[Taxacom] hominid challenge and Pavetta challenge

[Barry Roth]

Yes, and the further question is, how do we decide in what cases morphology is (or, more specifically, what characters are) reliable?  
 
I have enjoyed, for instance, comparing the results of molecular and morphological analyses of some groups of snails.  Some morphic characters long relied on by systematists were incongruent with the results based on DNA sequences.  Hmm, that was good to know, perhaps those are "shallow" morphic characters.  Others, that many of us had long written off turned out to sort rather congruently.  Ahh, worthy of a closer look; maybe we didn't parse those character-states well enough.  I consider this process of cut-and-fit to be sound, pragmatic systematics.
 
But should we accept one data set (e.g., molecules) as always determinative in this process?   Similarly, should geography tilt the balance between two otherwise equally weighty interpretations of relationship based on morphology?  To do so algorithmically in either situation is to take a fairly hard stance, it seems to me.
 
Barry Roth

--- On Wed, 9/23/09, Stephen Thorpe <s.thorpe at auckland.ac.nz> wrote:


> if we accept, on the basis of molecular results, that morphological evidence is unreliable
NOT ENTIRELY UNRELIABLE IN EVERY CASE! Morpho- evidence is likely reasonably reliable in most cases. A few cases where morpho- evidence (allegedly) gives the wrong answer doesn't mean it is totally unreliable in every case ...
________________________________________
From: taxacom-bounces at mailman.nhm.ku.edu [taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Barry Roth [barry_roth at yahoo.com]
Sent: Thursday, 24 September 2009 11:30 a.m.
To: Taxacom
Subject: Re: [Taxacom] hominid challenge and Pavetta challenge

It's not up to me to answer on behalf of John, but I take his question as a serious methodological one:  if we accept, on the basis of molecular results, that morphological evidence is unreliable, how can we turn around and, in a case where molecular data are unavailable, accept the available morphological evidence as trustworthy?

I suppose this could be justified as "you work with what you've got," and that is of course a familiar situation for paleontologists.  But if a whole modality of data is dismissed as unreliable, then you shouldn't be able to cherry-pick the situations where you accept and trust it.  At least not if consistency -- rather than special pleading -- is considered a virtue in phylogenetic analysis.

Barry Roth

--- On Wed, 9/23/09, Stephen Thorpe <s.thorpe at auckland.ac.nz> wrote: