The axial basins of the northern and central Rio Grande rift evolved since late Oligocene as a chain of half grabens between the Colorado Plateau on the west and the interior of the craton on the east. The basins range from 80 to 240 km in length and from 5 to 95 km in width, with an average width of approximately 50 km. Basin-fill sedimentary deposits range up to 5 to 6 km in thickness. The oldest dated volcanic rocks interbedded with syn-rift sediments are about 26 Ma. Extension in the rift was left oblique and increases southward from 8 to 12% in the San Luis Basin, to about 17% in the northern Albuquerque Basin, to at least 28% in the southern Albuquerque Basin, to an average of 50% in the Socorro area. The most rapid phase of regional extension was middle to late Miocene. Sedimentary deposits of this age are volumetrically dominant within the rift basins and on the rift flanks (Ogallala and Bidahochi Formations). Basin asymmetry shifts back and forth from east-tilted to west-tilted across complex northeast-trending accommodation zones that developed along preexisting transverse structural lineaments. These accommodation zones lie on small circles about the Miocene Euler pole of rotation for the Colorado Plateau relative to the stable craton. Rotation about this Euler pole in northeastern Utah caused synchronous episodes of Miocene extension and sedimentation in the Rio Grande rift and in the central Wyoming-northern Colorado area, largely by the collapse of Laramide uplifts. Late Cenozoic rotation of the Colorado Plateau was 1.0 to 1.5° clockwise. Structural subdivision of basin floors within the rift by longitudinal master faults increased with increasing extension until, at some point in excess of 28% extension, one or more longitudinal fault blocks was uplifted isostatically through the basin surface to form an intrabasin tilted-block mountain range that subdivided the basin. Faulting and subsidence tended to migrate towards the basin axis as newer master faults cut off or merged downdip with, earlier faults. This deactivated the outer longitudinal blocks and left them stranded as shallow suballuvial benches covered by relatively thin sections of the older syn-rift sediments. Recent paleobotanical studies indicate that elevations prior to rifting were similar to present elevations. This suggests that crustal thickening during Laramide compression resulted in regional uplift during the early Cenozoic. Exhumation of the escarpment along the Rocky Mountains-High Plains boundary by erosion of approximately 400 to 600 m of Cenozoic and Cretaceous strata may be a consequence late Cenozoic climate change.

Phanerozoic sedimentary and volcanic rocks exposed in the Joyita Hills area (Socorro County, New Mexico) record structures that developed in response to three deformational episodes: ancestral Rocky Mountain, Laramide, and Rio Grande rift. Additional exposures of Proterozoic gneiss in the core of the Joyita Hills reveal three preferred orientations of high-angle to vertical, gneissic to mylonitic foliations. These strike north, northwest, and east-northeast. Proterozoic amphibolite dikes parallel the east-northeast foliations. Comparison of Phanerozoic fault and dike orientations with attitudes of Proterozoic structures indicates that reactivation of basement flaws was a factor common to each Phanerozoic orogenic event. Reactivation of Proterozoic foliations during ancestral Rocky Mountain tectonism resulted in basins and uplifts of north and northwest trend. Normal, left-oblique slip on north-striking faults and normal slip on northwest-striking faults, when combined with the orientation and spatial distribution of basins and uplifts, indicates that late Paleozoic deformation in central New Mexico was the result of north-trending, divergent, sinistral wrench faulting. Laramide tectonism reactivated Proterozoic structures in a strike-slip sense. North-striking (synthetic) and east-northeast-striking (antithetic) faults define a north-trending, dextral wrench fault system. Other dextral wrench faults strike northeast, and parallel the Montosa and Del Curto fault zones. Thrust faulting and associated folds were a lesser component of Laramide deformation, and were apparently related to Laramide wrench fault systems. Extensional reactivation of Proterozoic structures influenced the development of basins and uplifts of the Rio Grande rift. Northwest-striking and east-northeast-striking normal fault and dike systems paralleled Proterozoic foliations and amphibolite dikes. North-striking foliations and ancestral Rocky Mountain faults observed along the East Joyita fault influenced the development of this major, down-to-the-east normal fault (approximately 3 km stratigraphic throw). Moreover, north-striking foliations and ancestral Rocky Mountain structures (the north-trending Lucero basin and adjacent uplift) influenced the location and orientation of the Albuquerque Basin segment of the Rio Grande rift.

