svp notes and other stuff

[Nick Longrich <nrlongri@midway.uchicago.edu>]

--Some SVP notes--

The Horned One
        Lots and lots of new Ceratosaurus material is finally coming out.
So you want to know the answer to the age-old question, riddle of the
sphinx itself: was the pubic boot present or absent? Sphinx answers: Yes.
It's an ontogenetic feature, with the boot absent in juveniles and present
in adults. Concievably the same thing was going on in abelisaurs. Whether
this is convergent on the tetanure boot remains to be seen I guess (the
Herrerasaurus boot seems built along wholly different lines). Another
question- so did young megalosaurs have pubic boots or not? What does this
mean about where the boot comes from and how it evolves and how we code it
in the matrices? Hold on, this is just one sign that the days of the
unchanging Types are done, we gotta learn to think riddle-of-the-sphinx
style: dynamic, algorithmic, ontogenetic, four legs in the morning, two at
noon and three in the evening... Evo/Devo is the wave and we're not talking
'80s music.
        The new Western Paleo specimen is gorgeous, and the teeth are just
monster- Dan Chure was saying that in anterior view the maxillary teeth jut
down below the dentaries, making them look like sabertooth cats. He has
described it as the best non-coelurosaurian theropod he's ever seen. It
lifts the spirit to see so much new bone and information coming out about
Ceratosaurus.

Kurzanov Vindicated
        Finally, the word is out- there have been some Japanese expeditions
to the Gobi and one of the things they've been digging up is additional
fossils of what Kurzanov was inferring as belonging to a single animal,
Avimimus. So, was Kurzanov right, or is Avimimus a chimera of several
different things? Well, apparently it turns out it's *not* a chimera.  If I
understand this right, they've got associated elements backing up all the
associations that Kurzanov was suggesting. What's more, the original
specimens that looked kind of screwed up- well, apparently aren't (well
they are still screwy but this is just the way the animal really was),
they've got new material showing the same things. The ulnar ridge is real
(although I still don't buy the feather attachment argument, since last I
checked nobody's pointed out anything similar in modern birds). The
carpometacarpus is real (but then Avimimus isn't the only thing that has
that). That weird skull with the bubblehead cranium is real.
        As to what avimimids are- the new stuff does show association of
the old elements but doesn't extend our knowledge too vastly, there isn't
for example a perfect skull or a full tail or anything, so far you're not
missing too much if you just stick to your Kurzanov monograph. They *do*
say that they've got teeth in the upper beak, seven, someone correct me but
was that seven per premax? They've also got a few elements out of the tail,
scapulocoracoid, some not-too-short tail vertebrae. I keep getting
avimimids falling out near(but outside) the caenagnathid-oviraptorid clade,
the details vary but that's been fairly stable over the past two years for
me. The pubis is very similarly shaped, the subequal pes digits II and IV
are typical of the group, the rodlike quadratojugal bar, pterygoids closely
appressed against the braincase, beaked premaxillae and dentaries,
downcurving paroccipital processes, etc. all argue that these things are
some weird offshoot of the oviraptor lineage but they apparently lack some
synapomorphies that would stick them in either Oviraptoridae or
Caenagnathidae...

Noasauridae... or no Noasauridae?
        If I recall, the epipophyses of the new Malagasy thing are supposed
to resemble Noasaurus in being really anteroposteriorly elongate, along
with those of one of the Indian theropods, and the neural spines are both
anteroposteriorly short. But if you look at Carnotaurus, you'll see the
neural spine is short, also the epipophyses are fairly elongate, it's just
the whole cervical is really short and they are very tall. Here's the real
problem. What the heck is going on with Carnotaurus cervicals? Well, the
single biggest reason they look so nightmarish is that the epipophyses are
really, really well-developed, and they've been extended anteroposteriorly
until you get them forming freakish anterior and posterior prongs. If you
look at the Noasaurus cervical arch you will see that although it is far
more low, elongate and lightly built, it's pretty much exactly the same
dang thing. Now maybe I'm completely misremembering the slides but I sure
don't think that's what the Malagasy thing had, which might have
implications for a Noasauridae.


