2. Toarcian extinction

Hallam's earliest studies were on the Liassic strata in and around the United Kingdom, and these allowed him to pioneer the investigation of topics such as trace fossils, facies analysis and sea-level change. Hallam's (Reference Hallam1961) paper on sea-level and faunal change in the Early Jurassic Epoch of Europe is especially important because it contains a species-level range chart for marine invertebrates constructed at the level of ammonite zones – a temporal resolution that has rarely been bettered even today. The chart reveals the loss of 64 of 66 benthic species at the base of the Toarcian Stage at a point where black shales, such as the Jet Rock in the UK, become widespread. The link between the two phenomena was immediately apparent, and Hallam concluded that there was a ‘mass disappearance followed by a complete renewal of forms [and] It can hardly be a coincidence that this striking change is intimately correlated with the widespread development of finely laminated bituminous shales, signifying bottom conditions unfavourable to benthic organisms.’ (Hallam, Reference Hallam1961, p. 154). Thus, was born the idea of a mass extinction linked with the spread of marine anoxia, decades before such ideas became commonplace.

Hallam's subsequent studies have added detail to the timing and nature of the Early Jurassic extinction losses. Thus, extinctions amongst the nektonic ammonites and belemnites were found to occur at the Pliensbachian–Toarcian boundary whilst the benthic losses were a little later, in earliest Toarcian time, at the point at which black shales become extensive (Hallam, Reference Hallam1967, Reference Hallam1976). Amongst a plethora of cogent observations, Hallam noted that the benthic bivalves show little evidence for ecological selectivity amongst the losses (Hallam, Reference Hallam1986). This discovery provides an interesting contrast with extinction selectivity during other crises such as the end-Triassic mass extinction discussed in Section 3 below.

As well as their temporal complexity, the Toarcian extinction losses also show geographic variation, with the timing of the extinction varying from region to region as first shown in Hallam's (Reference Hallam1972) study of the Iberian record. Black shales are weakly developed in this region, and some extinction losses in Portugal occur considerably after those in northern Europe. The Lower Jurassic of South America similarly has a dearth of black shales, and this region too shows an extinction history somewhat nonsynchronous with that of northern Europe (Hallam, Reference Hallam1986). Despite this temporal variation, the observation that transgression, anoxia and marine extinction went hand-in-hand in Early Jurassic time has been shown to be a generally recurring theme during other mass extinctions (Hallam, Reference Hallam1986, Reference Hallam1987, Reference Hallam1989; Hallam & Wignall, Reference Hallam and Wignall1999) and in some regards can be considered a ‘rule’ in global biodiversity studies (e.g. Hannisdal & Peters, Reference Hannisdal and Peters2011).

Much current research effort continues to be expended on the Toarcian extinction, with many workers especially concerned with the interpretation given to the substantial carbon isotope oscillations of early Toarcian time and in particular the sharp negative excursion at the same time that black shale deposition became widespread (Hesselbo et al. Reference Hesselbo, Gröcke, Jenkyns, Bjerrum, Farrimond, Morgans Bell and Green2000; Wignall et al. Reference Wignall, Hallam, Newton, Sha, Reeves, Mattioli and Crowley2006; Harazim et al. Reference Harazim, van de Schootbrugge, Sorichter, Fiebig, Weug, Suan and Oschmann2013). Oxygen isotope data have also shown that the interval was a period of warming (Bailey et al. Reference Bailey, Rosenthal, McArthur, van de Schootbrugge and Thirlwall2003), and radiometric dating reveals that the crisis coincides with the eruption of a large igneous province (LIP), the Karoo-Ferrar Traps of the Gondwanan continent (Pálfy & Smith, Reference Pálfy and Smith2000; Svensen et al. Reference Svensen, Planke, Chevallier, Malthe-Sørenssen, Corfu and Jamtveit2007).

The Toarcian crisis, thus, has many of the hallmarks of most Phanerozoic mass extinctions: LIP eruption, global warming and δ¹³C negative excursions (Hallam & Wignall, Reference Hallam and Wignall1997), together with transgression and marine anoxia – the two facets that were first identified by Hallam in 1961. The recognition of a temporal complexity to extinction losses also continues to feature as a key issue in understanding this event (e.g. Wignall, Newton & Little, Reference Wignall2005) and especially its relationship with basaltic eruptions (Caruthers, Smith & Gröcke, Reference Caruthers, Smith and Gröcke2013).

