumaroles” although some Show mottling. The explanation for the ap- parent absence of fumaroles is not clear. Because nearly all these deposits have a zone which contains vapor-phase minerals it is obvious that the vapors were there. Normally this zone does not extend to the top of the flow unit. The top still contains glassy ash and pumice and is, for the most part, still friable. It is in this part of the flow that the fumaroles Should occur. It is all the more peculiar that fumaroles are not apparent in the Valles Moun- tains deposits when it is considered that the vapor phase was more effective here in converting glass into crystalline materials than in the Crater Lake deposits. There was apparently very little devitri- fication and growth of vapor-phase minerals in the Crater Lake tuffs. This fact plus lack of welding would indicate a much lower temperature of em- placement and probably a lower volatile content than the Valles Mountains deposits. Williams found that the Signs of fumarolic ac- tivity disappeared about 10 miles from the summit of the former Mount Mazama. Thus distance from source might explain the absence of visible Signs of fumaroles in other tuff deposits, but this should only be true in those that Show no other evidence of vapor-phase activity. Many ash-flow tuff de- posits show Signs of this activity throughout their entire outcrop area, and this may be 20 miles or more. Fumaroles in more mafic ash flows should be more colorful because of the higher iron content yielding more iron oxide as a staining or cementing agent. If more colorful they would be more obvious. This might, in part, account for the differences between the Crater Lake flows and the Valles Mountains flows, aS the Valles Mountains ash-flow tuifs con- tain over 20 percent more Silica than the “scoria flows” of Crater Lake and about 10 percent more Silica than the dacitic pumice flows. Nearly all the Valles Mountains ash-flow tuffs contain less than 2 percent total iron, and some contain less than 11/2 percent, while the Crater Lake rocks range from 21/2 to more than 7 percent total iron. That is, the iron is inversely proportional to the Silica. How- ever, even if iron-oxide coloration was less obvious, fumarolic activity in the tops of the Valles Moun- tains ash flows should have caused induration be- cause of the release and redeposition of Silica and perhaps the development of clay minerals. The only explanation that seems to be left is that the differ- ence in age of the deposits may be great enough to have allowed changes in the appearance of the Valles Mountains ash-flow tuifs, obscuring features that were never too obvious in the'first place. This age difference is not very great as the Valles Mountains ash flows were formed in middle or late Pleistocene time and are so fresh in appearance that some of them could not be distinguished in hand Specimens from Katmai “sand flow” material. In summary, the “fossil fumaroles” can be rea- sonably expected to occur in the upper nonwelded parts of ash-flow tufi's, but they have not been re- ported from any prehistoric deposits of these rocks except the Crater Lake dacitic pumice and mafic scoria flows. Why they have not been found in relatively young deposits that were certainly poten- tial producers of such features is not clear. It may be that in rhyolitic ash-flow tulfs these fumaroles were never very obvious features and that only very careful scrutiny of prehistoric deposits will reveal their presence, if any trace remains. MICROSCOPIC CHARACTERISTICS The field characteristics of ash-flow tquS have been presented in the preceding section and are usually adequate for identification. The character- istics observable under the microscope are in gen- 32 ASH-FLOW TUFFS eral also an adequate means of identification, but some ash-flow tuffs are best studied by coordination of both methods. The results of microscopic studies will be presented in this section and coordinated with field studies. Characteristics observable under the microscope also present important information about genetic history and must be considered in connection with the following section on physical chemistry. The various physical"characteristics of ash-flow materials are shown in the illustrations, and discus- sion in the text will constantly refer to these. How- ever, to minimize so far as possible the need for readers turning from the illustrations to text, de- tailed descriptions have been prepared to accom- pany the figures. Each of the genetic episodes—formation of a glassy magma, vesiculation, eruption, explosive dis- ruption, transportation, emplacement, distortion, degrees of welding, devitrification, and vapor—phase mineralization—have imposed their distinctive ef- fects on the resulting tuffs. The geologist may be called upon to differentiate and evaluate these effects and, so far as possible, to determine the genetic history presented. The following section describes these characteristics in some detail, emphasizes those which seem most significant for purposes of identification, and discusses the situations where ambiguities may arise. PYROCLASTIC CHARACTER The origin of both ash-fall and ash-flow materials through the vesiculation and explosive disruption of glass has given these materials physical forms that are not greatly influenced by the chemical differ- ences in the magma from which they are formed. They vary Within certain limits in successive erup- tions, even from the same centers, but otherwise seem to be basically the same from region to region and have not varied through geologic time. The same shard forms seem to characterize both ash- flow and ash-fall tuffs. However, the subsequent history of these materials is so varied that distinc- tive differences are developed. The most important of these are the absence of sorting, and the presence of welding and devitrification in ash-flow tufi's. Tufl’ fragments vary in physical form, and these may, for convenience, be divided into types, but there are transitions from one type to another. One widely occurring type is derived from a vesicu- lated glass characterized by roughly globular bubbles as shown in figure 18. This material when explosively disrupted produces curved plates, as shown in figure 16, that represent fragments of the walls of these bubbles; cusp- and lune-shaped frag- ments as shown in figure 22; and difi‘erent forms representing the interstitial glass between several bubbles. These may be tricuspidate fragments bounded by arcs of circles, Y-shaped fragments formed in the same way and illustrated in figure 20, and which seem to be one of the most common char- acteristics of glassy tuff materials. More rarely, forms which represent cross sections of undisrupted spherical bubbles have been observed (figs. 18 and 29). The original plastic glass contained bubbles of different size, shape, and thickness of walls, and so there are many variations to the forms men- tioned. A slight modification of the type of ash derived from globular bubbles was the result of fracture of the walls of elongate bubbles which gave a large proportion of more nearly flat, but commonly slightly curved plates (figs. 17, 20, and 28), among the resulting shards. With these shards are U-shaped or much-elongated Y—shaped forms. The pumice type ordinarily occurs with other types, and more rarely characterizes an entire oc- currence. This material is made up of a fine ag- gregate of cells that are commonly tubular, but rarely are nearly circular. Pumice tuffs are illus- trated in figures 1, 11—13, and 42—48. . During any single eruption the chemical com- ' position of the parent magma and its gas content was about uniform; and at the time of vesiculation, , the temperature and the viscosity of the glass varied but little. For these reasons, eruption and vesicu- lation of the magma tended to produce shards show- ing marked uniformity during a single episode, ex- cept for associated dust and the pumice fragments to be discussed later. On the other hand, successive eruptions even from the same vent may differ greatly in the types of the shards produced. Some ash flows are made up of nearly uniformly sized shards, as illustrated in figures 28 and 71. More frequently there is a range in dimensions of a hundredfold or more. In other occurrences ash shards are enclosed in a matrix of dustlike materials as illustrated in figures 24 and 25. In some flows this dustlike material is no doubt due to attrition ' during fiowage. However, in some tuffs delicate cusplike shards and bubble walls have not been marred during flowage, even where enclosed in dustlike material as shown in figure 18. The trans- portation of the shards while enveloped in a film of gas would tend to isolate each shard from its neigh- bors, and reduce attrition to a minimum. More— MICROSCOPIC CHARACTERISTICS 33 over the shards were at a high temperature, and were still viscous, although viscosity was normally very high. For this reason the hot Shardswere very much less fragile than they would have been if cold. Rhyolitic glass has such a high viscosity that it may undergo either plastic yield or fracture, depending on the physical conditions. PUMICE FRAGMENTS The very common occurrence of pumice fragments as observable in the field has been described in the preceding section but studies under the microscope Show the presence of pumice fragments at least of ash Size in almost all ash-flow tuffs examined. Some flows are made up dominantly or wholly of pumice, and are best described as pumice flows. The char- acter and modifications of pumice are most clearly observable where it occurs as fragments large enough to be studied in the field, and so conclusions based on field studies need not be repeated here. However, some relations are clarified by microscopic studies and other significant ones are revealed. The pumice fragments which are normally as- sociated with shard materials, and which they greatly exceed in Size, are illustrated in figures 42— 48. Most commonly pumice fragments are made up of greatly elongated tubular pore spaces which give the specimen a fibrous structure, but more rarely they are made up of nearly equidimensional roughly spherical bubbles. Very porous pumice fragments are illustrated in figures 42—44. Pumice fragments in which there has been no collapse are shown in figures 42 and 43. The col- lapse of the cell structure in glassy pumice is illus- trated in figures 45—48, and in figure 13. Crumpling and collapse of pumice lying across the plane of compaction and pumice compressed against pheno- crysts are Shown in figures 47 and 48. There may also be collapse of the pumice and such thorough rewelding that all internal evidence of pumice struc- ture is lost as illustrated in figure 45. In the hand Specimen collapsed pumice may appear as dark blebs of glass as illustrated in figure 1, and Which have been discussed in the section on field character- istics. Pumice may lose its glassy character either by vapor-phase alteration or by devitrification (figs. 55—58). The difference in the Significance of these two processes have been discussed on page 44 and their effect on pumice fragments described. The effects of vapor-phase alteration are especially well Shown by the field relations. The occurrence of the vapor—phase minerals—feldspar and tridymite —iS shown in figure 66. Pumice fragments are subject to the same devitri- fication effects as are shards and the resulting minerals—feldspar and cristobalite—are the same. However, pumice fragments tend to be more readily devitrified, and to develop a much coarser grained aggregate of devitrification products than the as- sociated shard materials. In a few specimens the pumice has become devitrified while the Shards have remained glassy. The development of the coarse- grained aggregate commonly results in the complete destruction of structure within the pumice as illus- trated in figures 57, 58, 61, and 62. WELDING, DISTORTION, AND STRETCHING Welding and the accompanying distortion of pyro- clastic materials is one of the outstanding charac- teristics of ash-flow materials. Therefore, the structures formed during welding and criteria for their recognition, even after extreme distortion, present one of the outstanding problems in the study of ash flows. This includes the recognition of the progressive modifications of these structures as a result of increasing distortion. Welding and distortion can be observed both in the field and under the microscope for many occur- rences, but in others can be determined only under the microscope. Even where clearly observable in the field, relations are clarified by microscopic studies. All ash-flow materials are derived from glassy, vesiculated materials with a wide Size range, but .in general are dominantly of ash size, and all subse- quent changes which these materials may have un- dergone, have merely been imposed on these primary vesiculated glass structures. However great may be the effect of subsequent changes, they rarely completely obscure the pyroclastic origin of the materials. The genetic history of welded tuffs has given them very definite textural characteristics that are expressed in what has been termed “fab- ric.” Iddings (1909, p. 194) has defined fabric as follows: “The fabric, or pattern, of a rock as Shown on a surface or in a section, is that factor of its texture which is dependent on the relative sizes of its component crystals, [or glass fragments] on their shapes and arrangements with respect to one another.” This definition, as modified to include glassy materials, effectively outlines the microscopic characteristics which must be considered in a de- scription of welded tuffs. 34 ASH-FLOW TUFFS Welded tuffs usually have a foliate structure (eutaxitic) due to a parallel arrangement of origi- nally flattened plates, or to compaction and flatten- ing of glass shards and pumice fragments into more or less platelike units. The planar arrangement of these platelike units imparts a foliation which has, in many occurrences, been mistaken for lineation due to flowage. Welded tufi's may, however, show a marked lineation resulting from stretching and perhaps very rarely to incipient flowage of the welded, but still plastic material (figs. 40, 41, and 51). The fabrics developed during welding depend on: the primary form of the shards; the degree of their plasticity—a function of chemical composition and temperature; the compressive stresses acting upon them; and the rate of cooling. Fabric is shown most clearly in the welded tufl's that have remained glassy, but some of the devitrified tuffs also show very well the effects of distortion of the fabric. Varying degrees of distortion are shown in the frontispiece, figures 9, 10, and 24—41. In some ash-flow materials the shards are in con- tact with one another only at their points but With- out observable distortion. Commonly such tuffs have remained glassy, but more rarely even poorly consolidated tuffs have been devitrified as illustrated in figure 17. In some tufl's, especially poorly con- solidated ones, it may be difficult to determine whether incipient welding has preceded devitrifica- tion. However, many tufl’s show marked distortion of the structures that has preceded devitrification as illustrated in figures 51, 59, 60, and 68—70; in these, welding definitely preceded devitrification. The degree of distortion in welded tuffs ranges from those in which it is slight (figs. 20—29) to those in which flattening and stretching produced a close resemblance to flow rocks (figs. 30, 36, 40, 41, and 59). In many tuffs evidence of distortion may be observed only where the shards are molded against sharp corners of a phenocryst, or where they have been squeezed between two adjacent phenocrysts. In other occurrences there is slight distortion throughout the specimen, and very marked warping around phenocrysts (figs. 31 and 34). Most welded tuffs which have undergone com- paction show the greatest distortion perpendicular to the plane of deposition and much less distortion in those cut parallel to that plane, as illustrated in figures 68, 69 and 83 and 84. With extreme distortion welded tufl’s may come to resemble flow rocks. A welded tuff with marked stretching and distortion of the structure is shown in figure 36, and a flow rock with distorted flow lines in figure 37. The two illustrations showed marked similarity, but close comparisons show cer- tain dissimilarities, although these may not be great enough for definitely differentiating them. The discontinuity of the flow lines in the welded tuff (fig. 36), may be contrasted with the more con- tinuous flow lines as shown in figure 37. In the greatly flattened tufi' structures, groups of lines derived from distinct pumice fragments may butt against each other. or meet at an angle. Com- pressed Y- or U-shaped shards resemble flow lines which have been folded back upon themselves (figs. 51 and 94). Spherical bubbles have been com- pressed into very flat lenses as shown in figure 40. In some specimens the zones in which the original tufl‘ structure has become obscure will alternate with other zones in which the structure is clearly recog- nizable. J. Hoover Mackin studied and brought to the authors’ attention the welded tuff from near Iron Springs, southwestern Utah (figs. 40 and 41). The specimen shown in figure 41 has undergone such extreme stretching that it closely resembles a flow rock with lineation. The specimen shown in figure 40 is from the same ash flow, but from a locality where the stretching was not quite as thorough, and the ash structure, although extremely com- pressed is still discernible. The two are otherwise different in that the specimen represented in figure 41 is glassy whereas that in figure 40 has under- gone devitrification. Solovev (1950, p. 219) in his description of the “ignimbrites” of the Sikhote-alin volcanic region lying between the Amur River and the Sea of Japan, describes stretching of the tuff structures as follows: A mass of fragments settled during the movement and the particles—at least a certain portion of them—were com- pressed because of the high heat and also the pressure from the accumulating overlying pyroclastic material, and were welded in some places, especially in the lower horizons of the column. Such masses, existing in a viscous state, ac- quired in some localities a fluid character (and could, pos- sibly in some sections even become displaced somewhat to the sides under the weight of the overlying accumulation of fragments). Naturally, this condition could not have failed to be reflected in the structure of the rocks. In a number of places the pyroclastic character of the rocks was more or less completely efl"aced or even completely destroyed * * * a pyroclastic structure is clearly expressed at a number of points, whereas at other points it does not express itself, or its traces are completely destroyed. In the latter case, the main masses not infrequently acquire a fluid character * * * MICROSCOPIC CHARACTERISTICS 35 all transitions are present, so that these singular rocks have in some places the features of tufi‘s, and in others the features of lavas. PHENOCRYSTS AND FOREIGN MATERIALS Crystalline materials in ash-flow tuffs, other than the devitrification products, are of two types—those which represent phenocrysts which developed as a part of the genetic history of the rocks, and those rock and mineral fragments which are alien to it. Only very rarely have tuifs been observed which were almost free from phenocrysts, and very few which were without at least a few alien inclusions. The proportion of observed phenocrysts ranged from a fraction of 1 percent to nearly 60 percent. The obsidianlike welded tuff from southeastern Idaho illustrated in figures 9 and 10 contained only about 0.1 percent of phenocrysts, and the tuff from McKays, Calif, illustrated in figure 82, is uncom- monly rich in phenocrysts. The tuffs from Sumatra (fig. 26) are high in phenocrysts, whereas those from Peru (figs. 24, 25) are low. The rhyolitic welded tufl’s are always character- ized by feldspar phenocrysts, and commonly contain quartz. Biotite occurs in small amounts in many rhyolitic tuffs, but becomes abundant in some quartz latite, and rhyodacite tuffs. Magnetite is sparse but rarely, if ever, totally absent. Hornblende is present in a few tuffs. Augite is abundant in some andesitic tuffs, but is more rare in rhyolitic ones. The feldspars of the welded-tuff deposits of the Valles Mountains are almost wholly sanidine, and sanidine that shows a brilliant blue chatoyancy characterizes at least two welded-tuff units. Ortho— clase or microcline is rare and where present