ABSTRACT
The spectre of human-induced alteration of atmospheric composition and associated changes in climate has focused attention on how species, communities, and ecosystems respond to climate change (Solomon and Shugart, 1993). Such concern has prompted a wide range of research from short-term experiments which examine specific controls over the responses of individual tissues or plant parts, to the modelling of whole systems which predict ecological response at different levels of biological organization and spatio-temporal scales (Ehleringer and Field, 1993). While long-term, whole system experimental programmes in a broad range of biomes would be the ideal complement to efforts to understand the physiological response of individual plants, such idealism is constrained by funding and logistics. An alternative source of information about how species, communities, and ecosystems respond to environmental change is the palaeoecological record. Palaeoecology offers insights into the nature of climate-vegetation interactions that derive from the well-documented response of plant communities to environmental changes of the past: a record that is equivalent to natural, if unplanned, experiments (Davis, 1989). Since these natural experiments typically include conditions not observed in the twentieth century, the palaeo-record documents a substantially broader range of biotic responses to environmental variations than can be obtained from observational data.
