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

And on, and on... Re: Cladistics (was Sci. Am. - present)

At 03:40 PM 2/25/98 EST, Dinogeorge wrote:

>With regard to cladistic analysis, I'm not satisfied with what I have read
>about its ability to reproduce phylogeny successfully to the point where I
>would grant it priority over other methods and tests. For example, if a
>cladogram is presented that suggests there was biotic crossover between two
>geographically separated regions, but geological evidence indicates that there
>was no connection, I would be just as inclined to believe the cladistic
>analysis is incorrect as to believe the geological evidence is wrong.

Yes, if the geological evidence does indeed indicate that there was no
connection.  In some cases, it might well indicate such.

However, and this is where a LOT of people (including some colleagues and
friends) make a BIG error: paleogeography is based on a lot of guesswork!
The data points to reconstruct coastlines are very sparse for many, many,
many sections.

>Cladists, however, would take the cladogram as a priori evidence of a

Yes, subject to later falsification.

For example, pre-mid Cenozoic reconstructions of Siberia and Alaska suck.
So much of the terrain there is, well, terranes (microplates) which were
sutured there over various bits of time, that we have very, very little idea
of the previous coastline before geologically recent times.  On the other
hand, the coastlines at particular biozones (units less than a million years
or so) for the western Interior of North America are pretty well understood.
When your not dealing with cratons (stable continental interiors),
everything is pretty damn blurry.

See the text of Smith, A.G., Smith, D.G. & Funnell, B.M. (1994)  Atlas of
Mesozoic and Cenozoic Coastlines. Cambridge Univ. Press.  In particular pp.

Or to tie it in to something people here can appreciate, paleogeographic
maps are the equivalent of full color dinosaur life reconstructions.  They
may look impressive, but if you don't know what data went into them, you
might not realize how much imagination (hopefully grounded in reality) is
used to create them.

Some parts of a geologic maps would be like a full color reconstruction of
an adult Corythosaurus: something we know a LOT about: complete skeletons,
skin impressions, etc.  Others would be like full color reconstructions of
Cetiosaurus: parts are well known, others can be reasonably interpolated
from other taxa, some are plain guess work.  Others would be like full color
reconstructions of Rapator or Laevisuchus.  The Bering region for the
Cretaceous is somewhere between the last two: we have some idea of the
cratonic parts of Alaska and its position, but the western portions... Ugh.

So in this particular sort of case, fossils may be used as evidence (they
are observable things), and well resolved cladograms based on such might
also be used.  Poorly resolved cladograms or particular coastline maps are
much poorer bases for supporting or rejecting particular scenarios.

>Given a 15-20% error rate in cladograms, I would have to see something like
>30-40% of the taxa in the cladogram in both regions before I would agree that
>the analysis is providing evidence of a connection. Less than that is just

And these numbers are based on...?  Comparison with better documented
continental connections? (Say some of the Cenozoic ones?)  Statistical
analysis?  What?

Incidentally, some work by Vrba (various papers, including 1992. Mammals as
a key to evolutionary theory. J Mammology 73:1-28) suggests that habitat
restrictions are the key to what does and doesn't travel across
intercontinental bridges during such events.  This might have application to
dinosaur research (Mike Brett-Surman and I have talked about this with
regards to hadrosaur distributions every so often).

Thomas R. Holtz, Jr.
Vertebrate Paleontologist     Webpage: http://www.geol.umd.edu
Dept. of Geology              Email:th81@umail.umd.edu
University of Maryland        Phone:301-405-4084
College Park, MD  20742       Fax:  301-314-9661