Second Extinction Research Articles

The effect of environmental changes on the evolution and extinction of Late Devonian trilobites from the northern Canning Basin, Western Australia

Abstract The Frasnian–Famennian Virgin Hills Formation represents fore-reef facies deposited as part of the extensive Late Devonian reef system that fringed the SW Kimberley Block in Western Australia. It contains a rich trilobite fauna dominated primarily by proetids and, to a lesser extent, harpetids, phacopids, scutelluids and odontopleurids. To date, 49 taxa have been described, 40 of these being restricted to the Frasnian. Herein five Frasnian taxa are described, three in open nomenclature, and two the new species Telopeltis intermedia and Otarion fugitivum . Evolutionary trends in the Virgin Hills trilobites are dominated by a reduction in body size and eye size and, to a lesser extent, a reduction in exoskeletal vaulting. Although recording no sedimentological signature, the fauna was strongly affected by the two globally recognized Kellwasser extinction events. The first, at the end of conodont Zone 12, affected taxa at the species and genus level. The second, within Zone 13b, had a much greater impact on the fauna, causing extinctions at the familial and ordinal levels. Evidence is presented to suggest that evolutionary trends in the trilobites during the late Frasnian reflect selection for forms adapted to low nutrient conditions. The two intensive Kellwasser extinction episodes may reflect periodic massive inputs of nutrients from the terrestrial into the shallow-marine environment.

Famennian survivor turiniid thelodonts of North and East Gondwana

Abstract Microvertebrate samples from the Upper Devonian Hojedk section, southeastern Iran, and the Napier Formation, northwestern Australia, have yielded scales of agnathan thelodonts, dated as early/mid-Famennian ( crepida–marginifera/trachytera conodont zones). These scales are referred to Arianalepis megacostata , a new genus and species, and Arianalepis sp. indet., a second indeterminate species of this new turiniid genus. Further recorded scales of Australolepis seddoni from the Napier Formation confirm the age range for this taxon as extending into the late Frasnian. The new remains post-date the previously youngest thelodonts from Iran and Western Australia and provide the first evidence of thelodonts surviving the Frasnian–Famennian extinction events.

Revised correlation of the Frasnian–Famennian boundary and Kellwasser Events (Upper Devonian) in shallow marine paleoenvironments of New York State

The Frasnian–Famennian boundary (Upper Devonian) is exposed in a 200 km-long outcrop belt in New York State, with deeper paleoenvironments to the west and shallower ones to the east. Geochronology in the eastern end of the outcrop belt has been based primarily on lithostratigraphic correlation with western sections, which were dated using conodonts. We collected conodonts and brachiopods from several measured sections and numerous other localities, and these collections suggest that these east–west lithostratigraphic correlations require revision. We correlate the Wiscoy Formation with the upper Angola Formation and the Canaseraga with the upper Hanover. Thus, the Canaseraga Formation contains the Frasnian–Famennian boundary and Upper Kellwasser Event, and the dark shale above the Wiscoy is equivalent to the Pipe Creek Formation and Lower Kellwasser Event. These new correlations imply that the Lower Kellwasser Event had greater impact on the shelly benthos of New York than the Upper Kellwasser, at least for the subset of taxa examined here. All strophomenid brachiopods and rugose corals were extirpated at the Lower Kellwasser, along with numerous other brachiopods. The final species of atrypid brachiopod persisted to the Upper Kellwasser. • Frasnian–Famennian (Upper Devonian) biostratigraphy of New York State is evaluated. • Conodonts and brachiopods were collected from shallow-water, siliciclastic facies. • New correlations are proposed for the F–F boundary and Kellwasser extinction events. • The revised correlations will facilitate study of these extinction events. • The Lower Kellwasser extinction was severe for corals and brachiopods in New York.

