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

Re: W4MA (Wait for Mickey's Analysis)

Hey, I have my own acronym! ;-)  Undeserved, yes, but an acronym

Ken Kinman wrote-

>      I would much rather be viewed as Sereno-esque (a term Mickey used
> weekend).  I don't like Sereno's nomenclatural philosophy at all, but I
> think his analytical approach does have some definite advantages (in spite
> of the fact that I disagree with his placement of tyrannosaurs).  He may
> have done better with more accurate codings.

Okay, it had to come someday....
Disclaimer- The following criticisms of Paul Sereno's phylogenetic practices
are not meant to insult his character, as I have a lot of respect for him.
Paul has some really great qualities.  He's a very kind person and quite
responsive to e-mails.  He's willing to go to god-forsaken places like Niger
and Inner Mongolia on a regular basis, to collect more fossils.  He ran in
the Chicago Marathon to get funds (not sure if my laziness would let me do
the same).  There are some very detailed papers he wrote, particularily on
Herrerasaurus and Iberomesornis.  He's also very good at getting publicity
for dinosaurs, which certainly helps paleontology.
However, his phylogenetic "analyses" have a certain problem.....
Now, what a phylogenetic analysis is supposed to do, the whole reason we
need these computer programs and use a principle called parsimony, is to
find the hypothesis that best fits the data.  The one that is most
parsimonious, requires the fewest number of reversals and parallelisms.
This is how we test hypotheses, it is our measure of accuracy.  I mean, just
look at the name of the most commonly used program- PAUP = "Phylogenetic
Analysis Using Parsimony".
So Sereno comes up with a new idea, that alvarezsaurids aren't avians (which
seems true enough), but instead are arctometatarsalians closest to
ornithomimosaurs.  A fine hypothesis, I've noted the similarity myself,
particularily to Anserimimus.  He presents his cladogram and it seems to
support this view, it is most parsimonious given his data.  So do we say "ho
oray, change the official consensus to make alvarezsaurids
arctometatarsalians"?  No.  Because if you actually look at his data matrix,
something becomes evident.  Alvarezsaurids weren't given a choice to do
anything except clade with ornithomimosaurs.  Alvarezsaurids were placed
next to pygostylians in previous analyses because of the following
1. sternum longitudinally rectangular (Perle et al., 1993)
2. sternal carina (Perle et al., 1993)
3. prominant antitrochantor (Perle et al., 1993)
4. trochanteric crest (Perle et al., 1993)
5. fibula does not contact tarsus (Perle et al., 1993)
6. prominant cervicodorsal hypapophyses (Chiappe et al., 1996)
7. carpometacarpus (Chiappe et al., 1996)
8. ischium more than 66% of pubic length (Chiappe et al., 1996)
9. dorsal vertebral foramen more than 40% as tall as centrum (Chiappe et
al., 1996)
10. no ischial symphysis (Chiappe et al., 1996)
11. m. iliofibularis tubercle of fibula laterally projecting (Chiappe et
al., 1996)
12. quadratojugal not contacting squamosal (Chiappe et al., 1996)
13. no proximomedial fibular fossa (Chiappe et al., 1996)
14. no postorbital-jugal contact (Chiappe et al., 1996)
15. quadrate articulates with prootic and squamosal (Novas, 1997)
16. ulnar distal condyle present (Novas, 1997)
17. pubic symphysis reduced (Novas, 1997)
18. no obturator process (Novas, 1997)
19. humeral distal condyles mainly locvated in cranial aspect (Chiappe et
al., 1998)
20. distal end of ulna subtriangular (Chiappe et al., 1998)
21. no postorbital process on jugal (Forster et al., 1998)
22. ventral tubercle of humerus projects caudally and is separated from head
by capital incision (Forster et al., 1998)
23. pubic foot absent (Forster et al., 1998)
And these are just out of those papers written up to 1998, when Sereno would
have had time to incorporate them in his 1999 Science "analysis".  Looking
at his character list, how many of these are present?  The following- 6, 8
and 18.  How is Sereno testing his hypothesis if he doesn't let the other
side's evidence compete?  The obvious answer is, he isn't.  In every matrix
by Sereno I've seen, he excludes the vast majority of evidence against
topologies that disagree with his.  That's why his consistancy indices are
always so much higher than others', because he destroys any homoplasy before
it can exist.  This is not good practice, and it's why I never take his
results very seriously.  Perhaps it's unfair to label this type of
"analysis" Sereno-esque, as it's not confined to his papers.  The
"Scelidosaurus is an ankylosauromorph" paper by Carpenter I noted previously
is another example.  Actually, I see this kind of thing quite often, in
non-dinosaurian papers as well.  And I'm really quite saddened they make it
past the reviewers, as they aren't analyses.  You need to actually analyze
the problem by considering alternatives, not ignoring them.  This is why I
term your procedures Sereno-esque, because you say enigmosaurs aren't
monphyletic based on a few characters, that are outnumbered by other
characters which you subjectively ignore.  It's not something to be proud

>      Anyway, I still think a paraphyletic Enigmosauria and a polyphyletic
> Alvarezsauria are still likely alternative hypotheses, and that Mickey's
> analysis will probably show support for one or both of them (even if they
> end up occurring in trees that are a few steps longer).  If my hypotheses
> are too unparsimonious, I'll shut up.  We shall see.

You know, my analysis CAN'T show support for these if I have no characters
in the matrix that support them.  There are those few anti-enigmosaur
characters that I'll be including, but the two anti-alvarezsaurid characters
you provided won't do a thing.  So if you can come up with any more of these
characters, please post them.  Just don't use them to support an
unparsimonious topology.

Mickey Mortimer