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

Martin 2004 critique

Here's something I wrote on Spring Break while stuck away from my refs, and
decided to post...

Tim Williams wrote-

>Martin, L.D. (2004). A basal archosaurian origin for birds. Acta
>Zoologica Sinica 50(6): 978-990.

It's been a while since I've had the pleasure of critiquing an ABSRD paper.
This one is similar to the others- bringing up the same old criticisms yet
ignoring most recent papers, including those which have been written to
specifically address them.

"For nearly two decades, Deinonychus stood almost alone as a dinosaurian
proxy for a bird ancestor, but in recent years, smaller and more birdlike
theropods were discovered in the Early Cretaceous Confuciusornis fauna from
Liaoning, China."
Almost alone? Does Martin not know of all the well-preserved deinonychosaurs
described prior to 1998? Sinornithoides, Saurornithoides, Troodon,
Unenlagia, Velociraptor, Saurornitholestes, Dromaeosaurus, Adasaurus, etc..
Or the well-preserved oviraptorosaurs- Avimimus, Chirostenotes,
Microvenator, Oviraptor, Rinchenia, "Ingenia", Conchoraptor.

"The femoral head [of Archaeopteryx] turns forwards rather than extending
perpendicular to the shaft."
No it didn't (Hutchinson, 2001).

"The pelvis has an incompletely open acetabulum, and there is no
characteristic dinosaurian supra-acetabular shelf."
There actually IS a supracetabular shelf in Archaeopteryx (Paul, 2002),
unlike most other maniraptorans.

The above characters result "in a somewhat sprawling position for the femur
that is corrected at the knee joint, resulting in a functionally vertical
leg." No more so than other dinosaurs. Most non-avetheropods have the
femoral head orientation Martin claims for Archaeopteryx.

"Archaeopteryx lacks basic features of both modern bird and dinosaur
postures: and it is difficult to argue that it shared their peculiar backs
while running, or that the common ancestor of birds and dinosaurs had
already achieved such a running posture. In fact, confinement of the
articular surface to the front of the acetabulum results in a vertical
Martin earlier noted that modern birds have antitrochanters which produce
their slight outward femoral projection, unlike Archaeopteryx. This is
presumably the "basic feature of modern bird postures" Archaeopteryx lacks.
However, the dinosaur-bird dichotomy in pelvic structure Martin advocates
isn't real. There are representatives of every combination of acetabulum
closure, supracetabular shelves and antitrochanters.
Acetabulum partially closed, supracetabular shelf, no antitrochanter-
Acetabulum partially closed, supracetabular shelf, antitrochanter-
Acetabulum partially closed, no supracetabular shelf, antitrochanter- most
Acetabulum partially closed, no supracetabular shelf, no antitrochanter-
Acetabulum open, supracetabular shelf, no antitrochanter- most
non-coelurosaurian theropods, Stokesosaurus, Archaeornithomimus,
Acetabulum open, supracetabular shelf, antitrochanter- Shenzhousaurus,
ornithomimids, Patagonykus, Shuvuuia, Mononykus, Neimongosaurus,
Acetabulum open, no supracetabular shelf, antitrochanter- Tyrannosaurus,
Segnosaurus, Avimimus, Oviraptoridae, Troodontidae, Microraptor.
Acetabulum open, no supracetabular shelf, no antitrochanter- Gorgosaurus.
Even if Martin's dichotomy were true, it would not suggest the dinosaur-bird
common ancestor lacked a horizontal back. If the topology is (D(A,B)), and D
and B have a horizontal back, while A lacks one, it's just as parsimonious
to assume A reversed the trait as it is to assume it developed in parallel
in D and B. Paul's demonstration of a supracetabular shelf in Archaeopteryx
disproves Martin's claim the only acetabular articular surface was anterior,
so there is no evidence Archaeopteryx stood vertically anyway.

Interestingly, Martin brings up WAIR, but says it "seems better fitted for
modern birds with their peculiar leg positioning than Archaeopteryx..." Why
this should be is never explained and not obvious.

