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

Nemegtia and Oviraptorid Phylogeny (Part 1)

  Thanks to many people getting their hands on this paper and getting a copy to
me, I now have

  Lü J.-c., Tomida Y., Azuma Y., Dong Z.-m. and Lee Y.-n. 2004. New oviraptorid
    dinosaur (Dinosauria: Oviraptorosauria) from the Nemegt Formation of
    southwestern Mongolia. _Bulletin of the National Science Museum, Tokyo_,
    Series C, 30:65-130.

  The abstract has been previously posted.

  The holotype is GIN 100/2112, misprinted as GIN 10012112, which is considered
a type of translating a "/" into a "1" (it was provisionally considered
PC100/2112 while Lü studied it, so it has now been formally accessioned in the
Mongolian Geological Museum, Ulaanbaatar, Mongolia); Lü's previous publication
used the slash rather than the numeral. It comprises the skull, a complete
cervical series (fourteen cervicals) and partial dorsal series (dv1 and the
last five dorsals), sacrum (7 fused sacrals, though the authors listed eight
the last one appears to be a caudal and does not look as fused as the more
proximal sacrals in the synsacrum), and proximal 2 caudals, both ilia, the
proximal pubes and ischia, the proximal right femur, proximal end of the left
scapula and both distal humeri with a complete right radius; all except the
forelimb elements and dorsal series are articulated, though the gap between the
dorsals is significant.

  It derives from the Nemegt locality of Gradzinsky et al., as noted
previously, which would be "Nemegetu" in Mongolian typography as used by
Barsbold in various papers, and which lent its name to the valley and
formation. Thus the age is secured to the type inclusion of *Tarbosaurus*,
*Therizinosaurus*, etc.

  The mandible of an oviraptorid is subjected to CT for the first time, which
shows massive differences from caenagnathids in the absence of any medial
ornamentation or what I've tended to call the "sub-dental shelf", the platform
of bone on which teeth are situated. The mandibular symphysis is fused, with
eradication of portions of the "mandibular pillar" as seen in CT, causing
pneumatization of the intermediate bone; however, traces of the suture between
the mandibles is present on the rostroventral portion of the symphysis (though
less so the dorsocaudal) in a distinct "serrated" pattern, indicating that
fusion doesn't mean "obliteration of sutures." Arguments of fusion between
elements should involve some exploration of the nature of actual separation of
bones or loss of intermediate bone. Foramina bound the exterior of the
symphysis, and the symphysis has a ventral process or "chin" as present in
*Conchoraptor*, *Khaan*, *Heyuannia*, GIN 100/42 and *Oviraptor*, but not
*Citipati* or *Rinchenia*. Fusion of the surangular and angular is only caudal,
in contradiction of the typical pattern of complete obliteration of the suture
or fusion only in the rear of the bone as in other oviraptorids. A short,
robust retroarticular process is present in *Rinchenia* and *Citipati*, as well
as *Conchoraptor*, but tends to be longer with an expanded distal end and more
"slender" appearance in other taxa for which this region is known. The mandible
otherwise resembles *Conchoraptor* and *Khaan*.

  The frontals are extremely short along the midline and are unfused, and the
parietal is wholly the second longest bone in the skull, second only to the
pterygoid which is usually the longest when measured along the axis of the
basal skull length (in some oviraptorids the premaxilla's longest dimension,
usually height, exceeds other cranial measurements), whereas the parietal is
only about as long as the jugal; frontal brevity is likely due mostly to the
parietal overlying the frontal in a posterior scarf join, whereas normally it
is the nasals overlying the frontal; the posterolateral process of the frontal
is extended caudal to the parietal contact and forms the base of the
postorbital process. The parietal also becomes peaked in the midline and,
unlike other oviraptorids, appears to form a sagittal keel. The rear of the
skull appears to somewhat crushed causing the quadrate to be strongly bowed
cranial and the distal end of the paroccipital processes to strongly overhang
the skull caudally and nearly reach the ventral end of the quadrate. The
medial-ventral maxillary process are very short, a condition that occurs in
*Conchoraptor* and *Rinchenia*. The splenial is broad anteriorly and thins and
tapers caudally, a transition found in other oviraptorids but highly divergent
from the typical archosaur pattern. There is no separate coronoid apparent,
though if this means fusion or loss of the bone, is unknown. The nasals are
fused for their length, though intervention of the premaxillae separates the
rostrodorsal tips; the nasals extend for the length of the crest, and the
premaxillae extend caudally on the dorsal surface only for a few milimeters.
The premaxilla is truly "V"-shaped in profile, with one ramus extending only
dorsally, and the other caudally above the maxilla and lachrymal.