A Diverse Tetrapod Fauna at the Base of 'Romer's Gap'

The lack of fossil tetrapod bearing deposits in the earliest Carboniferous (‘Romer’s Gap’) has provoked some recent discussions regarding the proximal cause, with three explanations being offered: environmental, taphonomic, and collection failure. One of the few, and earliest, windows into this time is the locality of Blue Beach exposed in the Tournaisian deposits at Horton Bluff lying along the Avon River near Hantsport, Nova Scotia, Canada. This locality has long been known but, because the fossils were deposited in high energy settings they are almost always disarticulated, so the fauna has not been described in detail. Recent intensive collection has revealed a diverse assemblage of material, including for the first time associated elements, which permits an evaluation of the faunal constituents at the locality. Although not diagnosable to a fine taxonomic level, sufficient apomorphies are present to identify representatives from numerous clades known from more complete specimens elsewhere. The evidence suggests a diverse fauna was present, including whatcheeriids and embolomeres. A single humerus previously had been attributed to a colosteid, but there is some uncertainty with this identification. Additional elements suggest the presence of taxa otherwise only known from the late Devonian. Depositional biases at the locality favor tetrapod fossils from larger individuals, but indirect evidence from trackways and tantalizing isolated bones evidences the presence of small taxa that remain to be discovered. The fossils from Blue Beach demonstrate that when windows into the fauna of ‘Romer’s Gap’ are found a rich diversity of tetrapods will be shown to be present, contra arguments that suggested this hiatus in the fossil record was due to extrinsic factors such as atmospheric oxygen levels. They also show that the early tetrapod fauna is not easily divisible into Devonian and Carboniferous faunas, suggesting that some tetrapods passed through the end Devonian extinction event unaffected.

Ecologically and geologically relevant isotope signatures of C, N, and S: okenone producing purple sulfur bacteria Part I.

Purple sulfur bacteria (PSB) are known to couple the carbon, nitrogen, and sulfur cycling in euxinic environments. This is the first study with multiple strains and species of okenone-producing PSB to examine the carbon (C), nitrogen (N), and sulfur (S) metabolisms and isotopic signatures in controlled laboratory conditions, investigating what isotopic fractionations might be recorded in modern environments and the geologic record. PSB play an integral role in the ecology of euxinic environments and produce the unique molecular fossil okenane, derived from the diagenetic alteration of the carotenoid pigment okenone. Cultures of Marichromatium purpuratum 1591 (Mpurp1591) were observed to have carbon isotope fractionations ((13)ε biomass - CO2), via RuBisCO, ranging from -16.1 to -23.2‰ during exponential and stationary phases of growth. Cultures of Thiocapsa marina 5653 (Tmar5653) and Mpurp1591 had a nitrogen isotope fractionation ((15)ε biomass - NH 4) of -15‰, via glutamate dehydrogenase, measured and recorded for the first time in PSB. The δ(34) SVCDT values and amount of stored elemental sulfur for Mpurp1591 cells grown autotrophically and photoheterotrophically were dependent upon their carbon metabolic pathways. We show that PSB may contribute to the isotopic enrichments observed in modern and ancient anoxic basins. In a photoheterotrophic culture of Mpurp1591 that switched to autotrophy once the organic substrate was consumed, there were bulk biomass δ(13)C values that span a broader range than recorded across the Late Devonian, Permian-Triassic, Triassic-Jurassic, and OAE2 mass extinction boundaries. This finding stresses the complexities in interpreting and assigning δ(13)C values to bulk organic matter preserved in the geologic record.

