Reconciliation Ecology and the Future of Species Diversity
Alexander von Humboldt (1807) provided the first hint of one of ecology’s most pervasive rules: larger areas contain more species than do small ones. Many ecologists see that rule—the species–area relationship—as one of ecology’s very few general laws (e.g., Lawton 1999, Rosenzweig and Ziv 1999). In the past two centuries, ecologists have learned a lot about species–area relationships. I will explore that knowledge and show that we can already use it in the struggle to minimize extinction losses. It teaches us what proportion of diversity is truly threatened and how to prevent most losses by applying a new strategy of conservation biology. Olaf Arrhenius (1921) and Frank Preston (1960) formalized the species–area pattern by fitting it with a power equation: . . . S = Caz (1) . . . where S is the number of species, A is the area, and C and z are constants. For convenience, ecologists generally employ the logarithmic form of this equation: . . . log S = c + z log A (2) . . . where c = log C. (Note that I do not use the jargon term “species richness.” To understand why, see Rosenzweig et al. 2003.) The species–area power equation, or SPAR, can be fitted to an immense amount of data (Rosenzweig 1995). Ecologists are not sure why a power equation fits islands or continents. But we do have a successful mathematical theory for areas within a province. Brian McGill (personal communication) has deduced the species–area curve within provinces from four assumptions: • The geographical range of each species is independently located with respect to all others. • Species vary in abundance with respect to each other. • Species have a minimum abundance. • Each species’ abundance varies significantly across its own range, being relatively scarce more often than relatively common. (“Relatively” means with respect to its own average abundance.) Data support all four assumptions. From them, McGill shows that there is a species–area curve and that it approximates a power equation whose z-value ranges between 0.05 and 0.25 with a mean of about 0.15.