3. End-Triassic mass extinction

Was there a mass extinction at the end of the Triassic Period? Prior to the seminal work of Alvarez et al. (Reference Alvarez, Alvarez, Asaro and Michel1980) and its catastrophist message, the idea of an abrupt end-Triassic mass extinction had little currency. A major Triassic–Jurassic turnover of terrestrial vertebrates had been identified by the mid-twentieth century but it was thought to be a protracted affair rather than a short-lived crisis. The losses were viewed as ‘only a part of a large and extended sequence of events that makes the Triassic period a time span of great significance’ (Colbert, Reference Colbert1958, p. 973). In contrast, Newell (Reference Newell and Albritton1967) considered the end-Triassic to be one of several Phanerozoic mass extinction events that were marked by the rapid loss of many diverse groups. In this case, Newell especially highlighted the coincident loss of numerous ammonoids and reptiles at the end of the Triassic Period, pointing to a crisis on both land and sea. Today Newell's claims are uncontroversial, but at the time they were distinctly at odds with prevailing ideas and had little or no influence on contemporary studies. The idea of gradual change was deeply entrenched: ‘the transition from the Triassic to the Jurassic was not marked by sudden, simultaneous extinctions of large numbers of higher order taxa of vertebrates, but instead was a time of gradual faunal replacement’ (Olsen & Galton, Reference Olsen and Galton1977, p. 985). Indeed, the main changes amongst tetrapods were placed tens of millions of years earlier, within the Carnian Stage (e.g. Bakker, Reference Bakker and Hallam1977).

This appreciation profoundly changed with Hallam's evaluation of the marine fossil record (Hallam, Reference Hallam1981), and saw the end-Triassic crisis gain a mass extinction status. Hallam demonstrated that European bivalves show major losses in a geologically brief interval of latest Triassic time. Interestingly, unlike the Toarcian crisis, Hallam showed the extinction was clearly selective, with epifaunal groups showing much greater extinction losses compared with the sediment-dwelling infaunal groups (Hallam, Reference Hallam1981).

Hallam's paper provided clear data that showed the severity of the marine losses and invigorated the debate on the end-Triassic's mass extinction status. The main questions that developed in the 1980s and continue to be addressed to this day are: (1) How severe was this crisis, especially on land? (2) How quickly did it occur? (3) Was it in fact just a minor event in comparison with a much more intense Carnian extinction?

Within a few years of Hallam's Reference Hallam1981 paper, the end-Triassic mass extinction crisis was being ranked alongside the end-Cretaceous event and it even had its own meteorite-impact crater at Manicouagan in Quebec Province (Olsen, Shubin & Anders, Reference Olsen, Shubin and Anders1987). However, the crater is now known to be much too old to be implicated in an end-Triassic extinction, and for many palaeontologists the key extinction happened earlier. Thus, Benton has argued that both marine and terrestrial extinctions were at their most severe within the Carnian Age whereas the end-Triassic losses were both less intense and less sudden – very much a second-rate crisis compared with the earlier event (Benton, Reference Benton1986, Reference Benton1991). Late Carnian extinctions removed many taxa from species-rich tetrapod families (e.g. the kannemeyerids and the rhynchosaurs) but the end-Triassic event only removed species-poor families. Thus, Benton concluded that ‘The Carnian event unequivocally had greater impact than the end-Triassic event among terrestrial vertebrates’ (Benton, Reference Benton1991, p. 270) and ‘the end-Triassic extinction was a whimper’ (Benton, Reference Benton1991, p. 263). These views were very much in keeping with pre-1981 views and have echoes in the earlier opinions of Colbert (Reference Colbert1958) and Bakker (Reference Bakker and Hallam1977). More recent reviews have tended to acquiesce with Benton's (e.g. Tanner, Lucas & Chapman, Reference Tanner, Lucas and Chapman2004).

Despite these attempts of vertebrate palaeontologists to downgrade or dismiss the status of the end-Triassic terrestrial mass extinction, the marine record clearly shows substantial losses, albeit potentially spread over the last few million years of the Triassic (Hallam, Reference Hallam2002). More recent studies also suggest that the tetrapod extinction may indeed have been a severe but selective one. The diverse and successful pseudosuchians suffered severe end-Triassic losses leaving only crocodylomorphs, which radiated rapidly in Early Jurassic time (Toljagić & Butler, Reference Toljagić and Butler2013). In contrast dinosaur extinctions were minor (Brusatte et al. Reference Brusatte, Nesbitt, Irmis, Butler, Benton and Norell2010).

As well as exploring the nature of the marine extinctions, Hallam has also pioneered palaeoenvironmental studies of the Triassic–Jurassic transition interval. The end-Triassic crisis coincided with the onset of a phase of long-term sea-level rise that saw the flooding of extensive low-lying areas in central Pangaea (now western Europe). By the end of Early Jurassic time extensive epicontinental seas had developed, but in the latest Triassic Rhaetian Age this flooding had only just begun, and it is within Rhaetian strata that the extinction record is preserved. Thus, the relationship between sea-level change and extinction at the end of the Triassic Period is enigmatic. Hallam (Reference Hallam1981) considered Rhaetian eustasy to be the culmination of a first-order lowstand with a superimposed second-order transgressive–regressive couplet. The terminal Rhaetian regression is mostly clearly manifest as a karst surface developed atop reefs in the Austrian Alps (Satterley, Marshall & Fairchild, Reference Satterley, Marshall and Fairchild2006).