Calderas related to ash-flow sheets show a positive correlation between caldera area and ejecta volume; this correlation places constraints on magma drawdown during eruption and implies a systematic relationship between these parameters and magma volume of the chamber. Caldera areas range from 1 to 10 4 km 2 ; the volume of ejecta from caldera sites ranges from 1 to 10 4 km 3 , and the volume of the related magma chambers is thought to range from 10 to 10 5 km 3 . A tentative correlation between ejecta volume and the time required to produce that volume reveals an approximate production rate of 10 −3 km 3 /yr. This seems to hold for relatively small eruptions, such as the ash-flow and related pyroclastic eruptions of as little as 10 −3 km 3 that are associated with central-vent volcanoes, even though volumes below about 1 to 10 km 3 are not related to caldera formation. The correlation also holds for large eruptions up to the limits for cumulative ejecta volumes and composite batholith emplacements. Precaldera magma chambers probably all have physical and chemical gradients and measurable variations in chemistry and mineralogy. These variations are revealed in the pyroclastic deposits, if the eruption taps to the maximum eruptible level, and give insight into differentiation processes. Compositional contrasts are commonly greater in small-volume central-vent systems than in large-volume ring-fracture systems, but all systems tend to become more mafic with depth. Successive caldera-forming eruptions from the same system commonly become more mafic with time. This is possibly due to two effects: (1) decreasing thermal input and (2) progressive depletion in “residual” elements. However, new thermal inputs, probably in the form of mafic primitive magma from the mantle, take place intermittently throughout the volcanic life of the system and may extend, as a waning influence, into and beyond the crystallization stage of the pluton. Very strong resurgence of primitive magma may reactivate large-volume systems to begin new cycles of magmatic activity. In common stratovolcanic systems (Crater Lake, Oregon) that rarely fractionate to high-SiO 2 rhyolite, alternating progression and regression of chemical trends with time is dominated by major-element variations. In high-SiO 2 rhyolite magmas (Bandelier Tuff, New Mexico), minor elements may sometimes show striking alternating enrichments and depletions with successive eruptions from the most-fractionated boundary layer at the top of the magma chamber. In the Bandelier magma, many elements such as Nb, Ta, U, Th, Cs, Rb, Li, Sn, Be, B, W, Mo, F, CI, Pb, Zn, Sm, and the heavy rare-earth elements concentrated upward in the system, whereas other elements such as Ba, Sr, Eu, Ti, Cr, Co, Sc, Au, and Cu concentrated downward. For magmas in general, the direction and amount of concentration of both groups of elements are a function of both the initial composition of the parent magma and the effects of the dominant fractionation process operating at any given time. Thus the permissibility of mineralization by or ore formation involving those elements that are dependent on magmatic concentration may be directly related to specific stages in the volcanic history of any given system. Any ash-flow sheet holds basic clues for the behavior of certain elements in that system at a specific time and, seen in conjunction with the composition of magmas erupted from the system at other times, suggests concentration trends that provide insights into the controlling magmatic processes.

Many source areas for voluminous ash flows have histories of repeated catastrophic eruption and caldera-forming collapse within a few million years or less. Examples range in complexity from simple calderas related to eruption of a single cooling unit, through sequential collapses of the same caldera related to successive eruptions of separate cooling units, and nested calderas related to eruption of successively smaller cooling units between successively longer time intervals, to complexes of overlapping calderas related to eruption of several cooling units from overlapping source areas within a large volcanic field. Some caldera complexes are related to collapse of adjacent, simultaneously active ring-fracture zones by immediately successive ash-flow eruptions to form composite sheets. Despite this wide range in complexity, a basically similar primary sequence of events in each area reflects the formation of a large volume of magma, at least part of which rose to form shallow epizonal magma chambers. In some instances, only a single high-level chamber formed and produced ash-flow eruption and collapse one or more times. In others, several high-level chambers formed either simultaneously or in succession; where several chambers were active simultaneously, major ash-flow eruption and collapse at one chamber may have triggered events in an adjacent one. Variations in the pattern of ash-flow–caldera relations reflect variations in relative volumes of magma in the chambers, of tectonic controls on intrusions and the timing of eruptions, and of the continuity of magma generation beneath a volcanic region. These diverse patterns represent variations on R. L. Smith’s concept that voluminous ash-flow eruptions and related caldera formation are surficial expressions of the degassing, but not the total emptying, of large magma chambers at high levels in the Earth’s crust. Study of these systems offers unique insights into the physical and chemical processes and evolution of silicic magmas.