Did you know that Maniraptora was the "last dinosaurian clade to develop
(i.e. a crown group)"? I certainly didn't... Martin clearly knows his
dinosaurs and phylogenetic terminology.

Martin's first major qualm with BAD is the temporal paradox. He notes that
the fact Maniraptora is deeply nested in Theropoda leads to "an
inconsistancy in time of origin for Protoavis, a supposed Late Triassic
bird; and a similar problem exists for possible Early Jurassic "bird"
"Embedding birds in the maniraptoran crown group coupled with a Late
Jurassic age for Archaeopteryx creates a similar problem. Not only is
Archaeopteryx older than any credible maniraptoran fossil, but maniraptoran
diversification must be even older"
"Such discrete evolutionary bursts, if genuine, are more theoretically
important than the question of avian origins and should be tested with
evidence outside the bird/dinosaur controversy."
So just when undeniable maniraptorans are being found just a few stages
after Archaeopteryx (if not before, if the Daohugou is Middle Jurassic),
Martin's embraced Protoavis and controversial tracks as being avian in order
to push birds even further back. A shame Protoavis is thought to be avian by
so few people (one could bring up Eshanosaurus as an Early Jurassic
maniraptoran with at least the same amount of confidence), and that the
ichnological evidence is as ambiguous as the Middle Jurassic maniraptoran
dental records Martin finds uncredible.
As described below, in order to not hinder Martin's theory of bird origins,
Protoavis would actually have to cause longer ghost lineages. Nesting birds
within Maniraptora is only 'problematic' if Archaeopteryx is itself nested.
If Archaeopteryx is sister to other maniraptorans (with birds closer to
Archaeopteryx, as Martin might believe; or with birds closer to other
maniraptorans, as Paul believes), the temporal problem is solved. Of course,
both enigmosaurs and deinonychosaurs are known from undeniable specimens a
mere 10 Ma after Archaeopteryx, not to mention the earlier fragmentary
specimens. So the temporal paradox is weak to begin with.
Finally, the very studies Martin advocates HAVE been done, by Brochu and
Norell (2001). They concluded deriving birds from any non-dinosaurian
archosaur involved far more temporal inconsistancy and that the fossil
record actually fits BAD rather well compared to other accepted tetrapod

Martin's next section is titled "anatomical barracades". Instead of bringing
them up again, what Martin really needs to do is address Makovicky and Dyke
(2001)- Naive falsification and the origin of birds. For all these
barracades are are more instances of naive falsification. Birds and
dinosaurs are supposedly different in some characters, and those states
couldn't possibly have evolved from each other, so they must have a common
ancestor before either trait evolved in each line.

First he brings up the tooth implantation issue, which was addressed by
Currie and Zhao (1994). Martin says "... once one implantation pattern has
been achieved, it is hard to understand why it would be lost and replaced by
another." Regardless of how hard this may be, some deinonychosaurs have the
dinosaurian pattern, so evolution of this trait would have to happen, even
in Martin's phylogeny.
"The teeth in birds or crocodiles are bordered lingually by an extension of
the tooth-bearing bone (premaxilla; maxilla or dentary), while dinosaur
teeth are bordered by hypertrophied attachment bone (superpleurodonty:
Martin and Stewart, 1999). This suggests that the common ancestor of birds
and dinosaurs did not have socketted(sic) teeth."
I assume Martin is talking about interdental plates when he mentions
"hypertrophied attachment bone." The supposedly lost interdental plates of
most deinonychosaurs seem to be merely fused (and sometimes reduced)
instead. Crocodiles seem to be similar, from the reports of fused
interdental plates in various sphenosuchians. I wouldn't be surprised if
birds took the same evolutionary pathway, though we have no well described
intermediates between Archaeopteryx's unfused plates and Hesperornis'
'absent' plates at the moment. The sequence from Archaeopteryx to
Bambiraptor (with only some plates fused) to dromaeosaurids provides a
method to break Martin's barracade and show that birds needn't have branched
off so basally. See my comments on the term 'superpleurodonty' below.
"... although dromaeosaurs were originally reported to lack interdental
plates, a signature trait of dinosaurian tooth implantation. Their tooth
replacement has also been claimed to be birdlike (Currie, 1987). If they
had, or had had, the avian/crocodilian implantation and replacement pattern,
we might argue instead that they are not dinosaurs."
Currie's 1987 paper was on Troodon, not dromaeosaurs. Archaeopteryx has
interdental plates, which doesn't make sense in Martin's scheme. Maybe we
should argue that instead it is not a bird? ;) Bambiraptor and
Sinornithosaurus also have obvious unfused interdental plates. But these are
birds in Martin's view. I suppose ornithomimosaurs and alvarezsaurids are
birds too, since they lack obvious interdental plates?

