[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Parsimony

From: Dinogeorge@aol.com

 > In a message dated 96-11-06 03:01:52 EST, swf@ElSegundoCA.NCR.COM (Stan
 > Friesen) writes:
 > << The parsimony principle states that a phylogenetic reconstruction
 >  that requires fewer changes (evolutionary steps) is better than
 >  one which require more steps. ...
 > This is an interesting principle, of course, and it certainly has its uses.
 > But nobody has ever shown that nature obeys the principle of parsimony as
 > strictly as the cladists insist it does when it comes to phylogeny.

I quite agree.  Check my Web site (http://www.crl.com/sarima/dinosaurs) for
a brief look at my opinion on this subject.

One point that really bothers me about *strict* parsimony is the issue of
statistical significance.

Consider: cladogram A has 200 steps, cladogram B has 199 steps.
        Is this difference really signficant?
        Is it not likely that the difference is simply due to chance factors?

I would prefer to treat such "nearly most parsimonious" cladograms on the
same footing as the single most parsimonious one.  They should either all
be reported, or some sort of consensus cladogram should be generated from
them.  Unfortunately most of the cldogram generating programs I have access
to cannot deal with this - they can only make consensus cldograms from sets
of *equally* parsimonious cladograms. [I simply cannot get "the set of
cladograms within N steps of most parsimonious" out of any software I can
get runnning on my home system].

Will somebody give me a grant for developing a test for statistical
significance in step count differences? :-)

swf@elsegundoca.ncr.com         sarima@ix.netcom.com

The peace of God be with you.