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

Rauhut and Xu on Tugulusaurus and Phaedrolosaurus

  Rauhut and Xu have recently performed an analysis of two ambiguous theropods
from the Early Cretaceous of Xinjiang Province, China, in the Lianmugin
Formation (possibly from the Valanginian up through the Albian) at Wuerho,
*Tugulusaurus faciles* and *Phaedrolosaurus ilikensis*. Previously described by
Dong in 1973, they had not been adequately compared to taxa available at the
time and now, with more taxa available for study, an update is in order. As the
authors note, Sues in 1977 argued that *Phaedrolosaurus ilikensis* is based on
a single tooth, with some postcrania used to establish a comparative
description. However, this material cannot be compared to the tooth, since it
consists only of a partial hindlimb, and was not found in the same locality.
The material is comparable to the holotype of *Tugulusaurus faciles* but is
different in many features of the astragalus and proximal tibia, and thus
cannot be referred to *Tugulusaurus*. The authors consider the leg to be
diagnostic and comparable to other species, establishing the new taxon
*Xinjiangovenator parvus*. There are some problems I am having with this,
however, including the establishment of taxa based on a single autapomorphy, so
I am taking a closer, critical look. Also known from the same formation and
area, Dong has described the ?carnosaur? *Kelmayisaurus petrolicus* and the
stegosaur *Wuerhosaurus homheni*, as well as the dsungaripterid pterosaur
*Noripterus complicidens*.


  Dong Z.-M. 1973. [Dinosaurs from Wuerho. Memoirs of the Institute of
   Vertebrate Palaeontology and Palaeoanthropology.] _Academia Sinica_
   [in Chinese]
  Rauhut, O.W.M. and Xu X. 2005. The small theropod dinosaurs *Tugulusaurus*
   *Phaedrolosaurus* from the Early Cretaceous of Xinjiang, China. _Journal of
   Vertebrate Paleontology_ 25(1):107-118.


  *Tugulusaurus faciles* is based on IVPP V4025, consisting of four caudal
vertebrae, a fragment of a dorsal rib, left and right metacarpal I, and left
and right phalanges I-1 and both first articulated unguals (mdI-2), left femur
(215mm), right tibia (~240mm, rather fragmented and beaten up), left astragalus
and calcaneum, partial right astragalus, distal ends of metatarsals III and IV,
a pedal phalanx and a pedal ungual.

  The material is well preserved, but cracked and broken, especially the tibia,
obscuring some details of anatomy. Rauhut and Xu diagnose it as "Small theropod
(femur length c. 215 mm); proximal mid-caudal vertebrae with neural arch placed
only on anterior two thirds of centrum and centrum considerably broader than
high (ratio width/height ca. 1.5); caudal vertebrae rapidly increasing in
length distally; minimal length of metacarpal I less than width of this bone;
tibia with pronounced, semicircular lateral expansion of lateral malleolus."