Chemostratigraphy and magnetic susceptibility of the Late Devonian Frasnian–Famennian transition in western Canada and southern China: implications for carbon and nutrient cycling and mass extinction

Abstract We investigate the Late Devonian Frasnian–Famennian extinction interval in western Alberta and south China to shed light on the palaeoecological and palaeoceanographic conditions that characterize this biotic crisis. Both the Lower and Upper Kellwasser events are documented in western Canada. Only the Upper Kellwasser event has been evaluated in south China. Our multiproxy geochemical approach reveals that these events are characterized by positive δ 13 C and δ 15 N excursions and increasing magnetic susceptibility (Canada/China) and increases in detrital (Al, Si, Ti, Zr), productivity (Cu, Ni, Zn) and redox (Mo, U, V) elemental proxies (Canada). We interpret these trends as part of a systemic palaeoecological shift associated with the development of widespread terrestrial forests and their alteration of chemical–mechanical weathering patterns. Increase in detrital proxies is thus interpreted as resulting from pedogenically driven weathering on the continents that nutrified epeiric and continental margin seas. High biological productivity led to eutrophication and development of suboxic to anoxic conditions during both events and probably euxinic conditions during the Upper Kellwasser event in western Canada. Positive δ 13 C excursions are the telltale signature of excessive carbon burial, while redox proxies and δ 15 N records indicate suboxic–anoxic denitrifying conditions.

THE LAST GASP: TRACE FOSSILS TRACK DEOXYGENATION LEADING INTO THE FRASNIAN-FAMENNIAN EXTINCTION EVENT

ABSTRACT The Frasnian–Famennian boundary is correlated with one of several Late Devonian extinction pulses that resulted in a significant decrease in diversity as well as ecological restructuring. This event is recognized within the globally correlated Upper Kellwasser interval that is well exposed and biostratigraphically well constrained in shale units of western New York State. The ichnological and geochemical signals of the interval stratigraphically below the Upper Kellwasser event at these localities provides insight into the onset of this important extinction event. Detailed analysis of ichnogeneric composition, relative size of burrow populations, amount of bioturbation, and trace metal concentrations vary in concert. Deep-penetrating, pyritized Skolithos burrows terminate abruptly at a thin, laminated black shale interval with enriched Mo levels, up to 31 ppm, and are overlain by an interval of gray-green bioturbated shales dominated by Chondrites. These textural and chemical shifts reveal that b...

Changes of palaeoenvironmental conditions recorded in Late Devonian reef systems from the Canning Basin, Western Australia: A biomarker and stable isotope approach

Although the Late Devonian extinctions were amongst the largest mass extinction events in the Phanerozoic, the causes, nature and timing of these events remain poorly restrained. In addition to the most pronounced biodiversity loss at the Frasnian–Famennian (F–F) boundary and the end Famennian, there were also less extensively studied extinction pulses in the Middle to Late Givetian and the Frasnian. Here we used a combination of palynological, elemental, molecular and stable isotope analyses to investigate a sedimentary record of reef-systems from this time period in the Canning Basin, Western Australia.The acquired data generally showed distinct variations between sediments from (i) the time around the Givetian–Frasnian (G–F) boundary and (ii) later in the Frasnian and indicated a distinct interval of biotic stress, particularly for reef-builders, in the older sediments. Alterations of pristane/phytane ratios, gammacerane indices, Chlorobi biomarkers, δDkerogen and chroman ratios describe the change from a restricted marine palaeoenvironment with an anoxic/euxinic hypolimnion towards a presumably open marine setting with a vertically mixed oxic to suboxic water column. Simultaneous excursions in δ13C profiles of carbonates, organic matter (OM) and hydrocarbons in the older sediments reflect the stratification-induced enhancement of OM-recycling by sulfate reducing bacteria. Alterations in sterane distributions and elevated abundances of methyltrimethyltridecylchromans (MTTCs) and perylene indicate an increased terrigenous nutrient input via riverine influx, which would have promoted stratification, phytoplankton blooms and the development of lower water column anoxia.The detected palaeoenvironmental conditions around the G–F boundary may reflect a local or global extinction event. Our data furthermore suggest a contribution of the higher plant-expansion and photic zone euxinia to the Late Devonian extinctions, consistent with previous hypotheses. Furthermore, this work might contribute to the understanding of variations in Devonian reef margin and platform-top architecture, relevant for petroleum exploration and development in the global Devonian hydrocarbon resources.