Next, Martin discusses the semilunate and digital identity issues. He claims
dromaeosaurs have only two carpals, and says the dromaeosaur semilunate is
equivalent to avian distal carpal III! No, distal carpal III in dromaeosaurs
is fused to the base of metacarpal III, just like in the birds Martin
mentions (Gishlick, 2002). You might think Martin is renaming the distal
carpals to correspond to his thinking birds have digits II-III-IV. However,
he says distal carpal III is fused to metacarpal III in Jeholornis and
Confuciusornis, which is true of the standard distal carpal III, and not the
semilunate. And just because the ulnare is missing in many maniraptorans
doesn't mean it was absent (known in Caudipteryx, Heyuannia,
Sinornithosaurus and Microraptor, for instance). Because Archaeopteryx and
other Mesozoic birds have more than two carpals, Martin claims dromaeosaurs
are more likely to be derived from birds than vice versa. But even if
derived dromaeosaurs did have such a reduced carpus (physically implausible
as it may be), that could just be an apomorphy of the clade after it split
from the line leading to birds, inside Theropoda (since non-maniraptoran
theropods also all have four carpals or more, except maybe mononykines). So
Martin's barracade is only relevent to a phylogeny where birds evolved from
dromaeosaurids, which no one supports.
Martin brings up digital homology, assuming digital identity and analgen
identity are equivalent, without even a mention of frameshifts.

Martin then goes on about the astragalar ascending process and pretibial
bone, spending quite a lot of time trying to show Archaeopteryx had a
pretibial bone. He ignores Shenzhouraptor (= Jeholornis), which clearly
shows an ascending process (that was noted and illustrated by the authors to
help prove BAD), despite having referenced the taxon only two pages prior.
Rahonavis would also help here, though one must wonder if Martin thinks it's
a maniraptoran or a sauriurine, or both. He says, "We can speculate the
common ancestor of birds and dinosaurs lacked an ascending process of the
ankle, and that bipedality was achieved independantly." Um, what about the
ascending processes of non-avian maniraptorans? Don't they throw a wrench
into Martin's hypothesis?

The short postacetabular process of birds is said to be more primitive than
dinosaurs, which was "already elongated in the coelurosaurian dinosaur
Coelophysis from the Triassic." Coelurosaurian Coelophysis? Martin sure is
up to date on dinosaurs... In any case, one might bring up the large
postacetabular processes of most oviraptorosaurs and dromaeosaurs, as well
as ornithuromorphs, which yet again mean the barracade was broken through
somehow regardless of what birds are. Perhaps Martin thinks that
postacetabular processes can only grow, and not shrink, but how he could
defend such a view is beyond me.

Finally, Martin thinks the apparently absent external mandibular fenestra of
Archaeopteryx is symplesiomorphic with pre-archosaurian reptiles, thus
providing more evidence birds are outside the crocodile-dinosaur clade. The
presence of mandibular fenestrae in oviraptorosaurs and deinonychosaurs is
said to be convergent with dinosaurs+crocs. Jixiangornis, Omnivoropteryx,
confuciusornithids, Vescornis and the Spanish enantiornithine nestling all
have mandibular fenestrae too. Thus the distribution is not so simple, and
cannot be used to place birds so basally. Of course, even if birds
universally lacked mandibular fenestrae, their absence in compsognathids
proves the state can evolve from one where fenestrae are present.

