The structure of global species–range size distributions: raptors & owls


*Correspondence: K.J. Gaston, Biodiversity & Macroecology Group, Department of Animal & Plant Sciences, University of Sheffield, Sheffield S10 2TN, UK. E-mail:


Aims  To determine the shape of global species–range size distributions, the influence on these of island species, threatened species and patterns of latitudinal variation in range sizes, and the fit to logit-normal distributions.

Location  Global.

Methods  We take the spatial distributions of the raptors and owls of the world as exemplar data sets, document the shapes of their species–range size distributions, the influence of particular groups of species and of latitudinal variation in range sizes on these shapes, and the fit of these distributions to a variety of models.

Results  The global species–range size distributions of both raptors and owls are extremely right skewed on untransformed axes. They are not lognormally distributed, as has commonly been stated for species–range size distributions, nor logit-normally distributed as has been suggested might be the case. For raptors, departures from either a lognormal or a logit-normal are little mitigated by excluding groups of species that might be thought to distort the observed species–range size distribution, nor by the latitudinal gradient in geographical range size. For owls, the effects of excluding island and threatened species are more marked, with the fit of the species–range size distribution to a lognormal or a logit-normal becoming much closer.

Conclusions  A simple general description of the shape of species–range size distributions remains elusive. This constitutes a significant constraint on the development of theory as to how they are determined. Whilst in principle the fit of any given mechanistic model can be tested against one or more empirical data sets, whatever their form, a simple general mathematical description of species–range size distributions would make the process of rapidly testing the appropriateness of mechanistic models more straightforward.