is no doubt alien, probably being derived from invaded rocks. The essential feldspar grains are sharply euhedral in outline in a few tuffs, but more commonly they are subhedral. Some Show one side with a crystal face, and elsewhere Show irregular or fractured edges. Other feldspar grains are rounded or ir— regularly embayed as shown in figures 19, 34, and 38. In some tuffs, as in most of those from Sumatra, there has been extreme fracturing, so that most of the grains are sharply angular in outline (fig. 26). Quartz phenocrysts may be sharply euhedral as shown in figure 76, but more often Show rounding and embayment as illustrated in figure 31. Anthills are abundant in tuff areas in the Valles Mountains, N.Mex., and in these, quartz grains with lesser amounts of sanidine feldspar are dominant com- ponents. The quartz grains are about 2 mm in 552858a—61—rgal diameter and are euhedral or partly euhedral in outline and all show evidence of the hexagonal sym— metry of high-temperature quartz. A great variety of materials derived from rocks through which the tuff—forming magmas made their way to the surface, or that were picked up from the surface by the overriding flow, have been observed by the authors in the course of studies of many hundreds of thin sections of welded tuffs from different regions of the world. Alien inclusions are rarely absent and include granite fragments, oc- casionally fragments of metamorphic rocks, sedi- mentary rock material, and volcanic minerals and rock fragments that clearly differ in origin from those genetically related to the tuff. Some of these materials also occur as rock fragments large enough to be recognizable in the field or in hand specimens, but those of large size and those observable only under the microscope seem not to differ in character. Most commonly these foreign materials are those of andesitic rocks; andesiticdminerals and rock frag— ments are so ubiquitous in tuffs from the United States and other countries that their absence seems to be a rare exception. Andesitic eruptions have commonly preceded explosive rhyolitic ones, a long- recognized, almost normal geologic sequence, and so an abundance of andesitic materials in welded tuffs is to be expected. In general these rock or mineral fragments are fresh and unaltered. Augite and rare biotite and hornblende are normally per- fectly fresh except where subjected to vapor-phase activity. In one specimen examined, olivine in an included basalt fragment was nearly fresh. Very rarely there seems to have been partial fusion of an included fragment, probably representing material picked up at uncommon depths. Many tuffs show a wide range in the proportion of anorthite in the plagioclase crystals, and the more calcic ones are alien materials. The origin of the alien plagioclase is evidently the same as the commonly associated andesite rock fragments. The more calcic grains have tended to be out of equi- librium with their host rock and suffered altera- tion or re-solution. Biotite when present may be partly altered with the development of magnetite along its borders. Hornblende may be of a normal greenish color, or not rarely red brown. In some rocks both types of hornblende occur. Another type of alien material represents rock fragments that have been picked up from the surface by the over- riding ash flow, and that are conspicuous near the base of some flows. Some of the smaller fragments are observable only under the microscope, but most 36 of them are so large that field studies show their relations best. ' DEVITRIFICATION A large proportion of ash—flow tuffs have under- gone devitrification, and this process has been super- imposed upon fabric resulting from the welding and distortion of the pyroclastic material. These succes- sive events modify, in many occurrences profoundly, the primary glass shard structures. However, evidence of origin is retained in most devitrified tuffs. The physical characteristics of devitrified tufi’s are closely related to the identity of minerals formed, therefore these will be discussed before further consideration of textures. The slender 3' parallel intergrowths of the devitrification products (axiolitic structures) vary greatly in size, but are normally too fine grained to be identifiable under the microscope. The shards illustrated in figure 28 are marked by light-colored borders which, un- der crossed nicols, show a very slight grayness, but no resolution of mineral aggregate, even with very high magnification. In nearly all specimens how— ever, the materials are of such a size that under the microscope they have a rough surface, indicat- ing a parallel aggregate of minerals of greatly differing indices of refraction, as shown in figures 71—74, or as in the spherulites illustrated in figures 63—65. The mineral aggregates within the shards can not be definitely identified by means of the microscope. However, many tests of devitrified glass shards and pumice fragments of rhyolitic ash- flow tufi's from many localities have been made by means of X-rays, and these have invariably shown that the products are cristobalite and feldspar. Before the development of X-ray diffraction methods for the study of very fine grained ma- terials, the identification of minerals of devitrifica- tion in volcanic glasses was difl'icult, or impossible. Fenner (1948a, p. 883) states that small pores be- tween the shards of tuffs from the Valley of Ten Thousand Smokes and from Yellowstone National Park contained tridymite and orthoclase, and that the same secondary minerals occur in tuffs from the Arequipa region of Peru. Fenner’s identification of the tridymite in cavities is clearly correct, but he did not specifically consider the devitrification ma- terial in dense glass. Marshall (1935, p. 23, 24) described the develop- ment of very slender needlelike structures that develop in glass shards—the material that he called “pectenate” but that Zirkel (1876, p. 173, 174) had ASH-FLOW TUFFS previously named axiolites. Marshall also recog- nized feebly birefringent tablets which‘he believed to be tridymite, but did not distinguish between the two materials. The one was probably cristobalite, and the other undoubtedly tridymite as he believed. Mansfield and Ross (1935, p. 321) also concluded that the devitrification products in welded tuffs from southeastern Idaho were tridymite and feld— spar. Zirkel (1876, p. 165—168) reported tridymite in a rock from the Truckee Canyon region in west— ern Nevada, and restudy of his thin sections by the authors has shown that this rock was a welded tuff. Zavaritsky (1947, p. 12) reported cristobalite and tridymite from a welded tuff from Russian Armenia. Thus tridymite was recognized, but the very fine grained intergrowths within the shards remained unidentifiable until the application of X—ray methods. Tuffs from Katmai and Peru collected by Fenner, those from Idaho collected by Ross and Smith, and those from New Zealand collected by Ross were examined by means of X—ray, and in all these the characteristic devitrification product was cristobal- ite; and subsequent examination of many devitrified tuffs have shown cristobalite. The reason for con— cluding that devitrified tuffs were characterized by tridymite seems obvious. Tridymite developing in cavities occurs as crystals large enough to be readily identified, but the cristobalite cannot be so identified, although the microscope shows that the inter- growths contain a mineral of very low index of re— fraction. The presence of two distinct silica min- erals seemed improbable, and it seemed a logical assumption that tridymite, the mineral that was di- rectly observable, was representative of the other product of devitrification. In general the cristobalite-feldspar intergrowths that have developed entirely within a shard are compact, and without observable pore space. This is confirmed by the specific gravity of materials such as those illustrated in figure 68, which is simi- lar to that of crystalline flow rocks of like composi- tion. Devitrification tends to be more prevalent in tuffs that have undergone compaction and welding, but very porous ones may also be devitrified as in the ‘ tufl‘ shown in figure 17. The development of crys— tals during devitrification and vapor-phase crystal- lization has commonly brought about a hardening or induration of the tuff without the intervention of welding. Fenner (1948a, p. 884) has discussed such tufi's from Arequipa region of Peru and called the resulting material “sillar.” MICROSCOPIC CHARACTERISTICS 37 The aggregate of cristobalite and feldspar that develops during devitrification may assume very different patterns and these merit some discussion. In figure 51, the tuff structure is greatly compressed, but is still fully recognizable, despite the develop- ment of devitrification products whose pattern transects that of the shards. In figures 59 and 77 compression and devitrification have marred, but not entirely obscured, the tufi' structures, as shown under normal illumination. On the other hand, the same area under crossed nicols (fig. 60) shows only a coarse-grained crystal aggregate with the tufi" structure entirely obscured. The most characteristic and widespread type of devitrification product is that where the growth of the replacing feldspar and cristobalite began at the grain boundary, and the waves of crystal growth met at a central zone of discontinuity that forms a dark line, as illustrated in figures 71—74. This rep- resents the axiolitic structure of Zirkel (1876, p. 166, 167) and is illustrated in the reproduction of a drawing by Zirkel (fig. 95). In this type of devitrification each of the structural units is a replica of a glassy shard unit. This type of devitri- fication has a minimum effect in modifying shard structure as is shown in figures 71—74. This axio- litic structure is widespread in devitrified tuffs and is even discernible in the Precambrian welded tufi' from Morocco (fig. 98). Axiolitic structure has been observed only in ash-flow tuffs, and seems to provide excellent criteria for distinguishing between ash-flow and ash-fall tuffs. Possibly it could de- velop in fused tuffs (see p. 26) but has not been observed. In some tufl’s, devitrification began only after thorough welding and compaction, and the feldspar- cristobalite intergrowths show continuity across large areas of tuff structure. This may or may not have important effects in obscuring the original shard structure. In figures 69 and 70, this develop- ment of feldspar and cristobalite has not markedly obscured the tufi“ structures. In figure 51, the tufl" structure is less well defined, but this is due to, extreme stretching during weld- ing, rather than to the development of devitrifica- tion products. In figure 60, a very coarse grained parallel aggregate of feldspar and cristobalite is shown under crossed nicols. However, under nor- mal illumination of the same area, the tuf'f struc- ture, although greatly modified during welding, is still recognizable. In pumice grains in which there has been com- plete collapse, with elimination of pore space, but perhaps some re—solution of volatiles, there is a strong tendency for the development of an uncom- monly coarse grained aggregate of feldspar and cristobalite. In figures 57, 58, and 61 and 62, the structure in pumice grains has been almost com- pletely destroyed, but has been retained in the en- closing shards. The most severe destruction of tuif structures seems to be the result of the development of spheru- lites, with their radial aggregates of feldspar and cristobalite. In figures 55 and 56, radial groups have destroyed the structure over most of the area illustrated, but small lenslike areas retain tuff structures. Complete destruction of tuff structure is illus- trated in figures 63—65. In figure 63 a very large spherulite is illustrated on the right and a plumose structure has developed. In figure 64 the concen- tric structure characteristic of many spherulites is illustrated. These spherulitic structures developed in welded tufi's differ in no way from those occur- ring in many rhyolitic extrusive rocks. The identity of such rocks may be determined from geologic re- lations, or inclusions of materials of elastic origin may indicate ash—flow origin. On the other hand the origin of rhyolitic rocks with strongly marked spherulitic structure may be indeterminable. Not all spherulites cause destruction of pyroclastic struc- ture, as shown in figures 54, 67—69, and 92. The studies of many hundreds of thin sections from different parts of the world indicate that most ash- flow tuffs are readily recognizable by means of thin-section studies. However, the foregoing dis- cussion of devitrification structures indicates that a few may have lost all identifiable tufi‘ char- acteristics. Marshall (1935, p. 31—32) considered the possi- bility that devitrification could destroy all direct evidence of original ash structure and says, Obviously the question arises as to whether the rocks of some of the areas of spherulitic rhyolite, of relatively coarse texture, in which no remnant of ignimbrite origin is evident, have actually been formed in this way. This suggestion is strengthened by The frequent occurrence of spherulitic rocks in beds which are almost horizontal, without a scoriaceous surface, without obsidian selvages, with a columnar structure, and with a relatively crumbly nature * *. Studies of ash-flow tuifs from other localities raises the same question posed by Marshall. Lithophysae seem to be much less common in welded tufl’s than in the corresponding extrusive 38 ASH-FLOW TUFFS rocks, but differ in no principal character. How- ever, very abundant lithophysal cavities are es- pecially conspicuous in one widely occurring welded tuff in southeastern Idaho (figs. 49 and 52). These cavities are the result of the release of volatiles during devitrification and suggest that this par- ticular ash flow had retained an uncommonly large proportion of volatiles in solution in the glass. The so-called thunder egg shown in figure 67, represents an uncommonly large lithophysa which had de- veloped in a welded tuff from Oregon. The inner cavity was subsequently filled with chalcedony. PHYSICAL CHEMISTRY The concept of the emplacement of pyroclastic materials by flowage, and in many occurrences their subsequent welding into a compact rock with many of the relations of lava flows, is so remarkable that the full recognition of this mode of origin was ac- cepted but slowly. The seeming improbabilities of such a mechanism are so real that a very detailed consideration of the physical and chemical factors involved in ash flow, welding, and devitrification seems to be imperative. A consideration of the mechanisms of the forma- tion of welded tuifs and related materials includes only a part of the broader concept of igneous ac- tivity. Several factors involved in volcanism, such as the ultimate source of igneous materials and the mechanism of their migration from that source, will probably long remain controversial. A study of ash-flow tuffs is concerned with one phase of volcanism, that seems to be a late and superficial phase of igneous activity. However this late ac- tivity makes important relations available to direct observation, but others that occur in depths are not so observable. A fuller understanding will be desirable of erup- tive mechanisms in general, the volatiles that char- ‘ acterize them, the heat relations in an ash flow, and the flow mechanism of plastic ash shards suspended in anatmosphere of red-hot turbulent gas—the latter a unique flow medium. However, a synthesis of direct observation, geologic studies, thermody- namics, physical chemistry, and geophysical experi- ment seems to provide the basis for an attempt to understand the principal factors involved in the eruptions, vesiculation, explosive disruption, flow- age, emplacement, welding and devitrification proc- esses, and vapor-phase reactions that have produced welded tuffs. Not all geologists are agreed about the relative importance of differentiation in the formation of the different magma types. However, concentra- tion of volatiles accompanying progressive stages of differentiation and crystallization seems to be well established. The later products of differentia- tion, commonly contain a concentration of volatiles, and so provide the necessary mechanism for ex- plosive, pyroclastic eruptions. As volatile-laden magmas rise to regions of diminished pressure these volatiles tend to escape, and commonly escape ex- plosively. Variations in volatile content and pres- sures on the system at the time of eruptive break- through probably determine both the violence of the eruption and its nature, whether ash flow or ash fall. The physicochemical factors controlling the erup- tion and later history of welded tuffs, nonwelded tuffs and related pyroclastic materials are: (a) The chemical composition of the magma from which they were formed; (b) the proportion and character of the volatilesoriginally dissolved in the magma; (c) the temperature at the time of eruption and at suc- cessively later stages, until all reactions cease; and (d) the physical and chemical reactions at successive stages in the genetic history of the tuffs. Such physi- cal relations as thickness of the tuff deposit and the hydrostatic pressures developed in the system, total volume, slope on which they were deposited, and the tuff textures that affect retention of heat and vola- tiles are also essential factors in the formation of ash-flow deposits. The eruptions of the magma, its vesiculation, flowage as a turbulent mixture of ash and gas, welding after coming to rest, and final devitrification are all functions of the above re- lations. The initial difference between ash-fall and ash- flow eruptions is probably only the relative gas- controlled violence of the eruptions. However, this is the basis for fundamentally different posterup- tion processes. The ash-fall material has been blasted high into the air where its magmatic vola- tiles were largely dissipated and where it was quickly chilled and its magmatic history brought to an end. In contrast the ash-flow tuffs have been emplaced by mechanisms that conserved heat and at least some of the volatile constituents, Thus a series of physical and chemical relationShips was initiated which have left a partial geologic record. This section will explore the physical and chemical factors that seem deducible from that record. This means that the study of the origin of ash-flow tuffs is a geologic problem, buttressed by available di- rect observation, experiment, and chemistry and theories of mechanics and physics. PHYSICAL CHEMISTRY 39 HEAT A source of heat adequate for welding of ash-flow tuffs after eruption and travel for long distances has presented a major problem for geologists. Transportation through the air was long the only recognized mechanism for dispersal of ash from volcanic eruptions. However, this mechanism would involve such a loss of heat, that welding after wide dispersal seemed to be precluded, and this deterred geologists from accepting the welding of tuffs. The wide application of the term “nuées ardentes” and its equivalent “glowing clouds” to describe the dis- persal mechanism has tended to perpetuate the con- cept of cloud-borne materials in contrast with the ash-flow mechanism of dispersal. A dispersal mech- anism compatible with conservation of heat became understandable only when ash flowage, rather than cloud dispersal, came to be recognized. However, the entire problem of heat conservation during the formation and transportation and emplacement of ash-flow materials remains to be considered. Many measurements of the temperature of magmas at the time of extrusion have been made. Some of these have been summarized by Daly (1933, p. 68) and indicates a range from about 1,050°C to nearly 1,200°C. Perret (1950, p. 55) has described temperatures of 1,000°C to 1,150°C at Vesuvius. A large proportion of the recorded temperature meas- urements were made at basaltic or andesitic centers, and much less is known about the temperature in rhyolitic centers. However, the experimental work of Goranson (1932, p. 234) showed that a dry gran- ite would be completely molten at a temperature of 1,050°C. He (1932, p. 229) found that at 1,500 bars pressure and with 3.8 percent H20, the granite became completely molten in 3 hours at 900°C. Bowen and Tuttle (1953, p. 50) found that granite melts at about 720°C under 1,000 atmospheres pres- sure of water, and at about 640°C under 5,000 at- mospheres. Ingerson (1955, p. 346) remarks “These figures are much more in accord with field and