  The caudals have distinct lateral ridges in place of transverse processes,
indicating vestigial such processes, except for the first in the series which
has small triangular "spurs;" while the authors remark on the rapid increase in
length, three of the bones are nearly identical in length (23-25mm) while the
fourth is 34mm, but the distance between bones is unknown, so the state of
rapid lengthening is actually unknown. All caudals do come from the distal half
of the tail, given the shortness of the neural spines and transverse processes.
The forelimb material is diagnosed primarily by the extremely short first
metacarpal (shorter than broad), though such a condition exists in nearly
identical proportions in *Acrocanthosaurus* and *Suchomimus*. The manual
unguals do not curve greatly along their ventral margin. The femoral caput is
medially and slightly anteriorly oriented, is semi-spherical, and has a level
dorsal margin from medial end towards the trochanteric crest, but appears to
decline slightly in some views. The shaft is nearly circular, but crushed, and
has a slight cranial curvature but due to crushing it is unlikely this
represents its natural condition.  The tibial condyle is very short and higher
than long. The tibia is very difficult to discuss aside from the most proximal
and most distal ends. The proximal end preserves a short, rounded cnemial crest
which is deeper than long and does not project dorsally above the femoral
cotylus, and the fibular buttress on the lateral edge of the tibial caput is
reduced to a little ridge; the femoral condyles are distinct and there is a
pronounced groove between then, which are nearly of the same width. The
astragalar facet is relatively short and confined to the lateral half of the
shaft; the medial and lateral malleoli are distinct and pronounced with the
medial malleolus higher and narrower than the lateral malleolus, a condition
seen in some other species such as *Kakuru*. The astragalar ascending process
is broken, but was very narrow and is confined to the lateral half of the bone,
a condition also seen in some birds and non-coelurosaurian neotetanurans. The
calcaneum is discoidal and 1/5 the astragalar width; there is no lip capping
the calcaneum, unlike other coelurosaurs. By comparison, the calcaneum of
tyrannosaurs and *Compsognathus* are less than 1/5 the width of the astragalus.
Rauhut and Xu ran a matrix of 136 characters and placed *Tugulusaurus* as the
most basal coelurosaur. However, the material appears to be more consistent
with a non-coelurosaurian avetheropodan, as in carnosaurs.

  Mickey Mortimer wrote on *Tugulusaurus* based on the original paper by Dong
describing the taxa:


  His conclusions remain substantially close to my own given the new
description and Rauhut and Xu?s observations, regarding a very similar
relationship to both ?carnosaurs? and coelurosaurs, and a perhaps intermediate
or ?coelurosaur-like? avetheropodan. By preliminary conclusions are that
*Tugulusaurus* represents a possible avetheropodan prior to the development of
most coelurosaur apomorphies, and it?s features held in common above cannot be
accepted as exclusive apomorphies due to the mixture of ?carnosaur? and
coelurosaur-like features. Cladistically, it is a basal coelurosaur, and I
would advocate it as a coelurosaur based on the shared features, whereas
?carnosaur?-like features are plesiomorphic of the taxon.


  The type specimen is IVPP V4024-1, a tooth missing the apex, which is
apparently ?short, thick and solid? (Dong, 1973; translated by the Translation
Bureau of the Multilingual Services Division of the National Museums of Canada
? thanks to Mickey for providing this) in cross-section, as well as being
serrated along the entire distal carina but without serrations at the base of
the mesial carina. While Dong says it is not knife-shaped (ziphodont), the
tooth is described as having a flattened labial face halfway up the crown, has
almost no distal recurvature and only little mesial recurvature, there is some
distal curving of the crown present, and the crown appears to be only about
twice as high as the fore-aft basal length of the preserved crown. Other
material referred to IVPP V4024 has been catalogued as IVPP V4024-2, to which a
fibula catalogued under IVPP V4025 has been placed due to its connection with a
portion of articulated fibula. Rauhut and Xu determine the type material (as a
virtually non-diagnostic tooth) cannot be adequate to typify a species, and
remove what is considered to be diagnostic limb material into a new taxon
(below). In fact, the description of the crown is not entirely consistent with
theropod teeth, and aside from the tiny serrations, but these are present in
the Late Cretaceous *Goyocephale* as well.


  The holotype specimen is IVPP V42024-2, also comprising the fibula associated
with IVPP V4025. It includes a right tibia (tibiotarsal length, 312mm), right
fibula (152.5mm, proximal ?free? portion; 75.5mm, middle to distal ?free?
portion; and 32mm distal fragment articulating with tibia and proximal tarsals
- a combined length is likely close to 310mm, but a gap between the proximal
and median portions makes judgment weak), right astragalus (preserved height,
42mm; width, 37mm) and calcaneum (height, 18.6mm; width, 6mm) all articulating
with one another. The material is substantially better preserved than in
*Tugulusaurus* from the same locality.

