Re the issues Joe Morlan raised in his last installment (25 June) in=20
the BSC discussion/saga, I think there really are a few distinct=20
questions here that we should recognize as such:

1) Is reproductive isolation (with interbreeding as the other side of=20
this coin) a *valid* basis for a species concept?

2) Is reproductive isolation the *best* basis for a species=20
concept? =20

3) Are the hybridization and potential interbreeding criteria=20
of the classic BSC the best ways to diagnose reproductive isolation=20
and, more to the point, lack thereof? =20

The answer to #1 is "yes." I would answer "no" to #2 and #3. =20
Number 2 is a matter of preference: species-taxa are delimited=20
arbitrarily (but *objectively*, we hope). For phenotypically=20
homogeneous populations regarded under the BSC as=20
reproductively isolated, there will not be much argument; BSC,=20
PSC, ESC, etc. would agree these are distinct "species." =20
Differences arise when phenotypically distinct populations are=20
lumped under the BSC. I think species-taxa based on=20
demonstrated divergence are more useful than taxa based on=20
lack of repro. isolation. But again, that is my preference. =20
As to #3, I think that the BSC could do a better job of=20
diagnosing reproductive isolation/lack thereof, and hence species-taxa,=20
by discarding the hybridization and (especially) "potential=20
interbreeding" criteria. =20

When two recognizably different populations are lumped under the aegis of=
=20
the BSC on the basis of hybridization evidence, that does not indicate=20
that interbreeding between the two populations is so extensive that the=20
two populations will merge/become thoroughly mixed -- thus losing=20
distinctions now apparent -- and conversely, lack of observed=20
interbreeding does not mean two populations cannot/will not merge. =20
Hybridization simply does not, and cannot, predict whether or not=20
two populations will merge, nor even how likely that event is to occur. =
=20

But again, given effective population size and observed rates=20
of gene flow, it is possible to estimate how long it=20
would take for two populations to merge [maybe someone could=20
do this to estimate how long it would take for NW Crow to be=20
"assimilated" by Amer. Crow]. Why not simply use this kind of=20
estimate to assess *significance* of gene flow between=20
populations in question? I have to agree with Zink that if=20
populations that freely hybridize would nevertheless require=20
millions of years to become thoroughly mixed, they are effectively=20
isolated reproductively, regardless of the degree of=20
hybridization/interbreeding observed. Joe countered that this=20
kind of evidence merely indicates an error in the original assessment of=20
hybridization evidence. It will be interesting to see reassessments=20
based on comparing these two kinds of evidence. But why can't =92BSC
proponents use Zink'sapproach to argue that if two populations in=20
question could merge in (for example) a thousand years given observed=20
gene flow, they should be regarded as a single species despite their=20
being phenotypically dissimilar? That would be a lot more persuasive=20
than descriptions of hybrid zones. I don't have references to these=20
methods at hand, but you can probably find them in some of R. Zink's=92
recent work. =20

In addition to that approach, investigating cladistic/=20
genealogic relationships among populations in question=20
would be very useful for evaluating their effective independence,=20
or potential lack thereof. And cladistic "tree-thinking" methods=20
are now being brought to bear on questions of gene flow (e.g.=20
Edwards 1993, Evolution 47:1118-1137). Cladistic approaches=20
will also prove useful for the study of hybridization and=20
introgression in the context of divergence of populations,=20
and they will undoubtedly shed much light on the evolution of=20
prezygotic and postzygotic isolating mechanisms. BSC types should=20
embrace cladistic methods, not shun them simply because PSC types=20
use them. See DeQueiroz & Donoghue 1988 (Cladistics 4:317-338)=20
for a sympathetic treatment of reproductive isolation as a basis=20
for a species concept by a couple of dyed-in-the-wool cladists. =20
Ed Wiley's _The compleat cladist_ is a decent primer of cladistic=20
methods, but one can also get a good feel for cladistics simply by=20
exploring Maddison & Maddison's _MacClade_ application.

David Wright=20
dwright at u.washington.edu