mineralogical observations on granitic intru- sions than are anhydrous melting temperatures.” Explosive rhyolitic magmas are erupted under a sudden release of pressure, and so the foregoing figures do not directly apply to their temperature or volatile content. A knowledge is desirable of the temperature and pressure at which quartz pheno- crysts began to form, volatile content of the magma, and the pressure and depth at which vesiculation began. However, the temperature of a rhyolite magma was much lower than that of dry fusion of granite; a temperature of about 900°C, or perhaps even less seems probable. Clearly it was below 1,000°C. Some geologists have postulated reactions subse- quent to initial eruption that supplied exothermic heat, and of these, Fenner (1950, p. 600—601) be- lieves that exothermic reactions developed during the exsolution of volatiles from the molten magma are a fundamental factor in welded—tuff formation. He says, We are now prepared to consider what happens when fugi- tive constituents are given an opportunity to escape * * *. In these events each individual reaction between the infini- tesimal units is accompanied by a minute heat effect positive or negative, but the system as a whole is so immensely com- plicated, and there is so little information in the reactions of separate constituents, that no calculation of the net heat effect is possible. The evolution of heat might be very great * * *. Or, on the other hand, the net effect in the system as a whole might be negative. Fenner further considered field evidence of exo- thermic reactions, but this was based on relations in Katmai crater, rather than those of the “sand flow” in the Valley of Ten Thousand Smokes. The tuff of the “sand flow” shows no indications of solu- tion of phenocrysts or of included andesitic rock fragments and so there is no evidence of the super- heat that Fenner believed was indicated at Katmai. The study of ash-flow tuffs from many other loca- tions has shown that magmas from which welded tuffs are derived reached the surface with different, but in general with important amounts of pheno- crysts. Feldspar phenocrysts may be euhedral, or sharply angular, and fractured crystals rarely show the effect of rounding by re-solution. Quartz com- monly shows embayment but in no greater degree than in extrusive lavas of the same composition; this embayment probably occurred before eruption. Augite grains and rock fragments derived from andesitic rocks rarely ShOW significant reactions with an enclosing rhyolitic magma. Calcic feld- spar grains derived from the rocks traversed by the magma on its way to the surface, seem in some oc— currences to have been out of equilibrium with the magma, and show evidence of alteration or res— olution. The relations between included crystal grains and the magma are not primarily different from those observed in extrusive rocks. Thus, in general, tuffs show no evidence of superheat, or for any addition of a substantial amount of heat subsequent to erup- tion. The original heat of the magma seems to be 40 ASH-FLOW TUFFS all that is involved in the observable relations. A few welded tuifs are uncommonly low in pheno- crysts and may have been uncommonly hot. The adequate heat for welding within an ash flow must be sought in connection with other physical and chemical relations that tended to conserve this initial heat supply. VISCOSITY The viscosity of glasses with a large proportion of SiO2 and A1203 is so high that it has an important effect on the mechanism of explosive ash-flow erup- tions, and especially on welding and distortion of shards. Viscosity data for silicate melts have been discussed by Bowen (1934). Experiments on an albite glass indicated a viscosity of about 108 poises at 1,150°C. The viscosity of a rhyolite glass, al- though somewhat higher, would be of a similar or- der. An andesite glass with 51.00 percent SiO2 had a viscosity of about 3.1 X 104 poises at 1,200°C. Macdonald (1954, p. 172), by basing his calculations on the rate of flow, arrived at a viscosity of 2 X 104 poises for the 1950 Honoku flow at Mauna Loa, but thought that the figure was too great. These figures indicate the very high viscosity of rhyolitic magmas and the much lower viscosity of andesitic and basaltic ones. No quantitative data are available on the effect on viscosity of variations in temperature or the addition of H20. However, the experiments of Shepherd on the vesiculation of glass give some roughly qualitative information about the effect of volatiles on viscosity. Shepherd (1938, p. 340-341) states, For these ‘wet’ glasses (and the same would be true of the mother liquor of a crystallizing magma) a rapid change in the apparent viscosity is caused by relatively insignificant concentrations of volatiles. When a block of No. 5 [1.25 percent volatiles] reaches the vesiculating temperature it flows, i.e., froths, around and through small obstacles, or in a tube it froths all over the inside. But once volatiles are out, the septa and shreds of glass remain indefinitely rigid. It requires much time and a much higher tempera— ture before the mass begins to gather itself together again. This quick shift from fluid to rigid has obvious importance both for volcanology and petrology. Shepherd further states (1938, p. 343), The above-cited experiments on the relations of volatile con- tent to viscosity have an important bearing on the mech- anism of extrusion and lava flows. In the interval where vesiculation occurs the glass is notably fluid, the expanding pumice can froth around and between obstacles. But once fully expanded, that is to say, once the volatiles have been liberated, the foam becomes rigid. VOLATILE COMPONENTS Direct knowledge of the amount, volume, and the chemical character of the volatiles taking part in an explosive volcanic eruption is lacking, and only in- adequate indirect observations are available. Col- lections have been made from secondary vents, but it is not known how well they represent original composition and they tell little about the original proportion of volatiles. Moreover, most such collec- tions have been made from basaltic or andesitic lavas, and very much less is known about volatiles from rhyolitic ones. Volatile constituents are essential in the entire sequence of events resulting in welded tuffs. They probably take part in the rise of the magma into the explosive eruptive vents, and they are most obviously important in the vesiculation and erup- tion of this magma. They are part of the gas-ash mixture that constitutes the ash flow and they are partly released from solution in the glass shards during flowage. Some volatiles have remained in solution until after the ash flow came to rest where they were a factor in lowering the temperature of welding, and promoted devitrification and the for- mation of vapor—phase minerals in open cavities. Some of the primary volatiles probably remained in solution in the glassy parts of ash flows. The studies of volatiles in obsidian and associated perlite by Ross and Smith (1955) have indicated that most glassy volcanic flow rocks which have any degree of permeability have undergone postdeposi- tional hydration. The same detailed comparisons have been possible with glassy ash-flow tuffs, but the high permeability of nonwelded ash-flow tuffs indicates that they too have undergone hydration. This is confirmed by the similar water content in the two types of materials, that is, of pumice and obsidianlike welded tufl". Only a few welded tuffs studied by the authors seem to have escaped hydra- tion. One of these, a specimen from southeastern Idaho (figs. 9 and 10), contained 0.15 percent H2O, while another part of the same flow contained 3.06 percent H20. These relations indicate that pristine water characterized the one specimen while the other had undergone hydration. The retention of even 0.15 percent H30 is significant because this has been retained throughout all the volatile-con- trolled episodes that preceded cooling. Moreover, the physical eifect of volatiles is greatly dispropor- tional to their amount. PHYSICAL CHEMISTRY 41 WELDING ' Not all ash-flow materials have become welded, and others are welded only in some part of the flow. However, welding is one of the most significant features of ash-flow materials, and probably the one which has caused the most misunderstanding of ash—flow mechanisms. For this reason welding merits a detailed discussion. The factors to be considered in welding are as follows: (a) The initial heat of the magma; (b) dispersal by flowage; (c) insulation in thick ash flows; ((1) the effect of volatiles. Ash flows must come to rest with substantial amounts of heat. The dispersal by flowage, and the speed of movement are necessary factors in the conservation of the heat which is required for weld- ing after ash flows come to rest. N0 estimate, not even a vague one, of the speed of dispersal of the great ash flows of the past is possible, but the speed of some of the minor ones have been estimated as follows: Eleven to 26.6 meters per second, or 60 miles per hour at Pelée (Lacroix, 1904, p. 203) ; 33 meters per second, or 74 miles per hour at Pelée (Perret, 1937, p. 96) ; 100 miles per hour at Mount Lamington (Taylor, 1954, p. 86). The great flows that spread for many miles prob- ably moved even faster. Thus, movement is so fast that the time for cooling during fiowage is short. Rock materials are very poor conductors of heat, and would be still poorer with substantial amounts of pore space as in poorly consolidated materials. Thus, tuffs tend to remain hot long after they come to rest. In describing the eruptions at Pelée, Perret (1950, p. 95) says, Below the surface weeks after the passage of a nuée, a walking stick was burned at a depth of thirty centimeters or less * * So different was the more fully air borne ash which reached St. Pierre in the great nuée of May 8, 1902, that although its temperature on arrival is believed to be at least 800°C * * *. Negroes entered the ruined town in the afternoon only eight or nine hours later. Kozu (1934, p. 143) has given the results of tem- perature measurements on ash-flow materials of the eruption at Komagatake, Japan, of 1929. At a depth of only 40 cm, and 8 to 11 days after the eruption, the temperature at 6 different places ranged from 510° to 310°C and about 1 year later the temperature ranged from 210°C down to 28°C at the same depths. A temperature of 510°C was probably too low to allow welding, but perhaps not excessively low; and had it been a thick flow, weld- ing temperatures would probably have existed in some deeper part of the flow, and persisted for a long time. The reduction of the Viscosity of glass by the presence of small but significant amounts of vola- tiles discussed in the preceding section would be a factor in promoting the distortion and compaction which normally accompanies the welding of tuffs. Many pumice fragments in welded tuffs show evidence of lower viscosity and greater distortion than the associated ash materials. Figure 47 illus- trates a pumice area which was so plastic that it was distorted against an area of more rigid welded tuff. Figure 48 shows strong distortion of a pumice area. The complete elimination of porosity in many pumice fragments indicates the same relation, and that the vapor—phase material in the original pumice pores had been eliminated. The pumice was originally inflated by the exsolu- tion of its magmatic volatiles. The hot highly viscous glass walls of the pumice had a strength that a cold brittle glass would not have had. Therefore the vesicular pores of the viscous glass were prob- ably capable of retaining at least part of the original volatiles. Experiments show that the volatiles in- volved in the expansion of a pumice are about 0.1 to 0.3 percent and some part of these would remain trapped. When this pumice collapsed probably some part of these trapped volatiles would go back into solution in the glass of the pumice. This suggests a possible explanation of the lowered viscosity of pumice fragments, as pointed out by Shepherd (1938, p. 343) when he says, “A mass of foam trapped and subjected to heat and water at sufficient heat and pressure will, of course, collapse and be— come fluid again.” His study indicates that even 0.1 percent of H20 would have an appreciable effect on viscosity of a glass. The relations in the pumice differed from those in the associated ash in which the volatiles escaped into intergranular spaces when not trapped as they were in the pumice vesicles. Several geologists have studied the temperature of welding in ash-flow materials. Marshall (1935, p. 21) believed that because tridymite, with a mini— mum stability temperature of 870°C, was present in the New Zealand ignimbrites, this established that temperature as the minimum for their emplacement. He did not understand that both tridymite and cristobalite normally form as metastable minerals. Sosman (1927, p. 786) states, It might appear at first thought that the occurrence of one of the high temperature forms of silica in a rock is evidence that the rock was about 1470°C (cristobalite) or above 42 ASH-FLOW TUFFS 870°C (tridymite). But the tendency shown by silica to deposit in metastable form spoils this part of the argument. However, Marshall (1935, p. 22) noted that quartz was present and preceded tridymite in time of for- mation, and he was at a loss to explain this relation. Gilbert (1938, p. 1856) based his conclusions on temperatures in the Bishop tuff 0n the temperature of the conversion of part of the green hornblende to basaltic hornblende. This caused him to state, “The only safe conclusion is that the temperature of emplacement could not have been very much higher than 750°C, for if it had been, all the horn- blende should have been converted to basaltic horn- blende.” Shepherd (1938, p. 342) reports experiments on the fusion of rhyolitic glass from Cerro No Agua and states. “But for anything one would call flow in the field sense, temperatures above 600°C are needed * * *.” Boyd and Kennedy (1951, p. 327) experimented on the welding of rhyolitic pumice from Mono Craters and state that “The pumice was found to weld at a temperature ranging from 775° to 900°C.” Because the glass of ash-flow materials probably contained at least 0.15 to 0.3 percent of H20, the welding temperature would probably have been somewhat lower than this. A report on the investigations of the Carnegie Institution of Washington Geophysical Laboratory for 1953—54 cites the studies of Boyd (1954, p. 139) at Yellowstone National Park. Boyd concludes, A thermodynamic analysis of the eruptive process [of ava- lanches of fragmented viscous lava] suggested by field evi- dence has been undertaken. The temperature interval between the original magma temperature and the welding temperature of the erupted tufl“ will have a maximum value of 100°C; in the average case it will be less. The amount of heat lost during emplacement of a tufl’ avalanche will depend primarily on the thickness of the tuff flow, the manner in which the flow moves, and the time elapsed during emplacement. Reasonable value for all these quantities can be found from either field evidence or from data available on nuées ardentes. Assuming a collapsed thick- ness of 30 meters, an average time of emplacement of one hour, and turbulent flow in the moving avalanche, the loss of heat during emplacement will reduce the temperature by less than 10°C—a value well within the estimated limits. These figures are reasonable approximations for ideal ash-flow conditions, but many ash flows are probably not ideal. That is, the ash flows have been subject to mixing with air and their distal ends have come to rest at much lower temperature than the initial eruption temperature. These differences in temperature may be very real, as can be readily seen by the changes in amount and degree of welding that take place in many single ash-flow sheets be- tween source and distal end. Recent experimental work by Smith and Irving Friedman (written communication, 1956) has shown that welding of glassy rhyolitic ash and pumice can take place at temperatures at least as low as about 580°C. These experiments were carried out under a water-vapor pressure of about 300 psi (pounds per square inch) and a mechanical load of about 500 psi. They believe that pressure conditions of this order of magnitude could exist in ash flows having a precollapsed thickness of about 800 feet and a density of 1.5, or about 1,200 feet and a density of 1.0. The physical relations Within an ash flow may be restated as follows: 1. Rhyolitic tuffs are derived from a magma with an initial temperature not greater than 1,000°C, and probably below 900°C. An- desitic magmas probably reach a maximum of about 1,150°C. 2. The viscosity, especially of rhyolitic magmas, is high but not so high as to inhibit collapse and welding of shard materials under load. 3. Volatiles, at least in minor amounts, are re- corded in most welded tufi‘s. These reduce viscosity and lower the temperatures of weld- ing. Relations that bear on heat conservation are as follows: 1. Pyroclastic materials commonly escape to the surface without superheat, and show no evi- dence of later acquiring substantial amounts of heat. 2. Gas phenomena accompanying ash-flow erup- tions are powerful and the volumes of gas are large, but the mass is small compared with that of the ash. 3. Because the mass of gases that escaped during flowage is small their escape does not remove large amounts of heat from the system. 4. Ash flows move with great rapidity and the time effect on cooling is small. 5. The mechanism of dispersal is that of a gas- shard mixture of marked density and it rides close to the ground. The materials are not airborne as in ash falls. 6. Ash-flow materials are very poor conductors of heat. These relations seem to indicate that the mag- matic heat inherent in an ash flow is effectively conserved and adequate for welding. PHYSICAL CHEMISTRY 43 The physical chemistry of the welding process is complex and the relative effects of the variables are difi‘icult to evaluate. However, the relations in welded tuffs are similar to those known as sintering by the glass technologist, and some of the research on glass sintering helps the understanding of weld- ing in volcanic ash. The relations in sintering have been outlined by Eitel (1954, p. 1047) and are summarized as fol- lows: A minimum temperature must be attained for the initiation of the sintering reactions and atomic bridging. The mobility of the atoms or atomic groups must have attained a minimum intensity that is sufficient for a beginning of place exchange. The inference seems to be that the atoms and mole- cules must Vibrate so intensely around their struc- tural equilibrium positions that a large number of them are able to exchange places with surrounding atoms or molecules. Eitel (1956) published a “free translation” of a series of papers by Carl Kroger (1952—55), and in a supplemental discussion further amplified the mechanisms of sintering, which are abstracted as follows: The reactions in the solid system composed of individual units depend on the following factors: 1. A necessary threshold mobility of the atomic and molecular particles in the solids, along their boundary zones. 2. The temperature of a molecular exchange of place, under the action of thermal energy, is a func- tion of fusion points. In a glass system this implies an adequate reduction of viscosity, which is de— pendent on temperatures and volatile content. 