Determining the causes of variation in species diversity requires linking the scales at
which variation in diversity is measured to the scales at which the processes
hypothesized to affect diversity actually operate. Published analyses of the relative
effects of local versus regional processes on species diversity have failed to measure
diversity at spatial scales relevant to local processes. The effects of local processes,
such as competition, can only be detected at appropriately small local scales, and are
obscured by large samples that aggregate environmental heterogeneity. Relatively few
ecological and evolutionary processes can be identified as uniquely regional in scale,
and local processes are expected to produce regional-scale differences between regions
that differ consistently in environmental conditions that affect local processes. The
relative contributions of regional properties and local processes are hypothesized to
vary predictably along certain environmental gradients, and to produce locally
regulated patterns of species diversity at scales ranging from a few millimeters to the
entire globe.