often point to causes. In global mass extinctions such as the terminal Ordovician, Cenomanian-Turonian, and terminal Cretaceous events, tropical taxa, including reef communities, suffered preferentially. This is consistent with the idea that climatic cooling played a major role in extinction, but it may also reflect the typically narrow niche breadth of tropical taxa and the high degree of interdependence among species.
High-resolution stratigraphic and paleoenvironmental studies are crucial for understanding mass extinctions. Such studies reveal that a major extinction in deep-sea benthic assemblages at the end of the Paleocene, the most profound of the last 90 m.y. for this habitat, resulted from rapid warming of the deep oceans in conjunction with global warming (see Kennett and Stott, Chapter 5). Extinctions that removed between 35 and 50% of deep-sea taxa occurred in less than 2000 years, equal to the time required for the deep water to circulate through its basins. For a few thousand years, ocean circulation underwent fundamental changes that affected the deep-sea biota. High-resolution study of this event has illustrated, first, how events that are geologically brief but not instantaneous can strongly alter the ecosystem and, second, how such changes can be largely decoupled from events in other segments of the biosphere.
Patterns of extinction can point to particular causes of mass extinction. For example, the severe extinction of western Atlantic bivalve mollusks during the onset of the modern ice age seems to have eliminated all strictly tropical species of southern Florida; all survivors have broad thermal tolerances, ranging well beyond the tropics today. Here, a thermal filter seems clearly to have operated (see Stanley and Ruddiman, Chapter 7). Similarly, that climatic change was the ultimate cause of the previously discussed severe extinction of African mammals at the start of the modern ice age (-2.5 m.y. ago) is supported not only by the evidence that forest habitats shrank at this time but also by the fact that forest-adapted species were the primary victims (Vrba, 1985).
Some patterns of extinction have characterized higher taxa in more than one mass extinction. A striking aspect of the terminal Cretaceous extinction of planktonic foraminifera was the disappearance of species with large, complex, highly ornamented skeletons (see Keller and Perch-Nielsen, Chapter 4). Survivors were inherently small species or species that became dwarfed during the crisis. Deep-water planktonic species also died out first and in the largest numbers. These patterns must be taken into account in any analysis of the proximate causes of extinction. Most species of planktonic foraminifera that became extinct during the Late Eocene to Early Oligocene extinction were also complex, highly ornamented species. These were also largely warm-adapted taxa, which is compatible with evidence that climatic changes were the primary cause of this global crisis.
If there is one general pattern for extinctions, it is the rate of environmental change and not necessarily its magnitude that places most populations in jeopardy. This consideration is highly relevant to global changes predicted for the next century. If current models are correct, the magnitude of change will not be unusual on a geological time scale, but the rate of change may be.
Evidence of causation also comes from the nature of species that immigrate into a region or originate within it during or soon after a pulse of extinction. In other words, the disappearance of some species and the appearance of others during a brief episode of evolutionary turnover should offer compatible testimony about environmental change. Thus, not only did forest-adapted species of antelopes preferentially die out in Africa about 2.5 m.y. ago, but newly appearing species were virtually all adapted to grassy habitats. Simultaneously, the apparently semiarboreal gracile australopithecines gave way to early Homo, which had helpless infants and could not have climbed trees habitually for refuge (see Stanley, Chapter 14).