3. Sintering (welding) involves a real “creeping or flow factor.” That is, if two units have an inti- mate contact, particles from both solid parts will diffuse along the boundary surfaces. ,The rate of displacement will depend on the inner mobility (yield value viscosity) of the material, the surface tension of the material surrounding the medium, and the hydrostatic pressure under which the material is standing. Mobility of the particle material at the elevated temperature may mean a partial flow in a liquid phase. The above factors that control sintering apply primarily to crystalline materials, whereas in the ’welding of tuffs the materials are noncrystalline (glassy). The materials are also plastic, although the viscosity is so high that they will fracture under sudden stress. That is, glassy materials have prop- erties of solids such as to make these factors of sintering applicable to an interpretation of the weld- ing of tuffs. The glassy character would seem to simplify rather than complicate this relation. The factors that control welding are: (a) The presence of volatiles; (b) pressure within the sys— tem; (c) intimate contact of the individual par- ticles; (d) threshold mobility within the system (adequate temperature); (e) inner mobility (yield value viscosity and surface tension) ; and (f) creep- ing or flow factor. These factors are interrelated and not entirely independent. The foregoing outline may now be specifically applied to the consolidation and welding of tuifs. The tuffs come to rest in a heated condition, with welding initiated at about 650° to 700°C, and per- haps reaching about 900°C in the hotter tufi's. This provides them with a plastic yield, although the Viscosity is high in glasses of rhyolitic composition. They contain small amounts of dissolved volatiles that have an important effect in reducing viscosity. After coming to rest the tuffs are under a static load, depending on the thickness of the overlying « material. Porous glassy tuffs are exceedingly poor conductors of heat and remain hot probably for years. Shards and other glassy fragments, as formed by explosive eruption, are sharply angular and would come to rest touching one another only at points, and with a large proportion of interspaces. Under load, and with adequate initial mobility, there would be plastic yield accompanied by varying de- grees of distortion and accommodation of grain to grain. This plastic yield would, with time, increase the areas of contact between individual particles, and under optimum conditions there may be nearly complete elimination of pore space. In the pres- ence of volatiles some of the particles would be adsorbed on the grain surfaces even at temperatures of 700°C. The effect of all these factors would tend to promote a bridging effect and a transfer of ma- terials across surfaces of intimate contact. In the system here postulated volatile components were present, both as a vapor phase and in solution in the glass. The vapor phase may have been a factor in some transfer of materials, but it did not initiate any new phase (crystalline). A crystalline phase did not develop until after welding. Where viscosity is low and where there is contact only at points, welding may be confined to these small areas and recognition of welding under the microscope will be difi‘icult. On the other hand, a 44 ASH-FLOW TUFFS combination of adequate pressure and other factors promoting initial mobility may combine to permit thorough accommodation of grain to grain, and result in the welding of tuffs into a compact ob- sidianlike glass. The glass film that forms on the contact between two glass shards might be thought of as a secondary glass derived from the original glass. The inter- shard films are too thin to study in any detail, but differ slightly from the original or primary glass. In the glassy pumice from Argentina (figs. 83 and 84) the intershard film is colorless, While the origi- nal shards are brown. The lighter color is due in part to the nontransfer of pigmenting dust material, and indicates a slight difference in the composition of the two glasses. This does not seem, however, to constitute any discontinuity of phases. The relations shown in figures 59 and 60, and in figures 67—70, are especially significant. Here there was almost complete elimination of pore space so that shards had been brought into complete contact with one another previous to devitrification. The continuity of the glass structure had become so com— plete that crystal growth of feldspar and cristobalite took place directly across the zone of union between original shard boundaries. The originally frag- mental glass has become a structurally homogene- ous medium, that presented no impediment to crystal growth across the former zone of discon- tinuity. This would seem to represent welding into a single glass phase previous to the growth of crys- tals across the former zone of contact. The obsidianlike tuif illustrated in figures 9 and 10 has remained glassy, but the establishment of complete physical continuity seems evident. The collapsed pumice fragments have lost all trace of the original pumice structure (fig. 45) and here a complete continuity of structure has been~estab- lished. DEVITRIFICATION Not all ash-flow tufi‘s have undergone devitrifica— tion; many are devitrified only in part, and have remained glassy at the top and base. Nevertheless, devitrification is one of the most distinctive charac- teristics of ash-flow tufi's, hence the physical and chemical factors involved in devitrification are of especial interest. The factors controlling devitri- fication are: (a) Chemical composition of the tuff; (b) chemical composition of the accessory volatiles; (c) rate of cooling; ((1) temperature of devitrifica- tion; (e) identity of the minerals formed; and (f) stability relations of these minerals. The chemical composition of the tuff may be closely determined, although it may have been very slightly modified during devitrification in the pres- ence of volatiles. The amount and chemical composition of the volatiles retained in a glass would probably not be greatly different from that discussed in connection with welding. This similarity indicates that the volatiles were present in the glass to the extent of not more than a few tenths percent. However, even this proportion would reduce Viscosity and pro- mote crystallization. The presence of volatiles dur- ing devitrification is indicated in figures 49, 52, and 67. In these specimens devitrification was related to the formation of vesicular gas cavities. Rhyolitic tuffs seem to be more commonly de— vitrified than andesitic tuffs. The reason for this is not clear because high-silica extrusive flows show a much greater tendency to remain glassy than do andesitic ones. The difference may be due to a more effective retention of volatiles in high-silica glass shards than in andesitic shards. The temperature of devitrification has a wide range, probably beginning immediately after weld- ing and continuing to some problematical minimum temperature. However, once crystallization starts, it probably proceeds rapidly. X-ray examination of ash-flow tufi‘s has shown that the devitrification products of dense shards are a fine-grained intergrowth of cristobalite and sani- dine feldspar. On the other hand, tridymite, and more rarely cristobalite, have formed in vesicles and uncollapsed pumice fragments. The factors that determine whether cristobalite or tridymite develop are of interest. Where compact homogene- ous glass is devitrified there is a diffusion of ions within the glass over very small distances, but there seems to be no vapor-phase transfer, although vola- tiles are in solution in the glass and help to promote devitrification. The same minerals form in col- lapsed pumice where free vapors do not collect. On the other hand, Where the pumice does not collapse or where vesicles form, tridymite and feld- spar form in the presence of vapors. There is vapor-phase transfer of materials at least for short distances. The formation of tridymite in a gas cavity is illustrated .in figure 66. In rhyoliticfii welded tufi's tridymite and feldspar are the typical / vapor-phase minerals, but more rarely biotite, '\ amphiboles, and zeolites may form. PHYSICAL PROPERTIES 45 The metastable character of cristobalite and tridymite means that these minerals tend to invert to quartz, the stable allomorph of SiO,. The geo- logic conditions that allow tridymite or cristobalite to persist, or that may promote inversion to the stable form are not accurately known. Cristobalite and tridymite seem to have resisted inversion to quartz in all the very abundant ash- fiow tufi's of younger Tertiary and Quaternary age that have been available for study, the only excep- tion being those which have undergone alteration. Tufi's of Mesozoic age are less well known, and cristobalite has not been definitely recognized in tuffs of Paleozoic or older age. Upper Carbonifer— ous tufi's from Corsica described by Bodenhausen (1955, illus. 21, 22) show evidence of fine-grained quartz. A specimen of welded tuff of Precambrian age from Sweden described by Hjelmqvist (1956) is characterized by quartz (examination by the authors). Welded tufi's of Precambrian age from the Anti-Atlas region of Africa have been described by Bouladon and Jouravsky (1954, p. 37—48), and samples have been made available for study. The shards illustrated in figure 98 still retain typical axiolitic structures, but these now contain only quartz and feldspar. PHYSICAL PROPERTIES INDICES 0F REFRACTION George (1924) has provided useful data for de- termining the approximate silica content of a vol- canic glass by means of indices of refraction. The study of the effects of water in glasses by Ross and Smith (1955, p. 1080—1085) included their optical properties, and the effect of water in raising the indices of refraction. This showed that the pres- ence of 1 percent of water would lower the indicated SiO, content (compared to a water-free glass) by about 2 percent. Ash-flow tuifs commonly contain foreign material and these may give an incorrect indication of the character of the primary eruptive magma. Bono— rino (1944) has described “ignimbrites” from near Malargue Mendoza, Argentina, and says, “The crys- tals are acid to basic plagioclase, and clinopyroxene, the two latter often associated as to suggest their common origin from a basaltic rock.” A restudy of this material submitted by Bonorino (figs. 83 and 84) has indicated that the calcic plagioclase and clinopyroxene did have a common origin, but that these materials seem to have been derived from an invaded rock, probably an andesitic one. The index of refraction of the glass fraction (1.508 to 1.512) indicates that the eruptive material of the glass shards was much more silicic than the composition of the rock as a Whole. For this reason the silica content of the glassy groundmass part of this rock was determined by the US. Geological Survey. The SiO2 content of 67.20 percent indicated that the rock, without inclusions, was probably dacitic. An important variable affecting the index of re- fraction of a glass is the state of oxidation of the iron. The total iron content of volcanic glasses may be as much as 7 percent or more, and in unaltered glasses is present in both the ferrous and ferric state. Such iron in solution in a glass gives it varying intensities of dull-brown tints, except in very low iron rhyolites, in which the glass may be nearly colorless. The collapsed pumice fragments shown in figure 1 appear black in the hand specimen. Under the microscope this glass is nearly colorless while the groundmass glass (red in the hand specimen) is slightly brown. The colorless glass contains magnet- ite microlites abundant enough to permit a perfect separation from the nonmagnetic groundmass. The presence of magnetite microlites in the colorless glass is also observable under the microscope. In this glass the iron that was originally in solution was reduced to magnetite and segregated into dis- tinct granules by reactions that took place during the collapse and welding of the pumice fragments. The elimination of part of the iron from solution in the glass and its collection as discrete microlites of magnetite has lowered the index of refraction of the colorless glass below that of the brown glass. The different states of oxidation of iron, and the resulting changes in a glass are shown in the frontis- piece where reproduction in color depicts relations that black and white illustrations could not reveal. The iron in the glassy tuff illustrated in the top photograph was oxidized and colloidally dispersed, as suggested by Iddings (1899, p. 407), probably during emplacement in the presence of oxygen from the air. A later change in the environment resulted in the partial reduction of the ferric iron, and its segregation as finely disseminated dustlike magnet- ite. Thus iron was removed from colloidal disper- sion in the glass, leaving it nearly colorless, with an index of refraction abnormally low for a glass of this composition. The dull-brown glass in the central part of the two large shards shown in the bottom photomicro- graph seems to retain the normal ferrous-ferric iron relations of a glass of that composition; the index 46 ASH-FLOW TUFFS of refraction is intermediate, and characteristic of a glass of that composition. On the other hand, oxidation of most of the iron to Fe203, and its colloi— dal dispersion in the glass, has raised the index of refraction above that of the normal brown glass. As may be seen from the foregoing discussion, several factors may have a marked effect on the indices of refraction of a glass, and caution should be exercised in determining the composition of a glass from its index of refraction. The glass frac- tion may differ in composition from the overall composition of the rock because of the presence of phenocrysts and the inclusion of alien minerals and rock fragments. The glass of tuffs has usually undergone hydration, which increases the index of refraction. If the glass has any appreciable pro- portion of iron the different states of oxidation will have a marked effect on the indices of refraction. The phenocrysts of a volcanic rock may give an incorrect indication of the chemical composition as pointed out by Rittman (1952, p. 76) who states, The phenocrysts by themselves give very often a wrong idea of the real composition of the whole rock, because quite different kinds of minerals than those forming the pheno- crysts may be present in the optically indeterminable micro- crystalline groundmass, or may be potentially present in the glass. For example, many dacites, or even rhyodacites have been called andesites because their contents in quartz and sanidine occult in the groundmass were not taken in account. Rittman shows the necessity of extreme reservation in assigning an andesitic character to a tuff on the basis of observable minerals. The rock in bulk may be much more silicic than suggested. Andesitic rock and mineral grains are so ubiquitous in welded tufl’s throughout the world that their complete absence is rare. Where they occur in rhyolitic welded tuffs, their alien character is obvious; but in ash-flow tufi's of intermediate composition they may lead to incorrect identifications. SPECIFIC GRAVITY Most welded tuffs retain at least some pore space, and only a few specimens permit a determination of the true specific gravity. However, the specific gravity of four virtually pore—free rhyolitic welded tuifs is given as follows: ‘ Specific gravity of dense rhyolitic tufis Specific Character of material Locality gravity Glassy welded tufl" ................ Valles Mountains, N.Mex __ 2.38 Obsidianlike welded tufl’ ...... Southeastern Idaho ______________ 2.34 The same tufl" hydrated ________ Southeastern Idaho ___________ 2.40 Dense devitrified tuff ____________ Oregon ................................. 2.47 The small difference in the specific gravity of glass (2.34, 2.38) and of dense devitrified tuff (2.47) might seem surprising, but this is due to the large proportion of cristobalite (specific gravity 2.32) in the tufl". POROSITY Ash-flow tufl's range from nonwelded, unconsoli— dated materials, to thoroughly welded tuifs with a density equal to that of the corresponding normal glass or crystalline rock. This range in porosity is illustrated in columnar sections of tufl" deposits from three localities. Marshall (1935, p. 29) gives porosities in a typi- cal occurrence of the New Zealand ignimbrites. Fourteen nearly equally spaced samples were col- lected from a 250-foot cliff at Te Toki Point, Lake Taupo, New Zealand. Porosity of acid rocks, Taupo-Rotorua, New Zealand, volcanic district Sample Porosity (percent) Sample Porosity (percent) 1 ______________________________ 28.6 2 .............................. 25.2 3 ______________________________ 23.4 ...... 22.0 5 ______________________________ 20.0 6 ______________________________ 19.8 7 .............................. 23.1 Marshall notes that samples 5 and 9 contained surface-deposited secondary silica. Gilbert (1938, p. 1843) gives the following table of densities of a characteristic section of the Bishop welded tuff of California: Porosity of Bishop tufi Vertical distance from top of tuff Porosity Specific (feet) (percent) gravity 0 37 1.35 175 i 25 __________________________________________________ 17 1.9 250 i- 25 __________________________________________________ 7.9 2.2 300 -_+- 25... 4.6 2.26 400 i 25 __________________________________________________ 3.6 2.35 The welded tuff at Battleship Rock, Valles Moun- tains, N.Mex, was studied by the authors, and is shown in figures 42 and 43. ASH-FLOW MATERIALS 0F INTERMEDIATE COMPOSITION Porosity of Battleship Rock, N.Mex., ash-flow tufi” Distance below Porosity Distance below Porosity Sample the top (feet) (percent) Sample the top (feet) (percent) 69 8 ________________ 175 24 58 9 ________________ 185 33 51 10 ________________ 200 48 33 11 ________________ 205 51 29 12 ________________ 215 59 _, 25.5 13 ,,,,,,,,,,,,,,,, 255 70 7 ________________ 145 24 Only a single unit is represented at the Battleship Rock locality and exposures present an excellent opportunity for detailed study. The porosity of the Battleship Rock ash-flow unit ranges from about 70 percent near the top, to 24 percent, about one- third the distance above the base, and then back to 70 percent at the base. In another locality, a part of the same ash-flow unit had undergone complete welding and elimination of pore space. Its specific gravity was 2.36 and this value was used in calculat- ing the porosity of the section studied in detail. The variations in porosity are shown in the curve in figure 15. Tufi‘ units with maximum welding at or near the base, like those described by Marshall and by Gil- bert, have been reported by others engaged in the study of ash-flow materials. The position of maxi- 47 mum load is commonly believed to be the reason for this position of maximum welding. On the other hand, the position of the greatest conservation of heat is also a factor in effecting maximum weld- ing. For this reason many ash flows show relations similar to those in the Battleship Rock ash-flow tuff. Here maximum welding occurs about one- third of the distance to the top of the unit or about 80 feet above the base (fig. 3). Most of the ash-flow tuffs of the Valles Moun- tains show similar relations. This same relation has been observed in the Rattlesnake fermation of Oregon (T. P. Thayer, oral communication, 1950), and has been reported in Salvador by Weyl (1954b) and in some of the Aso flows of Japan (Williams, 1941b, p. 279—280). ASH-FLOW MATERIALS OF INTERMEDIATE COMPOSITION The preceding discussion has applied to ash-flow materials in general. However, in the Western United States these are dominantly rhyolites, dacites, quartz latites, or rhyodacites in composition, and for this reason a large proportion of the tuffs discussed have been rhyolitic in composition. The physical characteristics of rhyolitic and andesitic ash-flow materials are mostly the same, but the andesitic ones differ in some details and merit a 250 2 /i° ,/ 200 ‘ f/ '— LIJ LIJ LI. 3 3 E 150 5'1 LL] 0% 6 < 7 g 100 E 8 9 (2 N-ON O 50 \glle Eu 0 ‘13 o 10 20 30 4o 50 60 7o 80 90 100 POROSITY, IN PERCENT FIGURE 15.APorosity of Battleship Rock ash-flow tuft, Valles Mountains, N. Mex. 48 ASH-FLOW TUFFS brief discussion. The basaltic types of ash-flow ma- terials seem to be comparatively rare, therefore they are considered together with those of inter- mediate composition. Ash flows of intermediate type have been described under a variety of names, including andesite, quartz andesite, latite, some so- called dacite, and in some regions, trachyte. Ash flows of intermediate composition are domi- nant types in several regions, and have been de- scribed from Japan, Costa Rica, El Salvador, Argentina, Russian Armenia, and they occur locally in the Western United States. Specimens from the above localities, except El Salvador, are illus- trated in figures 79—86. Basaltic “scoria” flow materials have been de- scribed by Williams (1942) from Crater Lake, Ore., where they overlie dacitic “pumice flows.” These are related to the eruptions that formed the caldera of Crater Lake. Ash-flow tuffs of intermediate composition seem to be dominantly glassy, and in those available for study devitrification has been rare. The tuffs from Costa Rica, Argentina, Russian Armenia, and most of those from Japan, are glassy. The “schmelz- tuffe” from El Salvador described by Weyl (1954b) are glassy. An examination of figures 79—86 seems to indicate that the glass shards of intermediate composition tend to be platelike. In several of the figures the shards are almost shredlike. Few of them show forms derived from the shattering of rounded bubbles. The disparity between the mineral grains recog- nizable under the microscope and the composition of the groundmass material has been described for the so—called andesite welded tuffs from Argentina (p. 45). The same disparity is strongly suggested by microscopic studies of other welded tuffs of inter- mediate type. ANDESITIC ASH FLOWS OF COSTA RICA A description presented by Williams (1952, p. 173—176) is the most detailed available of a typically andesitic welded tuff. His petrographic description of this rock is a valuable contribution to an under- standing of andesitic welded tufl's, so will be quoted in some detail. One of the specimens (figs. 85 and 86, this report) described by Williams was collected by Gabriel Dengo, transmitted to the authors, and later divided with Williams to supplement his own collections. In his description Williams says, One of Sefior Dengo’s specimens is a magnificent example of a thoroughly welded tufl“. The unaided eye sees abundant frayed disks of black glass, up to an inch in length and one- quarter as thick, set in an almost fluidal base of pale-gray glass shards, generally smaller, but equally drawn out and sinuous. Peppered throughout, and serving as nuclei around which the glassy lenses curve, are varicolored chips of andesite, up to half an inch in diameter, and smaller but more numerous phenocrysts, mainly of feldspar. A Rosiwal analysis shows that crystals total about 15 percent of the tufi‘. Among them, basic andesine-acid labradorite (12 per- cent) occurs in phenocrysts up to 2 mm long; smaller prisms of pale green diopsidic augite (2V : 55°) comprise 2 percent of the whole, and hypersthene forms less than 1 percent. Hornblende, biotite, and olivine are lacking. Lithic fragments of varitextured pyroxene andesite amount to 5 percent. The remainder of the tufi" is made up of sinuous, tightly adpressed shards of clear glass varying in color from pale lilac and buff to deep brown, the darker ones corresponding to the black lenses seen by the unaided eye. None of the glass is devitrified, but some of the minute pores between the shards carry spheroids of cristobalite. Especially noteworthy is the wide range in the refractive indices of the glassy particles. In the palest wisps the in— dex is as 1.510 t .002; in the darkest, it increases to ap- proximately 1.550. Most of the glass varies in index from 1.530 to 1.540. Even more remarkable is the fact that al- though most of the individual shards are composed of uniformly colored glass, some reveal a fine banding of pale and dark glass of notably different indices. Williams (1952, p. 175) also states that another tuff in the area consists of 10 percent andesite frag- ments. In discussing the different types of glass in the tuff Williams says (1952, p. 176), =*>.’