Reply to the comment on Kaiho et al., “A forest fire and soil erosion event during the Late Devonian mass extinction” [Palaeogeogr. Palaeoclimatol. Palaeoecol. 392 (2013): 272–280

Kaiho et al. (2013) showed that high values of organic molecule indices of combustion, soil erosion, and euxinia occurred at the Frasnian–Famennian transition at Sinsin, Belgium. Marynowski and Racki (2014) commented on the paper. Their issues mainly address (1) the low resolution of the data, (2) the reliability of the proxies, (3) weathering as a cause of the peaks of the proxies, (4) the discrepancy between rare charcoal occurrence and the high combustion proxy, and (5) the stepwise collapse of the huge global reef ecosystem ending well before the F–F boundary. In response to comment (1), the positive excursion of δ13Ccarb marking the F–F transition in the Sinsin section indicates that the Sinsin section has enough strata to detect environmental changes during the culmination of the Late Devonian stepwise mass extinction. On comment (2): soil erosion was indicated not only by dibenzofuran but also by a spike of cadalene. Dibenzothiophene should have increased when euxinic water developed on the surface of sediments, which is supported by the coincidence of isorenieratane and dibenzothiophene on the middle shelf in the late Permian. We show the combustion index at the Cretaceous–Paleogene transition, indicating that this ratio is useful as a proxy of combustion. On comment (3): we cannot explain the high and low combustion, soil erosion, and euxinia index values by weathering because of the consistently well-preserved state of the rock samples. On comment (4): we need to detect charcoal during the F–F transition and early Famennian, corresponding to maxima of high sea levels, to clarify the discrepancy. On comment (5): our model is only for the F–F transition event marked by the extinction acme; it may not explain the collapse of the reef ecosystem that ended before the F–F boundary, but it can explain the extinction acme at the F–F transition.

Geochemical evidence for euxinia during the Late Devonian extinction events in the Michigan Basin (U.S.A.)

Several mass extinction events occurred in the Late Devonian, but the trigger for these events remains elusive. In this study, geochemical evidence in the Late Devonian Antrim Shale, Michigan Basin, U.S.A., records episodic euxinia contemporaneous with these extinction events. Diagnostic changes in iron proxy data and elevated trace metal enrichments correspond to the Kellwasser Crisis. In this study, carbon, sulfur, iron and trace metal geochemistry preserved in the Antrim Formation validates the establishment and expansion of euxinic conditions associated with the Kellwasser Crisis and the Frasnian–Famennian boundary. The strength of the sequential extraction iron mineral data presented here, in concert with trace metal and sulfur isotope proxies, provides definitive signatures of euxinia when other data may be more ambiguous in regard to paleoredox conditions. During the time of the Frasnian–Famennian boundary extensive sulfide oxidation at the chemocline, the result of Fe-limiting conditions within the basin, provides an alternative explanation for the oceanic decline in δ34SSO4 during, and following, the Frasnian–Famennian event. Our geochemical evidence, indicating the presence of anoxia in the Michigan Basin, is consistent with data from other globally distributed locations. Euxinia should be considered a key driver for these global extinction events, and possibly others such as the Hangenberg Event in the Late Devonian.

When the invasion of land failed: the legacy of the Devonian extinctions

Preface1. The Evolution of Life on Land2. The Plants Establish a Beachhead3. The First Animal Invasion4. The First Catastrophe and Retreat5. The Second Animal Invasion6. The Second Catastrophe and Retreat7. Victory at Last8. The Legacy of the Devonian ExtinctionsNotesReferencesIndex

