Re: Cladistics (was Sci. Am. - present)

[Stanley Friesen <sarima@ix.netcom.com>]

At 08:29 AM 2/23/98 -0600, chris brochu wrote:
>
>Not entirely.  Huelsenbeck and some of his coauthors have devised
>likelihood methods based on stratigraphic data,

I think I remember some of this work.  Or at least I read some articles on
doing MLE estimates of stratigraphic ranges, and also some article on using
range confidence limits to constrain phylogenies.  Is that what you are
talking about here?

> and ML has been used for
>morphology before.
>
For phylogenetic reconstruction? or for clustering?

I certainly do not remember ever seeing the former published. Though, given
my non-pro status, I could easily have missed it.  If so, I would
appreciate some references.
>
>Those already exist, in large number.  Have a look at the Templeton 
>test,
>or the compare-2 test, or the PTP test, or the T-PTP test, 
>or.........
>
These sound interesting.  I will try to hunt down references to these.  Do
you have any immediately handly?

>(These last two are actually tests of nodes rather than trees, and 
>the
>first is more a test of character distributional difference, but 
>they do
>address your point.)
>
Node testing might be adequate.  In fact it might simplify things.  Instead
of having to compute consensus trees from the non-significantly different
trees, one could just drop non-significant nodes.

--------------
May the peace of God be with you.         sarima@ix.netcom.com
                                          sfriesen@netlock.com