[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
At 07:43 PM 3/28/97 -0500, George Olshevsky wrote (qouting me):
>Wrong-o. It was >not< lost in the common coelurian ancestor and then
>redeveloped (whatta convoluted history that would be). Keeled sternum
Note that I meant (and I understand that this was unclear) "ossified
(in adults) keel" on an "ossified (in adults) sternum". Clearly, the sternum
is primitive for Archosauria. The ossified keel is derived. Whether a
cartilagenous [sic?] keel was present in most theropods may be unknowable.
Since the current distribution of that character is confined to "the
'headless' wonder", _Mononykus_, and various ornithothoracine birds
(assuming mony isn't one), and other charcters presumably suggest that "the
'headless' wonder" is not a bird (sorry, George, as scientists, we must
proceed from the hypothesis which best fits the evidence; "BCF" does not),
then the ossification of the sternal keel is most likely to have happened at
This does not preclude the use of this character as a synapomorpy
for _Mononykus_ and birds more advanced than _Archaeopteryx_, if other data
also support this conclusion (which they apparently do).
And where does "coelurian" come from, btw? I do not recall a taxon
Coeluria being defined in a modern context (or in any another one, for that
>perhaps somewhere among primitively flying ceratosaurians [...]
Until there is a phylogeny which supports this statement, it is not
useful for evaluating character distributions.
>coelurian ancestor (a flying animal). It's still there (retained) in
>_Mononykus_ and in _Archaeopteryx bavarica_, too.
Does _Archaeopteryx bavarica_ have a keel? I do not recall this
being the case, but I've only read secondhand reports.
><< And yes, vestigialization, commonly occuring or not,
> is still a character. >>
>But one subject to massive convergence and therefore liable to greatly
>confuse character analysis.
How can you say this *BEFORE* you have done your character analysis?
If you do not code a character because you believe it to be "subject to
massive convergence", then you do not allow for the possibility that your
initial interpretation of convergence is wrong! If you want to start with an
assumed phylogeny, then discard characters until your result fits the
assumtion, *of course* you will end up supporting your origional phylogeny.
Your analysis will NOT, however, be scientific.
>> possible. The answer is not to abandon our methods, but to seek better data
>> which might guard against this.
>Not quite. Something like 95% of the time, we >can< discover what nature has
>done using the character data we have.
How do you know this?
>It's the other 5% of the
>time--signaled by a manifestly absurd result, such as segnosaurs in Theropoda
I should point out that Dong and Russel determined that _Alxasaurus_
was a theropod using *traditional* criteria in the field before ever doing a
>and Mononykus in Avialae--that we need to take a closer look at methodology.
And *how*, pray, are we to know when this "5% of the time" is
occurring? Certainly we may evaluate the cladograms at our disposal and
choose which is the most likely (as Benton has done), but how are we to
compare a scientifically derived cladogram to our own spurious notions of
"what makes sense" without some equally scientifically derived alternative?
>> d) Further, you have yet to demonstrate that nature, when viewed in
>> toto, is not parsimonious. If we had every scrap of data we could ever hope
>Alas, this assumption is completely untestable.
If this assumption is untrue, then science is a fool's errand.
>But there is a much, much, much larger number of ways for nature to be
>unparsimonious< than for nature to be parsimonious.
That is a tautology. As usual, you miss the point. If Woden came
down from Valhalla and gave you *the truth* of phylogeny, the only means you
would have to evaluate that information is parsimony. Since the phylogeny
you got would be *the truth*, it would contain the least number of
unsupportable assuptions (zero) and the
least number of untestable suppositions (zero, if we accept asking Woden as
a test). Obviously, this is a silly example, but to extend the principle...
*Evolution* is obviously not constrained to being parsimonious. I
don't even know how evolution can "be parsimonious". By minimizing
homoplaisy? We've already gone over that issue...
The fundamental principles of scientific investigation tell us that
by accumulating evidence and applying the principle of parsimony to that
evidence, we can most closely approach understanding the processes of
nature. This does not imply that nature functions parsimoniously, it implies
that nature acts according to internally consistant processes. This does not
imply that their effects are consistant. It does imply that, utilizing a
rational approach to the study of nature, we can come close to uncovering
By accumulating evidence (data), and processing that data, and in
doing so make the fewest number of unsubstantiated assumptions and
untestable assertions (through cladistics), we hope to come close to
phylogeny. We also perform science. If you wish to eschew this principle,
you discard science, and any assertion you make is no more useful in trying
to discover the path of evolution than the random speculations of the man on
>I'd go with the numbers.
Jonathan R. Wagner, Dept. of Geosciences, TTU, Lubbock TX 79409
Web Page: http://faraday.clas.virginia.edu/~jrw6f