The long and the short of ulnae
        The SUNY people presented a much more extensive case for making
Rahonavis one animal, especially based on how the stuff was buried, nice to
be able to add that ulna into the matrix with more confidence, I wish they
would have told us exactly what their character evidence was though. Anyone
ever notice that in Archaeopteryx, Confuciusornis- heck, as far as I can
remember, any theropod from there on down through towards the roots of the
trees, no exceptions come immediately to mind- the ulna/radius are *always*
shorter than the humerus... the hand is also pretty long too.
        Implications for the wingspan of Rahonavis are kind of interesting,
go back and look at how they restored the humerus and hand in the paper
while keeping typical Archaeopteryx arm proportions in mind...

Venomous theropods?
        Don't reject this out of hand, it's a very, very weird tooth. It
has to be seen to be believed. Instead of denticles it's got a series of
pits on the carinae, towards the distal end of the tooth these pits joined
together into a groove. Really weird, it pretty much blew away all the
people who looked at it. One supposes that the grooves might be useful for
venom transport (as was suggested at the meeting), say from the saliva
although if there were actual glands you'd expect the grooves to extend
farther down, wouldn't you? Another interesting thought- what if these pits
served to collect bits of flesh, bacteria, morning mouth and general
nastiness to inflict into whatever hapless animal the thing bit into, as
has frequently been suggested for theropods here and elsewhere, a strategy
apparently employed by the Komodo dragon. Something else that intrigues me-
the little pits sort of reminded me of the KGB Killer Umbrella incident,
where back during the Cold War the Russians used an umbrella to shoot a
tiny BB thing with ricin (one of the world's most potent poisons) into a
defector who was living in England (no, I'm not making this up, and yes, he
died). The BB had tiny pits/holes drilled into it where the poison was
stored; so I suppose pits might be a good analogous way to store a
toxin(infectious bacteria), which could be put into the prey by the groove.
When you bit down, blood would run down the tooth because of capillary
action and mix with whatever was in the pits, and probably some of that
blood would then be drawn back out of the groove when the tooth was
removed.
        Anyways what bugs me about the Morning Mouth of Death Theory is
that not much actually does this. Predators that inject venom include
shrews, snakes, spiders, octopi, gila monsters, scorpions, wasps, and cone
shells, platypi inject it defensively with spurs and then various things
like toads and frogs have skin toxins, some millipedes use cyanide (which
gives them that lovely almond smell), and there are even birds with skin
poisons if I am recalling a recent Jared Diamond article correctly. Not
much is known besides the Komodo that uses bacteria. Why is this? The one
difference I can think of is that  the Komodo is going after big prey, this
may make this strategy feasible. For one thing, it might take a lot of
venom to bring down big animals; this may make it difficult to inject
enough to reliably subdue the animal and expensive to produce a lot. The
cool thing about using bacteria as your toxin is that unlike venom
molecules the little suckers are self-replicating. You just need to get a
little in and they reproduce until there is enough to kill the animal. The
other thing is that the venoms of most predators are used to rapidly subdue
prey to prevent it from escaping. This just might be really difficult with
large animals- consider that even a substance as lethal as rattlesnake
venom might take something like an hour to start affecting an animal as
large as a human, and I assume it's even harder to subdue, say, a cow-sized
young duckbill. So maybe venom doesn't really have an advantage over
bacteria as a toxin when you are subduing large animals, since in either
case it is going to take a while to subdue them.


        Also got a chance to look at the Labocania material. It's pretty
hard to figure out but it was possible to identify a maxilla (shattered
remains therof), incomplete portions of a dentary, surangular, chevron, and
ischium, and a really good frontal. It's pretty clear what this animal was.
More or less... more on that later. Operative word being "later".