The ash-fall and outflow sheets of the 0.7-m.y.-old Bishop Tuff represent >170 km 3 of compositionally zoned rhyolitic magma emplaced during collapse of the Long Valley caldera, California. Field, mineralogic, and chemical evidence agree that tapping of the thermally and chemically zoned chamber was continuous, without interruptions sufficient to permit mixing or phase re-equilibration. Fe-Ti oxide temperatures for 68 glassy samples increase systematically with eruptive progress from 720 to 790 °C; this increase corresponds well with the stratigraphic sequence, but the temperatures in no way correspond to the degree of welding. Ubiquitous quartz, sanidine, oligoclase, biotite, ilmenite, titanomagnetite, zircon, and apatite change composition progressively with temperature. The uniformity within every sample of each mineral species (irrespective of size and whether discrete or as inclusions) is not compatible with protracted crystal settling. Euhedral allanite (ρ > 4, La + Ce > 16% by weight) occurs in all early-erupted samples (720 to 763 °C) but in none erupted later. Despite this, whole-rock La + Ce values increased threefold during the eruption. Pyrrhotite, hypersthene, and augite appear abruptly at 737 °C and occur in all later samples. These sharp isothermal interfaces indicate lack of any extensive history of crystal settling. Whole-rock major-element gradients were modest, but many trace-element concentration gradients were very steep despite a drop of only ∼2% by weight SiO 2 within the magma volume erupted. Enrichment factors (the ratio of the value in the early-erupted samples to the value in the late-erupted samples) are Ba, 0.02; Sr, <0.1; Mg, <0.1; P, 0.17; Eu, 0.12; La, 0.3; Yb, 2.35; Mn, 1.6; Sc, 1.65; Y, >2; Ta, 2.5; U, >2.5; Cs, 3.8; Nb, >5; Rb/Sr, 22; Mg/Fe, 0.1; Ce/Yb 0.2; Eu/Eu*, 0.07; Zr/Hf, 0.65; Ba/K, 0.02; and Ba/Rb, 0.01. These can neither have been established by transfer of any reasonable combination of phenocrysts nor inherited from progressive partial melting. Abundant bulk and phenocrystic data further exclude large-scale assimilation, liquid immiscibility, and contamination by underplating mafic magma. The compositional and thermal gradients existed in the liquid prior to phenocryst precipitation and developed largely independently of crystal-liquid equilibria. Within water- and halogen-enriched high-silica roof zones of large magma chambers, chemical separations take place through the combined effects of convective circulation, internal diffusion, complexation, and wall-rock exchange to develop compositional gradients, which are linked to gradients in the structure of the melt and are controlled by the thermal and gravitational fields of the magma chamber itself. Liquid-state differentiation through processes of convection-driven thermogravitational diffusion probably requires progressive establishment of a stable density gradient in order to retard convective re-mixing of the zoned upper part of the system. These processes evidently produce differentiated tops on magma bodies that represent a wide range in initial bulk composition. The degree of enrichment within a given chamber probably reflects the repose time between eruptions, the volatile flux, and the rate of energy transfer from the mantle more than it does the bulk composition of the unerupted dominant volume. Dikes and stocks injected above such a system will be either barren of or enriched in elements susceptible to subsequent hydrothermal concentration to ore grades; enrichment or barrenness depends on the timing of emplacement relative to cycles of enrichment and eruption. Crystal-liquid equilibria probably predominate during initial generation of the dominant magma volume as well as during its ultimate plutonic consolidation. Thermogravitationally generated, strongly differentiated capping magmas commonly erupt, but some crystallize as alaskites, leucogranites, and granite porphyries, and some may be resorbed by the dominant volume during the waning stages of a pluton’s magmatic lifetime.