< in the Costa Rica tufl's the individual shards are usually of one color or another. This confused mingling of minute varicolored shards of widely differing refractive indices, presents a difficult problem for which a satisfactory answer has yet to be found. It cannot be attributed to successive eruptions of different magmas; on the contrary, magma of heterogeneous character seems to have effervesced simultaneously from the feeding vents. There is no sign here of solution or even dismemberment of the included lithic fragments * * If the heterogeneity of the Costa Rica tuffs resulted from contamination, the process must have taken place at depth and proceeded so far as to leave no indubitable evidence. DISTRIBUTION IN TIME Some of the notably large ash-flow deposits of the world, and others that have been most important in the development of an understanding of them, have been mentioned (p. 14, 23). A compilation of known localities would provide useful references for several lines of geologic study, but is not within the scope of this paper. Moreover, papers reporting newly discovered occurrences of welded tuffs, or the extension of previously recognized areas, are appearing in ever-increasing numbers, and any list of known localities would soon be out of date. On the other hand, some consideration of distribution DISTRIBUTION IN TIME 49 of welded tuifs in time seems to have significance, and for this reason a discussion of known pre- Tertiary occurrences of welded tufi's will be helpful. Volcanic activity has persisted throughout geo- logic time, and probably welded tuflfs have always been involved in that activity. A compilation of all recognized occurrences would show a very great preponderance of those in Tertiary and Pleistocene time with a remarkable decrease in their number in older rocks. No doubt their recognition will be precluded in many of these older rocks, but there is increasing evidence that detailed studies, backed by a knowledge of significant characteristics, will reveal a welded-tuff origin in many rocks that might otherwise be misinterpreted. The increasing in- terest in the geologic occurrence of welded tuffs has led to the recognition of several Precambrian oc- currences within the past few years. One of the major objectives of this report has .been to present the characteristics of welded tufi's, as seen in those rocks in which there has been little or no postdepositional modification, and to trace significant characteristics into tuffs whose origin has been obscured by different changes. The au- thors feel that an intimate knowledge of significant features will lead to identification of increasing numbers of welded-tufi' occurrences in older rocks. MESOZOIC The Brisbane tuff of Australia of Triassic age has been described by Richards and Bryan (1934‘, p. 51). A suite of specimens of the Brisbane tufl" was received through the courtesy of Bryan and one of these is shown in figure 75. A great thickness of welded tuffs of Triassic age occurs in Inyo County, Calif., and has been studied by Ward Smith, U.S. Geological Survey (written communication, 1950). A Cretaceous age is assigned to glassy welded tufi's from Montana, which were studied by Barksdale (1951). Welded tuffs studied by Solovev (1950, p. 211) from the southern Sikhote-alin region of Si- beria are said to be Late Cretaceous and early Ter- tiary in age. PALEOZOIC Tuff deposits of “Pelean” origin have been de- scribed from the Ordovician of Snowdon in Wales by Dakyns and Greenly (1905), and by Williams (1927). Oliver (1954) has described and given excellent illustrations of “welded tufts” in the Lake district of England and in Wales. Bodenhausen (1955) has given detailed descrip- tions of welded tuffs from the lower Carboniferous of Corsica. His illustrations present many of the relations that characterize devitrified welded tuffs. Some of these show a large proportion of lenslike areas with coarse-grained devitrification products, clearly representing areas that were originally pumice. Ross (1958) has studied volcanic materials that occur in drill cores from a deep well (4,000 feet) in Clinch County, southeastern Georgia. One horizon in the well is characterized by welded-tuff frag- ments representing materials derived from some densely welded and devitrified tuff body. Above that is about 35 feet of material which was origi- nally a glassy welded tufl" that has been replaced by laumontite. Applin (1951, p. 9) concludes that these tufi's are probably of early Paleozoic age. Vol- canic materials are widespread in southeastern Georgia and Florida, and welded tuffs will probably be recognizable in other localities of the region. Weyl (1954b, p. 28, 29) has considered the possi— bility of welded tuffs in central Europe and says, The wide distribution and large quantity of younger welded tuffs pose the question of whether such rocks are not in existence also in the volcanic series of central Europe, but have escaped identification so far. Judging by the mag- matic-tectonic occurrence of the younger Welded tuffs and their chemistry, similar formations would be expected first of all in the region of an acid subsequent volcanism. Thus, they could be perhaps encountered in the system of sand— stones, shales, and conglomerates of the lower Permian (Rotliegendes) in the central mountains of Germany. [von Gaertner called attention to tufi‘s of this system in the Forest of Thuringia] * i‘. The porphyritic tuffs of the ‘Rotliegendes’ system include rocks, the description of which—mostly written in older times—fit completely the character of welded tufi's. Weyl also calls attention to the streaked texture, coarse bedding, platelike jointing, and poorly delin- eated stratification in the tuffs of the Sturmheide porphyry and Kickelhahn porphyry. He has also mentioned the porphyry tuifs of the Black Forest and Oden Forest that suggest derivation from flows. Nuées ardentes were mentioned in connection with the origin of Trass in Brohltal. PRECAMBRIAN Rocks of Precambrian age that retain evidence of welded-tuff origin are of special significance. For this reason Bouladon and Jouravsky (1954, 1955) have made an outstanding contribution in present- ing the first clear-cut evidence of ignimbrites of Precambrian age from the Anti-Atlas region of Morocco. The descriptions and illustrations pre- sented by these geologists show good welded-tufi“ structures. Through the courtesy of Bouladon the authors received an excellent suite of these rocks. 50 The feldspar phenocrysts have been altered, but the shard structures show uncommonly good preserva- tion. Traces of the parallel growth of devitrifica- tion structure in the individual shards (axiolitic structure) are still observable. A photomicrograph of one of these specimens is illustrated in figure 98. Ignimbrites of Precambrian age have been de- scribed from the Dalarna region of central Sweden by Hjelmqvist (1956). His illustrations present excellent examples of welded-tuif structures. Sev- eral of these show marked distortion, but without loss of clear evidence of their true origin. The dis- torted welded tufl‘.‘ illustrated in figure 78 is derived from this same group of welded tuffs of Pre- cambrian age (Hjelmqvist, written communication, 1956). The recognition of the character of this specimen caused Ross to communicate with Prof. P. Geiger concerning the occurrence of welded tuffs in Sweden. (See footnote by Hjelmqvist, 1956, p. 9). Ross has identified welded tufi's from two deep ASH—FLOW TUFFS wells in Texas, one, No. 4—B Masterson well, from near the Canadian River, Potter County, at a depth of 1,952 feet; and the other from the Phillips No. l—A Stevens well, Bailey County, Tex., at a depth of 8,226 feet. A valuable paper on the basement rocks of Texas and southeast New Mexico has been pub- lished by P. T. Flawn (1956). This intimate knowl- edge of the basement rocks of Texas enabled Flawn to give the authors information about the geologic position of the aforementioned welded tuffs. He wrote (written communication, 1956) : I believe that the rhyolitic rocks penetrated in both these wells are late Precambrian age and part of an extensive late Precambrian volcanic terrane which I have called the Panhandle volcanic terrane. I suspect that in a rhyolitic volcanic terrane as extensive as this one you would expect to find welded tuffs *. Thompson and Williams (1956) have described glowing avalanche deposits of early Precambrian age in the Sudbury basin of Canada. SELECTED REFERENCES Abich, Herrmann, 1882, Geologische Forschungen in den kaukasischen Landern (Geologie des Armenischen Hoch- landes): Wien, Alfred Holder, v. 2, 210 p. Akkersdijk, Ir. M. E., 1941, Volcanic phenomena during the months of October, November, and December 1939: Netherlands Indies Volcanol. Survey Bull., no. 90, p. 55—67. Anderson, C. A., 1933, Tuscan formation of northern Cali- fornia, with a discussion concerning the origin of vol— canic breccias: California Univ. Pubs. Dept. Geol. Sci. Bull., v. 23, no. 7, p. 215—276. 1950, The 1940 E. W. Scripps cruise to the Gulf of California, pt. 1, Geology of islands and neighboring land areas: Geol. Soc. America Mem. 43, 53 p. Anderson, C. A., and Russell, R. D., 1939, Tertiary forma- tions of northern Sacramento Valley California: Calif. Jour. Mines and Geology, v. 35, no. 3, p. 219—253. Anderson, Tempest, and Flett, J. S., 1902, Preliminary re- port on the recent eruption of the Soufriére in St. Vincent, and of a visit to Montagne Pelée in Marti- nique: Royal Soc. [London] Proc., V. 70, p. 423—455. 1903, Report on the eruptions of the Soufriére in St. Vincent in 1902, and on a visit to Montagne Pelée, in Martinique. Part 1: Royal Soc. [London] Philos. Trans., ser. A, v. 200, p. 353—553. 1908, Report on the eruptions of the Soufriere in St. Vincent in 1902, and on a visit to Montagne Pelée in Martinique. Part 2, The changes in the districts and the subsequent history of the volcanoes: Royal Soc. [London] Philos. Trans., ser. A, v. 208, p. 275—302. Applin, P. L., 1951, Preliminary report on buried pre— Mesozoic rocks in Florida and adjacent States: U.S. Geol. Survey Circ. 91, 28 p. Bailey, E. B., and others, 1924, Tertiary and post-Tertiary geology of Mull, Loch Alene and Oban: Scotland Geol. Survey Mem., Expl. Sheets 43, 44, 51, and 52. Barksdale, J. D., 1951, Cretaceous glassy welded tufi's— Lewis and Clark County, Montana: Am. Jour. Sci., v. 249, p. 439—443. Bemmelen, R. W. van, 1930, The origin of Lake Toba (North Sumatra): Pacific Sci. Cong., 4th, Batavia-Bandoeng (Java) 1929, Proc., V. 2a, p. 115—124. 1932, Geologische waarnemingen in de Gajolanden (North Sumatra): Jaarb. mijnwezen Ned.—Indié, 59 Verh. 3, p. 71—94. 1939, The volcano—tectonic origin of Lake Toba (North Sumatra): De Ingenieur Ned-Indie, v. 6, no. 9, p. 126—140. 1949, The geology of Indonesia, v. 1: The Hague, Government Printing Office, p. 191—203. ’(Blyth, F. G. H., 1940, The nomenclature of pyroclastic de- posits: Bull. volcanol., ser. 2, v. 6, p. 145—156. Bodenhausen, J. W. A., 1955, Geologie de la region au nord du golfe de Porto (Corse): Amsterdam Univ. Geol. Inst. mededeel., no. 222, p. 11—98. Bonorino, F. G., 1944, Nota sobre la presencia de ignimbri- tas en la Argentina: La Plata, Univ. Nac. Mus., Notas v. 9, Geol. no. 35, p. 577—590. Bouladon, Jean, and Jouravsky, Georges, 1954, Les ignim- brites du Precambrian III de Tiouine et du sud maro- cain: Maroc, Service Geol., Notes et Mém., no. 120, p. 37—59. 1955, Les Gisements de manganese volcanogenes de Tiouine (Infracambrien du sud marocain): Maroc, Service Geol., Notes et Mém., no. 127, p. 5—180. Bowen, N. L., 1934, Viscosity data for silicate melts: Am. Geophys. Union Trans., 15th Ann. Mtg., pt. 1, p. 249— 255, Natl. Research Council. Bowen, N. L., and Tuttle, O. F., 1953, Beginning of melting of some natural granites: Carnegie Inst. Washington Yearbook no. 52, p. 50. SELECTED REFERENCES 51 Boyd, F. R., 1954, The Yellowstone rhyolite plateau: Car- negie Inst. Washington Yearbook no. 53, p. 139—140. Boyd, F. R., and Kennedy, G. C., 1951, Some experiments and calculations relating to the origin of welded tuffs [abs.]: Am. Geophys. Union Trans., v. 32, p. 327—328. Callaghan, Eugene, 1939, Volcanic sequence in the Marys- vale region in southwest-central Utah: Am. Geophys. Union Trans., 20th Ann. Mtg, pt. 3, p. 438—452. Conway, Martin, 1907, Discussions: Geographica Jour., v. 30, p. 101—102. Curtis, G. H., 1954, Mode of origin of pyroclastic debris in the Mehrten formation of the Sierra Nevada: California Univ. Pub. Dept. Geol. Sci. Bull., v. 29, no. 9, p. 453—502. Dakyns, J. R., 1900, First fruits of a geological examination of Snowdon: Geol. Mag. [Great Britain], 4th ser., v. 7, p. 267—273. Dakyns, J. R., and Greenly, Edward, 1905, On the probable Pelean origin of the felsitic slates of Snowdon, and their metamorphism: Geol. Mag. [Great Britain], 5th ser., v. 2, p. 541—549. Daly, R. A., 1933, Igneous rocks and the depths of the earth: New York, McGraw-Hill Book Co., 598 p. Dearnly, Ray, 1954, A contribution to the geology of Lod- mundarfjtirdur: Acta Nat. Islandica, v. 1, no. 9, p. 5—31. Dell’Erba, Luigi, 1892, Considerazioni sulla genesi del piperno: Gior. Mineralogia, Cristallografia and Petro- grafia, v. 3, p. 23—53. Diller, J. S., and Patton, H. B., 1902, The geology and petrography of Crater Lake National Park: U.S. Geol. Survey Prof. Paper 3, 167 p. Easton, N. Wing, 1894, Een geologische verkenning in de Toba-Landen. Bijdrage tot de geologie van Sumatra: Jaard. mijnwezen Ned. Oost-Indié, 23 wetensch. ged., p. 99—163. 1896, Der Toba—see, Ein Beitrag zur geologie von Nord-Sumatra: Deutsche geol. Gesell. Zeitschr., v. 48, pt. 3, p. 435—467. Eitel, Wilhelm, 1954, The physical chemistry of the sili- cates: Chicago, Univ. Chicago Press, 1592 p. 1956, Solid state reactions in silicate synthesis and glass production: Toledo Univ. Pub. Inst. Silicate Re- search Inf. Circ. no. 9, 13 p. [Free translation of series of papers by Carl Krtiger, 1952—55 as follows: Uber die Geschwindigkeit der zur Glaschmelze fuhrenden Reak- tionen: Glastechniche Ber., 1952, v. 25, p. 307—324; 1953, v. 26, p. 346—353; 1954, v. 27, p. 199—212; 1955, v. 28, p. 51—57, 89—98. Also supplemental discussion by Eite1.] Enlows, H. E., 1955, Welded tufi‘s of Chiricahua National Monument, Arizona: Geol. Soc. America Bull., v. 66, no. 10, p. 1215—1246. Escher, B. G., 1933a, On a classification of central eruptions according to gas pressure of the magma and viscosity of the lava: Leidsche geol. Mededee1., v. 6, p. 45—49. 1933b, On the character of the Merapi eruption in central Java: Leidsche geol. Mededeel., v. 6, p. 51—58. Feitler, Stanley, 1940, Welded tufl" resembling vitrophyre and pitchstone at Bare Mountain, Nevada [abs]: Geol. Soc. America Bull., v. 51, no. 12, pt. 2, p. 1957. Fenner, C. N., 1920, The Katmai region, Alaska, and the great eruption of 1912: Jour. Geology, v. 28, p. 569—606. 1923, The origin and mode of emplacement of the great tufl“ deposit in the Valley of Ten Thousand Smokes: Natl. Geog. Soc., Contrib. Tech. Papers, Katmai ser., no. 1, 74 p. Fenner, C. N., 1926, The Katmai magmatic province: Jour. Geology, v. 35, no. 7, pt. 2, p. 673—772. 1937, Tufl’s and other volcanic deposits of Katmai and Yellowstone Park: Am. Geophys. Union Trans., 18th Ann. Mtg., pt. 1, p. 236—239. 1948a, Incandescent tufl.‘ flows in southern Peru: Geol. Soc. America Bull., v. 59, p. 879—893. 1948b, Pleistocene climate and topography of the Arequipa region, Peru: Geol. Soc. America Bull., v. 59, p. 895—917. 1950, The chemical kinetics of the Katmai eruption: Am. Jour. Sci., v. 248, no. 9, pt. 1, p. 593—627; no. 10, pt. 2, p. 697—725. Fisher, N. H., 1954, Report of the subcommittee on vol— canology, 1951: Bull. volcanol., ser. 2, v. 15, p. 71—77. Flawn, P. T., 1956, Basement rocks of Texas and southeast New Mexico: Texas Univ., Bur. Econ. Geology Pub. no. 5605, 261 p. Fritsch, K. V., and Reiss, W., 1868, Geologische beschrei- bung der Insel. Tenerife: Winterthur, Verlag Von Wurster & Co., 494 p. George, W. 0., 1924, The relation of the physical properties of natural glasses to their chemical composition: Jour. Geology, v. 32, p. 353—372. Gilbert, C. M., 1938, Welded tufl" in eastern California: Geol. Soc. America Bull., v. 49, no. 12, p. 1829—1862. Gilluly, James, 1956, General geology of central Cochise County, Arizona: U.S. Geol. Survey Prof. Paper 281, 169 p. Goranson, R. W., 1932, Some notes on the melting of gran- ite: Am. Jour. Sci., 5th ser., v. 23, p. 227—236. Grandjean, J. B., 1931, Bijdrage tot de kennis der Gloed- wolken van den Merapi van Midden-Java: De. Mijnin— genieur, v. 12E, no. 12, p. 219—223. Griggs, R. F., 1917, The Valley of Ten Thousand Smokes: Natl. Geog. Mag, v. 31, p. 12—68. 1918a, Are the Ten Thousand Smokes real vol- canoes?: Ohio Jour. Sci., v. 19, p. 97—116. 1918b, The great hot mud flow of the Valley of Ten Thousand Smokes [Katmai district, Alaska]: Ohio Jour. Sci., v. 19, p. 117—142. 1918c, The Valley of Ten Thousand Smokes [Katmai district, Alaska]: Natl. Geog. Mag, v. 33, p. 115-169. 1919, The character of the eruption [of Katmai] as indicated by its effects on nearby vegetation: Ohio J our. Sci., v. 19, p. 173—209. 1921, Our greatest national monument [Katmai Na- tional Monument, Alaska]: Natl. Geog. Mag., v. 40, p. 219—292. 1922, The Valley of Ten Thousand Smokes [Alaska]: Washington, Natl. Geog. Soc., 341 p. Gutzwiller, Emil, 1912, Petrografische beschrijving der erup- tief gesteentan van het Goemaigebergte: Jaarb. mijn- Wezen Ned. Oost—Indié, v. 41, Verh., p. 50—86. Henderson, J., 1913, The geology of the Aroha Subdivision, Hauraki, Auckland: New Zealand Geol. Survey Bull. 16, new ser., p. 69—73. Hjelmqvist, Sven, 1956, On the occurrence of ignimbrite in the pre-Cambrian: Sveriges Geol. Undersfikning, Arsbok 49, no. 2., ser. 0., no. 542, p. 3—12. Holmes, Arthur, 1920, The nomenclature of petrology: Lon- don, Thomas Murby & Co., 284 p. 52 Holthausen, E., 1925, Beitrag zur Kenntnis der Petrographie des Geibiets des Tobasees in Nerd-Sumatra: Munster Univ. Dissert., 45 p. Hudson, F. S., 1951, Mount Lincoln-Castle Peak area, Sierra Nevada, California: Geol. Soc. America Bull., v. 62, p. 931—952. Iddings, J. P., 1885—86, Obsidian Cliff, Yellowstone National Park: U.S. Geol. Survey 7th Ann. Rept., p. 249—295. 