--Some notes that relate to recent list discussions on phylogeny--

Clintistics
        Life is messy, deal with it: sometimes we just don't have the
evidence to get a nice, fully resolved tree. In these cases, if we don't
have the evidence for a nice neat answer but we keep working until we get a
nice neat answer, there's a very good chance we got the wrong answer. If we
then become attached to that answer, when new evidence arrives and the real
one does happen along, maybe we'll be too busy with the wrong one to give
the right one the time of day. Sometimes I think we can become bad coders
in the search for ever more resolution, finding answers and codings where
there really aren't any. I know the temptation and I do succumb, and I
frequently end up introducing error and then kicking myself for not showing
a little more reserve and just waiting for more information. I think it's
easy to underestimate the importance of "?" in your matrix.
        Basically we've got to know when to code a character and when to
just go out for a beer (for paleontologists, the beer part at least should
be fairly easy to grasp). As Dirty Harry says, "A man's got to know his
limitations."


Mine Is Bigger Than Yours
         Which would you rather have: one hundred characters which have
been accurately and competently coded from the original fossils by someone
who's done their comparative anatomy homework, which show easily
distinguishable states, have well-understood relationships with ontogeny
and allometry, and don't vary like crazy within species, *or* two hundred
characters which have been poorly coded from literature and illustrations
and appropriated uncritically from other matrices by someone who has a poor
grasp of the anatomy, with difficult to distinguish states, with dimly
understood relationships with ontogeny and allometry, and which vary like
crazy within the species ( like my latest matrix...)? Personally I'd go for
the one hundred and take quality over sheer quantity. Data is not
necessarily information. I mean, if I could travel back in time and get the
snout-tail length of every single dinosaur that had ever lived, however
many trillions- well, I'd have tons of data, and with respect to phylogeny,
virtually no information. And though character quality issues might not be
so much a problem in places where quantity can diminish and cancel out
errors you can't fall back on quantity where you have a high amount of
missing data, as we too often have. Where you have very little information,
one or two miscoded characters have the potential to wreak a lot of havoc
in your tree and give you really weird things.
        Also, there may be some inverse relationship between quality and
quantity, insofar as the more characters you have in your matrix, the
harder it becomes to make sure you understand each of them and code them
accurately, and the harder it becomes to proofread your matrix for the dumb
typos (Tyrannosaurus rex: teeth absent), which we are all going to make.
Gauthier had 18 taxa if you include outgroups and 85 characters, this means
1530 codings. That's a lot of codings to do accurately but I doubt he made
too many typos. A 40 x 275 matrix for theropoda would hardly be out of the
question today with so many more animals, new material, and new
understanding of their anatomy, this will generate eleven thousand codings.
That is one heck of a lot of codings, even if a large percentage are "?".
It becomes that much more difficult to proofread that matrix than
Gauthier's and I suspect that it would contain a proportionately larger
number of typological inaccuracies as well as things the author just got
wrong. Finally I think that despite some shortcomings (assumed monophyly of
Carnosauria and Ceratosauria being the big ones) the results of Gauthier's
85-character study are remarkably close to those produced by some more
recent studies which include many more taxa and characters, e.g. Makovicky
and Sues 1998 which is sort of "Gauthier's Topology Part II: The Revenge".
Or check out the consensus tree produced in Tom Holtz' 1999 SVP
presentation.
        Bottom Line: when it comes to character lists, there is more to it
than length.