Rare-earth-element abundances have been determined for 26 rhyolite obsidians and glass separates of both peralkaline and subalkaline chemical character for which a history of marked differentiation had been inferred from very low Sr, Ba, Mg, and minor transition-element contents. All but five of the specimens possess very strong negative Eu anomalies, with values of Eu/Eu* of 0.06 or less. Eight have values of Eu/Eu* of 0.02 or less, and subalkaline rhyolites from the Fish Creek Mountains Tuff, Nevada, have Eu/Eu* of <0.003. Calculations using coefficients appropriate for the distribution of Eu between Ca-poor feldspar and silicic melt suggest that, if the Eu anomalies are the result of crystal fractionation, many of the glasses represent 10% to 20% or less of hypothetical unfractionated parent liquids. Distribution coefficients for Eu between Ca-poor feldspar and peralkaline silicic melt are significantly lower than for subalkaline systems; this implies high degrees of differentiation for peralkaline rocks having even modest negative Eu anomalies. A number of the specimens studied represent relatively large rock units, and some are from the lower parts of large, compositionally zoned ash-flow sheets. The mechanism—whether crystal fractionation or thermogravitational diffusion—called upon to produce large-scale vertical compositional variations within bodies of highly silicic magmas must be able to explain the profound depletion of large volumes of magma in such elements as Sr, Eu, Ba, Mg, Co, and Ni. The moderate negative Eu anomaly of glass from the Summit Lake Tuff, Nevada, can be explained by crystallization of the anorthoclase phenocrysts present in the rock. This suggests that this compositionally unzoned ash-flow sheet was produced from rhyolitic magma that possessed little or no Eu anomaly before it rose to a high-level chamber where it subsequently underwent appreciable intratelluric crystallization, but little or no crystal segregation, prior to eruption. In contrast, the strong Eu depletion (Eu/Eu* = 0.05 to 0.15) of crystal-rich specimens of the Fish Creek Mountains Tuff indicates that considerable differentiation had taken place before growth of the present generation of phenocrysts. Peralkaline specimens from the northern Great Basin and Snake River Plain have lower La/Lu ratios (2.4 to 6.2) than do specimens from the southern Great Basin (La/Lu = 6.4 to 15.7), which suggests that they may have been derived from source regions that have significantly different rare-earth-element abundance patterns.

Analysis of 30 individual pumice blocks, together with bulk samples from the ash-flow member of the Los Chocoyos Ash within the Quezaltenango Valley, Guatemala, demonstrates that prior to its eruption, its associated magma-chamber was zoned. Eruption of a high-K (K 2 O/Na 2 O > 1), crystal-poor, biotite-bearing rhyolite with crystal equilibration temperatures of less than 800 °C produced the widespread H-tephra member and the initial phases of the ash-flow member. As the ash-flow eruption continued, a more-heterogeneous, low-K, crystal-rich, cummingtonite- and hornblende-bearing rhyolite became predominant; its phenocrysts had equilibrated at temperatures of about 950 °C. The water content of the high-K rhyolite was several percent, whereas the low-K rhyolite was much drier. Bulk samples of the ash-flow member are homogenized mixtures of matrix shards that represent either the high-K or low-K rhyolite magmas; the overall ratio for the ash-flow member is 60% high-K and 40% low-K type. The 87 Sr/ 86 Sr ratios for both high-K and low-K magma types are identical and average 0.70405 ± 0.00003. This value is nearly the same as all basaltic, all andesitic, and most rhyolitic Quaternary volcanic rocks tested in Guatemala so far. The 87 Sr/ 86 Sr ratios for bulk samples of the ash are significantly higher and more variable (0.70426 ± 0.00009), probably because of xenocrystic contamination. Detailed mixing and Rayleigh calculations using observed mineral phases in the ash show that the concentrations of 8 major and 17 minor elements in the ash are consistent with the derivation of high-K rhyolite from low-K magma by crystal fractionation at shallow depths. The time required for such fractionation is at least 10 4 yr. The absence of a continuum of compositions from low-K to high-K rhyolite and the differences in p H 2 O and temperature suggest that the two magmas were separated during fractionation. The Los Chocoyos Ash is the most silicic major Quaternary unit in the Guatemalan Highlands; the volume of magma from which it was derived is far greater than that of all other Quaternary volcanic rock units in the area.

Upper Pleistocene rhyolitic ash-flow and air-fall tuffs, erupted from several centers, were sampled in 23 pumice-filled basins over an area of 16,000 km 2 . Fifty pumice-matrix samples were analyzed for as many as 20 trace elements. Ba, Fe, Hf, Rb, Sm, Sr, Th, Ti, and Zr were particularly useful in “fingerprinting” correlations between basins and in corroborating the stratigraphy previously established within individual basins. On the basis of similar trace elements, a tephra and an overlying ash-flow sheet (together, a unit here named the Los Chocoyos Ash) appear to have formed from a multiphase eruption. The tephra, whose volume exceeded 100 km 3 , blanketed an area greater than 1 × 10 6 km 2 . The second phase of the eruption produced an ash flow of greater than 200 km 3 . Areal geochemical patterns within the ash-flow sheet are probably related to sequentially less explosive eruptions of progressively more mafic ash flows. Changes in chemical composition, size of pumice and lithic fragments, thickness, and elevation all suggest a source for the Los Chocoyos Ash in the Lake Atitlán cauldron. Chemical data suggest correlation of the H-tephra member of the Los Chocoyos Ash with the most prominent D layer of the Worzel ash of the equatorial Pacific.

