[Taxacom] Phylogenetic Classification?

[Richard Zander]

One might at this point remark that it is possible to find expressed
trait differences between exemplars representing molecular variants, BUT
such differences may be, if evaluated alone, be simply biotypes, minor
adaptations, or chance combinations of traits. Corroboration of a
molecular difference at the species level must be that ALONE the variant
is acceptable as a good, different taxon. If there is only one or a few
exemplars, then recognition of a molecular variant as a good new taxon
requires more than discovery of a couple minor traits. The molecular
differences may be solely gradually accumulated mutations because the
populations are isolated though remaining static in expressed traits.

By chance alone, I mean:  suppose a pair of exemplars are molecular
variants. By chance alone you are going to find differences if you look
for them, when in any other case you would ignore them except that now
you have a molecular difference that you are trying to shore up with
morphological differences. This is a statistics problem called "multiple
comparisons." Certainly when you have molecular differences you check
for taxonomic differences, but corroboration, again, requires such
morphological differences to stand alone.

-----Original Message-----
From: taxacom-bounces at mailman.nhm.ku.edu
[mailto:taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Bob Mesibov
Sent: Thursday, August 06, 2009 11:15 PM
To: winston.t.edwards at gmail.com
Cc: TAXACOM
Subject: Re: [Taxacom] Phylogenetic Classification?

Winston Edwards wrote:

"Would it be possible for you to include a reference to a paper that
names the molecular variants as different species using the molecular
phylogeny only?"

I've seen so many papers with something like this in the title:
Molecular evidence for cryptic species/speciation in blahblah, but I
haven't bothered to follow any of these up to see if the authors
actually *named* their new species on molecular data alone. I'm not sure
the authors would have bothered, either, since they typically weren't
funded to do taxonomy. The multitude of unpublished cryptic species is
one of the real-world problems that make Richard Zander's comments worth
noting. 

The only paper in my little specialty that comes to mind for
molecular-only diagnosis is this one:

http://www3.interscience.wiley.com/journal/119111642/abstract

It appeared in the Biol J Linn Soc in 1998 and the species diagnoses
derive not from a phylogeny but from allozyme differences.

A not quite so bold example is this one:

http://www.publish.csiro.au/?paper=IS07036

which diagnoses a new species based on a single-base polymorphism in 18s
rRNA, a single indel in 28S rRNA and some mainly quantitative
morphological differences - the kind which if found in a single widely
distributed species might be regarded by a morphotaxonomist as
geographical variation. One of the two species is paraphyletic in the
COI tree published in that paper. At a conference last year I heard that
this problem had been 'fixed' by additional exemplar sampling, and
reanalysis.
-- 
Dr Robert Mesibov