222 PALAEOZOIC/End Permian Extinctions and that their samples came from below the extinction level Later in 2001, Kaiho and colleagues described Permian–Triassic sediment grains that were supposedly formed by impact, as well as geochemical shifts that they interpreted as indicative of a huge impact However, their data are far from conclusive and were rapidly and severely criticized by other geochemists In 2003, Asish Basu and colleagues revived interest by claiming to have found 40 tiny (50–400 mm) unaltered fragments of meteorite in a sediment sample from a terrestrial Permian–Triassic boundary section in Antarctica Although Basu and colleagues were quick to dismiss contamination as a source of these grains, meteorite experts were immediately sceptical of the findings, as meteoritic metals are highly reactive and, in terrestrial settings, oxidize extremely quickly Doubtless scientists will continue to produce evidence of extraterrestrial impact at or near the Permian–Triassic boundary, if only because such hypotheses readily appeal to the wider public and journal editors alike However, current data are highly controversial and lack the key criteria of major impact that have been recorded time and again in Cretaceous–Tertiary sections the world over Independent replication of results is crucial for scientific acceptance, especially when the data are unusual and controversial So far, all the evidence proposed for an impact at the Permian–Triassic boundary has failed this necessary test Eruption The largest outpouring of continental flood basalts in the Phanerozoic occurred in Siberia during the Permian–Triassic interval Including both the Siberian Platform basalts and the newly discovered coeval deposits buried in the West Siberia Basin, described by Marc Reichow and colleagues, the flood basalts covered an area of 1.6 Â 106 km2 to maximum depths of 3.5 km (Figure 1) If all other igneous rocks, such as pyroclastic flows, are included, then this coverage increases to 3.9 Â 106 km2 Dating the top and bottom of the lava pile shows that the eruptions occurred over a relatively short period of time, maybe just 600 000 years Was this huge volcanic event a cause of the Permian–Triassic extinction crisis? Radiometric dating is the only way to answer this question because no fossils have yet been collected from sediments interbedded with the basalts that provide correlation with other regions Early efforts at dating the Siberian Traps produced an array of ages, from 160 Ma to 280 Ma By contrast, more recent results, by different scientists using a variety of methods, cluster around 250 Ỉ Ma In the late 1990s, this was considered to be exactly the date required, and the flood basalts were thought to be the primary trigger for the catastrophic extinction However, the recent redating of the Meishan beds now implies that the Permian–Triassic event occurred before 253 Ma, and the Siberian Traps are therefore several million years too young On a Phanerozoic time-scale, the excellent correlation between extinction episodes and flood-basalt provinces means that it is difficult to accept the possibility that the Siberian Traps played no role in the end-Permian extinction event There is still a chance The oldest date that Reichow’s team have for the onset of volcanic activity (dating the emplacement of intrusive gabbros) is 253.4 Ỉ 0.8 Ma Another intriguing consequence of the 253 Ma age for the Permian–Triassic boundary is that the Emeishan flood basalts of western South China now need to be considered Dating by Ching-Hua Lo and colleagues (published in 2002) shows that the main eruptions, which were largely marginal marine events, occurred around 251–253 Ma, with an initial phase of activity at around 255 Ma; thus, they predate the Siberian Traps and are dated close to the new dates for the Permian–Triassic crisis Interestingly, Ching-Hua Lo’s new dates preclude the possibility that the Emeishan flood basalts played a role in the end-Guadalupian crisis (at about 256–259 Ma; see above) Mei-Fu Zhou and colleagues had championed this particular correlation following their dating of the Emeishan basalts to 259 Æ Ma, which was published just a few months before Ching-Hua Lo’s results in the same journal! Absolute dating is a difficult business (see Analytical Methods: Geochronological Techniques) Global Warming Animals and plants not curl up and die just because a volcano is erupting on the other side of the world Species become extinct when their population falls below viable levels Population decline in natural systems is a response to local changes, such as loss of habitat, drought, and temperature change, many of which are ultimately driven by climate What climate changes are recorded around the extinction interval and how might they be related to the flood-basalt volcanism? Global cooling was proposed early on as a cause of the Permian–Triassic event; this largely hinged on the identification of latest-Permian glacial deposits in Siberia and eastern Australia However, reanalysis of the biostratigraphy proved that these deposits are,