1890, On a group of volcanic rocks from the Tewan Mountains, New Mexico: U.S. Geol. Survey Bull. 66, 34 p. 1899, Geology of the Yellowstone National Park: U.S. Geol. Survey Mon. 32, pt. 2, chap. 10, p. 356—430. 1909, Igneous rocks, v. 1: New York, John Wiley & Sons, 464 p. Iki, T., 1899, Report on the geology of the Volcano Aso: Earthquake Inv. Comm. Rept. no. 33, p. 3—40 [in Japanese]. Ingerson, Earl, 1955, Methods and problems of geologic thermometry: Econ. Geology, 50th anniversary volume, 1905—55, pt. 1, p. 341—410. . Ishii, K., Tanida, K., and Aoki, K., 1956, The preliminary report on the Yabekei welded tuff in Oita Prefecture:~ Japanese Assoc. Mineralogists, Petrologists, Econ. Geol- ogists Jour., v. 40, no. 4, p. 131—136. Ishikawa, Toshio, and Minato, Masao, 1955, Age of the welded tuf’l’s in northern Japan: Internat. Quaternary Assoc., 4th Conf., Rome-Pise, 1953, v. 1, p. 198—205. Jenks, W. F., and Goldich, S. S., 1956, Rhyolitic tuff flows in southern Peru: Jour. Geology, v. 64, p. 156—172. Jicha, H. L., Jr., 1954, Geology and mineral deposits of Lake Valley quadrangle, Grant, Luna, and Sierra Counties, New Mexico: New Mexico Bur. Mines and Mineral Re- search Bull. 37, 93 p. Kemmerling, G. L. L., 1932, De controverse uitgeschoten gloedwolken (nuées ardentes d’explosion dirigée) of lawinen gloedwolken (nuées ardentes d’avalanche): De. Ingenieur, v. 47, no. 13, p. 129—137. Kemp, J. F., 1896, Handbook of rocks: New York, D. Van Nostrand Co., 176 p. King, Clarence, 1878, Systematic geology: U.S. Geol. Explor. 40th Parallel, v. 1, 803 p. Kozu, Shukusuke, 1934, The great activity of Komagatake [Japan] in 1929: Tschermaks Mineralog. u. Petrog. Mitt., v. 45, p. 133—174. Kriiger, Carl, 1952—55 (See Eitel, Wilhelm, 1956, trans- lator). Kuch, Richard, 1892, Geologische studien in der Republik Colombia 1. Petrographie. 1. Die Vulkanischen Ges— teine: Berlin, Verlog von A. Asher & Co., p. 84—86. Kuellmer, F. J., 1954, Geologic section of the Black Range at Kingston, New Mexico: New Mexico Bur. Mines and Mineral Research Bull. 33, 100 p. Kuno, Hisashi, 1941, Characteristics of deposits formed by pumice flows and those by ejected pumice: Tokyo Univ. Earthquake Research Inst. Bull., v. 19, pt. 1, p. 144—149. 1950, Geology of Hakone volcano and adjacent areas, pt. 1: Tokyo Univ. Fac. Sci. Jour., sec. 2, v. 7, pt. 4, p. 257—279. 1952, Cenozoic volcanic activity in Japan and sur- rounding areas: New York Acad. Sci. Trans., ser. 2, v. 14, p. 225—231. Lacroix, Alfred, 1903a, L’éruption de la montagne Pelée en ASH-FLOW TUFFS janvier 1903: Acad. Sci. [Paris] Comptes rendus, v. 136, p. 442—445. Lacroix, Alfred, 1903b, Principaux résultats de la mission de la Martinique: Acad. sci. [Paris] Comptes rendus, v. 136, p. 871—876. 1904, La Montagne Pelée et ses éruptions: Paris, Masson et Cie, 662 p. 1906, Les conglomérates des explosions vulcaniennes de Vésuve: Acad. sci. [Paris] Comptesrendue, v. 142, p. 1020—1022. 1908, La Mortagne Pelée aprés ses éruptions, avec observations sur les éruptions du Vésuve en 1879 et en 1906: Paris, Masson et Cie, 136 p. 1930, Remarques sur les matériaux de projection des volcans et sur la genése des roches pyroclastiques qu’ils constituent: Soc. géol. France, Livre Jubliaire Cente- naire, 1830—1930, v. 2, p. 431—472. Lasky, S. G., 1947, Geology and ore dep0sits of the Little Hatchet Mountains, Hidalgo and Grant Counties, New Mexico: U.S. Geol. Survey Prof. Paper 208, 101 p. Macdonald, G. A., 1954, Activity of Hawaiian volcanoes during the years 1940—1950: Bull. volcanol., ser. 2, v. 15, p. 119—179. Macdonald, G. A., and Alcaraz, Arturo, 1956, Nuées ar- dentes of the 1948—1953 eruption of Hibok-Hibok: Bull. volcanol., ser. 2, v. 18, p. 169—178. MacGregor, A. G., 1938, The Royal Society expedition to Montserrat, British West Indies; The volcanic history and petrology of Montserrat, with observations on Mt. Pelée in Martinique: Royal Soc. [London] Philos. Trans., ser. B, v. 229, 90 p. 1946, Nuées ardentes and ignimbrites: Nature, v. 157, p. 305. 1952, Eruptive mechanisms—Mt. Peléc, the Soufriére of St. Vincent [West Indies] and the Valley of Ten Thousand Smokes [Alaska]: Bull. volcanol., ser. 2, v. 12, p. 49—74. 1955, Classifications of nuée ardente eruptions: Bull. volcanol., ser. 2, v. 16, p. 7—10. ‘IsMackin, J. Hoover, 1952, Hematite veinlets in an ignimbrite in the Iron Springs district, southwestern Utah [abs.]: Geol. Soc. America Bull., v. 63, no. 12, pt. 2, p. 1337— 1338. Mackin, J. Hoover, and Nelson, W. H., 1950, Early Tertiary welded tufl’s in the Iron Springs district, Utah [abs.] :' Geol. Soc. America Bull., v. 61, no. 12, p. 1528. Mansfield, G. R., and Ross, C. S., 1935, Welded rhyolitic tutfs in southeastern Idaho: Am. Geophys. Union Trans., 16th Ann. Mtg., pt. 1, p. 308—321, Nat. Research Coun-, cil, Aug. 1935. Marshall, Patrick, 1929, Building-stones of New Zealand: New Zealand Dept. Sci. Indus. Research Bull., no. 11, 37 p. 1932, Notes on some volcanic rocks of the North Island of New Zealand: New Zealand Jour. Sci. Tech- nology, v. 13, p. 198—202. 1935, Acid rocks of Taupo—Rotorua volcanic district: Royal Soc. New Zealand Trans., v. 64, pt. 3, p. 323—366. [Off—print pagination 1—44.] Marsters, V. P., 1912, The physiography of the Peruvian Andes with notes on early mining in Peru: New York Acad. Sci. Annals, v. 22, p. 225—258. SELECTED REFERENCES 53 Mathews, W. H., 1951, A useful method for determining ap- proximate composition of fine grained igneous rocks: Am. Mineralogist, v. 36, p. 92—101. Matumoto, Tadaiti, 1943, The four gigantic caldera vol— canoes of Kyusyu: Japanese Jour. Geology and Geog- raphy, v. 19, spec. number, 57 p. Matumoto, Tadaiti, Ishikawa, T., and Minato, Masao, 1956, Some problems of welded lava and welded tufl" related with sunken calderas in Japan: Pacific Sci. Cong., 8th, Philippines 1953 Proc., v. 2, pt. 1, p. 130—138. Moody, C. L., 1949, Mesozoic igneous rocks of northern Gulf Coastal Plain: Am. Assoc. Petroleum Geologists Bull., v. 33, p. 1410—1428. Moore, B. N., 1934, Deposits of possible nuée ardente origin in the Crater Lake region, Oregon: Jour. Geology, v. 42, no. 4, p. 358—375. Oliver, R. L., 1954, Welded tuf’fs in the Borrowdale volcanic series, English Lake district, with a note on similar rocks in Wales: Geol. Mag. [Great Britain], v. 91, no. 6, p. 473—483. Osborne, G. D., 1950, The Kuttung vulcanicity of the Hunter-Karuah district, with special reference to the occurrence of ignimbrites: Royal Soc. New South Wales Jour. and Proc., v. 83, pt. 4, p. 288—301. Perret, F. A., 1937, The eruption of Mt. Pelée 1929—1932: Carnegie Inst. Washington Pub. 458, 126 p. 1950, Volcanological observations: Carnegie Inst. Washington Pub. 549, 162 p. Pirsson, L. V., 1915, The microscopical characters of vol- canic tuffs—a study for students: Am. Jour. Sci., ser. 4, v. 40, p. 191—211. Ransome, F. L., 1898, Some lava flows of the western slope of the Sierra Nevada, California: U.S. Geol. Survey Bull. 89, 74 p. Reynolds, D. L., 1954, Fluidization as a geological process, and its bearing on the problem of intrusive granites: Am. Jour. Sci., v. 252, p. 577—613. Richards, H. C., and Bryan, W. H., 1934, The problem of the Brisbane tufl": Royal Soc. Queensland Proc., v. 45, no. 11, p. 50—62. Richey, J. E., and Thomas, H. H., 1930, The geology of Ardnamurchan North-west Mull and Coll: Scotland Geol. Survey Mems. 51 and 52, p. 370. Rickmers, W. R., 1913, The Daub of Turkestan: Cambridge [England] Univ. Press, 563 p. Rittman, A., 1952, Nomenclature of volcanic rocks: Bull. volcano]. ser. 2, V. 12, p. 75—102. Rosenbusch, H., 1896, Mikroskopiche Physiographie der Massigen Gesteine, v. 2, 3d ed.: Stuttgart, E. Schweizer- bart’sche Verlagshandlung, p. 766. Ross, C. S., 1941, Origin and geometric form of chalcedony- filled spherulites from Oregon: Am. Mineralogist, v. 26, p. 727—732. 1950, The dark—field stereoscopic microscope for min- eralogic studies: Am. Mineralogist, v. 35, p. 906—910. 1958, Welded tufl" from deep-Well cores from Clinch County, Georgia: Am. Mineralogist, v. 43, nos. 5 and 6, p. 537—545. Ross, C. S., and Smith, R. L., 1955, Water and other vola- tiles in volcanic glasses: Am. Mineralogist, v. 40, p. 1071—1089. Rutley, F., 1886, Notes on alteration induced by heat in cer- tain vitreous rocks: Royal Soc. London Proc., v. 40, p. 430—441. Shepherd, E. S., 1938, The gases in rocks and some related problems: Am. Jour. Sci., 5th ser., v. 35-A, p. 311—351. Sollas, W. J., and McKay, Alexander, 1905, Rocks of Cape Colville Peninsula, Auckland, New Zealand, v. 1: Well- ington, John McKay, p. 123. Solovev, S. P., 1950, The principal features of the complex of young acid effusives and ignimbrites in southern Sikhote-alin, and its petrochemical characteristics: Vses. Mineralog. Obshch. Zapiski, v. 79, no. 3, p. 211—222 [in Russian]. Sosman, R. B., 1927, The properties of silica: New York, The Chemical Catalog Co., 856 p. Stearns, H. T., Crandall, Lynn, and Steward, W. G., 1938, Geology and ground-Water resources of the Snake River Plain in southeastern Idaho: U.S. Geol. Survey Water- Supply Paper 774, 268 p. Stehn, Ch. E., 1935, Volcanic phenomena during the months of October, November, and December 1934: Netherlands Indies Volcanological Survey Bull. no. 70, p. 119—133. Swineford, Ada, and Frye, J. C., 1946, Petrographic com- parison of Pliocene and Pleistocene volcanic ash from western Kansas: Kansas Geol. Survey Bull. 64, pt. 1, p. 1—32. Taylor, G. A., 1954, Volcanological observations Mount Lamington, 29th May 1952: Bull. volcanol., ser. 2, v. 15, p. 81—97. ‘ 1956, An outline of Mount Lamington eruption phenomena: Pacific Sci. Cong., 8th, Philippines 1953, v. 2, pt. 1, p. 83—88. Thompson, J. E., and Williams, Howel, 1956, The myth of the Sudbury lopolith [abs]: Internat. Geol. Cong., 20th, Mexico City, Resumenes de los trabajos presentados, p. 300. Thorarinsson, S., 1955, Discussions: Bull. volcanol., ser. 2, v. 16, p. 11—13. Tilley, C. E., 1922, Density, refractivity, and composition relations of some natural glasses: Mineralog. Mag., no. 96, v. 19, p. 276—294. Tryggvason, Tomas, and White, D. E., 1955, Rhyolitic tufis in Lower Tertiary plateau basalts of eastern Iceland: Am. Jour. Sci., v. 253, p. 26—38. Tsuya, Hiromichi, 1930, The volcano Komagatake, Hokkaido, its geology, activity, and petrography: Tokyo Univ. Earthquake Research Inst. Bull., v. 8, p. 238—270. Turner, H. W., 1894, The rocks of the Sierra Nevada: U.S. Geol. Survey 14th Ann. Rept., pt. 2, p. 435—495. 1896, Further contributions to the geology of the Sierra Nevada: U.S. Geol. Survey 17th Ann. Rept., pt. 1, p. 521—762. Verbeek, R. D. M., 1884, Verslagen over de Uitbarsting van Krakatau: Batavia, Landsdrukkerij, 546 p. Volsing, K., 1910, Der Trass des Brohltales: Jahrb. Konig, Preuss. Geol. Landesanst., v. 28 (1907). Washington, H. S., 1896, Italian petrological sketches, pt. 2: Jour. Geology, v. 4, p. 826—849. Wentworth, C. K., and Williams, Howel, 1932, The classi- fication and terminology of the pyroclastic rocks: Natl. Research Council Bull. 89, p. 19—53. Westerveld, J., 1942, Welded rhyolitic tuffs or “ignimbrites” in the Pasoemah region, West Palembang, South Sumatra: Leidsche Geol. mededeel., v. 13, p. 202—217. 1947, On the origin of the acid volcanic rocks around Lake Toba, North Sumatra; a further contribution to the knowledge of Welded rhyolite tufi’s deposited along 54 ASH—FLOW TUFFS the Sumatran longitudinal fault-trough system, and a critical examination of data concerning the formation of the great Toba cauldron: Koninkl. Akad. Neder- landse Wetensch., Verh., Afd. Natuurk., sec. 2, v. 43, no. 1, p. 1—51. Westervelt, J., 1952, Quaternary volcanism on Sumatra: Geol. Soc. America Bull., v. 63, p. 561—594. 1955, Phases of Neogene and Quaternary volcanism in Asia Minor [abs.]: Internat. Geol. Cong, 20th, Mexico City, Resumenes de los trabajos presentados, p. 21—22. Weyl, Richard, 1954a, Glutwolken, Gluttufl‘e und Schmelz- tufl‘e: Natur u. Volk, v. 84, no. 8, p. 275—282. 1954b, Beitrage zur geologie E1 Salvadors, V. Die Schmelztufl‘e der Balsamkette: Neues Jahrb., Geologic u. Palaontologie Abh., v. 99, pt. 1, p. 1—32. Wichmann, Arthur, 1897, Petrographische studien uber den Indischen Archipel [abs.]: Neues Jahrb., Mineralogie, Geologie, u. Palantologie, v. 1, p. 282—284. Wilkinson, W. D., 1950, Welded tufl’ member of the Rattle— snake formation [abs.]: Geol. Soc. America Bull., v. 61, no. 12, p. 1534. Williams, Howel, 1922, The igneous rocks of the Capel Curig district, (North Wales): Geol. Soc. Liverpool Proc., v. 13, p. 166—206. 1926, Notes on the characters and classification of pyroclastic rocks: Geol. Soc. Liverpool Proc., v. 14, p. 223—248. 1927, The geology of Snowdon (North Wales) : Geol. Soc. London Quart. Jour., v. 83, p. 346—431. 1941a, Calderas and their origin: California Univ. Pub., Dept. Geol. Sci. Bull., v. 25, no. 6, p. 239—346. 1941b, Volcanology: Geol. Soc. America 50th anni— versary volume, p. 367—390. 1942, The geology of Crater Lake National Park, Oregon, with a reconnaissance of the Cascade Range southward to Mount Shasta: Carnegie Inst. Washing— ton Pub. 540, 162 p. 1952, Volcanic history of the Meseta Central Occi- dental, Costa Rica: California Univ. Pub. Dept. Geol. Sci., v. 29, no. 4, p. 145—180. ' Williams, Howel, Curtis, G. H., and Juhle, Werner, 1956, Mount Katmai and the Valley of Ten Thousand Smokes, Alaska (a new interpretation of the great eruptions of 1912) [abs.]: Pacific Sci. Cong, 8th, Philippines 1953, Proc., v. 2, pt. 1, p. 129. Williams, Howel, and Meyer-Abich, Helmut, 1955, Volcanism in the southern part of E1 Salvador—with particular reference to the collapse of basins of Lakes Coatepeque and Ilopango: California Univ. Pub. Dept. Geol. Sci. Bull., v. 32, no. 1, p. 1—64. Wilson, I. F., and Rocha, V. S., 1948, Manganese deposits of the Talamantes district near Parral, Chihuahua, Mexico: U.S. Geol. Survey Bull. 954—E, p. 181—208. Wilson, I. F., and Veytia, Mario, 1949, Geology and manga— nese deposits of the Lucifer district northwest of Santa Rosalia, Baja California, Mexico: U.S. Geol. Survey Bull. 960—F, p. 177—233. Wolf, Theodor, 1878, Der Cotopaxi und seine letzte Erup- tion an 26 Juni 1877: Neues Jahrb. Mineralogie, Geol- ogie u. Palantologie, p. 113—167. Work, L. T., and others, 1949, Dynamics of fluid-solid sys- tems. Fifteenth Annual Chemical Engineering Sym- posium: Indus. and Eng. Chemistry, v. 41, no. 6, p. 1099—1250. Wyllie, P. J., and Tuttle, O. F., 1957, The effect of volatile materials on the melting temperatures of silicates [abs.]: Am. Geophys. Union Trans, v. 38, no. 3, p. 413—414. Yagi, Kenzo, 1956, Welded tufi" of North Japan: Geol. Soc. Japan Jour., v. 62, no. 730. Yamaguti, Kamaji, 1938a, Petrological study of “ash stone” around the bay of Kagosima, pt. 2: Geol. Soc. Japan Jour., v. 45, p. 227—246. - 1938b, On the several rocks found at the walls of A50 Somma: Volcanol. Soc. Japan Bull., v. 4, p. 34—44. Zambonini, F., 1919, Il tufo pipernoide della campania e i suoi minerali: Italy, R. Comitato Geologico Mem., v. 7, pt. 2, p. 1—130. Zavaritsky, A. N., 1946, On the Quaternary volcanic tuffs of Armenia: Akad. Nauk SSSR Doklady, v. 53, no. 8, p. 729—731. 1947, The ignimbrites of Armenia: Akad. Nauk SSSR Izv., Ser. Geol., no. 3, p. 3—18 [In Russian with English summary]. Zies, E. G., 1929, The Valley of Ten Thousand Smokes; 1, The fumarolic incrustations and their bearing on ore deposition; 2, the Acid gases contributed to the sea during volcanic activity: Natl. Geog. Soc., contributed Tech. Papers [Katmai ser.], v. 1, no. 4, 79 p. Zirkel, Ferdinand, 1876, Microscopical petrography: U.S. Geol. Explor. 40th Parallel Rept. (King), V. 6, 297 p. 1893, Lehrbuch der Petrographie: 2d ed., v. 1, Leip- zig, Verlag von Wilhelm Englemann, 845 p. FIGURES 16—98 ASH-FLOW TUFFS FIGURE 16.—Glassy ash-flow tuff from South Sumatra collected by Westerveld. Characterized by platelike, slightly distorted shards. Note the absence of fracturing of the slender branches of the large shard near the right. An unusually massive shard lies in the upper third of the figure. FIGURE 18,—Glassy ash-flow tufl‘ from the Ammon quadrangle, southeastern Idaho, collected by Mansfield. The tuff is virtu- ally nonwelded, but there has been slight warping of the shards. There is an uncommon number of unbroken nearly circular glass—bubble walls. There is a wide range in the size of the shards and interstitial dust is present. Phenocrysts are completely absent. PHOTOMICROGRAPHS OF NONWELDED FIGURE 17.—Ash-flow tufi‘ from the Ammon quadrangle, south- eastern Idaho, collected by Mansfield. Composed of large platelike shards, which are completely nonwelded and with a large proportion of pore space (small white areas). Very coarse grained devitrification products (feldspar and cristobalite) have developed. A large feldspar phenocryst is on the right margin. FIGURE 19.—An ash-flow tufE from Iceland, received from Tomas Tryggvason. There may be slight welding, but there is» no distortion of the structures. A large euhedral plagio- clase crystal at the bottom of the figure has been embayed on its right‘ margin. The rounded area near the upper left corner is a pumice fragment with nearly round vesicles. Secondary alteration has resulted in the formation of a little bright-green celadonite. OR SLIGHTLY WELDED TUFFS ASH-FLOW FIGURE 20.~Ash-fiow tuff showing alignment of shards due to orientation during emplacement by flowage rather than to compression. Several Y-shaped shards are observable. At the extreme right is a large twinned plagioclase crystal, embayed along its left margin. TUFFS FIGURE 21.—The large central area represents a fragment of an older welded tuff enclosed in a younger tufi‘ material. In the older tuflf there is thorough welding and marked distortion of the shards. The younger tufi‘ is poorly welded, but its shards show evidence of alinement. The small white areas represent pore spaces. FIGURE 22.—Glassy welded tufi from South Sumatra collected by Westerveld (1942). This tuff is characterized by sharply angular shard fragments, many of which represent glass walls at the junction of several bubbles. The tuft is almost without phenocrysts. FIGURE 23.—Glassy ash-flow tuff from Barley Canyont Valles Mountains, N.Mex., collected by Ross and Smith. The tufE shows only slight welding, but platelike shards on the ex- treme right are bent against the fragment of quartz. A large irregular bubble wall on the left encloses a ground- mass of tuff fragments. PHOTOMICROGRAPHS OF SLIGHTLY WELDED ASH-FLOW TUFFS C7! ASH-FLOW TUFFS FIGURE 25,—Glassy welded tufl‘. Disruption during flowage or FIGURE 24.—Glassy tufl with a wide range in the size of the by explosive violence has resulted in a large proportion of shards and tmth much Interstltlal dust. Near the upper fragments that lack characteristic shard forms, but a shard left 15 a blOtlte crystal, but the Arequxpa hills are, for the representing the glass walls between three bubbles lies in the upper part of the figure. The dark area is one of most part, very low in phenocrysts. many fragments of andesitic rock. Very calcic plagioclase, also derived from alien invaded rocks, is present. FIGURE 26.—Collected at Kampong, Getang W. of Hoetambolon FIGURE 27.—Collected on the road to Sibolga near the water- Samosier Peninsula, Lake Toba. A glassy welded tufi' with fall Toroetoing Lake Toba region North Sumatra. A compressed shards molded against the phenocrysts. Most 1 t if .fl'] 1 I' h ld‘ ’ . t . f of the feldspar and quartz occurs as fractural angular gassy u WI ony 513 t we mg and no dls ortion 0 This speci- the shards, grains, but some show rounding by resorption. men illustrates an uncommonly crystal-rich welded tufi‘. Biotite (dark plates) is characteristic of the Sumatra tuffs. PHOTOMICROGRAPHS OF WELDED TUFFS FROM PERU AND SUMATRA collected by Fenner. These are the type of tufE called sillar by Fenner. Specimens for figures 24 and 25 are from the Arequipa region, Peru, Collected by N. Wing Easton in 1892 and part of a suite of Specimens for figures 26 and 27 are from the Lake Toha region, North Sumatra. specimens received from Zeiljlmans V. Emmichoven, and A. L. Simons of the East Indian Geological Survey. PHOTOMICROGRAPHS OF WELDED TUFFS SHOWING ONLY SLIGHT DISTORTION 552858761—i5 ASH-FLOW TUFFS FIGURE 28.