Abundance and Body Size in British Birds: Reconciling Regional and Ecological Densities

Abstract

Considerable attention has recently focused on the use of species body size as a predictor of species abundance, with consensus lacking over whether or not size is a useful predictor (Blackburn et al. 1993, Cotgreave 1993, Currie 1993, Damuth 1993, Blackburn and Lawton 1994). Initial studies of the relationship between body mass and abundance across species, plotted on logarithmic axes, suggested that the two were highly correlated (Damuth 1981, 1987, Peters 1983, Peters and Wassenberg 1983, Peters and Raelson 1984). Body mass generally explained around 60% of the variation in abundance across species. The strength of the size:abundance correlation gave rise to the view that a species' abundance was set by its body size, and therefore that body size is a good predictor of abundance (Damuth 1981, Peters 1983). This view has been criticised (Brown and Maurer 1986, 1987, Lawton 1989, 1991, Blackburn et al. 1993, Blackburn and Lawton 1994). Negative size:abundance relationships have generally been shown in data compiled from singleor few-species studies (e.g. Damuth 1987), and thus may not be representative of the situation in whole animal assemblages. The contribution to these relationships of rare species and low density populations of common species, both of which are rarely studied, is likely to have been underestimated (Brown and Maurer 1987, Morse et al. 1988, Lawton 1989, 1991). Subsequent studies sampling whole assemblages of taxonomically similar animals revealed different, polygonal relationships between log body size and log abundance, with intermediate-sized species having peak abundance (Brown and Maurer 1987, Gaston 1988, Gaston and Lawton 1988, Morse et al. 1988, Cotgreave et al. 1993). These polygons generally showed either no overall correlation between body size and population abundance (Morse et al. 1988) or a weak negative relationship, with body size generally explaining less than 10% of the variation in abundance (Brown and Maurer 1987, Gaston and Lawton 1988, Blackburn et al. 1993). Whole assemblage studies gave rise to an alternative view, that species body size is a poor predictor of species abundance (Blackburn et al. 1993). Conclusions from whole assemblage samples, however, have in turn been criticised (Damuth 1991, Currie 1993). Abundances in samples from assemblages tend to be veiled (Nee et al. 1991a). Many species will be represented in abundance samples by a single individual. This is unlikely to be the true abundance of the species, but rather the result of incomplete sampling. Therefore, the true abundances of the rarest species in these assemblages are not revealed by the samples (Currie 1993). In addition, samples may include transient individuals from species not otherwise resident in the assemblage, inflating the number of apparently rare species observed (although the presence of transients did not significantly change the size:abundance relationship in the one assemblage where their contribution has been tested (Gaston et al. 1993)). One major problem with reconciling the results of these different studies is that assemblage studies generally report species abundance using a different measure to compilation studies (Damuth 1991). Compilation studies generally compare species densities in suitable habitat ('ecological' density), whereas many, although not all (e.g. Gaston and Lawton 1988), assemblage studies compare species abundances in samples from a given area ('regional' or 'crude' density; Damuth 1991). Ecological density gives information about the abundance attained by a species within the habitat it occupies, but takes no account of how that area is distributed. Thus a species with a high ecological density might be very rare if it is adapted to a geographically restricted habitat. A comparison of regional densities, on the other hand, is a comparison of species abundances within a fixed area (e.g. birds within North America; Brown and Maurer 1987), and so