Hi all,
When the BSC/PSC went around on BirdChat, I saved this excellent post by
Dr. Futuyma - a well-respected researcher in evolutionary biology.
Mark
__________________________________________________________________________
[from Doug Futuyma:]
Chatters,
I'm prompted by the extensive (and probably unending) discussion
of species to add a few words. I'm professionally involved with the
subject, both in my research on insects and in covering it for the
revised edition of my textbook on evolutionary biology, nearing completion.
A definition is a convention, so there are no scientifically
"correct" definitions, only those that are more or less useful - which
depends on the intended purpose. The purposes of a species concept differ
to some extent between biologists who study evolutionary processes and
those who study evolutionary history and are concerned chiefly with
classification and phylogenetic relationships among groups of organisms.
Mayr's biological species concept (BSC) continues to be favored by
most process-oriented evolutionary biologists (including me). It defines
a species as a population or group of populations that is actually or
potentially reproductively isolated from other such groups.
"Reproductively isolated" means that two groups do not substantially
exchange genes in nature, because of **biological** differences that
prevent gene exchange, such as behavioral isolation (they don't mate even
if they encounter each other) or sterility of hybrids. Mere geographic
isolation does not define groups as species. This raises the great
problem of whether we suppose such groups are *potentially* reproductively
isolated, a problem I'll return to.
The "recognition concept", to which some chatters have referred,
is not substantially different from the BSC, except that some of
its advocates accept only behavioral isolation, not hybrid sterility, as a
basis for species distinction. It has not been widely accepted.
The major current contenders with the BSC are several versions of
a "phylogenetic species concept" (PSC), which are gaining quite a few
adherents among systematists. Some (e.g., ornithologist Joel Cracraft)
define a phylogenetic species as a *diagnosable* population or group of
populations "among which there is a pattern of ancestry and descent"
(i.e., gene transmission across generations). Under this definition, any
population in which most members share a distinctive feature of any kind
that distinguishes it from other populations is a species. It could be a
single morphological feature, a single base-pair difference in DNA,
or whatever. Advocates of the PSC argue that this definition recognizes
the **historical products** of evolution (i.e., genetically divergent
populations), rather than the future evolutionary potential of populations
(i.e., whether or not they will some day interbreed), which is often
unknowable. The emphasis here, then, is on what the history of evolution
has produced, rather than on the process of divergent evolution as such.
The difference between these concepts of species stems from a
philosophical difference in what "work" a species concept is supposed to
do - and since the advocates want it to do different work, they want
different definitions. For birders, the main consequence is that
the number of named species (taxa with different binomial names) will
increase substantially if taxonomists incline toward the PSC and if
organizations such as the AOU follow their revisions. There are two
major contexts in which the species recognized under the PSC will differ
from those recognized under the BSC.
One situation, of which we know a few examples and will
probably discover many more, is when a biological species, A, is found
(usually by molecular data) to be phylogenetically more closely related
to certain populations of biological species B (say B1) than to
other populations of B (say B2). That is, A and B1 stem from a more
recent common ancestor than B1 and B2. For instance, mitochondrial DNA
analysis by John Avise et al. suggests that the American Black Duck is
more closely related to American populations of the Mallard than the
American and Eurasian populations of Mallard are to each other. This can
be determined only if there are diagnosable differences between American
and Eurasian Mallards, of course. Under the PSC, the American and
Eurasian Mallards would be named as different species (even if they
potentially interbreed - as we might find them to do freely in a zoo -
and even if they actually interbreed somewhere, such as perhaps in
western Alaska).
The much more common situation is found with allopatric
populations. For a process-oriented evolutionary biologist, potential
interbreeding is interesting (after all, geographic ranges change, and
geographically isolated forms may meet, with or without human
assistance), but unfortunately the potential for interbreeding is usually
very hard to assess. There is no good way to *apply* the BSC in such
cases, even though it *conceptually* embraces such situations. Thus,
many allopatric forms are named as subspecies, many are named as species
if they show more pronounced differences, many have wandered between
subspecific and specific status over the years, and there is no way to
say which is right. (Charles Sibley, in the current Birding and earlier,
calls them "allospecies," which is fine; unfortunately, you can't tell
from a binomial whether a form is an allospecies or a "good" biological
species.)