Igneous rocks of the Sierra Madre Occidental have been studied along two traverses across the range. One is at lat 24°N between Mazatlán and Durango City, where contiguous mapped areas extend across the Sierra; the other is near lat 28°N, where several separate areas west and north of Chihuahua City have been mapped. In these regions the Sierra contains two vast and largely coextensive igneous sequences, both calc-alkalic and both including ignimbrites. The older sequence of rocks, which ranges in age from 45 m.y. to at least 100 m.y., is characterized by abundant batholithic as well as volcanic rocks and is dominantly intermediate in composition. The younger sequence is dominated by rhyodacitic to rhyolitic ignimbrites erupted from large caldera complexes, generally accompanied by rhyolite flows and domes and small outpourings of mafic lavas. Intermediate rocks are rare, and volcanism was largely confined to the interval 34 to 27 m.y. ago, although some activity persisted until 23 m.y. ago. Mid-Tertiary volcanic rocks are exposed eastward across Chihuahua between the Sierra Madre Occidental and the alkalic volcanic province of Trans-Pecos Texas. Volcanic rocks of this region are chiefly calc-alkalic in chemistry, but their alkalinity is intermediate between that of the Sierra Madre Occidental and of Trans-Pecos Texas. Ages are both similar to and older than those in the Sierra. The chronology of both igneous sequences of the Sierra Madre Occidental fits some of the known sea-floor–spreading patterns. The older sequence is similar in age and composition to batholithic terranes of the western United States that were emplaced during Cretaceous–early Tertiary convergence along the western margin of the North American plate. The younger sequence was also emplaced during a period of convergence. Volcanism declined 27 m.y. ago at about the time of reorientation of an east Pacific spreading center and early ridge-trench interaction at the nearest continental margin. The gap in magmatism between 45 and 34 m.y. and the abrupt onset of volcanic activity 34 m.y. ago, however, do not match known sea-floor–spreading events. Furthermore, the brief age span and the bimodal calc-alkalic composition of the younger sequence are not typical for an igneous province at a continental margin.

Subaerial pyroclastic deposits are of three genetic types: (1) fall, (2) surge, and (3) flow. Although pyroclastic flows include a wide range of volume magnitudes, only small-scale eruptions have been observed. Flows are composed of a dense, basal avalanche and an overriding cloud of entrained particles that rise by convective buoyancy. Theoretical models explain the generation of pyroclastic flows by gravitational collapse of an eruption column. The high apparent mobility of pyroclastic flows is largely due to the dissipation of kinetic and potential energy imparted at the time of eruption. The regular stratigraphic succession of fall, surge, and flow deposits is compatible with a column-collapse model. The textures, morphology, and distribution of pyroclastic flow deposits suggest emplacement as thick, dense beds that may be quasi-fluidized near their source, but which flow mainly in a laminar mode toward their distal reaches. Smaller flows may move entirely as single laminar beds. Large primary deposits may show a facies change controlled by the gradation from an inflated proximal stage into a deflated distal stage. Some deposits may undergo secondary (or tertiary) flowage following welding to produce textures and structures similar to lava flows. The various degrees of welding and secondary crystallization are principally controlled by emplacement temperature, thickness, and composition.

