For every method of tree construction known to us, we can think of a
situation in which the method will be inconsistent and misleading. And all
these situations do not contradict what we expect from "descent with
modificatio" and what we know from "molecular evolution", thus these
situations seam to be realistic.
Especially the so called Felsenstein-zone is such a situation. But whether
two long branches cluster because of common ancestry or by pure
chance is not a priori decidable. If those two taxa are not closely related,
MP will probably cluster them incorrectly. But if they are closely related
(which would not be too improbable to be unrealistic, since they both
share relatively high evolution-rates) ML will probably separate them
incorrectly.
Siddall and Whiting called that phenomenon "Long-Branch-Abstraction" (see
Siddall & Whiting 1999, Cladistics 15, 9-24).
After all, when observing identical character states at a specific position
within a given sequence alignment, I would think that they are identical
due to a shared selection pressure and not because of multiple hits and
chance... But how to interpret the shared selection pressure in terms of
phylogenetic information content is a different topic.
Lars Vogt
Universitaet Bielefeld
Fakultaet fuer Biologie
Zoomorphologie und Systematik
PF 10 01 31
D-33501 Bielefeld
Germany
-- ANNELIDA
Discuss = annelida at net.bio.net = talk to all members
Server = biosci-server at net.bio.net = un/subscribes
Archives = http://www.bio.net/hypermail/annelida/
Resources = http://biodiversity.uno.edu/~worms/annelid.html
--
