Re: stratocladistics

[chris brochu <cbrochu@fmppr.fmnh.org>]

Dinogeorge@aol.com wrote:
> 
> In a message dated 8/31/99 9:03:21 PM EST, cbrochu@fmppr.fmnh.org writes
> (replying to me):
> 
> << >
>  > (3) Isn't any kind of a sample, uniform or not, better than no sample at
> all?
>  > Even a poor sample can falsify some hypotheses.
> 
>  Actually, a poor sample can positively mislead - many simulations have
>  shown this. >>
> 
> Which brings up the meaning of the term "mislead." Do you mean that
> incorporation of stratigraphic data from a poorly sampled fossil record will
> somehow render a cladistic analysis incorrect? 

Exactly - and this isn't a problem unique to stratigraphy.  ANY poor
sample can strongly support a false signal.  This, many of us now think,
is why some molecular data sets strongly pointed to guinea pigs not
being rodents - lots of base pairs, good support, but pathetic species
sampling.

Stratigraphy potentially compounds the problem because the missing
information is more likely to be strongly biased.  





Or do you mean that a poorly
> sampled fossil record may in some way conflict with a cladistic analysis? On
> what basis would you resolve this kind of conflict/misleading analysis? That
> is, how would you know which data are the misleading ones?\


In principle, you can never *know* which is the misleading signal. 
Congruence between data sets is generally seen as strong evidence that
they preserve the same signal.  But simply mixing stratigraphy in with
morphology is not the answer, as it prevents us from using one as a
predictor of the other.


chris


-- 
----------------------
Christopher A. Brochu
Department of Geology
Field Museum of Natural History
Roosevelt Road at Lake Shore Drive
Chicago, IL 60605

voice: 312-665-7633  (NEW)
fax: 312-665-7641 (NEW)
electronic:  cbrochu@fmppr.fmnh.org