The Wall Mountain Tuff was deposited about 36 m.y. ago in paleovalleys extending from west of Salida to beyond Castle Rock, Colorado, a distance of at least 140 km. The Gribbles Run paleovalley, 16 km northeast of Salida, is exceptionally well exposed by modern dissection and reveals a complex interplay of primary and secondary structures formed during deposition of the Wall Mountain Tuff. Deposition occurred in laminar boundary layers between the bottom and sides of the channel and ash flows passing above. The overlapping sequence of events during deposition was (1) agglutination and incipient collapse of glassy particles, (2) laminar shearing of the compacting and welding mass to form a primary foliation analogous to flow banding in lavas, (3) expulsion of gases from the collapsing spongy mass and concentration of these gases along shear planes, (4) formation of gas pockets in places where the volume of gases expelled exceeded that which could be accommodated on shear planes, (5) elongation of gas pockets and pumice to form a primary lineation in the plane of the foliation, (6) statistical alignment of the long axes of solid particles parallel to the direction of flow and imbrication of the long axes so that they dip sourceward relative to the foliation planes, (7) development of primary flow folds with axes perpendicular to the lineation, and (8) end of forward motion. By the time forward motion ceased within a given layer, the tuff had the rheological properties of a rhyolite lava. The high viscosity of the welded tuff preserved open cavities and prevented differential compaction over lithic fragments or primary folds. More rapid deposition along the sides of the paleovalley than along its axis caused inward accretion of welded tuff with steep primary flow foliation to form a U-shaped cross-channel profile. Secondary folds, whose axes parallel the lineation, and concurrent growth faults formed locally by creep toward the valley axis. Spectacular internal unconformities developed where undeformed tuff was deposited over primary or secondary folds, yet all the tuff welded together to form a remarkably uniform simple cooling unit. Episodic downstream movements of a few metres, as the Wall Mountain Tuff adjusted to its bed and to its rising center of gravity, opened swarms of tension cracks along certain horizons. These tension cracks dip steeply in the downstream direction and provide a useful indicator of flow direction. Other structures useful in determining flow direction are imbricated crystals, streamlined ridges and grooves in lineated gas cavities, and upstream dip of axial planes on asymmetric primary folds. Emplacement temperature is the dominant factor in determining whether a tuff undergoes primary or secondary welding. If the temperature is well above the softening point, the glassy particles will agglutinate and collapse during deposition in the laminar boundary layer, and the tuff will show megascopic laminar flow structures (primary welding). If the temperature is below the softening point, deposition occurs as loose ash, and welding is a postemplacement process (secondary welding). Most tuffs undergo secondary welding and only the preferred orientation of solid particles and the grain-size distribution remain as evidence of laminar flow during deposition. The type of welding may vary both laterally and vertically within some ash-flow sheets and impose a significant facies variation.

Welded tuffs formed by air fall, rather than by pyroclastic flow, are a common type of volcanic rock. Examples of welded air-fall tuffs are the Thera and Therasia welded tuffs, Santorini volcano, Greece, and parts of the ejecta of the 1875 eruption of Askja, Iceland. The Thera and Askja welded tuffs cover a few square kilometres, and the Therasia welded tuff covers at least 6 km 2 . The Santorini tuffs and parts of the Askja deposit grade laterally and vertically into nonwelded, coarse-grained Plinian air-fall pumice deposits. The nonwelded equivalents of the welded tuffs have grain-size characteristics and depositional features typical of airborne ejecta, such as good sorting (σ < 2.0), internal stratification, distinguishable fall units, and a systematic decrease in thickness and maximum and median grain size away from the source. Zones of dense welding, partial welding, and no welding are similar to zones in ignimbrites. A zone of marked color change was also observed in the nonwelded part. Lateral zonation is well developed. The tuffs are most densely welded where the thickness is the greatest. Zone boundaries are parallel to isopachs for the uncompacted thickness. In the zone of dense welding of the Askja tuff, the tuff is welded from bottom to top, although no significant overburden has been removed since 1875. Vertical profiles of strain ratio, porosity, and density were determined for the thickest parts of the three tuffs. There are systematic increases in strain ratio and density and decrease in porosity toward the central parts of the welded zone in the Santorini tuffs and the lower central part of the Askja tuff. The Askja and Thera tuffs show notable fluctuations of strain ratio, porosity, and density superimposed on the general compaction profile. These fluctuations are attributed to the heterogeneous nature of the stratified air-fall ejecta and may be one way of distinguishing tuffs of air-fall and pyroclastic flow origins. Two air-fall units were traced laterally from the zone of dense welding to the zone of partial welding. The strain ratio of the clasts and the porosity and density of tuff change in a rapid linear manner over a distance of only 70 m. Marked lateral changes in strain and welding, normal to isopach contours, are considered evidence that rapid accumulation is critical in producing welding in air-borne ejecta. Historical records indicate that the Askja tuff was erupted in less than 1 h. An accumulation rate of 20 cm/min is estimated for the area of densest welding, before compaction occurred. All three of the welded tuffs show evidence of physical mixing of contrasted magma compositions: the Askja tuff contains rhyolitic and basaltic components, and both Santorini tuffs contain dacitic and andesitic components. In all three, the volume of the mafic component is minor, and superheating of the silicic component probably contributed to produce welding.