For advocates of the PSC, allopatric populations aren't a problem:
if you can find a difference between them, no matter how subtle, call them
species. Rigorous application of the PSC will *easily* double the number
of bird species; a very large fraction of current subspecies will assume
species status, and many populations that don't differ in morphology but
which do show molecular differences will be named as species. This may be
a logical consequence of a perfectly defensible species concept, but it
may well prove burdensome (especially for birders who already fret about
distinguishing Gray-cheeked from Bicknell's Thrushes, which at least have
some field characters). Even the concept of subspecies has long been
criticized by some biologists, such as the paleontologist/evolutionary
biologist George Gaylord Simpson, who wrote in Principles of Animal
Taxonomy (1961), "The single species of pocket gophers Thomomys umbrinus
has 213 currently recognized and named 'subspecies' in southwestern United
States and northern Mexico, and those who enjoy that game may well go on
until every little colony of those gophers sports its own Linnaean name."
I hope "species" don't come to suffer such scorn.
Molecular data play a big role in modern systematics, and there
seems to be some confusion among some birders on what such data tell us
about species status. First, we have to distinguish the use of such data
(or any data) in inferring phylogenetic relationships among organisms
(i.e., which populations or species have more recent versus more ancient
common ancestors) from the role such data may play in diagnosing
populations as being different species or not. DNA data are exceedingly
useful in phylogenetic analysis; they may or may not be useful in
diagnosing species. Whether or not they are depends on whether the
populations are allopatric or are in contact with each other.
Before getting into that, I have to note that whether the DNA used
for study is an adaptively important gene or is silent, noncoding,
so-called "junk" DNA is irrelevant. The DNA variations are used as
**markers** (indicators) for the degree to which genes in general are
being exchanged between populations (if the populations are in contact),
or are used as an index of the overall degree of genetic divergence, a
rough index of how long it has been since the populations ceased
exchanging genes (whether due to geographic isolation or to biological
isolating barriers).
A DNA difference that consistently distinguishes two populations
will be grounds for recognzing them as species under the PSC - but so
would a morphological difference. The amount of difference in DNA
sequence is likely *not* to be a good guide to whether or not the
populations would potentially interbreed, so it doesn't much help in
naming allopatric populations as species if one uses the BSC. In the
Great Lakes of eastern Africa, for example, there are many sympatric,
clearly different biological species of cichlid fishes that differ in
morphology, behavior, and ecology, but can't be distinguished by the DNA
sequences that have been examined so far. Conversely, geographic
populations that do freely interbreed where they abut each other commonly
show substantial differences in at least the proportions of different DNA
variants (human populations are a clear example). So the test cases in
which we do know about reproductive isolation tell us that we can't judge
potential reproductive isolation from DNA in the allopatric cases.
The situation is different if two populations are sympatric or
paraptric (have abutting ranges). Molecular markers can provide a better
indicator than morphological features to quantify the amount of
interbreeding (gene exchange) between sympatric forms (e.g., work by Gill
and others on Blue-winged and Golden-winged Warblers). For parapatric
forms, such as Baltimore and Bullock's Orioles, DNA markers can provide an
indication of whether there is no gene exchange, or a little, or a lot.
If the DNA variant characteristic of each such form penetrates at low
frequency only a short way into the range of the other, it is likely that
something, such as low reproductive success of hybrids, is preventing free
genetic interchange. (At least for the genes examined; other genes can
well show different patterns. This is a very complex subject.) If there
is a very broad "hybrid zone," with each form's distinctive DNA variants
penetrating at low frequency far into the other's range, there are
probably weak barriers to interbreeding, if any, and an advocate of the
BSC would be inclined to name subspecies rather than species (e.g.,
"Yellow-shafted" and "Red-shafted" Northern Flickers). But the *pattern*
of the clines (frequency changes) in the molecular markers is what makes
the difference. The mere fact of an average difference means little.
Thus, if Plain Titmouse populations west and east of the Sierras differ in
DNA (see DeBenedictis's Gleanings in most recent Birding), that only means
that those populations diverged while formerly isolated, and perhaps don't
disperse freely enough across the Sierras to rapidly homogenize their gene
pools - it doesn't necessarily mean that they are reproductively isolated
and will continue to diverge. (If differences in vocalizations, behavior
and ecology actually do reduce gene exchange, that's quite another matter,
which can be inferred either by studying those features themselves, or by
analyzing the pattern of geographic variation in molecular markers.)
Well, that's probably more than enough. My main points are that
two species concepts that differ in philosophical foundation and in
practice are now in use by researchers who have different research goals;
that allopatric forms usually can't be unambiguously defined as species or
not under the BSC; that the PSC, if widely applied, will also have some
awkward practical consequences; that DNA differences in themselves will
not generally help to decide the status of allopatric populations under
the BSC; and that the pattern of DNA differences, not the mere existence
of differences, is critical for establishing the status of parapatric
hybridizing populations. I agree with most of what Sibley wrote in his
article on species in the current Birding, but wanted to expand on some of
his points.
Doug Futuyma
Dept. of Ecology and Evolution
State University of New York
Stony Brook, NY 11794-5245
------------------------------------------------------------------------
|