Martin ends the section with the kind of flawed logic Makovicky and Dyke
argued so firmly against. Just because dinosaurs and birds (supposedly) show
different character states, their common ancestor had to lack the states,
since they can't evolve from each other. Martin knows how states must/did
evolve. A pretibial bone can't evolve from an ascending process, so both
have to evolve from a tarsal arrangement without a proximally extended
component. Etc.. In addition, Martin constrains states to evolve in only one
direction. Mandibular fenestrae can't close, sterna can't become
cartilaginous, femoral heads can't angle anteriorly, sickle claws can't
reduce. It's all just assumed, and is so often contradicted by exceptions in
each case.
Figure 2 is *gasp* an ABSRD cladogram of bird origins. It has the following
`--+--+--Longisquama, etc.
| `--+--maniraptorans
| `--birds
Notice Martin's basically advocating non-archosaurian birds. Of course,
birds are by definition archosaurs, but in this scheme birds are placed
outside the normal concept of Archosauria (including thecodonts, crocodiles
and dinosaurs), and are descended from protorosaurs like Cosesaurus ands
avecephalans (all normally outside Archosauria). It's nice to see Martin
agreeing with current phylogenetic thinking that avicephalans and Cosesaurus
are outside the Crocodylia+Dinosauria clade. However, he still places them
as archosauromorphs higher than prolacertiformes, because of their supposed
antorbital fenestrae. When more characters are analyzed (e.g. in Senter,
2004), they fall outside Sauria. Note the precise placement of Sauria in
Martin's cladogram is unknown since he doesn't include lepidosaurs.
Also interesting is Martin placing thecodonts as avemetatarsalians, because
of their 'superpleurodont' teeth. His use of this term doesn't make sense.
Dinosaurs are not pleurodont- their teeth are not fused to the dentigerous
bones and they are in distinct sockets. In any case, Martin previously said
that "dinosaurs retain the primitive pattern of tooth replacement", and
indeed taxa as basal as Mesosuchus and Captorhinus share it. So this
character can't support placing birds outside traditional Archosauria.
Anyone know what kind of tooth replacement sphenosuchians have?
Finally, Martin uses the elongated postacetabular process as a character to
unite ornithischians and Coelophysis, but this character is not present in
basal saurischians (Eoraptor, herrerasaurids) or sauropodomorphs. The
elongation developed convergently in theropods and ornithischians. Martin
could have used one of the many actual dinosaurian synapomorphies which are
known (no postfrontal, cervical epipophyses, perforate acetabulum, prominent
anterior trochanter, etc.), but maybe their presence in basal birds
discouraged him.

Martin finds "such a wide distribution of feathers or hairy coverings" (as
seen in coelurosaurs and pterosaurs) improbable. I don't see what's so
improbable about one more clade (in addition to mammals and
coelurosaurs/birds) having filamentous integument. Martin is noncommittal in
regards to non-bird filamentous integument though.

"Protarchaeopteryx has typically avian teeth and there is no reason to doubt
that it is avian." I suppose if maniraptorans are birds in Martin's mind,
that's true enough.
The long primaries of Caudipteryx are said to "so greatly reduce the
usefulness of the hand for grasping that it must imply derivation from an
arboreal flyer/glider." Someone hasn't read Gishlick (2001) I see.
"An arboreal stage is further supported by a reflexed hallux on the foot in
all of these forms." Someone hasn't read Middleton's work either.
"The most recent cladistic analyses of oviraptorosaurian relationships (Lu
et al., 2002; Maryanska et al., 2002) show oviraptorosaur(sic) as flightless
birds more derived than Archaeopteryx." Convenient how he picks out the two
oddball analyses, eh? What about all the versions of the Theropod Working
Group's analysis? Holtz's? Sereno's? Rauhut's?
There are a lot of unsupported assumptions about microraptorians. The
hindwings were "clearly a gliding adaptation". "In order for the femur to
function, they must be able to extend laterally." The "animal could barely
walk, let alone run." This is also supposedly supported by the small toes
and reversed hallux. I doubt the toes are any more reduced than
ornithomimosaurs and troodontids, and the hallux is not reversed.

