"uncompromising, single-minded"

[Richard Zander]

See below . . .

_______________________
Richard H. Zander
Missouri Botanical Garden
PO Box 299
St. Louis, MO 63166-0299
Email: richard.zander at mobot.org

> -----Original Message-----
> From: Richard Jensen [SMTP:rjensen at SAINTMARYS.EDU]
> Sent: Monday, November 18, 2002 3:53 PM
> To:   TAXACOM at USOBI.ORG
> Subject:      Re: [TAXACOM] "uncompromising, single-minded"
>
> Thomas Lammers wrote:
>
> >
> > hypothesis.  Cladistics is just a way of creating (and I suppose
> testing)
> > hypotheses.
> >
>
> I would be careful here.  The process used to create the hypothesis, e.g.,
> a
> cladistic algorithm, cannot be used to test the hypothesis.  A cladistic
> algorithm can, given a set of data, yield a hypothesis of relationships.
> But,
> this hypothesis can only be tested in the face of new data.
>
COMMENT: Yes, new data is helpful, but you can use hidden data to do
immediate testing. You need to use constraint trees to see what data
supports the most immediate contrary arrangements of branches. If the
support is small, you got a reliable phylogeny, if comparatively large, you
don't. The null is that any three closest lineages are sister groups and all
support is equal and random

R. Zander

>  Using the same
> algorithm with the new data included would be a test of the original
> hypothesis;
> using a different algorithm on the original data would not be a test of
> the
> hypothesis.  You would simply have two competing hypotheses; if both
> algorithms
> yield the same hypothesis, then they are consistent for the data at hand,
> but I
> don't believe wither can be viewed as a test of the other.
>
> Dick
> --
> Richard J. Jensen              | tel: 574-284-4674
> Department of Biology      | fax: 574-284-4716
> Saint Mary's College         | e-mail: rjensen at saintmarys.edu
> Notre Dame, IN 46556    | http://www.saintmarys.edu/~rjensen