In fragmented tropical landscapes, among the most pervasive causes of ecological change are edge effects – diverse ecological alterations associated with the abrupt, artificial boundaries of forest fragments (Laurance & Bierregaard 1997, Lovejoy et al. 1986, Turner 1996). A striking edge effect in fragmented Amazonian forests is chronically elevated tree mortality (Ferreira & Laurance 1997, Laurance et al. 1998a). Large (> 60 cm diameter) trees are especially vulnerable to fragmentation, dying three times faster within 300 m of edges than in forest interiors (Laurance et al. 2000). Elevated tree mortality alters canopy-gap dynamics, promotes a proliferation of disturbance-adapted successional species (Laurance et al. 1998b), reduces above-ground biomass (Laurance et al. 1997), and accelerates litter production (Didham & Lawton 1999, Sizer et al. 2000) and carbon cycling (Nascimento & Laurance, in press).

Understanding the determinants of the amount of biomass in various forests is presently a global imperative because forests may (e.g. Phillips et al. 2002) or may not be carbon sinks (Clark 2002). Presently, the assumption for modelling and empirical studies is that forest biomass and net primary production (NPP) are in equilibrium. For example, Beerling & Woodward (2001) modelled world-wide plant biomass on the basis of how net primary productivity (NPP) and annual transpiration rates affect tree size. They predicted that the largest stands of organic carbon are in the wet tropics. Similarly, in a recent textbook, Chapin et al. (2002) indicate that amongst forests, tropical forests have the greatest biomass and greatest NPP. In contrast, Midgley (2001) drew attention to the negative correlation between basal area and disturbance rates in some tropical forests. Also, Enquist & Niklas (2001) demonstrated that biomass is not correlated with latitude. Indeed, many indigenous forests with exceptionally large total basal areas and thus total standing stem biomass occur in the cool-temperate areas. For example, the Pacific Northwest redwood and Tasmanian mountain-ash forests have a total basal area of 300+ m2 ha−1, which exceeds the mean of many tropical forests (i.e. 35 m2 ha−1) (see Midgley 2001 and references therein).