Figure 3 includes a portion of the TWG phylogeny, to show how primaries
imply a volant ancestor for oviraptorosaurs and eumaniraptorans. This would
suggest Martin agrees oviraptorosaurs are outside Eumaniraptora, but see

Martin claims Sereno (1991) called Scleromochlus a dinosauromorph, but he
did not. He says that only "birds, some pterosaurs, and certain basal
archosaurs" have scapulae parallel to the vertebral column. Rather
inconvenient for him the 'basal archosaurs' he uses as bird ancestors aren't
included in that list. Longisquama has the primitive oblique position
similar to most dinosaurs, drepanosaurids have scapulae perpendicular to the
dorsal column, and Cosesaurus lacks a scapular blade altogether. Cosesaurus
is claimed to have a furcula, contra Sanz and Lopez-Martinez (1984) and
Peters (2000). Megalancosaurus is said to as well, which I don't recall
being reported anywhere else. Martin now correctly believes Longisquama
lacks a mandibular fenestra, and says it has thecodont dentition, including
teeth with constricted bases. Other avian characters are listed, but are of
course, also found in non-maniraptoran theropods- accessory antorbital
fenestrae (also in averostrans), pointed snout (coelurosaurs), elongate
postorbital (how is this birdlike?), strap-like scapula
(ceratosaurs+tetanurines), furcula (theropods), elongated penultimate
phalanges (theropods) and feathers (coelurosaurs).

Martin says "although we might expect flight feathers to have originated
from scales along the edge of supporting structures", the fact primary
feathers originate from the middle finger of birds and the middle of the
back in Longisquama "would only make sense if they were derived from
overlapping scales." I don't follow this logic. Besides the non-sequitur,
there are a couple of additional problems. Martin's Triassic bird Protoavis
has more quill knobs on its third metacarpal than its second, so might
suggest basal birds developed primaries on the outer edges. Also, because
maniraptorans had arched non-pronated hands, their 'outer edge' IS
metacarpal II. Martin goes on to ignore all the developmental data showing
feathers did not evolve from scales and are only homologous at a very basic
level. He speculates Longisquama "may have expanded its ribs much as in the
prolacertian rib gliders." Which gliders he's talking about is uncertain, as
I would expect him to include drepanosaurids in his 'Longisquama, etc.'
group, which is closer to dinosaurs than prolacertids in his scheme. Does
Martin think drepanosaurids are prolacertiformes which developed their
birdlike characters convergently? In any case, Martin also ignores the
criticisms of Reisz, Sues and Senter which show most of the feather-like
aspects of parafeathers are illusions (sheath, rachis, barbs). He
furthermore claims that Confuciusornis' retrices have their odd morphology
because the sheath stays fused to the barbs and rachis. I doubt this
arrangement would work, and find it more probable the barb ridges simply
aren't separated early in ontogeny (which I think is how modern birds with
similar feathers develop).

Figure 4 includes terribly inaccurate reconstructions of Longisquama's skull
and furcula. The furcula looks like it was just altered from Archaeopteryx
or Bambiraptor, instead of the more U-shaped structure in Longisquama with
its pointed distal tips. The cranial reconstruction assumes a posterior
section of the skull was broken off, which is contradicted by the cervical
vertebrae extending to the preserved posterior edge. The frontoparietal
crest is nowhere to be seen. The naso-fronto-lacrimal area has been
redesigned to look theropodan. The ventral mandibular edge has been changed
to expand beneath the orbit. The Archaeopteryx is inaccurate too, of course,
with no ventral squamosal or postorbital processes.

