Hemispheric asymmetries in biodiversity--a serious matter for ecology.

Abstract

Penguins have been receiving a lot of bad press lately. They are considered somehow counter, spare, strange. Unlike most plant and animal groups, they do not show a peak of species richness towards the equator and a decline towards the poles. This more conventional spatial pattern is conveniently known as the latitudinal diversity gradient because of the strong covariance of richness and other measures of biodiversity that it describes. It is one of the most venerable, well-documented, and controversial large-scale patterns in macroecology (Willig et al. 2003). Equatorial peaks in species richness have characterised the planet since the Devonian (408–362 million years ago) (Crame 2001) and are typical of a wide range of both terrestrial and marine plants and animals (Gaston 1996; Willig et al. 2003). Despite the fact that this pattern has been documented since the late 1700s, sustained interest in both the regularity of the pattern and its likely underlying mechanisms is relatively modern. The realisation that human activity is posing substantial threats to biodiversity has quickened the pace of this interest (Willig et al. 2003). Where the peaks in richness lie (biodiversity hotspots), how these peaks relate to centres of endemism (areas that support large numbers of species that occur nowhere else), and how these patterns are likely to change through time, especially in the face of major environmental change, are major concerns. Without such knowledge, conservation is unlikely to succeed. Although spatial patterns in biodiversity, and particularly the latitudinal gradient, are increasingly well documented for a range of taxa, the proposed mechanisms underlying these gradients remain controversial. In essence, the multitude of mechanisms proposed to explain diversity gradients can be reduced to three categories: historical, ecological, or null. Most significant in raising the temperature of recent discussions is the question of the relative importance of each of these major categories. Historical mechanisms are those that suggest that earth history (e.g., the opening of the Drake Passage and the cooling of Antarctica) and phylogenetic history have played major roles in generating current patterns in diversity, and tend to emphasise regional (and especially longitudinal) differences therein (Qian and Ricklefs 2004; Ricklefs 2004). Explanations involving ecological mechanisms often downplay the significance of such regional differences and give most attention to covariation between current diversity and variables such as energy and water availability, and to the ultimate mechanisms underlying this covariation (Hawkins et al. 2003; Currie and Francis 2004). By contrast, null models, and specifically the geometric constraints model, argue that the expected pattern of latitudinal variation in richness is not a uniform one, but rather a mid-domain peak, which is almost inevitably the outcome of the random placement of a set of variable species ranges within a bounded domain (Colwell et al. 2004; but see also Zapata et al. 2003). It is deviation from the mid-domain expectation that is then argued to be of most interest. In many cases the historical and ecological mechanisms might be difficult to disentangle, such as the historical effects of the establishment of the Antarctic Circumpolar Current, and its consequences for energy availability in the region today (Clarke 2003). Nonetheless, juxtaposing these three major mechanisms raises several questions that could substantially inform the debate in many ways, but have enjoyed far less attention than debating the relative merits of each of them. The geometric constraints model suggests that, to the extent that there is symmetry in the continuity of land (or water) about the equator, declines in richness from the tropical peak should also be symmetrical, with any asymmetries in the latter matching those in the former. Indeed, most texts and reviews dealing with latitudinal diversity gradients only briefly mention hemisphere-related differences and focus instead on the general decline of diversity away from the tropics in both directions (e.g., Brown and Lomolino 1998; Willig et al. 2003). However, that diversity gradients in the two hemispheres might in many cases be highly asymmetric has long been appreciated (Gaston 1996). Although several historical hypotheses suggest reasons why this asymmetry should exist (reviewed in Brown and Lomolino 1998), differences in present ecological factors, such as temperature gradients and rainfall variation, might also explain such asymmetry. If ecological factors are important, then these asymmetries should show up not only in diversity patterns, but also at other levels in the ecological and genealogical hierarchies. From the perspective of ecological explanations for such spatial variation, the questions, then, are how common and strong are such asymmetries, how common are they in patterns of diversity, and what, if any, might be the ecological, rather than null or historical, mechanisms responsible for them?