Oxic facies and the Late Devonian mass extinction, Canning Basin, Australia

The close association of anoxic or dysoxic sedimentary rocks and the major Late Devonian (Frasnian–Famennian) mass extinction has focused considerable attention on anoxia as the major cause or as a major factor in a multicausal scenario. The record of the Late Devonian biotic crisis in the well-known reef complexes of northwestern Australia (Canning Basin), in contrast to many localities elsewhere, does not display sedimentological evidence of anoxia through the Frasnian–Famennian boundary interval. Analysis of continuous drill core through this interval has yielded three positive δ 13 C isotopic excursions, only one of which coincides with total organic carbon (TOC) maxima in our data. Multi-element geochemical proxies suggest that TOC maxima preceding positive shifts in δ 13 C most likely resulted from higher productivity caused by nutrient influx from continental weathering, given the close association between TOC maxima and regional relative sea-level falls. Our interpretation supports the view that anoxia was not a fundamental driver of mass extinction and stresses the importance of integrated data sets and understanding regional controls on environmental changes and/or stresses.

Comment on the Kaiho et al., paper “A forest fire and soil erosion event during the Late Devonian mass extinction” [Palaeogeography, Palaeoclimatology, Palaeoecology 392 (2013): 272–280

Kaiho et al. (2013, Palaeogeography, Palaeoclimatology, Palaeoecology 392 (2013): 272–280) interpreted the occurrence of elevated concentrations of high molecular weight polycyclic aromatic hydrocarbons and dibenzofuran as indicators of wildfires and enhanced run-off near the Frasnian–Famennian (F–F) boundary. We argue that other processes, including weathering or hydrothermal oxidation (not discussed by Kaiho et al.) led to the observed increase in the concentration of these compounds and also changed their distribution. Kaiho et al.'s evidence for soil erosion and eutrophication-induced euxinia is also weak in the case of the investigated Belgian sections. Finally, Kaiho et al. rather unfortunately omitted a great wealth of important data published elsewhere, choosing instead to include only those which support their ideas and interpretations.

A new model for the Kellwasser Anoxia Events (Late Devonian): Shallow water anoxia in an open oceanic setting in the Central Asian Orogenic Belt

The Frasnian–Famennian mass extinction event devastated tropical marine ecosystems and ranks in the top six in taxonomic and ecological severity. The close stratigraphic association between the extinction and the Kellwasser Anoxia Events support a link between oceanographic anoxia and extinction. The Upper and Lower Kellwasser horizons have been identified in epicontinental and basinal settings in Laurussia, Gondwana, Siberia, and South China. The Hongguleleng Formation (Late Devonian) in northwestern Xinjiang, China, contains both the Frasnian–Famennian boundary and the rebound from the Frasnian–Famennian extinction event in a highly fossiliferous shallow marine setting associated with a Devonian oceanic island arc complex (part of the Central Asian Orogenic Belt, or CAOB). Here we show that the Hongguleleng Formation also records the Upper Kellwasser Anoxia Event through analysis of multiple geochemical proxies. In contrast to previous studies asserting that the Kellwasser Events were restricted to epicontinental seaways and basins, our results indicate that it occurred not only along the shallow continental margins of the closing Rheic Ocean, but also in shallow water in the open oceanic part of Paleotethys. Previous explanations for the Kellwasser Events from epicontinental margins and basins call for the migration of deep anoxic bottom water into shallow water environments as a kill mechanism for shallow marine ecosystems or attribute it to sea level rise and subsequent stagnation. There is no evidence that the Devonian oceans completely overturned during the Kellwasser Events; similarly, many transgressive events in the Devonian are not associated with black shales. We therefore suggest an alternative mechanism for the Kellwasser Events based on new evidence from the CAOB, where anoxia is driven by episodic eutrophication of surface waters.

The Late Devonian (Frasnian/Famenian) mass extinction: a proposed test of the glaciation hypothesis

It is argued in this paper that late Frasnian global cooling was the first step in the onset of the Late Palaeozoic Ice Age, and that the late Frasnian extinctions are analogous to the early Oligocene (Oi-1) extinctions that took place in the first step in the onset of the Cenozoic Ice Age. It is argued that the physical evidence for the existence of late Frasnian glaciation, like the Oi-1 glaciation, is largely geochemical: the sharp increases in δ 18 O values and positive δ 13 C anomalies that occurred in the late Frasnian and earliest Famennian. In addition to the geochemical evidence, also like the Oi-1 glaciation, stratigraphic calculations indicate a major sea-level fall occurred during the late Frasnian and early Famennian, a sea-level fall that is argued to have been glacially produced. It is here proposed that the best possible independent physical evidence for the existence of late Frasnian glaciation, other than the geochemical and sea-level evidence, would be the discovery of ice-rafted debris in marine sediments of late Frasnian age similar to the ice-rafted debris found in Oi-1 marine sediments (Zachos et al., 1992; Ehrmann and Mackensen, 1992).

