cycle, the species category is something like a natural kind. In counting
species and comparing the species richness of different regions, we can, roughly, compare like with like.6
2.4 species and biodiversity
The living world is organized into phenomenological species: recog-
nizable, reidentifiable clusters of organisms. This fact makes the pro- duction of bird and butterfly field guides, identification keys for inver- tebrates, regional floras, and the like, all possible. In this chapter we have embraced the common biological wisdom that phenomenological species richness captures a crucial dimension of biodiversity. There are many routes through which one population can become demographi- cally isolated from, and hence evolutionarily independent of, popula- tions that were once sources and sinks of its own genes. But the fact of isolation and evolutionary independence is of immense importance to the fate of local adaptation in such populations. So the phenomenologi- cal species richness of a region is, in an importance sense, a catalogue both of phenotypic variety and of the potential evolutionary resources available in that region. There is an important difference, on this pic- ture, between a single widespread and phenotypically variable species (like the common brushtail) and a set of closely related species. The available phenotypes, population sizes, and ecological roles might be exactly the same. But one set of phenotypes will be entrenched by spe- ciation mechanisms, and hence will survive minor ecological changes that increase migration rates across the landscape; the other set is much more fragile in the face of relatively minor ecological change. So it does not just matter what phenotypes are present; how they are bundled into species is also important. In effect, we have defended a version of an evolutionary species concept, and we accept that the collection of in- dependently evolving lineages in a region is a key component, perhaps the key component, of that region’s biological diversity. That said, we need to add some important qualifications. The iden- tification of phenomenological species with metapopulations in partial stasis holds good only for some chunks of the tree of life. It may not fit plants. It clearly does not fit microbes. Frederick Cohan has devel- oped an account of bacterial species that identifies those species with the units of bacterial evolution.7 But this involves abandoning the idea that phenomenological species are typically important units of stasis and change; in Cohan’s view, phenomenological bacterial species are amalgams of many evolutionary units (Cohan 2002). Even if we set that aside, phenomenological species do not represent equal amounts