Abstract

The rapid assessment of tropical plant biodiversity has become an important tool for the quantitative investigation of regional scale (between location) biogeographical patterns. However, the analysis presented here of local scale (within location) variation in a Mexican tropical dry forest suggests that a significant part of regional scale variation in this forest type may be an artefact of the undersampling of sites within locations. Such undersampling is common but is shown to have potentially serious implications for regional conservation assessment in tropical dry forest. It is argued that several sites within each location need to be sampled before the plant diversity at those locations can be meaningfully compared.