  Rauhut and Xu diagnose *Xinjiangovenator* as: ?Small theropod (length of the
tibia plus proximal tarsals: 312 mm); lateral femoral cotyle (?fibular condyle?
of Rauhut and Xu) of tibia extending farther posteriorly than lateral side of
proximal end of this bone; fibula with longitudinal groove on anterior side of
proximal end.? Almost assuredly, more material would produce a more adequate
description. The diagnosis largely hinges on the unique feature of the cranial
groove on the fibula, in my opinion.

  The position of the lateral femoral cotyle is further caudal than the medial
femoral cotyle. The distinct nature of the tibiae compared between the two taxa
seems to be more inclined towards the greater length of the cnemial crest in
proximal view in *Xinjiangovenator* than in *Tugulusaurus*, as well as the less
blunted or rounded aspect of crest in dorsal view, which is ?sharper? and less
laterally oriented in *Xinjiangovenator* than it is in *Tugulusaurus*. The
cnemial crest is trapezoidal with distinct proximal, anterior, and distal
?edges? and the head slightly projects more proximally than the femoral
cotylus; the anterior margin of the crest is nearly parallel to the shaft, and
there is a sharp crest on the proximal end extending the length of the cnemial
crest caudally onto the femoral cotylar surface and to the lateral cnemial
crest laterally. The tibial shaft is straight, and becomes narrower
cranio-caudally below the mid-point, but not as much as in ornithomimosaurs or
other coelurosaurs with tibias as slender as in *Xinjiangovenator*. A dorsal
groove separates the condyles of the astragalus, but is relatively shallow as
compared to non-coelurosaurs or *Piatnitzkysaurus*, or example. The base of the
astragalar ascending process is broad across the width of the astragalus, and
the preserved portion indicates the process was relatively tall, though a
distinct fossa for its position on the anterior surface of the tibia is not
present to indicate how tall it was. The calcaneum is discoidal, and is less
than 1/5 the width of the astragalus, but there is now overhanging lip or
embayment in the astragalus to receive the calcaneum. The fibula has what is
considered a distinct groove on the anterior end of the proximal portion of the
bone; the iliofibular crest on the fibula is a broad crest that projects
further laterally than anteriorly, but is relatively small compared to some
similarly oriented taxa, such as segnosaurs. The medial surface preserves a
large and distinctly emarginated fossa in the proximal end, extending for most
of the anteroposterior length and from the proximal end down nearly to the
level of the iliofibular crest. The median fibular shaft is largely oval and
appressed to the tibia based on its shape, but is not medially concave; the
distal end is broad compared to the more proximal shaft, and it contacts the
astragalus only on the lateral surface rather than forming a corollary facet on
the dorsal side of the bone, as in other coelurosaurs; the same is also true of

  Mickey Mortimer discussed *Xinjiangovenator* using Dong (1973) and updating
the information regarding the new paper discussed here:


  Mickey notes constraining the tree to make *Bagaraatan* a tyrannosauroid,
whereupon *Xinjiangovenator* follows suit. Though *Bagaraatan* is largely
tyrannosauroid-like due to the jaw anatomy, comparison with *Xinjiangovenator*
and tyrannosauroids would perhaps clarify some anatomical differences. For
example, the cnemial crest of tyrannosaurs is much more laterally oriented and
the semi-lunate, helical astragalar groove above the condyles is less
pronounced. A large calcaneum, broad and expanded distal fibula, and low
cnemial crest appear to be shared between *Xinjiangovenator* and
tyrannosauroids, and the cranial orientation of the astragalar condyles with an
apparent ?inflection? cranially of the distal tibia cranially is also seen in
ornithomimosaurs, dromaeosaurids, troodontids, *Bagaraatan*, birds, and some
oviraptorosaurs, but not other coelurosaurs, a complex of features that would
seem to advocate for a less ?derived? position for *Xinjiangovenator* than
Rauhut and Xu advocate and perhaps even a more ?tyrannosauroid? position.