  The Quadratojugal lacks an ascending process and is not fused to the
quadrate; as in other oviraptorids, a medial socket receives a peg from the
lateral quadrate to lock the two in place. The quadrate is extremely broad and
there is virtually no ventral process to demarkate the ventral edge of the
pterygoid flange, which essentially ends ventrally at the distal quadrate
condyles, unlike other oviraptorids. Erosion of the squamosal reveals the
pneumatic and slender proximal head of the quadrate. Extensive braincase fusion
results in one of the largest of oviraptorid orbitosphenoids, and an
extensively anterior position of the laterosphenoid. The vomer has the typical
"stickman" shape as described by Barsbold for GIN 100/42, and the ventral
premaxilla possesses two distinct ridges that abut the maxillary ventral
"prongs", with two further lateral ridges and finally the narrow "C"-shaped
premaxillary tomia, which is highly vascularized and forms a festooned margin
on the surface but not large denticle-like serrations. The ectopterygoid and
palatine are both nearly straight in lateral view and narrow in ventral view,
and much smaller than in other oviraptorids of large size (*Rinchenia*,
*Citipati*, GIN 100/42 or *Oviraptor*). Normally the ectopterygoid is biramal
(has only two ends) and "dumbell"-shaped with an "elbow" in the middle in
lateral view, and the palatine is narrowly triradiate, but in this case the
palatine more closely resembles the biramal ectopterygoid and is constricted in
the middle. *Nemegtia* has an unsually narrow palate for an oviraptorid.

  The external naris is irregular and it's clear it's margins are not entirely
natural and were thus much smaller than shown. Unprepared matrix or crushed
bone in the whole of the premaxilla may indicate the external naris was more
oval and with the dorsal margin further caudal than illustrated. However, shape
of the premaxilla is unique, so shape of the naris may also be unique. The
antorbital fenestra is large and ovate, and the maxillar fenestra is tiny. The
orbit is large and circular, unlike *Rinchenia* or *Oviraptor*, but is longer
dorsoventrally and rostrocaudally, unlike other oviraptorids. The lower
temporal fenestra is taller than rostrocaudally long and roughly rectangular in
appearance, with a narrower dorsal margin than ventral; the supratemporal
fenestra is elongated, roughly "egg"-shaped, and each is narrower than the
paired and fused parietals.

  The cervicals are the first well-illustrated oviraptorid cervicals, as those
in *Khaan* were very summarily described and thus few details are available for
comparison. They are preserved from the posterior half of the axis to the 12th
cervical, which are articulated to what Lü et al. identify as the first dorsal;
the axis is said to be fused to the third cervical. The centra are much longer
than in *Khaan* but become nearly quadratic or about as long as high in the
thoracic transition. A distinct rectangular neural spine is present on nearly
all cervicals except where erosion in the middle of the series as removed them.
Cervical ribs are short and triangular without an elongated distal process; the
capitulum is very short. Postzygapophyses do not extend substantially caudal to
the centra, and thus do not really overhang the caudal vertebra. Epipophyses
are weak and "nubbin"-shaped, and are present on the fourth to last cervical,
while the first dorsal is not sufficiently preserved to compare. All cervicals
bear pleurocoels as preserved (the axis is not sufficiently preserved to
determine this, though in *Khaan* pleurocoels are evident in the axis).
Anterior cervicals are opisthocoelous, and all neural arches are "X"-shaped in
dorsal view, permitting use of this character to support caenagnathoid
monophyly. Dorsal vertebrae are mostly preserved as neural arches, with centra
of the last three dorsals preserving large round pleurocoels, short trapezoidal
neural spines, and neural arches deeper than centra. The last dorsal (first
presacral) is preserved between the anterior blades of the ilia. Preservation
of the elements prohibits knowledge of their fusion, making assignment of the
last element a sacral dubious, especially given the broad and apparently
unfused nature of the neural arches and spines to the posterior ilia; the ilia
terminate as preserved at the 7th sacral and the 8th? is entirely free; the 7th
bears saral ribs with are tightly connected to the ilia, but it is impossible
to infer the presence of the last "sacral" as such without fusion or contact
with the ilia, or preservation of haemal arches, thus lending doubt to the
assignment. There are thus between 7-8 sacrals, and the last "sacral" may be
the first caudal. Lü's caudals are preserved only as neural spines and are not

  The humerus is especially unique in preserving a distinct brachialis fossa,
unlike any other oviraptorosaur; there is no anconeal (rear of the distal end)
fossa and the distal end of the triceps fossa is preserved as with the distal
extent of the deltopectoral crest, indicating preservation of more than half of
the humerus; it is, nonetheless, very short. This condition can be used to
consider more avian affinities, as supported by Lü previously and Maryanska et
al., but it is not clear the nature of the fossa's impact on avian affinities
since most basal birds such as *Archaeopteryx* lack it. The radius is long,
robust, and bears a distal crest that would abut the ulna when articulated; the
proximal end is broader than the distal. Radius and humerus were most likely
nearly the same length, with the humerus longer by less than 15% or so. It was
thus about half the length of the ilium, if not slightly more than half.