Martin's conclusion has much praise for Paul's work, but misrepresents it as
being equivalent to his own. Neoflightless maniraptorans are not a
compromise between BAD and ABSRD, since Paul still thinks maniraptorans and
birds are dinosaurs, while Martin thinks neither are. Martin claims Paul's
placement of various maniraptorans closer to birds than Archaeopteryx solves
the problem of "numerous anatomical anomalies indicating that these special
dinosaurs are too derived to be the ancestors of birds." Yet Paul has
maniraptorans ancestral to birds in the same way Gauthier and almost
everyone else does. He only differs in the order of pre-pygostylian
offshoots (Archaeopteryx more basal, oviraptorosaurs more derived, etc.).
Another interesting fact- "The monophyly of the Dinosauria has always been
difficult to support." Apparently Martin hasn't read any of the relevent
literature since the mid 1980's.
Martin still believes in the Sauriurae-Ornithurae split. He notes that
ornithurines "all have rod-like quadratojugals", while "sauriurines all have
reduced L-shaped quadratojugals." Another symplesiomorphy used to support
Sauriurae?! Who would have guessed? I suppose that makes Cosesaurus,
Protoavis and Shuvuuia ornithurines. It will be interesting to see what the
quadratojugal of the 'basal ornithurine' Yanornis looks like.
Martin discusses maniraptorans in the sauriurine section (they do have
L-shaped quadratojugals...), but says they "may form a more exclusive clade
united by the mandibular fenestra and fan-like tail feathers terminating on
an elongate tail." He then says Confuciusornis may be related to
oviraptorosaurs, while Jeholornis "seems closer to dromaeosaurs". So one is
left confused as to exactly how Martin thinks maniraptorans relate to birds.
Are maniraptorans sister to Ornithurae+Sauriurae, sister to other Sauriurae,
polyphyletic within Sauriurae? He supports each of these hypotheses
somewhere in this paper.

One has to wonder if Martin has thought through the consequences of
embracing Protoavis as a bird (the following comments assume Chatterjee's
identifications are correct). The acetabulum is completely open and the
femoral head medially projected, which would not be expected in a Triassic
bird if Archaeopteryx's supposed "somewhat sprawling posture" is primitive.
It also has an astragalar ascending process, large postacetabular process
and external mandibular fenestra. This would support these as
symplesiomorphies of birds, contra Martin's hypothesis. Manual digit III
(digit IV according to Martin) lacks elongate penultimate phalanges, unlike
digit IV of Longisquama; the clavicles are unfused and there are no
accessory antorbital fenestrae, both less birdlike than Longisquama. Of
course, this all assumes Protoavis is actually helping the temporal paradox
by being a bird more basal than maniraptorans, sauriurines and ornithurines.
The first five characters listed above could support Protoavis as a
maniraptoran in Martin's scheme. Sauriurine characters seem to be absent
(based on Hou et al., 1996), while some 'ornithurine' characters are seen
instead (e.g. rod-like quadratojugal; slender furcula; concave facet for
scapula on coracoid; procoracoid process; large sternal keel placed
anteriorly). Other 'ornithurine' characters are absent though (e.g. elongate
sternum; coracoid sulci on sternum; distally fused metatarsus; tarsal cap).
So Protoavis either helps the temporal paradox but disproves most of Martin'
s proofs that birds branched off prior to theropods; or it makes the
temporal paradox worse by requiring ghost lineages of 65 Ma for sauriurines
and 80 Ma for ornithurines, and makes Martin's phylogeny less credible by
either having dinosaurian/maniraptoran characters in a basal ornithurine or
ornithurine characters in a maniraptoran. The number of additional ghost
lineages that would be needed if it is maniraptoran is uncertain pending
precisely how Martin thinks they relate to (other) sauriurines.

In conclusion, Martin uses the standard BAND debating tactics without
addressing the meta-issues raised by BAD supporters, doesn't realize
implications for his points which argue against them, shows an ignorance of
dinosaur knowledge, misrepresents anatomy and misunderstands phylogeny
reconstruction and character distribution, and uses several non-sequiturs.
On the plus side, we finally get an explicit BAND hypothesis, although now
the exact relationships and membership of the maniraptorans is unclear.

Mickey Mortimer
Undergraduate, Earth and Space Sciences
University of Washington
The Theropod Database - http://students.washington.edu/eoraptor/Home.html