Subject: BRYONET: Paraphyly
Date: Fri, 7 Nov 2008 16:10:38 -0600
From: Richard Zander <Richard.Zander@mobot.org>
To: bryonet bryonet <bryonet-l@mtu.edu>
BRYONET
The thread of discussion seems to be wavering into a "what is a
species?" exchange. Let me lob one more query about monophyly and
paraphyly.
Take an extreme paraphyletic situation just for discussion purposes:
((AB)C . . . where A and B are terminal on a clade and C is just below
them (as nearest neighbor). A and C are exemplars of one species. B is
an exemplar of a different family.
How shall we classify this admittedly extreme example? A and C make a
paraphyletic group. (1) We could make A, B and C all families, which
makes three monophyletic groups. (2) We could sink family B into the
single species represented by A and C and make one monophyletic species.
(3) We could describe A and C as different "cryptic species" and reduce
B to a third species, all three in the same genus. (4) Something else?
Three different genera?
How does this get across evolutionary information?
Is this different from the common case when A and C are the same genus
and B is a different family or even a different genus? How? If species
are so sacrosanct and different from genera that they cannot be used in
the above example, then why are they commonly split as "cryptic"?
*****************************
Richard H. Zander
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis, MO 63166-0299 USA
richard.zander@mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
*****************************
4 comments. Please write yours here.:
-------- Original Message --------
Subject: RE: BRYONET: Paraphyly
Date: Fri, 7 Nov 2008 23:59:34 +0100
From: Volker Buchbender
BRYONET
In such an extreme paraphyletic situation we should perhaps not think so
much about how to classify A, B and C in the first place but to go back
to morphology and re-evaluate the morphological characters that had been
used so far to classify the groups. Quite some morphological characters
used so far are more correlated to e.g. environmental conditions than to
relationships between the groups (e.g. Peristome reduction or length of
the seta, just to name a few). I see the molecular based phylogenies as
a chance to uncover the (un-)reliable characters. Perhaps A,B and C are
not that extremely differing any more when we ignore the unreliable
characters?
Cheers
Volker
*******************
Dipl.-Biol.
Volker Buchbender
Clara-Viebig-Str. 5
01159 Dresden
Germany
------- Original Message --------
Subject: RE: BRYONET: Paraphyly
Date: Sat, 08 Nov 2008 14:58:54 +0200
From: Johannes Enroth
BRYONET
Volker Buchbender's view of molecular phylogenies as a way of testing
the (un)reliable morphological characters closely echoes my ideas. And
it seems most of the "traditional" generalizations of which characters
are reliable must go. Given that those generalizations have been more
or less paradigmal what we are witnessing is in fact not an "overhaul"
but a total revolution of ideas. I consider it a privilege to play a
small part in that. To me, as a moss taxonomist and systematist, the
science is right now more rewarding than ever.
Yours truly,
Johannes Enroth
Lainaus "Volker Buchbender" volker.buchbender@gmx.de:
--
Dr. Johannes Enroth
PhD, University Lecturer, Bryologist
Dept. of Biological and Environmental Sciences
P.O. Box 65
FIN-00014 University of Helsinki
Finland
Certainly we should re-examine the morphological basis of taxa when
apparent homoplasy or paraphyly occurs. On the other hand, splitting any
group of relatively few taxa, say five genera, will create two groups
that are describable. The new "uncovered" or "emergent" traits that
distinguish the two groups may simply be no better than any other traits
that allow diagnoses of two new groups when splitting the five taxa
randomly.
This is called a "multiple tests" or "multiple comparisons" problem in
statistics. Finding a set of traits that allow morphological splitting
to match molecular splitting is NOT corroboration. You must, instead,
demonstrate that the traits you are proposing to base the new
morphological split on are indeed much better than any other combination
when the group of five taxa is split any other way. Only when an
reliable morphological diagnosis is available for the morphological
split can this be corroboration for the reliable molecular split. I have
not seen any such evaluation done though finding morphological traits
are commonly reported as "supporting" molecular splits.
*****************************
Richard H. Zander
Voice: 314-577-0276
Missouri Botanical Garden
PO Box 299
St. Louis, MO 63166-0299 USA
richard.zander@mobot.org
Web sites: http://www.mobot.org/plantscience/resbot/
and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
*****************************
Statistics is certainly not my strongest area, but if I understood
correctly what Richard means with "multiple tests" or "multiple
comparisons", they are sort of bootstrapping. Of course homoplasy in
nucleotide sequences does occur, but isn't it precisely because of
that that we should have several sequences instead of one? And the
results even then must pass rigorous statistical tests before they can
be considered reliable. Some of the species-groups we have ended up
with in the Neckeraceae would never have been made on morphological
grounds but what is interesting is that they are, almost without
exception, biogeographically very meaningful.
Johannes E.
--
Dr. Johannes Enroth
PhD, University Lecturer, Bryologist
Dept. of Biological and Environmental Sciences
P.O. Box 65
FIN-00014 University of Helsinki
Finland
______
Post a Comment