Regarding genera, families, and other higher taxa, it is true that the
*ranks* "genus," "family" etc. are human constructs, but the taxa, the
clades (monophyletic groups) that are labeled with these names are quite
*real*. Clades exist independent of our ability to recognize them. It is
the notion of *equivalent* levels ("categories" or "ranks") in the
hierarchy of taxa that is artificial. As Bill Wimley alluded, the
Linnaean categories that we still use today predate Darwin and Wallace by
a hundred years (i.e., they are pre-evolutionary). The notion of
equivalent ranks grossly distorts the way we view the evolution of
biodiversity. Biologists for the most part pay lip service to the
realization that rank is artificial, but nevertheless many comparative
analyses at the "genus" or "family" or whatever "level" will be published
this year. We use the nomenclatural system that was around when Mozart
was a boy to describe biodiversity today, despite fundamental changes in
how we view the genesis of that biodiversity.


Bill Wimley recommended that we shift away from the Linnaean system and
move to one based on "quantitative genetic relatedness." I agree that we
should dump Linnaeus, but these are really two different issues -- method
of nomenclature vs. method of analysis of relationships. It is not
necessary to move to a system based on genetic data in order to jettison
Linnaean taxonomy. Moreover (but without wanting to completely open
another can of worms...), it is not clear that a system based on (e.g.)
genetic *distance* (e.g. DNA-DNA hybridization, UPGMA) even represents
clades. DNA sequence data can be (and sometimes are) analyzed
cladistically, though (but this is not genetic distance). Morphological
characters can be (and nowadays usually are) analyzed cladistically, too.
(And interestingly enough, cladistic analyses of the two kinds of datasets
have the same ranges of degree of internal consistency, a yardstick of the
"goodness" of the data.) My point here is that the method of phylogenetic
analysis is more important than the kind of characters analyzed --
molecular vs. morphological. To get really robust cladograms we should
look at both kinds of characters (and note that by including morphology
you can include fossils and absolute time). Some people equate "Linnaean
taxonomy" with "morphological systematics" but that is simply wrong. But
regardless of the method of discovering clades it is desirable to *name*
them once discovered so they can be conveniently referred to. Obviously
it is possible to name clades without assigning them to the old Linnaean
categories. A good paper that covers the Linnaean nomenclature issue (and
why this taxonomy forces "splitting" and "lumping") is de Queiroz and
Gauthier 1994. Trends in Ecol. & Evol 9:27-31.


On an issue Joe Morlan raised, it is often said that the species is the
only "natural" category in the Linnaean hierarchy. It is true that
species can be identified objectively, but which populations are
recognized as species depends on which Species Concept is used to identify
them. The Biological Species Concept (BSC) has held sway in evolutionary
biology for about 50 years now. It is based on reproductive isolation.
The Phylogenetic Species Concept (PSC) is younger. It is based on
recognizing distinct *clades* (evolutionary "lineages"). Both of these
species concepts are objective. Both are also arbitrary in that we must
define the criteria used to identify which populations we call species.
Which of these Species Concepts (or another, and there are others) you use
depends on what properties you believe are most important. (For example,
if you want a species concept that deals with asexual as well as sexual
organisms you would choose the PSC over the BSC). Species as such are
arbitrarily delimited and thus are, ironically enough, less "natural" than
"higher" taxa (in the pure sense of clades, not categories). Species thus
have more in common with the other Linnaean categories than might be
apparent at first glance. A paper worth reading if you into pondering
such issues is de Queiroz and Donohue. 1988. Cladistics 4:317-338.


A practical matter that is affected by the arbitrary nature of species
definitions/concepts is the Endangered Species Act. Many of the
populations currently regarded as "subspecies" using the BSC (including
extrapolations such as "potential interbreeding") are simply "species"
using the PSC. My understanding of the current ESA is that it protects
populations, not just "full species." By doing so it simply (and wisely)
sidesteps the issue of alternative species concepts.


David Wright
dwright at u.washington.edu
Seattle