EXTINCTIONS IN THE FOSSIL RECORD

Direct comparison of ancient extinctions to the present-day situation is difficult, because quantitative palaeontological data come primaril y from marine invertebrates, fossilized species are usually drawn from the more abundant and widespread taxa, and time resolution is rarely better than 10(3)-10(4) years. A growing array of techniques permits q uantitative error estimates on some of these potential biases, and all ows calculation of species extinction intensities from genus-level dat a, which are more robust. Extensive as today's species losses probably are, they have yet to equal any of the Big Five mass extinctions. Bac kground extinction patterns are potential sources of insight regarding present-day biotic losses; over 90% of past species extinction has oc curred at times other than the Big Five mass extinctions. Mean duratio ns of fossil species vary by more than an order of magnitude even with in clades, rendering uninformative any global average for background e xtinction. Taxon-specific variation is evidently related to intrinsic biotic factors such as geographic range and population size. Approache s to extinction analysis and prediction based on morphological variety or biodisparity should be explored as an adjunct or alternative to ta xon inventories or phylogenetic metrics. Rebounds from mass extinction s are geologically rapid but ecologically slow; biodiversity recovery and the re-establishment of some communities typically requires 5-10 m illion years.