Abstract

Following publication in 1980 of the celebrated paper in Science (208, 1095–1108) by Luis Alvarez and his colleagues, a widespread interest developed in the possibility that bolide impact was a major if not indeed the most important cause of mass extinctions in the Phanerozoic. Of particular interest was the claim of a 26 Ma extinction periodicity in the last 250 Ma, which led to suggestions of ultimate astronomical control. Intensive research subsequently casts serious doubt on such a periodicity, and the only really persuasive evidence associating impact with mass extinction is that at the Cretaceous-Tertiary boundary.

It is generally accepted that there were five major episodes of mass extinction in the Phanerozoic record: end-Ordovician, late Devonian (Frasnian-Famennian boundary), end-Permian, end-Triassic and end-Cretaceous. In addition a number of lesser events have been widely recognized, most notably at the Devonian-Carboniferous, Cenomanian-Turonian and Paleocene-Eocene boundaries, and in the early Toarcian. The extinction record is based essentially on marine invertebates, and the evidence from terrestrial vertebrates is far less clear, with the exception of dinosaurs at the Cretaceous-Tertiary boundary. Although it has been maintained that terrestrial plants were not subjected to mass extinction episodes there is abundant evidence for at least extensive regional extinction events, most notably at the Permian-Triassic, Triassic-Jurassic and Cretaceous-Tertiary boundaries.

Only a small number of factors can plausibly be invoked to account for environmental perturbations on a global scale, of the sort that could significantly increase extinction rate: bolide impact, volcanism, climatic cooling, climatic warming, marine regression and marine transgression with the accompanying spread of anoxic bottom waters. Good evidence for bolide impact, in the form of significant iridium anomalies on a global scale, and shocked quartz coincident with mass extinction, is only available for the end-Cretaceous event, and a likely impact crater has been identified in the Yucatan Peninsula of Mexico. Microtektites have been recognized both in Upper Devonian and Upper Eocene strata, but their horizons do not coincide with mass extinction events. Claims for impact at the end of the Triassic, based on purported shocked quartz, are decidedly dubious. The only good correlation with large-scale flood basalt volcanism is for the end-Permian and end-Cretaceous events, respectively the Siberian and Deccan Traps, but any causal correlation is more likely than not to involve climate. Climatic cooling is strongly implicated only for the latest Ordovician and late Eocene to Oligocene, and climatic warming episodes, as in the earliest Silurian, Triassic and Eocene, were probably at least as important. Marine regression caused by sea-level fall correlates well with extinction events in the late early Cambrian, end-Ordovician, late Permian, end-Triassic and end-Cretaceous. More important than any other factor, anoxia and transgression appear to be strongly associated with events in the late early Cambrian, late Cambrian, earliest Silurian, Frasnian-Famennian and Devonian-Carboniferous boundaries, Permian-Triassic, Cenomanian-Turonian and Paleocene-Eocene boundaries, and early Toarcian.

Often, of course, more than one factor may be implicated, one of the most striking associations being between episodes of marine transgression-anoxia and climatic warming. Such environmental changes could clearly affect both marine and terrestrial organisms simultaneously. Even if one accepts bolide impact as a major causal factor for the end-Cretaceous extinctions, its environmental effects are still controversial and difficult to disentangle from those caused by phenomena intrinsic to our planet. Arguments persists among palaeontologists about how catastrophic the mass extinction was.