[Taxacom] Felsenstein lecture available on-line

[Dr. David Campbell]

> There's a nice sub-text in Richard Zander's posts, too, suggesting
> that long and close experience with a group puts you in a better 
position to make judgements about evolutionary relationships. That 
certainly runs counter to the notion that any raw graduate student can 
grab a few exemplars, get a few sequences out of them and generate a 
credible history relating those exemplars - within a few months, at 
most.<

In fact, this applies equally to molecular data.  The clades that 
persist with support through many analyses (preferably using more than 
one analytical technique, too) with different taxa, different genes 
and/or morphological character suites, etc. deserve greater credence 
than the ones that only emerge out of my latest run.

Experience also helps you evaluate the existing models.  For example, 
in the Bivalvia there seems to be a particularly strong tradition of 
examining a single character suite and ignoring all other evidence to 
produce a classification.  Occasional convergence, reversion, 
orotherwise misleading results are likely to occur with any character, 
molecular or morphological.  Also, many older taxa were explicitly 
envisioned as paraphyletic grades, and it should come as no surprise if 
they aren't monphyletic when examined with new data.  Molecular data 
are especially vulnerable to errors from misidentification, 
contamination, etc., which must be in mind if an unusual result 
appears.  Aligning or at least checking alignments by hand is a 
valuable way to catch anomalous sequences, whereas feeding raw 
sequences into an analysis that internalizes the homology model makes 
anomalous sequences easy to overlook.  

-- 
Dr. David Campbell
425 Scientific Collections Building
Department of Biological Sciences
Biodiversity and Systematics
University of Alabama, Box 870345
Tuscaloosa AL 35487-0345  USA