Explanations and morphology

[Richard.Zander at MOBOT.ORG]

Clarification:

I guess what I meant was that theory (I think) implies that although
molecular processes do exhibit convergence, parallellism, and other
homoplasy produced during random generation of traits, this is amenable to
analysis and may be discounted (revealed) in various ways, e.g. by
independent tests that show different patterns of
homoplasy/convergence/parallelism for different, independent data (assuming
there is such).

Morphology has its (comparatively few) traits acted upon by selection (in
addition to drift, founder, gene duplication, whatnot) and such selection
may act upon all the often few morphological traits available, thus no
independent test is possible to reveal convergence and parallelism among
fairly closely related species (except from additional molecular data and
twinges of uncomfortable intuition) or in cases of great morphological
reduction.

Don, you have evidence that your simulations of evolution of haploid genomes
match, approximately, the amount of parallelism demonstrable in
morphological study? Has this been published? I assume cladograms showing
parallelism would not be the same from haploid genome and any associated
morphology but would be independent, since generated differently. Since you
mention selection pressure as a factor, I assume the simulations are
restricted to cDNA, and ignore non-coding DNA.




-----Original Message-----
From: Don.Colless at csiro.au [mailto:Don.Colless at csiro.au]
Sent: Saturday, February 25, 2006 8:03 PM
To: Richard.Zander at MOBOT.ORG
Cc: taxacom at listserv.nhm.ku.edu
Subject: RE: Explanations and morphology



Richard's commment seems to suggest that "neutral" mutations will be less
affected by "convergence". Insofar as they are different, this doesn't apply
- or only applies slightly - to "parallelism". I have a programme for
"evolving" taxa -  actually, models of haploid genomes - on a random basis.
It can allow for various phenomena, such as selection pressure, evolutionary
stasis, etc.; but "real" and estimated phylogenies always showed
considerable parallelism - around the 15-20% commonly found with real
morphological data. I had to increase the number of characters enormously to
reduce this. I suspect that there is a simple statistical explanation.

Don Colless,
Div of Entomology, CSIRO,
GPO Box 1700,
Canberra. 2601.
Email: don.colless at csiro.au
Tuz li munz est miens envirun