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

Re: Dinosaur Genera List corrections #67

From: "Jonathan R. Wagner" <znc14@ttacs1.ttu.edu>
 > In fact, it should contain *more* homoplasies, as the hypothesis
 > generated in the paper assumes a *minimal ammount of homoplasies*...
 >  Thus, any other tree should include a
 > *greater* number of homoplasies.

One of the problems with the traditional sort of maximum parsimony
analysis is that it tends to ignore the uncertainty factor. In
particular, even if the best tree is only one or two steps longer
than the next dozen or so trees, and those extra steps are all in
characters that are homoplasies or reversals in the "best" tree,
these other trees are ignored.  Yet, a proper statistical analysis
would almost certainly show that the differences in parsimony among
the entire set are *insignificant*.  That is, for a statistical point
of view, it is frequently the case that the several nearly most
parsimonious trees are statistically equivalent.

Hopefully I will be able to find this article Sunday, but I am
doubtful that my library has the right journal, and even if it does,
the current issue may not be on the shelves yet.

So, does this article use some method like consensus trees or
presentation of multiple alternative trees to show the range of
variation amoung the best trees?  If not, I would be highly skeptical
of their results.

Another point that I just thought of:  what non-avian species are
included in the cladogram?  Does their sample range as widely as,
say, the Ceratosauria??

Also, unless their character matrix includes any characters used
in prior studies that place these animals outside of Avialae,
or they show why those characters cannot be used, then any prior
cladograms are, IMHO, just as valid as theirs.  Just because theirs
is the newest doesn't make it the best.

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

The peace of God be with you.