There is, thus, a temporal link between regression and extinction at the end of the Triassic Period and the notion that the two phenomena are causally linked has a long pedigree. It was first explicitly proposed by Newell (Reference Newell and Albritton1967). The idea derives from MacArthur and Wilson's species-area effect and it suggests that, as shallow seas retreat and shallow-marine habitat areas are lost, marine extinction rates increase (although the concept clearly cannot be invoked to cause terrestrial losses). The general link between sea level and diversity is a recurrent and disputed theme in biodiversity studies (e.g. Hallam, Reference Hallam1977; Hallam & Wignall, Reference Hallam and Wignall1999; Hannisdal & Peters, Reference Hannisdal and Peters2011; Smith, Reference Smith2007). The link at the end of the Triassic Period is complicated because the terminal Rhaetian regression was swiftly followed by the rapid spread of anoxic bottom waters, a phenomenon that could have been also bound up in the extinction and its aftermath (Hallam, Reference Hallam1981, Reference Hallam1995; Richoz et al. Reference Richoz, Van De Schootbrugge, Pross, Püttman, Quan, Lindström, Heunisch, Fiebig, Maquil, Schouten, Hauzenberger and Wignall2012). Summing up the difficulties of disentangling the key factor, Hallam noted that: ‘Because such transgressions normally follow quickly after major regressions, it is not always clear from analysis of extinction events what the critical causal factor was, although in both cases [regression and transgression-with-anoxia] there would have been a reduction in benthic and nektobenthic habitable area.’ (Hallam, Reference Hallam1989, p. 443).

Subsequent studies by Hallam and other workers have added to, and to some extent, clarified the possible culprits for the end-Triassic mass extinction. Analysis of the rate and regional variation of sea-level change has provided further, indirect evidence of a potential smoking gun. The Rhaetian sea-level changes seen in Europe, with their regressive–transgressive couplet around the end of the Triassic Period are only weakly manifest in North America (Hallam & Wignall, Reference Hallam and Wignall2000) and are not seen at all in South America (Hallam, Reference Hallam1989) nor in the Perigondwanan sections of southern Tibet (Hallam et al. Reference Hallam, Wignall, Yin and Riding2000) where the story is one of gradual sea-level rise across the Triassic–Jurassic boundary (Hallam & Wignall, Reference Hallam and Wignall1999).

Rates of sea-level change across the Triassic–Jurassic boundary in western and central Europe (but not further afield) are too fast to be attributed to normal eustatic drivers such as changes in mid-ocean ridge spreading rates (Hallam, Reference Hallam1997). Instead, they can be linked with North Atlantic tensional tectonic activity ‘where it is associated with substantial basaltic activity’ (Hallam, Reference Hallam1997 p. 777) and widespread deformed horizons that probably formed as a consequence of this activity (Hallam & Wignall, Reference Hallam and Wignall2004).

Studies of the fragments of a flood basalt province now found in Morocco, Spain, Brazil and the United States have borne out Hallam's claim of ‘substantial basaltic activity’. This volcanism is now recognized as the Central Atlantic Magmatic Province, or CAMP, and it is one of the largest of all LIPs with an eruption onset that appears to closely coincide with the mass extinction (Marzoli et al. Reference Marzoli, Renne, Piccirillo, Ernesto, Bellieni and De Min1999; Wignall, Reference Wignall2005; van de Schootbrugge et al. Reference van de Schootbrugge, Quan, Lindstrm, Pttmann, Heunisch, Pross, Fiebig, Petchick, Rhling, Richoz, Rosenthal and Falkowski2009; Pálfy & Kocsis, Reference Pálfy, Kocsis, Keller and Kerr2014; Bachan & Payne, Reference Bachan and Payne2015, this issue).

4. End-Permian mass extinction and its aftermath

Whilst the role of widespread marine anoxia during the end-Triassic crisis is enigmatic, there is a much more clear-cut link between anoxia and the end-Permian mass extinction. Hallam's work in collaboration with one of us (PBW) was the first to show the close synchrony between the spread of marine anoxia and the marine extinction losses (Wignall & Hallam, Reference Wignall and Hallam1992, Reference Wignall and Hallam1993). There are, however, close parallels between Triassic–Jurassic and Permian–Triassic events: sea-level changes for both show a regression–transgression couplet (Hallam & Wignall, Reference Hallam and Wignall1999). The contrast is that whilst the end-Permian losses occurred during the transgressive–anoxic phase of the sea-level cycle, the end-Triassic losses were during the preceding regression. Summing up, Wignall & Hallam (Reference Wignall and Hallam1992, p. 43) concluded that, ‘the disappearance of Permian faunas at the end of the period is abrupt, taking place not at the level of regression but shortly above, associated with a major transgression [and] dysaerobic to anaerobic conditions’.