Characters or Taxa
        Another thing is an emphasis on characters versus taxa. Given a
choice between the two, which would you rather have more of- additional
characters or additional taxa? Generally speaking, more *taxa* rather than
more characters is apparently preferable. The reason is that new taxa can
have novel character combinations that link together otherwise dissimilar
taxa. This is sort of like what they used to call "missing links".
Patagonykus being an excellent example. It shows enough weird similar
characters (ball and socket caudals, weird thumb claw, funky metacarpal and
proximal phalange, huge olecranon, etc.) that it's pretty clearly related
to Mononykus. And it also displays a lot of more conventional coelurosaur
characters such as hyposphene-hypantrum articulations in the centra, a
pubic boot, a pubic apron, a third metatarsal that participates in the
ankle, such that suddenly it's not so hard to get Mononykus in with your
garden-variety coelurosaurs. A may show no similarity to C, but A shows a
lot of similarities to B, and B shows similarities to C, so you can link
them together. Also, remember that the other big thing that forced
reassessment of the alvarezsaur phylogenetic position was the addition of
yet another taxon to the problem, Confuciusornis; it provided evidence that
things like the lack of pubic symphysis and lost postorbital bar in
alvarezsaurs might be convergent, since features like the pygostyle argued
that it was closer to modern birds than alvarezsaurs but it was lacking
those supposedly birdlike features seen in alvarezsaurs. And suddenly
alvarezsaurs are falling out of Avialae left and right.


--Some more specific stuff on theropod phylogeny

"megalosaurs" and Megalosauridae
        I have real problems with scattering the megalosaurs up and down
the tree in a pectinate fashion. Although I haven't done a ton of work on
these, I've spent a lot of time looking at Afrovenator and have managed to
dig up a few similarities that appear to unite it with Megalosaurus and
Torvosaurus, some of them apparently unique to these guys; in addition
published illustrations do not show things exactly as they are, actual
Afrovenator bones are more Megalosaurus/Torvosaurus-like in some crucial
respects. Brooks Britt also has some synapomorphies he uses to unite
Torvosaurus and Megalosaurus. So I think that Megalosauridae is a real,
monophyletic taxon uniting these three and maybe more. Which would be damn
cool considering how long it's been a junkbasket for anything and
everything even though Megalosaurus is the first named dinosaur. I don't
know if technically Torvosauridae has priority, but if it does, the hell
with it. In (much delayed) honor of Megalosaurus, the name should be
"Megalosauridae" (closer to Megalosaurus bucklandi than birds). If you
can't break arbitrary human conventions when you really need to, what good
are they?

"Compsognathidae"
        Compsognathidae, on the other hand, is a major problem. I still
find the characters used to unite Compsognathus and Sinosauropteryx
unconvincing. Partly this is because there weren't very many.  Partly this
is because few of them seem really unique or limited in distribution. For
example, unserrated premaxillary teeth can also be seen in Byronosaurus,
Pelecanimimus, and Sinornithosaurus,  and I could be wrong but I don't
recall there being any serrations in premax teeth in Caudipteryx or
coelophysids. I think this may be a fairly homoplastic character;
unserrated teeth may simply be useful for picking up small prey items among
other things. So I'm skeptical about its use for higher-level analyses of
theropod relationships. Lack of anterior serrations on the maxillary teeth
is similarly problematic as well as a character coding problem- if you code
it as "posterior serrations but not anterior serrations on maxillary teeth"
then you've got at least some of the troodontids with this as well. If you
consider anterior and posterior serrations as independent characters, then
your character is "serrations absent on anterior edge of maxillary teeth"
and so you also have that in some troodontids, alvarezsaurs, Pelicanimimus,
and the anterior maxillaries of Irritator. At any rate it's pretty clear
that presence/absence of tooth serrations in various parts of the jaw is
all over the place and I'd guess it rapidly evolved to suit the various
feeding habits of a given theropod. There just may not be too many
different useful ways of producing teeth to pierce, hold and cut, so you
probably just hit on the same designs over and over. I suspect that very
little will ultimately be learned about large-scale problems in theropod
phylogeny by studying presence or absence of tooth serration. Another part
of the Compsognathidae problem is that some of the proposed similarities
weren't well-illustrated, e.g. fan-shaped dorsal neural spines. The
Sinosauropteryx manus is well-known from one specimen (but unfortunately
poorly illustrated for those of us lacking the ability to closely examine
it); but then the Compsognathus manus isn't especially well-understood,
though it has been recently reinterpreted by Gauthier and Gishlick.
Personally, the reconstruction didn't remind me much of the freakish thumb
apparently present in Sinosauropteryx. Anyways, Sinosauropteryx is just
really poorly known. We have the type/counterpart, which is almost
impossible to make anything out on. I've seen it up close on several
occasions and tried. I've seen it *really* up close and still failed to
learn much. And Compsognathus is also very difficult to make anything out
of. Two-dimensionally preserved fossils often look very pretty but in fact
may allow far less information to be retrieved than more fragmentary
fossils which have been three dimensionally preserved and/or fully prepared
out. Finally, GMV 2124 (the one with the mammal lunch) likely isn't even in
the genus Sinosauropteryx, or even related to it other than that they're
both probably tetanurines. I don't know how people are doing it, but if you
code GMV 2124 (likely maniraptoran), Compsognathus (coelurosaur?), and
Sinosauropteryx (megalosaur, allosaur, basal coelurosaur, kangaroo?) all as
one OTU then I think you are going to get a screwy result. In recent matrix
runs, I have found these three animals scattered up and down the tree with
no particular relationship to each other. If we are to have a
Compsognathidae, the burden of proof is on people to show that
Compsognathus is actually related to anything else, and it has not been
met. Ditto for putting GMV 2124 in Sinosauropteryx.