—Welded tuif from 40 miles west of Crestone, Saguache County, Colo. The shards are largely glassy, but the narrow white margin marks incipient devitrification, whose materials are too fine to be resolved by the microscope. The shards are only slightly distorted, yet there has been slight plastic yield and very thorough accommoda- tion of each shard to its neighbor. Phenocrysts are sparse. IT} 1T1 FlGURE 29.—Tutf of the Rattlesnake formation from central Oregon, collected by William T. Pecora, US. Geological Survey. The individual shards show slight flattening, due to compression, but uncommonly perfect retention of the primary structure. Many of the unbroken glass bubbles are slightly distorted, and are now oval in outline. The shards are dominantly glassy, but local areas show partial devitrification. 59 AS H-FLOW FIGURE 30.—An obsidianlike welded tuft with perlitic cracks. There is only a vague retention of tuflf structure, now rep— resented by discontinuous horizontal lines, that suggest, but are not adequate proof of, welded tufi’. However, the geologic relations confirm such a derivation. FIGURE 32—17mm Leach Canyon near Iron Springs, Utah, collected by J. Hoover Mackin. Thorough welding and strong compression have only slightly impaired the shard structure. Devitriflcation has followed welding. A nearly euhedral biotite phenocryst lies in the upper left corner and one of plagioclase in the lower right. TUFFS FIGURE 31.—From Jamarillo Creek. Glassy welded tufE show- ing evidence of plasticity of the glass during welding and extreme distortion. Glass material has been squeezed into a deep embayment in a quartz crystal through an opening about 0.2 mm in diameter. FIGURE 33.—Welded tuft from Granite Mountain near Iron Springs, Utah, collected by Ross and Smith. The distortion is similar to that in figure 32, but the specimen in figure 32 is devitrified while this specimen has remained glassy. PHOTOMICROGRAPHS SHOWING DISTORTION OF WELDED-TUFF STRUCTURES Specimens for figures 30 and 31 from the Valles Mountains, N. Mex, were collected by Larsen, and Ross and Smith, respectively. ASH-FLOW TUFF‘S 61 FIGURE 34.—A fine-grained glassy welded tuff showing extreme FIGURE 35.wGlassy welded tufi from southeastern Idaho, col- compression and molding against an embayed plagioclase lected by Mansfield. Thoroughly welded tufi‘ with marked phenocryst. The shard structure is recognizable, although compression of the shards, but with good retention of the it simulates the flow structure of an extrusive rhyolite. structure. Note the flattened Y-shaped shard, upper left. FIGURE 36.—Glassy welded turf with extreme compression and FIGURE 37.-—A glassy HOW-banded rock from an intrusive dike, distortion of the shards. These are molded against the collected by Wilbur Burbank, U.S. G8010gi081 SUI‘VEY, from phenocrysts until there is a marked simulation of flow struc- the Twin Lakes tunnel, Lake County, 0010. Burbank be- ture, very similar to the specimen shown in figure 37, lieves that this represents a feeder for the immediately However, this specimen shows recognizable tufi structures overlying Welded tufis of the area. There is no vesiculation at the extreme right, and a distinct discontinuity of the and the flow lines, although nearly distorted, are more stretched shards that resemble flow lines. Represents a continuous than those in figure 36. tufl' with abundant phenocrysts. PHOTOMICROGRAPHS SHOWING COMPARISON OF WELDED TUFF AND FLOW STRUCTURES 62 ASH—FLOW TUFFS FIGURE 38.—Photomicrograph of welded tuff from the Valles Mountains, N.Mex., collected by Ross and Smith. Well- developed parallel arrangement of the shards in a thor- oughly welded tufi is shown. The shards were originally platelike and so the elongation is probably not due to stretching. FIGURE 40.—Photomicrograph of devitrified welded tufi‘ that has undergone very marked compression, but retains recog- nizable shard structure. Just below the upper right corner there is a greatly flattened bubble (marked with an arrow). To the right is a flattened Y-shaped shard. FIGURE 39.—Photomicrograph of glassy welded tufE from Lake Toba, northern Sumatra, between Pagaran Psang and Sekanan, along the road from Taroetoeng to Sibolga. Col- lected by N. Wing Easton, 1892. A thoroughly welded glassy tulf with large pumice fragments that have collapsed and all pore space eliminated, but with a retention of pumice structure. Dark areas represent biotite plates. FIGURE 41.—Fhotomicrograph of glassy specimen derived from the same welded-tuft" horizon as specimen shown in figure 40. Slightly greater stretching has eliminated all clearly recognizable shard structure, and the specimen resembles a flow rock. However, the lenses lack the continuity of typical flow rocks. Compare with figures 34 and 36. EXTREME DISTORTION OF THE WELDED-TUFF STRUCTURES Specimens for figures 40 and 41 were collected by J. Hoover Mackin from near Iron Springs, southeastern Utah. ASH-FLOW TUFFS FIGURE 42.—Photograph of a hand specimen showing nonwelded, uncollapsed pumice fragments that stand in relief on the weathered surface. The pumice retains about 70 percent of pore space. Compare With figure 1, which represents a completely welded. horizon from the same flow. FIGURE 43.——Photomicrograph of thin section of the same tufi shown in hand specimen in figure 42. Several noncollapsed glassy pumice fragments are shown. The white areas are uncommonly large glass shards. This tuft contains a large proportion of alien rock fragments and crystal grains. FIGURE 44,—Photomicrograph of glassy welded tufi' from un- known locality in the Western United States. On the right is a thoroughly welded shard structure. 0n the left is a large pumice fragment in which there has been complete collapse of the pore spaces and welding into a compact glass, but with retention of the pumice structure. Compare with the noncollapsed pumice shown in figure 43. PUMICE IN ASH-FLOW TUFFS Specimens shown in figures 42 and 43 are from the basal part of the Battleship Rock ash flow 5 miles north-northeast of Jemez Springs, N.Mex. 63 64 ASH-FLOW FIGURE 45.—Specimen from the Valles Mountains, N.Mex., collected by Ross and Smith. The large light-colored area represents a pumice fragment in which there has been complete collapse and elimination of internal pumice struc- ture, but with traces of the original pumice structure re- tained in the taillike projection extending toward the lower right corner and molded against a feldspar grain. Compare with the frayed ends of collapsed pumice frag- ments in figure 13. FIGURE 47.—From Redondo Peak, Valles Mountains, N.Mex. In the lower left corner is an area of welded tuff with much flattened shard structure (diagonal). The pumice that formed the remainder of the figure was much more plastic than the welded shards, and was distorted and molded around the welded-tuff part. The original pore space of the pumice has been completely eliminated but the pumice structure is retained. TUFFS «.y o :s‘. ‘ I $3.51 m m 3‘ FIGURE 46.AA glassy welded tuf‘E from Kuinimi, Kaga, Japan, U.S. National Museum specimen 61728. The left two-thirds of the figure represents ash material, and the right third a pumice fragment. The tufl’ has been welded into a com- pact glass and subsequently perlitic fractures have devel- oped. The collapsed pumice structure gives this fragment the appearance of a flow rock. FIGURE 48.~—From Redondo Peak, Valles Mountains, N.Mex. A large area of pumice that has completely lost its original porosity and has been molded against the quartz pheno- crysts and at the same time greatly distorted. PHOTOMICROGRAPHS SHOWING PLASTICITY OF SHARD AND PUMICE MATERIALS ASH—FLOW TUFFS FIGURE 49.—Hand specimen of a ledge-forming facies of the welded tufi‘ that crops out over an extensive area 8 to 10 miles east-southeast of Idaho Falls, Idaho, in the Ammon quadrangle. This facies (6 to 8 ft thick) is everywhere characterized by abundant lithophysal cavities that developed after complete welding and distortion of the shard structures. In this specimen the lithophysae are 2 to 3 cm in diameter. This bed is not recognizable as a welded tufi' from field relations, and only by microscopic study is its true character revealed. FIGURE 50.—Photomicrograph of thin section cut from a dense area in the specimen represented in figure 49. The tufE structure is clearly evident but seems to have been distorted while still plastic by the development of the closely spaced vesicular cavities shown in figure 49. FIGURE 5L~Photomicrograph of thin section of a specimen from the same horizon as figure 49, but collected 8 or 10 miles farther southeast. The vesicles are larger and more elongated (as much as 4 cm) and much more widely spaced. The tuff structure has undergone extreme flattening, but is still recognizable. Note flattened shards that were originally Y shaped. The structure of devitrification products cuts across the shard structure. DISTORTION OF TUFF STRUCTURE 65 66 ASH-FLOW TUFFS 2 l___...._.._l FIGURE 52.—--Photograph of a large hand specimen of a welded tuft. The dark deposition and complete welding: of the tuft. Their presence indicates the ash retained volatiles in solution that were released after welding and (luring partial devitrification. Two circular areas near the lower left part of the figure show where lithophysae with concentric shell structure have broken out. X 2/3. areas are vesicular cavities which developed after FIGURE 53.—Photomicrograph of thin section cut from the specimen shown in figure 52. The rock is dominantly glassy, but the dark circular area represents a spherulite that has FIGURE 54.—Photomicrograph of part of a very large spherulite. The differences in the devitrification products have caused three distinct concentric zones. The inner zone seems almost grown out from a plagioclase phenocryst acting as a nucleus. The feldspar-cristobalite intergrowth of the spherulite is not actually dark colored, but appears dark in the photograph because of the very strong dispersion of light resulting from the fine-grained intergrowth of materials with very different mean indices of refraction (K-Na feldspar:1.524, cristobal- ite: 1.486). black because of the very marked dispersal of transmitted light. Outside of that is a lighter zone where the ash-shard structure is observable, but it also shows in the third and outer zone. SPHERULITES AND LITHOPHYSAE SPECIMENS Collected by Mansfield in southeastern Idaho. ASH-FLOW TUFFS FIGURE 55.—Specimen collected half a mile south of Sulphur Springs, Jemez Springs quadrangle, Valles Mountains vol- canic region, New Mexico, by Ross and Smith. The dark areas retain the shard structures, but in the lighter areas radial aggregates of feldspar and cristobalite have developed and completely destroyed the shard structure. The left third of the figure represents a large pumice fragment. FIGURE 57.-——A large pumice fragment imbedded in shard ma- terial in which the structure is still recognizable. The devitrification of the pumice grains has destroyed the original fibrous structure. Specimen collected by Ross and Smith, Valles Mountains, N.Mex. FIGURE 56.—The same area shown in figure 55, but under crossed nicols. This shows the coarse-grained intergrowth of feldspar and cristobalite in areas that have lost their shard structure (light) and the much finer structure in those areas where the structure has been retained (dark). FIGURE 58.—The same area shown in figure 57, but under crossed nicols. This shows the tendency, observable in almost all welded tui‘fs, for coarse-grained devitrification products (feldspar and cristobalite) to develop in the pumice fragments; and much finer grained ones in the shards. PHOTOMICROGRAPHS SHOWING DEVITRIFICATION 0F PUMICE 5’2.35876177*6 67 68 ASH-FLOW TUFFS FIGURE 59,—Welded tuflf that has undergone welding and FIGURE 60.~Same area as illustrated in figure 59, but under extreme distortion of the still-recognizable shard struc- crossed nicols. This shows a coarse-grained radial ag— tures. gregate of feldspar and cristobalite laths which have de- veloped independently of the original shard structure. FIGURE 61.——Welded tuE from southeastern Idaho, collected FIGURE til—Welded tufi' from Redondo Peak, Jemez Springs fi 1d. All th h '1; -fi d, b t th 1- ht quadrangle, Valles Mountains, N.Mex., collected by Ross by Mans e e s ards are devx r1. e hllh e 1g and Smith. Welding and extreme flattening of the ash area represents a pumice fragment m w m coarser structure is represented at the extreme right. In the grained intergrowth of cristobalite and feldspar has de- central part of the figure (white) the development of veloped and where the pumice structure has been partly uncommonly coarse grained .devitrification products has completely destroyed the original structure. If the entire destroyed. rock had undergone the same loss of structure, evidence of derivation from a tufif would be lost. PHOTOMICROGRAPHS SHOWING DESTRUCTION OF STRUCTURE BY DEVITRIFICATION Specimens for figures 59 and 60 were collected by Ross and Smith from the Valles Mountains, N.Mex. ASH-FLOW TUFFS 69 FIGURE 63.—Welded tufl? from the Valles Mountains, N.Mex., FIGURE 64-—Welded tuff from the Valles Mountains, N.Mex., collected by Ross and Smith, showing a Very large spheru- Showing the development of groups of spherulites, with con- lite in which the radial aggregates of feldspar and cristo- centric banding due to variations in grain size of the balite are well represented. At the left the secondary aggregates of cristobalite and feldspar. minerals have developed a plumose structure. All direct evidence of tufE structure has been destroyed. FIGURE 65.wWelded tulf from the Lake Toba region, Sumatra, FIGURE 66.—Welded mm? from the Valles Mountains, N.Mex., showing spherulitic areas developed in a welded tufi‘. collected by Ross and Smith. An open cavity in a welded tufl‘ in which platelike crystals of tridymite have developed. PHOTOMICROGRAPHS OF DEVITRIFICATION MATERIALS Figures (33, 64, and 65 illustrate the complete destruction of the original tufi‘ structure. ASH-FLOW FIGURE (ST—Photograph of a polished section through a so- called thunder egg that represents a type of hollow spheru- lite which commonly develops in glassy volcanic rocks. Vola- tiles originally in solution in the glass collect at the center during the development of anhydrous minerals which form the gray borderA—the material shown in figures 68 and 69. The originally hollow central part was subsequently filled with chalcedony. X 1 FIGURE 69,-Photomicrograph of the outer rim material of figure 67. The section represented in this figure was cut parallel to the plane of deposition. Figure 68 and this figure shows the marked difference in flattening and distortion of the shard structure in the two planes; in both planes the shard structures are well preserved. Compaction and weld- inier was complete before devitrification and the parallel growth of feldspar and cristobalite (diagonal structure in this figure) developed independently of the original shard structure. Compare with shards in figures 717774, in which the development of feldspar and cristobalite started at the grain boundary of each shard. TUFFS FIGURE 68.—Photomicrograph of the outer rim material of figure 67. The section represented in this figure was cut perpendicular to the plane of deposition. FIGURE 70,—Photomierograph of welded tuff from southeastern Idaho, collected by Mansfield. The larger shards seem to have undergone very little distortion, but the others are molded against them giving a marked parallel structure. Welding was complete before devitrification and the feldspar- cristobalite formed parallel aggregates cutting across the shard structure (horizontal). DEVITRIFICATION CHARACTERISTICS Figures 67, 68, and 69 represent material from central Oregon, collected by Gen. J. S. Hatcher, U.S. Army, retired. ASH-FLOW TUFFS 71 FIGURE ‘71.—Thin section of tul'f showing welding and devitrifi- _ . . FIGURE 72.—An ' f h'n t' h ' fi 7 cation, but Without distortion of the shard structure. area rom t 1 sec ion 5 own in gure 1 under high magnification. The parallel intergrowth of feld- spar and cristobalite (axiolitic structure) is well repre- sented. The large shard fragment represents the walls of several bubbles. , tr 0.4 mm FIGURE 73.—Devitrified welded tuff. Most of the thin sec- FlGURE 74.—Devitrified welded tuflf from an unknown locality tion shows little distortion of the original shard structure, in the Western United States showing uncommonly perfect but at the right side of the figure there is marked molding axiolitic structure. Although the devitrification products of the shards around a feldspar phenocryst (white). (feldspar and cristobalite) are uncommonly coarse grained, the original forms of the shards are perfectly retained. PHOTOMICROGRAPHS OF WELDED TUFFS SHOWING AXIOLITIC STRUCTURE Specimens for figures 71, 72, and 73 are from the Valles Mountains, N.Mex.; collected by Ross and Smith. Axiolitic structure is the result of crystal growths, beginning at the shard boundaries, and meeting at a line of discontinuity (dark central line). The high relief is due to the large differ- ence in index of refraction of the devitrification productsicristobalite and feldspar. Compare with figure 95 which is a reproduction of a sketch by Zirkel illustrating the structure he named “axiolitic.” ASH-FLOW TUFFS (‘J FIGURE Fri—waded tug fmm Ausm‘ha’ confined. by E‘Chards FIGURE 76.—Devitrified welded tuff from Glashutte, Schermitz, and Bryan (1934). The shard structure is discernible, al- . . . , ‘ . . . . Hungary; U.S. National Museum specimen 36270. I‘he shard though it has been partly obscured by devntrlfication. Other . _ f- th B“ b t [1‘ 'h t 1 f structures are obscured, but not entirely lost. An euhedral, spec1mens 10m e “S ane u s b ow even grea er 055 0 high-temperature quartz phenocryst is near the bottom. structure. FIGURE 77.—Devitrified welded tuff from Pusan, Korea, col- FIGURE 78,—Thin section from an old German collection of Iected by Allen Nicol, US. Geological Survey. The shard slides, which was marked as representing a rock from Elf- structure is only dimly retained, but axiolitic structures are dalen Daloria, Sweden. Hjelmqvist states that these welded observable at the left. tufi‘s are sub-Jotnian (sub-Keweenawan) in age. The shard structure has undergone stretching so that parts of the thin section are indistinguishable from a flow rock, but, locally, shard structures are preserved. PHOTOMICROGRAPHS OF WELDED TUFFS FROM FOREIGN LOCALITIES ASH-FLOW FIGURE 79.7A large glassy pumice fragment imbedded in shards FIGURE 80.hFine-grained andesitic tuf‘f. Welding is moderate (lower left). Welding is slight, but the pumice fragment but there has been warping and accommodation of shards to seems to have undergone distortion. one another. The slide shows many rock fragments and crystals derived from an earlier andesitic rock. The large crystal (upper right) is augite. FIGURE 81.7Andesitic welded mi? from ASO caldera, Kyusyu, FIGURE 82.——Welded tul‘f from near McKays, Calaveras County, Japan, collected by C. G. Johnson, U.S. Geological Survey. Calif., collected by Smith. Illustrates the type of biotite- A thoroughly welded tuff, with abundant andesitic phenO- augite latite described by Ransome. The tulf structure is Cl‘yStS and l'OCk fragments. Augite crystals lie near the poorly preserved. This specimen represents a welded tuff center and lower right corner; an euhedral plagioclase crystal uncommonly rich in crystal fragments. Some of these ma- is near the lower left. terials represent fragments derived from older andesitic rocks. PHOTOMICROGRAPHS OF WELDED TUFFS OF INTERMEDIATE COMPOSITION Figures '79 and 80 are from Erevan, Russian Armenia; U.S. National Museum specimen 52079. 74 ASH-FLOW TUFFS FIGURE 83.—Photomicrograph of specimen collected from the FIGURE 84.—Photomicrograph of a thin section cut from the bank of the Malargue River at La Valenciana Mendoza, same specimen as figure 83, but perpendicular to the plane Argentina, by F. G. Bonorino, and described by him as of deposition. Note the greater flattening of the structure. andesite. The thin section was cut parallel to the plane of deposition. FIGURE 85.~Photograph of a hand specimen. The black areas FIGURE 86.~Photomicrugraph of a thin sction of specimen represent the larger pumice fragments which have collapsed, shown in figure 85. The dark lenses represent pumice frag- and lost the original pumice structure. X 1%. ments in which the original pumice structure has been largely destroyed. ANDESITIC WELDED TUFFS The tuflf shown in figures 83 and 84 is glassy and shows very complete welding and marked flattening of the shard structure. A large proportion of the crystal and rock fragments represent alien materials. The specimen for figures 85 and 86 illustrates andesitic welded tufi‘s from Costa Rica. collected by Gabriel Dengo, from the Nucstro Arno power plant, Virilla River. ASH-FLOW TUFFS 7 FIGURE 87.