What can we learn about ecology and evolution from the fossil record? Jeremy B.C. Jackson 1,2 and Douglas H. Erwin 3 1 Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0244, USA 2 Smithsonian Tropical Research Institute, Box 2072 Balboa, Republic of Panama 3 Department of Paleobiology, MRC-121, National Museum of Natural History, Washington DC, 20013-7012, USA The increased application of abundance data embedded within a more detailed and precise environmental context is enabling paleontologists to explore more rigorously the dynamics and underlying processes of ecological and evolutionary change in deep time. Several recent findings are of special theoretical inter- est. Community membership is commonly more stable and persistent than expected by chance, even in the face of the extreme environmental changes of the Ice Ages, and major evolutionary novelties commonly lie dormant for tens of millions of years before the ecological explosions of the clades that possess them. As we discuss here, questions such as these cannot be adequately addressed without the use of the fossil record. Introduction ‘Our ability to reconstruct past communities and the history of life hinges on thinking like biologists doing field work in ancient environments.’ [1] The fossil record is a uniquely rich source of information about the patterns of past biological diversity: the waxing and waning of various clades and their component species, of novel morphologies and of ecological communities unlike those alive today. From rigorously documented patterns, paleontologists can also infer much about basic ecological and evolutionary processes that cannot be revealed by the study of living organisms alone. Some of the best examples of the synthesis of paleobiological pattern and process include the influence of Quaternary [since 1.8 million years ago (Mya)] climate change on the distribution and composition of temperate North American plant commu- nities [2], the causes and consequences of the end- Cretaceous mass extinction (65.5 Mya) [3] and the Paleo- cene–Eocene thermal maximum (55.8 Mya) [4], and the environmental setting of the spectacular radiation of animals during the Ediacaran–Early Cambrian (575– 509 Mya) [5]. In each case, paleontologists have worked with other geologists to obtain extensive new collections of fossils tied to detailed and independent records of changes in climate, oceanography, tectonic events and other aspects of the physical–chemical environment through geochemical proxies. Such abiotic forcing drives many ecological and evolutionary processes, so careful attention to reconstruct- ing environmental scenarios independent of the biological record is essential. The examples that we discuss here reveal how ecological and evolutionary processes can be reconstructed from fossil data. These new insights provide a deep-time perspective to longstanding and fundamental ecological and evolutionary questions for which biological data are inadequate, and have inspired paleontologists to address more vigorously, and with greater focus, the roles of intrinsic versus extrinsic drivers of evolutionary change, a question that Stephen Jay Gould described as one of the ‘eternal metaphors’ in understanding the history of life [6]. Here, we discuss three fundamental ecological and evolutionary questions for which breakthroughs based on the fossil record are occurring because of a more holistic paleontological perspective, the formulation of testable hypotheses about process, and new programs of field work and data collection specifically designed to test them. Testability and an emphasis on process have long been hallmarks of research in other areas of ecology and evolution and must be more widely adopted by paleobiol- ogists to ensure continued progress toward our under- standing of the history of life. Ecologists developed an approach to experimental design and testability that became more rigorous during the 1960s in response to what many believed was an overly descriptive approach and has set the standard for most empirical ecology ever since. However, with the advent of macroecology and the need to confront pressing problems of human impacts, ecologists are once again increasingly faced with deducing process from pattern without recourse to experimental manipulation, the situation that paleontologists have always faced. But paleontologists still have much to learn from ecologists about the nature of sampling and the data required to test alternative hypotheses about the processes driving biodiversity [7,10,54]. Until recently, most paleontological data collected was used to document new species, to understand the geographical, environ- mental and temporal extent of existing species, and to document and understand ecological relationships. New hypotheses commonly require new data and, as paleontol- ogists more rigorously define hypotheses and better understand how to test them, they will have to better Corresponding author: Jackson, J.B.C. (jbjackson@ucsd.edu). ARTICLE IN PRESS Review TRENDS in Ecology and Evolution Vol.xx No.xx Monthxxxx www.sciencedirect.com 0169-5347/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tree.2006.03.017