Re: numerical taxonomy (plants)

[John Alroy <jack@xxxxxxxxxxxxxxxxxxxxxxxx>]

Una Smith <una@doliolum.biology.yale.edu>:

> I think we might come to agree on the characters to use to resolve the
> higher systematics of angiosperms before we manage to resolve the higher
> systematics, as we learn what characters are consistently informative
> about relationships among subsets of the group!  Some ~300 angiosperm
> families are traditionally grouped into subclasses and orders within two
> classes, the dicots and the monocots.  We're starting to find evidence
> that most of these taxa above the family level are paraphyletic and/or
> polyphyletic.  It's a big mess, and phylogenies based on molecular data
> are crippled by computational barriers so huge that interest is rapidly
> returning to traditional morphological and anatomical characters, and to
> fossils.

With all due respect, I don't think that having too many taxa to allow a  
branch and bound analysis is a reason to give up on the data set! Nor  
would be the fact that certain authors may not have done a thorough  
heuristic search, or made other mistakes. If the "advantage" of a  
morphological data set would be that the "taxa" were a manageable number  
of representatives of the above-mentioned subclasses and orders, we'd be  
right back to square one because the assumption of monophyly for those  
groups still would be untenable and untested. This would be true  
regardless of whether said representatives were real species or  
hypothetical ancestors. Molecular data have problems, but so do anything,  
and too much of a good thing isn't a bad thing in _this_ case.