—Ph t gra hi d t' f a colored draw- FIGURE 88.—Photomicrograph of one of Iddings‘ thin sections ing by Iddingso $188§78g 11:19:: ugglog) 0 While this is not a of rocks from Yellowstone National Park. This evidently _ _ represents the same material shown in figure 87, but may welded tuf’E It 15 presented as a part Of the development 0f not have been the same thin section used by Iddings for the Iddings’ concept of welded pumice. drawmg Show“ in figure 87‘ FIGURE 89.—Photomicrograph slightly enlarged and copied from FIGURE 90.—Photomicrograph of a thin section from a group an illustration from Iddings’ study (1899, pl. 50, fig. 100) of f ' 1 ' . ' ~ . . rhyolitic rocks from Yellowstone National Park. The illus- o lspecimens 1:01 ected 1n the Valles Mountains, NMex, by tration is also shown in another report by Iddings (1909, Homes and owell and tuined “V“ to Iddings for study. fig. 22, p. 332') and bears the title “Welded pumice with The specimen from which this thin section was cut was de- an: 03:: internist art'sxgofsozmaifizi; scribed as “from Jew and welding, some distortion of the shards, molding against the ‘S from the same reglon now under study by the authors. quartz phenocryst, and devitrification; all with good retention of the shard structure. ILLUSTRATIONS SHOWING SOME OF THE ORIGINAL MATERIAL DESCRIBED BY IDDINGS AS WELDED PUMICE Figures 87 and 88 are from Obsidian Cliff, Yellowstone National Park, and represent some of the material called welded pumice by Iddings. This is evidently a brecciated and rewelded pumice from a rhyolitic lava flow. 76 ASH-FLOW TUFFS FIGURE 91.—Welded tuflf from Fall River basin, west slope of mountain north of the head of Conant Creek (Iddings, 1899, p. 378; specimen 1958). Glassy welded tuff with a large inclusion of andesite that seems to be typical of the Yellowstone National Park welded tui'fs. FIGURE 93.—Zirkel collection, specimen 22297; U.S. National Museum specimen 350. Locality described as “southeast from Wadsworth, Nevada.” Glassy welded tuif. Near the left margin there is bleaching of the outer borders of the shards. The platelike character indicates that the thin sec- tion was cut diagonally to the plane of deposition. FIGURE 92.—Iddings’ thin section 2150, not described by him. It belongs to a group from the northeastern corner of Yellowstone National Park. The dark part shows incipient devitrification and the light part is glassy. Thorough weld- ing and slight distortion of the shards is evident. FIGURE 94.—Zirkel collection, specimen 21834; U.S. National Museum specimen 499. Locality described as “Mopung Hills, W. Humboldt, Nevada.” The figure shows extreme flattening and stretching, but with good evidence of shard structures. The lenses show marked discontinuity, and in many places there are nearly parallel branches of a lens, due to flattening of a Y-shaped shard. Areas where very flat lenses are folded back upon themselves is a characteris- tic of much—compressed welded tuifs. PHOTOMICROGRAPHS OF WELDED TUFFS OF HISTORIC INTEREST Figures 91 and 92 are from Iddings’ original set of thin sections of “pumice tufts” from Yellowstone National Park. Figures 93 and 94 are re- produced from a very large collection of rocks studied by Zirkel as a part of the 40th Parallel Report; a large proportion of this collection proved to be welded tufis, and so these thin sections made about 80 years ago have now provided new evidence of the abundance of welded tuffs in the Basin and Range region. ASH-FLOW TUFFS FIGURE 95.—Photographic reproduction of a drawing pre- sented by Zirkel (1876, pl. 7, fig. 4). Zirkel did not recognize the full significance of the welded tulfs he ob- served, but this illustration shows that he recognized the essential petrographic characteristics. The parallel inter- growths of feldspar and cristobalite, for which he proposed the name “axiolitic” are well illustrated. Compare with these structures shown in figures 87790. FIGURE 97,—Photomicrograph of welded tufE from Taupiri Gap, Waikato River, North Island, New Zealand, collected by Ross. This represents the materials which were the basis for Marshall’s (1935) outstanding contribution to an understanding of ash-flow tuffs. This ml? is welded and devitrified, and there has been dimming by devitrification, but little distortion of the tufl’ structure. The new Zea- landttuifs contain abundant phenocrysts of feldspar and quar z. FIGURE 96.—Photomicrograph of ash-flow tuff from the Valley of Ten Thousand Smokes, collected by Zies. Specimens of these tufi‘s available to the authors are nonwelded, but platelike shards which were bent against phenocrysts have been observed. Little induration is shown and the porosity is high. There is a wide range in size of materials, with dust, shards, and pumice fragments (extreme right) all present. A circular unbroken glass bubble lies near the center and another near the upper right corner. In the upper part of the figure there is a large irregularly shaped glass shard. FIGURE 98.—Photomicrograph of thin section of a welded and devitrified tuff cut from one of several specimens submitted by Jean Bouladon. These specimens were collected by Bouladon and Jouravsky, and described by them (1954), from the Anti-Atlas region of Morocco. Their description represents the first recorded recognition of a welded tufi‘ of Precambrian age. As can be seen from this thin sec- tion, the structure is remarkably good and traces of the axiolitic structure are recognizable. Compare with this structure in figures 71774. CONTRIBUTIONS TO THE WELDED-TUFF CONCEPT 77 Agglomerate lava Alaquez massif . Altaville, Calif . Amphiboles Anderson, Tempest, and Flett, J. S., quoted 15 Andesite ________ 10, 35, 39, 44, 45, 48, 58, 73, 74, 76 Arequipa region, Peru 14, 28, 36, 58 Ash, definition ________________ 3 Ash, volcanic, definition 3, 4, '7 Ash fall, definition ......... 3 Ash-fall materials, welding 5 Ash-fall tufi', occurrence _. 21 recognition 23, 37, 38 Ash flow, definition ,,,,,,,,,,,, 3 nonwelded ______ . 17, 18, 27 study of .. 1, 2, 15 temperature 16, 21 velocity Ash-flow deposits, thickness _ 2, 20, 22 used as stratigraphic marker ________________ 14, 23 Ash-flow deposits in foreign localities, Africa Argentina Asia Minor . Australia __ Queensland Canada, Sudbury Basin central Europe Corsica .............. Island of Scandola Costa Rica ....................... 48, 74 Ecuador 16 El Salvador England Hungary __ ,, Iceland ..... 2, 15, 56 Indonesia 5 Italy 9 Phlegrean Fields ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 7 Japan ...................... 2, 5, 7, 47, 48, 73 Kuinimi, Koga . ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 64 Kyushu 3, 4 Korea ..... 72 Mexico ,, 2 Guerrera .. 29 Morocco 2, 37, 49, 77 New Zealand, 2, 13, 14, 23, 24, 28, 36, 41, 46 North Sumatra, Lake Toba region ,,,,,,,,,,,,,,,,,,,,,,,, 2, 28, 58, 62, 69 Peru ,,,,,,,,,,,,,,,,,,, . 2, 5, 35, 36 Russian Armenia 8, 23, 28, 48, 73 Siberia ............... A 7 ,7 49 South Sumatra 56, 57 West Pasoemah region 2 Sumatra . 23, 35 Sweden . 50, 72 Wales _. ,,,,,,,,,,,,,,,,,,,,, 49 Ash-flow materials, criteria for recognition , 2, 18 dispersal by fiowage 41 iron content A 31, 45 nomenclature 2-3 origin ............. 2, 12, 16, 27 INDEX Ash-flow materials—Continued welding .................................................... 5, 24—26 Ash-flow tufi, areal extent 2, 13, 14, 16, 17, 22—23, 29 cooling history ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 24, 26 crystal content 8, 56, 57 degree of sorting ,,,,,,,,,,,,, 19 definition . . 3 density _ 24 development of concept 9, 75 erosion patterns 20, 21, 24, 29, 30 gentle dips _____ 23, 24 joint pattern mapping ,,,,,,,,,,,,,,,,,,,,,, 2 mechanisms of deposition 15, 17, 38, 39 mineralogy ,,,,,, 21, 27 origin ......... 2, 8, 17 percentage of rock fragments 8, 18, 21, 24 permeability ,,,,,,,,,,,,,,,,,,,, 40 porosity ...... 24, 25, 26, 36, 3‘7, 46—47, 57, 62, 63, 64 pyroclastic character ____________ 18—19, 32, 33, 37 sorting ,,,,,,,,,,,,,,,,,,,,,, 21, 22 stratification 16, 19, 20—22 texture _. ______________ 21 thickness 13, 20, 26 volume . 14, 17, 20 welded, occurrence . 17, 24 recognition . 18, 24, 30 welding-__ 12, 17, 24—26, 33—35 Aso caldera . 3, 4, 23 Aso lava, areal extent _ definition ______________ _ 3 petrologic characteristics 24 use of term _________________________ 4, 14 Augite ..................................... Auvergne region, France . Avalanche, volcanic, definition Avalanche deposits, glowing 1, 4, 7, 14, 17 Axiolite, definition ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4 Axiolitic structure, description and definition ,,,,,,,,,,,,,,,,,,,,,,,,,,, 4, 9, 36, 71 B Bandelier formation .................................................. 30 Bare Mountain, Nev ................................. 14 Battleship Rock, N. Mex W 19, 28, 46, 47, 63 Biotite ,,,,,,,,,,,,,, 30, 35, 44, 58, 60, 62 Bishop tufi' 13, 14, 20, 23, 28, 42, 46 Block and ash flow, use of term '7 Blocks, definition .............................. , 4 Blyth, F. G. H., quoted _ 3, 4, 5 Boyd, F. R., quoted ,,,,,,,,,,,,,,, 42 Breccia cemented fused welded ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Brisbane tufi‘ ,,,,,, Bubble fractures . 72 C Calderas, Aso __________________________ collapse, Krakatau type .14 relation to welded-tuflf formation ._ 17, 18 Canary Islands, Tenerife ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 9 Celadonite Chalcedony 70 Chiricahua National Monument, Ariz ,,,,,,,,,,,,,,,, 30 Collapse structures, linear characteristics 17 Cotopaxi eruption 16 Crater Lake National Park .................................... 14 Crater Lake region, volcanic rocks ____________________________ 11, 13, 14, 23, 30 Cristobalite ................ 27, 33, 36, 37, 41, 44, 45, 46, 56, 66, 67, 68, 69, 70, 71. 77 Crystallization, stages ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 10 vapor-phase ,,,,,,,,,, 5, 18, 21, 22, 24, 26, 30, 36 Crystals, concentration ,,,,,,,,,,,,,,,,,,,,,, 25, 26, 36, 73 Curtis, G. H., quoted ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 5 D Dacite _____________________________________ 10, 11, 26, 30, 31, 45, 46 Devitrification , ,,,,,, 9, 13, 18, 21, 24, 26—28, 31, 33, 34, 36—38, 44—45, 56, 59, 60, 62, 65, 66, 67, 68, 69, 70, 71, 72, 76, 77 Dikes, rhyolite .................................................... 17, 26 Diller, J. S., and Patton, H. B., quoted ._ 11 Directed-blast hypothesis _______________________________ Dust, volcanic, definition 4 volcanic, occurrence .. 18, 19 Dust clouds, definition .............................................. 4 E Ejecta, volcanic .......................................... . 12 Eruptions, Cotopaxi, Ecuador . 16 Katmaian ............................ 9, 13, 14 observations _ ........ 16, 3] Pelean 9, 17, 41 producing ash-flow tufi‘s , 14, 16, 38 Eutaxite, definition ,,,,,,,,,,,,,,,,, 4, 10 Exfoliation .................................................................. 30 F Fabric ,,,,, 33 Faults or rift zones ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 17 Feldspar _, ,,,,,,,,,,, 27, 33, 35, 36, 37, 39, 44, 45, 56, 58, 60, 66, 67, 68, 69, 70, 71, 73, 77 Fenner, C. N., quoted .................... 5, 6, 8, 16, 39 Fiamme, definition and description 4, 7, 9, 24 Flow banding ..................... 10, 26, 30, 61 Flow breccia, definition 7 ,,,,,,,,,,,,,,,,,,,,, 12 Flow structures Fluidization Fluorine, in late-stage minerals Foreign materials ,,,,,,,,,,,,,,,,,,,,,,,,,, 40th parallel, exploration Fossil fumaroles ,,,,,,,,,,,,,,,, Fusion, occurrence _ temperature ,,,,,,, See also tufi', fused. 17, 24, 25, 34, 61, 72 , 30, 31 G Glass, brecciation color ______________ obsidianlike , viscosity Glowing clouds Glutwolken Graded bedding ,,,,,,,,,, Griggs, R. F., quoted 11, 24, 26 11, 24, 25, 45 H Heat conservation ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 41, 42 Hibok-Hibok volcano ,, 17, 20 Hornblende __ 35, 42 80 I Iddings, J. P., quoted ........................................ 4, 11 Ignimbrite, age 50 definition 4 microscopic characteristics 34, 37, 45 porosity recognition in the field study of _________________________ l, 14 temperature thickness use of term Incandescent tuflf flow, definition . description ____________________________________________________________ 14 origin use of term Inclusions, elastic Independence Pass region, Colorado Indices of refraction ,,,,,,,,,,,,,,,,,,,,,,,,,, Iron content of ash-flow materials Iron Springs area, Utah J Joints ................................................ 21, 24, 26, 28—29 K Kaolin 31 Katmaian eruption , 13, 14, 16, 36, 39 Komagatake volcano . ........ 14, 17, 22, 31 Krakatau caldera ...................................................... 14 L Lacroix, Alfred, quoted Ladu deposits ,_ Lahar, definition field characteristics __ 21 Lake Toba, Sumatra, pyroclastic deposits _ 14, 28, 58, 62, 69 Lamington volcano Lapilli, composition in ash-flow tufi's 5 definition . 5 deformation , 25 La Soufriere, eruption ,,,,,,,,,,,,,,,,,,,, 4, 5, 12, 13, 15 Latite ‘ 10 biotite-augite 10, 73 quartz 30, 35 Lithophysae, occurrence 19, 26, 27, 28, 37, 65, 66 Load compaction ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 24, 26, 28 M Magmatic differentiation _ Magnetite ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Mansfield, G. R., and Ross, C. S., quoted ,,,,,,,,,,,, 8 Mapping, of ash-flow tuif. See ash-flow tufi'. Marshall, Patrick, quoted Matumoto, Tadaiti, quoted McKays, Calif Merapi volcano Microcline _ Mineralogy _ , 21 Minerals, devitrification _ 27, 36, 44, 45 late-stage ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 12, 69 vapor-phase 14, 18, 21, 26728, 30, 31, 33, 44 Mount Katmai, eruptions ________________________________ 12, 13 Mount Pelée, eruptions __________ Mudfiow, volcanic, definition volcanic, field characteristics water and gas content See also lahar. Mud Lake region, Idaho ............................................ 14 Mud lava, definition Mud spates, description _ N National Geographic Society Expedition Nomenclature, ash-flow materials welded tuifs __ Nomlaki tuflf ..... North Island, New Zealand, rhyolitic tuifs Nuée ardente, cause of lahars .................... INDEX Nuée ardente—Continued definition 5 dual character . 6 flowing properties occurrence _ processes producing 0 Obsidian, colored bands Obsidian Cliff rhyolite flow Obsidian spindles _________________ 35 Olivine Opal 20, 31 P Pectinate structure . 4 36 12, 13, 17, 5—6, 7, 12, 13, 15—16, 41 6, 16, 41 35, 39, 46, 58, 60, Pelean cloud deposit Pelée, eruption _________ Perret, F. A., quoted Phenocrysts, composition 61, 64, 71, 72, 73, 77 effect on distortion ____________________________ 33, 34, 58 number 17, 22, 26, 33, 35, 39, 56, 57, 59, 61 Phonolite 10 Physicochemical factors __________________________________________ 38 Piperno __________ 4, 14 definition 7, 10 origin ................... 9, 12 Pipernoid structure 7 Pirsson, L. V., quoted ____________________________________________ 8 Plasticity, glass 8, 13, 33, 34, 41, 59, 60, 64, 65 magma fragments Pressure, effect on erupted materials 39 Pumice, collapsed 8, 11, 19, 24, 27, 37, 45, 64, 73, 74, 77 color ________ _ 24, 25, 45 definition 7 deformation during welding ................ 25, 26, 27, 28, 67, 77 ejected . _ 14, 16, 22 glassy microscopic characteristics origin recrystallization size of particles sorting __ viscosity welded, definition 8 recognition use of term ..... Pumice and scoria flows , Pumice fall, definition 7 Pumice-fall tufi', occurrence 21 Pumice flow, banding 22 definition 33 lithologic characteristics 14, 19, 22, 23, 30, 31, 33 origin __ 12, 1‘7, 30, 31 thickness 22 Purpose of report __ 2 Pyroclastic, definition _ 7 Pyroclastic materials, definition 7 nomenclature origin ______________ size ’classification sorting ___________________ transporting mechanism welding __________________________________ Pyroclastic rocks, distribution ............................. 1 Pyroxene 45 Q Quartz .................... 35, 39, 42, 45, 57, 60, 64, '72, 77 R Ransome, F. L., quoted _________________________________________ 10 Rattlesnake formation .. Rhyodacite Rhyolite domes __________________________________________________________ Rhyolite flows ........ 9, 10, 11, 17, 31, 37, 39, 61, Rhyolite flows—Continued feeder dikes .......................................... 17, 18, 61 sheared zones ...................................................... 11 Richards, H. C., and Bryan, W. H., quoted ______ 13 Rickmers, W. R., quoted ’ Ring fractures _______ Rittman, A., quoted . River channels S Sand flow, area] extent ,, definition origin __________ use of term Sanidine San Juan Mountains, Colo , Sedimentary deposits, characteristics Shepherd, E. S., quoted .......................... Sikhote-alin volcanic region ................... Silica, percentage in ash-flow materials , redeposition ._ Sillar, definition origin use of term a 5, 14 Sintering 43 Slate, felsitic ______________________________________________________________ 12 Snake River Plains area, Idaho . Snowdon, Wales, felsitic slates . rhyolites Solovev, S. P., quoted Sosman, R. B., quoted _____________ Specific gravity, rhyolitic tufis _ 46 Spherulites .............. 14, 26, 27, 28, 37, 66, 69, 70 Structure, axiolltic.. 4, 9, 36, 37, 45, 71, 72, ‘75, 77 columnar _ ______________________ 20, 28, 29 eutaxitic .. _ 9, 10, 24, 29, 30, 34 rhyolitic ............................................................... 9 T Temperature, ash flows ................ 16, 21, 26, 31, 39 lava 16, 39 magma 41 mudflow _ piperno pumice ,,,,,,,,,,,,,,,,,,,,,,,, pyroclastic materials _ sand flow welded tuflf 14, 26, 28, 34 Tent rocks Tephra, definition 8 Terminology, diversity 2 Thunder eggs ................ , 14, 38, 70 Topography, effect on ash-flow deposition __ . 20, 22, 23, 28 Trachyte 10 Tridymite ,,,,,,,,,,,,,,,,,,,, 27, 33, 36, 41, 42, 44, 45, 69 Truckee Canyon region, Nevada ,,,,,,,,,,,,,,,,,,,,,,,,,,,, 36 Tufi‘, crystal, definition definition ,,,,,,,,,,,,,,,, devitrified ,,,,,,,,,,,,,,,,,,,,,,,,, distinguished from lava _ felsite _, fused lithic, definition _________ nonwelded, definition erosion forms occurrence ....... origin _____________________________________ similarity with sand flow Pelean recrystallization . 14 rhyolitic, composition 13, 14, 17, 44, 46 temperature ......... .. 42 viscosity __ vitric, definition 8 recognition 18, 24, 25, 26 weathering characteristics 21, 22, 24 Tui’f—Continued welded, age ..... andesitic _. 45, 48—50, 77 14, 17, 35, ‘73, 74 areal extent ................. 29 crystal content _. 58, 60, 73 definition . 8 erosion forms _ 30 glassy 14, 25, 28, 29, 33, 34, 35, 40, 56, 57, 58, 60, 61, 62, 76 jointing ................................................ 28, 29 mechanism of deposition .................. 6, 39 occurrence ________ 14, 17, 18, 28, 29, 36, 38 origin ........................ 9, 12, 14, 17, 18, 20, 33—35, 39, 41—44 recent descriptions ............................ 14—15 recognition 25, 26, 30, 65, 68, 69, 71, 77 reexamination ................................ 9, 10, 11 rhyolitic 35, 75 study of . 1, 2, 13 temperature of formation 14, 26, 28, 39 thickness ........................................ 20, 26, 29 INDEX Tufl’, welded—Continued use of term _____ Tuflz‘ flows, incandescent Tufi' lava, definition 1, origin Turner, H. W., quoted _ Tuscan formation _________ V Vallecito, Calif __________________________________________________________ 10 Valles Mountains, N. Mex., ash-flow tufi's .___2, 17, 18, 19, 20, 21, 22, 23, 28, 30, 31, 60, 62, 64, 67, 69, ‘75 Valley of Ten Thousand Smokes, pyrcolastic deposits 5, 12, 13, 14, 16, 20, 23, 31, 36, 39, 77 Vents, volcanic ____________________ volcanic, fissure type .. on volcanic domes . open craters Vesiculation ____________ 17, 39, 40, 61 81 Viscosity 33, 40, 42 Vitrophyre ............ 10 Volatile content 28, 31, 37, 38, 39, 40, 41, 44, 66 W Welding, experimentally produced . .4 14 fabrics developed ................ .. 34 Welding processes ______ 11, 13, 18, 22, 24—26, 33—35, 41—44 Wentworth, C. K. and Williams, Howe], quoted ....... 3, 4, 5, 7 Weyl, Richard, quoted 49 Williams, Howe], quoted . 31, 48 Wineglass dacite flow _________ 11 Y Yellowstone National Park, volcanic rocks ............ 11, 13, 20, 36, 42, 75, 76 Z Zeolites 44 Zirkel, Ferdinand, quoted 4, 9 Zoning, in flow units ____________________ 20—22, 24, 26, 66 11.5. GOVERNMENT PRINTING OFFICE: l95l