Temporal trends of predation resistance in Paleozoic crinoid arm branching morphologies

The rise of durophagous predators during the Paleozoic represents an ecological constraint imposed on sessile marine fauna. In crinoids, it has been suggested that increasing predation pressure drove the spread of adaptations against predation. Damage to a crinoid's arms from nonlethal predation varies as a function of arm branching pattern. Here, using a metric for resilience to predation (“expected arm loss,” EAL), we test the hypothesis that the increase in predation led to more predation-resistant arm branching patterns (lower EAL) among Paleozoic crinoids. EAL was computed for 230 genera of Paleozoic crinoids and analyzed with respect to taxonomy and time. The results show significant variability among taxa. Camerates, especially monobathrids, display a pattern of increasingly convergent and predation-resistant arm morphologies from the Ordovician through the Devonian, with no significant change during the Mississippian. In contrast, the mean EAL among cladids follows no overall trend through the Paleozoic. Regenerating arms are known to be significantly more common in camerates than in other Paleozoic taxa; if regeneration is taken as a proxy for nonlethal interactions with durophagous predators, this indicates that nonlethal predation occurred more often among camerates throughout the Early and Middle Paleozoic. In addition, frequency of injury among camerates is inversely correlated with EAL and positively correlated with infestation by parasitic snails. From this we conclude that decreasing EAL signals a selective pressure in favor of resistance to grazing predation in camerates but not in other subclasses before the Mississippian, with an apparent relaxation in this constraint after the late Devonian extinctions.

A forest fire and soil erosion event during the Late Devonian mass extinction

The Late Devonian mass extinction occurred in a stepwise manner and culminated close to the Frasnian–Famennian (F–F) boundary (372million years ago). Organic-molecular indices from marine sedimentary rocks at the Sinsin section, Belgium, indicate that the sequence of combustion of land vegetation, soil erosion, and anoxia–euxinia occurred close to this boundary. The increased concentrations of biomarkers indicating forest fire and soil erosion measured in the Sinsin section suggest that fire became widespread at this time, leading to various damaging consequences (increased runoff and oceanic anoxia) that caused marine extinctions. Magnetic susceptibility data in the Sinsin section indicate a relatively dry climate spanning the F–F boundary, which would have encouraged forest fires. The study of organic biomarkers presents several lines of evidence to link forest fire and soil erosion to the Late Devonian mass extinction.

40Ar/39Ar ages of flood basalt provinces in Russia and China and their possible link to global faunal extinction events: A cautionary tale regarding alteration and loss of 40Ar∗

The hypothesis that the Permo–Triassic boundary (PTB) mass extinctions were caused by flood basalt volcanism in Russia (Siberian Traps) and/or China (the Emeishan Traps) is investigated from the point of view of time of occurrence (40Ar/39Ar ages). Numerous published ages in the literature are rejected as good estimates of the time of crystallization. The filters applied in this respect are (a) statistical reliability of plateau/isochron sections of stepheating data and (b) the alteration state of the material that was dated. Alteration appears to be ubiquitous, unsurprising since most of the material dated was used without acid leaching – a procedure that is effective in yielding fresh(er) samples. Of ∼70 ages in the literature for the main pulse of Siberian Trap volcanism, less than ten prove to be reliable ages. Similar techniques applied to 40Ar/39Ar for the Emeishan Traps, leaves only a single reliable age for the magmatic episode. These ages are compared to both published and new 40Ar/39Ar ages for the PTB as based on analysis of minerals from critical ash beds in China. There is good overlap in the ages (PTB – 250.0±0.1Ma, Siberian Trap lavas – 250.1±0.4Ma), lending credence to a genetic link between the formation of the Siberian Traps and the faunal extinction event at the PTB. A similar link for the formation of the Viluy Traps (Russia) and the Late Devonian extinction event is investigated; only a single reliable 40Ar/39Ar age is available for the Viluy Traps, and falls close to the interpolated age for the Frasnian–Fammenian boundary. The use of the unspiked K–Ar technique to yield accurate ages for such (altered) samples is questioned.A review of U–Pb data pertinent to these problems suggests a close temporal link between the formation of the Siberian Traps and the PTB. Comparison of U–Pb and 40Ar/39Ar ages for the PTB, raises questions about the accuracy of high precision sanidine ages, possibly resulting from very slow leakage of 40Ar∗ from this mineral.

New 40Ar/39Ar and K–Ar ages of the Viluy traps (Eastern Siberia): Further evidence for a relationship with the Frasnian–Famennian mass extinction

We present here new ages, combining both K–Ar and 40Ar/39Ar techniques, reinforcing the possible link between the Viluy traps (eastern Siberia) with the Late-Devonian mass extinctions. The Viluy traps (also referred to as the Yakutsk large igneous province) are associated with the breakup of the eastern margin of the Siberian platform as a triple-junction rift system, with volcanic outcropping along the Lena, Markha and Viluy rivers and an initial volume of volcanic products of about one million cubic kilometers. K–Ar ages obtained on seven samples from the Viluy rift display values ranging from 366±5 to 381±5Ma, with a mean age of 374±5Ma. Our new 40Ar/39Ar results display well-defined step-heating plateau ages ranging from 362.3±3.4 to 379.7±3.5Ma. Although most K–Ar and 40Ar/39Ar ages agree within about 1% uncertainty, which reinforces the validity of the K–Ar ages obtained with the Cassignol and Gillot unspiked technique (even in the Paleozoic), a slightly clearer scenario appears when only the 40Ar/39Ar ages are considered. Effectively, they cluster in two different groups with weighted mean ages of 364.4±1.7 and 376.7±1.7Ma. These ages support a multi-phases emplacement of the Viluy traps, as is the case for many continental flood basalts. Because the age of the first volcanic phase is undistinguishable from recent determinations of the age of the Frasnian–Famennian boundary, our results further strengthen the hypothesis of a causal relationship between the two events. The second phase of emplacement of the Viluy traps occurred at the end of the Devonian, apparently slightly before the Devonian–Carboniferous boundary. Finally, except for the Ordovician–Silurian extinction which remains to be associated with a volcanic province, all major Phanerozoic mass extinctions have been correlated with the emplacement of a CFB, thereby reinforcing the possible causal link between these events.

Onshore expansion of benthic communities after the Late Devonian mass extinction

A detailed ichnological analysis of the Upper Devonian–Lower Mississippian Bakken Formation of sub-surface Saskatchewan and the partially coeval Exshaw Formation of Alberta indicates the presence of an anomalous ichnofacies gradient. The distal Cruziana Ichnofacies, which in rocks of other ages is restricted to lower-offshore facies, here ranges from this setting to the lower shoreface. No archetypal Cruziana Ichnofacies is present in these deposits. This pattern is interpreted as resulting from the differential effects of the Late Devonian mass extinction in shallow-water ecosystems. The onshore expansion evidenced by ichnological data is consistent with the pattern displayed by the body-fossil record, which indicates a re-invasion of shallow-water environments by the Palaeozoic evolutionary fauna during the Late Devonian and into the Early Carboniferous. The ichnofauna studied is overwhelmingly dominated by deposit feeders, with suspension feeders being notably absent, further underscoring the importance of trophic type as a selectivity trait during mass extinctions.