Osteomorphological Evidence on the Phylogenetic Position of the Gracile
Theropod Cursor Deltadromeus agilis: What's up with that?
        What is Deltadromeus? Who the heck knows but no way is it
maniraptoran, and I have serious reservations about even placing it in
Avetheropoda. Unfortunately the level of missing data here is simply
atrocious; so in my trees it's been jumping between basalmost Coelurosauria
and basalmost Tetanurae, the latter of the two being far more plausible in
my eyes. The thing is, the femoral head in Deltadromeus and Bahariasaurus
(which is probably related if not the same thing) is *anteromedially*
directed! The lesser trochanter is also more like that seen in allosaurs,
spinosaurs, megalosaurs etc. than the one seen in tyrannosaurs and
maniraptoriformes. Deltadromeus also  has a humongous triangular *ischial*
boot (the "pubic boot") which looks like something off a ceratosaur.
Furthermore, despite their large size the coracoids are what's known in the
scientific jargon as "frickin' primitive", more like those of megalosaurs
and ceratosaurs than avetheropods and not remotely like
oviraptorid/dromaeosaur style coracoids. The humerus is pretty weird, it
doesn't much resemble that of megalosaurs and avetheropods but is in some
ways sort of Ceratosaurus like. Although there are some ceratosaur-like
features present, such as the Carnotaurus-like coracoids and the ischial
boot, the winglike lesser trochanter is classic tetanuran, so my guess:
Deltadromeus is "megalosauroid" or something a bit more primitive than
megalosaurids.


How to Win Friends and Influence People: Work on Alvarezsauridae
        A nice, happy-fuzzy subject that all us rational paleontologists
can agree to disagree about civilly. I think they probably are farther from
birds than Archaeopteryx, but they are maniraptoran. They've got the
trochanteric crest that articulates with an antitrochanter like in a
dromaeosaur or oviraptorosaur. They've got a strongly reduced brevis fossa
like those guys as well. There is a tarsometatarsus in at least some
specimens, and big hypapophyses, short caudal prezygs, nice long fourth
toe, downcurved posterior iliac blade, very proximal transition point, etc.
These things argue that these guys are about as far beyond ornithomimids as
the oviraptorosaurs. As for how they fit relative to birds, I think the
current evidence puts them farther from birds than the dromaeosaurs. For
instance, they lack that posterior tapering of the iliac blade seen in
dromaeosaurs and basal birds. The proximal caudal transverse processes are
also much more dorsally placed as well. Dromaeosaur pubic boots are also
far more like those Archaeopteryx and Confuciusornis than are those of
Patagonykus. Furthermore, digit II hyperextensibility seems to be present
in dromaeosaurs, Archaeopteryx, and Confuciusornis, and alvarezsaurs don't
have that either.
        Oh yeah. And they probably liked to eat ants and termites.

-n, ms




                the simple you things you see are all complicated...