The concept of an ‘abrupt’ end-Permian mass extinction was, in the early 1990s, counter to the widely held view of a protracted extinction spanning several million years (e.g. Teichert, Reference Teichert, Kauffman and Walliser1990). It has subsequently proved to be a rather non-controversial claim. It is certainly less contentious than the notion that there was an abrupt end-Triassic extinction. Most studies of the past 20 years have viewed the end-Permian crisis to have taken place in a geologically short period of time, probably only a few tens of thousands of years at most (e.g. Kaiho et al. Reference Kaiho, Chen, Kawahata, Kajiwara and Sato2006): the latest radiometric dates suggest 60 ka (Burgess, Bowring & Shen, Reference Burgess, Bowring and Shen2014). However, detailed study of marine sections has revealed that the crisis in South China can be resolved into two separate extinction events straddling the Permian–Triassic boundary (Song et al. Reference Song, Wignall, Tong and Yin2013). It may be that the discrete extinction intervals were much shorter than this (a few thousand years?) and it is their spacing that is measured in tens of thousands of years.

The parallels between the end-Permian and the end-Triassic and Toarcian extinctions are manifold, but it is the aftermath of the crisis that marks the end-Permian crisis out as a uniquely severe event. Hallam investigated the fortunes of bivalves with Miller and revealed the peculiar nature of the Early Triassic aftermath fauna. Rather than marking a recovery and diversification phase, there was instead a prolonged phase of low diversity throughout Early Triassic time followed by the reappearance of many bivalves in Middle Triassic time that had not been seen since Permian time (Hallam & Miller, Reference Hallam, Miller and Larwood1988). Hallam returned to this theme in 1991 in a short but influential paper in which he suggested that the long-delayed recovery was caused by the prolonged duration of the harsh conditions (such as marine anoxia) that had triggered the mass extinction (Hallam, Reference Hallam1991). Subsequent study has confirmed that the Early Triassic Epoch was indeed marked by one of the most prolonged and intense phases of Phanerozoic oceanic anoxia (Isozaki, Reference Isozaki1997; Wignall et al. Reference Wignall, Bond, Kuwahara, Kakuwa, Newton and Poulton2010, Reference Wignall, Bond, Sun, Grasby, Beauchamp, Joachimski and Blomeir2015, this issue).

As with all major topics, the nature of the Early Triassic world, its biota and the role of environmental constraints on radiation have been the subject of intense debate. There are currently three distinct viewpoints:

1. (1) The Early Triassic world was a harsh one that inhibited the recovery of all but a few hardy groups (Hallam's original idea). The recognition of extremely hot conditions at this time (Sun et al. Reference Sun, Joachimski, Wignall, Yan, Chen, Jiang, Wang and Lai2012), alongside the widespread anoxia, adds credence to this stance as do studies of the recovery record in South China (Song et al. Reference Song, Tong, Wignall, Luo, Tian, Song, Huang and Chu2015, this issue).

2. (2) The Early Triassic world was normal but the preceding extinction had caused such devastation that it took a long time for the biota to even start to recover. This viewpoint can be traced back to a highly influential paper by Schubert & Bottjer (Reference Schubert and Bottjer1992). They identified the extraordinary abundance of stromatolites in Early Triassic time and argued that the dearth of grazers, such as gastropods, following the mass extinction allowed cyanobacteria to flourish.

3. (3) The Early Triassic world was normal as was the recovery, which proceeded untrammelled. For this ‘nothing unusual’ view see the recent study of Hofman et al. (Reference Hofman, Hautmann, Wasmer and Bucher2013) on the post-extinction record in the US Rockies.

Clearly these are all discordant views and the debates continue, but it was Hallam's characteristically perceptive thoughts that sparked this research field. To some extent, the most recent studies suggest some compromise between these alternatives with an initial early harsh environmental phase followed by the first hints of recovery in early Olenekian time only to be pegged back by a resultant crisis around the Smithian–Spathian boundary (e.g. Song et al. Reference Song, Wignall, Chen, Tong, Bond, Lai, Zhao, Jiang, Yan, Nin, Chen, Yang and Wang2011). But the original observation of Hallam remains pertinent – alpha diversity in Early Triassic environments was exceptionally low. It remains to be seen if a consensus develops or whether, like the end-Triassic extinction, the same themes are still being debated in the decades to come.