Re: Dromomeron and Lagerpeton query

["David Marjanovic" <david.marjanovic@gmx.at>]

Having strange problems with Outlook Express, so here goes, using the clumsy 
webmail interface...

-------- Original-Nachricht --------
> Datum: Sun, 19 Jul 2009 05:06:43 -0500
> Von: David Peters <davidrpeters@charter.net>
> Betreff: Re: Dromomeron and Lagerpeton query

> Exactly my point. The pterosaurs are ALSO ingroup taxa and, as you
> suggest, the Nesbitt/Irmis studies are "getting random garbage as a
> result. "

Obviously, what you need to do to test this possibility is _add a new outgroup_ 
which, with enough certainty, lies outside the smallest clade formed by the 
entire ingroup.

Plain as day, no?

> I followed Nesbitt/Irmis changing nothing but adding taxa in a TEST of
> their work.

That wasn't a test.

> > Remember: if you don't explicitly specify an outgroup, PAUP* uses
> > the first (topmost) OTU as the outgroup.
> >
> > In this case, you'd either have to use *Huehuecuetzpalli* as the
> > outgroup or add a non-crown-diapsid outgroup to the matrix.
>
> Which means you have to KNOW something about the entire spectrum of
> amniote relations before embarking on making your taxon list. Not just
> follow the paradigm.
>
> The paradigm needs testing.

Yes.

To root a cladogram by an outgroup, you need to assume that the ingroup is 
monophyletic with respect to the outgroup. This assumption can only be tested 
by an analysis with a bigger sampling, where both the ingroup and the outgroup 
of the smaller analysis are part of the ingroup of the bigger analysis. To find 
out if Nesbitt et al. screwed up by assuming the pterosaurs are part of their 
ingroup, you need to conduct a bigger analysis, say with the crown of Diapsida 
as the ingroup and the araeoscelidians as the outgroup, for example.

This, BTW, is why finding the root of the tree of life is so difficult and 
proceeds using arguments from gene duplications and the like.