  The ilium is both longer and deeper in the posterior blade than the anterior,
while the anterior blade lacks preservation of the ventral lobe of the blade,
preserved as in *Nomingia*. The ilia contact the sacral neural spines, but are
more laterally than dorsally inclined, unlike *Avimimus*. The pubic peduncle is
deeper than the ischiadic, while the latter is broader than the former. The
ilia are equally wide at anterior and posterior blades as preserved, and since
crushing in the posterior blade narrows the preserved width, it was likely
muchg broader; preserved distal end of the anterior blade also suggests it was
broader, so widths are approximate and neither may be wider than the other.
Ridges on the posterior blade lead dorsally into the thickened dorsal rim of
the ilium, as in *Microvenator* and *Caudipteryx*, but such features are absent
in other oviraptorids.

  The preserved proximal femur lacks division between and anterior and greater
trochanters, and is preserved only partially distal to the caput. There is a
surface groove that shows the division between anterior and greater
trochanters, but a dividing slot is absent, unlike in *Avimimus*. The dorsal
trochanteric crest is rounded and centrally placed along the cranio-caudal
length, as in other oviraptorids. Preservation obscures further details. Lü
reports that there is a neck between the caput and trochanters, and that the
caput is oriented medially at 90 degrees to the femoral shaft. Breakage shows a
highly avian "chambers and struts" interior surface, a term for which I am not
aware, since it is neither clearly "camellate" nor "camarate".

  Taxonomic typos abound in the paper, reflecting a nightmare in typography for
editioralship. The name is always *Nemegtia barsboldi* or *N. barsboldi* or
*Nemegtia* in all figures and tables and the cladistic analysis used at the end
of the paper, and is the spelling in the abstract and systematics section, so I
doubt any question of priority useage will arise. However, in the main text
misspellings are abundant (pg. 108 - Nemrgtia barsholdi; Nemrgtia harsholdi;
Oviraptor- philoceratops; Nemegtia harsholdi; pg. 109 - Nemegtiu barbsoldi;
Citiputi osmolskue; Nemrgtiu barsboldi; Khaun mckennai; pg. 110 - Khuun;
Nemegtiu barsboldi; Khaan mckennui; Nemegtia barsholdi; Nomingiu gobiensis;
Nemegtia hursholdi). No other typos in taxon names are noted.

  Lü et al. offer an analysis of 200 characters and 20 taxa, with almost even
balance between cranial and postcranial features, modified from Maryanska et
al. [Sigh]. Using PAUP and MacClade, a single most parsimonious tree resulted
with 481 steps and a CI of 0.5073. Previous criticisms of Maryanska's choice of
characters, which tended to favor oviraptorosaurs and birds, did not change
whereas basal non-maniraptoran taxa were largely ommitted due to internal
oviraptorosaur comparisons. The authors included several more recent taxa than
did Maryanska et al.

        |  `--Archaeopteryx lithographica
        `--+--Incisivosaurus gauthieri
           `--+--Caudipteryx zoui
              `--+--Microvenator celer
                 `--+--Avimimus portentosus
                    `--+--Chirostenotes pergracilis
                       `--+--Khaan mckennai
                          `--+--+--GIN 100/42
                             |  `--Oviraptor philoceratops
                             `--+--+--Heyuannia huangi
                                |  `--+--Nemegtia barsboldi
                                |     `--Citipati osmolskae
                                `--+--Conchoraptor gracilis
                                   `--+--Ingenia yanshini
                                      `--+--Rinchenia mongoliensis
                                         `--Nomingia gobiensis
  I will need to spend some time analyzing the support in the matrix.
Criticisms of the marix made 4 years ago by Mickey and myself may tend to cloud
my judgement, however. Aside from the unillustrated caudals, inability to
determine posterior skull or proximal/distal radial or distal humeral
morphology, this taxon whould be reasonably possible to code into Holtz, though
I will predict it will only confuse the matrix and force the apparently
universal (except in Lü, Lü et al., and Maryanska et al) inability to resolve
Oviraptoridae cladistically. I think this may be a tendency of not including
several oviraptorid-only cranial and postcranial features, of which I have some
30+ of, that while they have some distribution in other taxa, may help resolve
the affinities of these taxa. At heart include the position of the "ingeniine"
*Heyuannia*, and as a friend noted, the lack of consistency among crested
versus un-crested oviraptorids, long-handed or short, robust or gracile, etc.

  Since this post is "Part 1," "Part 2" will be any emmendations or
observations I make following the analysis of the matrix.


Jaime A. Headden

"Innocent, unbiased observation is a myth." --- P.B. Medawar (1969)

Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around