Statistical Analysis

  Rauhut and Xu used an analysis of 17 taxa and 136 characters, of which only
11 (8%) could be coded for *Xinjiangovenator* and 25 (18%) for *Tugulusaurus*.
They found the following strict concensus topology (from 10 trees, 246 steps,
CI of 0.618, RI of 0.719, RC of 0.444, and HI of 0.382):

                    |  `--Therizinosauroidea
                       |  `--Xinjiangovenator

  Variation occurs in two polytomies marked by the basal coelurosaurian
radiations and perhaps contributed by the number of coded characters for
*Coelurus* (38; 27%), comparable only to the similarly uncertain position of
*Nqwebasaurus* (42; 31%) and *Bagaraatan* (27; 20%); in this matrix, 1/5 to 1/3
of the codable matrix may show a less than concise placement for these taxa. I
have not tried to run these taxa in Holtz? updated GAIA matrix yet, though I
may get around to this at some point. The authors of course questioned the
referral of these taxa to where the tree places them, but advocate for a
coelurosaurian relationship for both based on combinations of characters:

  *Tugulusaurus* is a coelurosaur based on (1) the medial surface of metacarpal
I produces a sharp crest where it contacts metacarpal II; (2) no extensor pit
on the dorsal surface of mcI; (3) a reduced, laterally oriented facet on the
astragalus for contact with the distal end of the fibula (thus no socket for
its reception); (4) a semi-lunate groove at the base of the astragalar
ascending process offsetting it from the anterior margin of the condyles
(forming, in fact, a dorsal margin for the condyles apart from the ascending
process); (5) and absence of a horizontal groove across the anterior surface of
the astragalar condyles.

  *Xinjiangovenator* is a coelurosaur based on (1) a high ascending process of
the astragalus that arises from the complete breadth of the astragalar body;
and the shared features with *Bagaraatan* of the proximal tibia: (2) lateral
femoral cotylus extends further caudally than the medial and (3) the reduced
incision between medial and lateral cotyles.

  Bootstrap support values for the *Bagaraatan* + *Xinjiangovenator* clade is
weak (47%), supported only by the combined (possibly related) characters of the
posteriorly placed lateral femoral cotyle and the reduction of an incision
between it and the medial femoral cotyle. Due to extensive homoplasies rampant
in theropod analyses (Holtz, 1999 [abstract and poster], 2002), bootstrap
values are relatively low, but this also has a lot to do with the extensive
smapling of characters compared to other dinosaur groups, and such homoplasious
values in other analyses, such as mammals, are comparable if not prone to more
extensive changes in topology from one data set to another.

  However, the tibiae of these two taxa (*Bagaraatan* and *Xinjiangovenator*)
differ greatly in the proximal end, for which *Bagaraatan* compares more
closely to more Avialan taxa than to *Xinjiangovenator*; the proximal tibia is
wider than anteroposteriorly long, and the cnemial crest is almost entirely
mediolaterally oriented rather than projecting anteriorly; the lateral cnemial
crest is massive and projects strongly laterally, forming a groove between the
condyle and cnemial crest; the distal fibula has almost no contact with the
astragalus and is much narrower on the shaft, with a crescentic cross-section
(medial surface is concave); the fibula doesn?t have a distinct broad,
basin-like fossa but a narrow, valley-like one and the proximal fibula is
constricted below the proximal end more than 50% of its anterorposterior
length; the posterior process of the proximal fibula is longer than the
anterior, whereas the opposite is true in *Xinjiangovenator*. The iliofibular
crest does not project at all cranially as it does in *Xinjiangovenator*. A
groove on the anterior condyles of *Xinjiangovenator* is absent in *Bagaraatan*
as in *Tugulusaurus* and other coelurosaurs, but this channel is very shallow
and the eroded nature of this region of the tibiotarsus leaves this feature in
some doubt as to its authenticity. I recoded the matrix of Rauhut and Xu
(thanks to Mickey for saving me a bunch of time in transcribing it into a nexus
file for me) for the few differences I saw between Rauhut and Xu?s work and my
own observations. Primarily I dealt with the *Xinjiangovenator* and
*Tugulusaurus* codings, preferring not to involve myself into a more detailed
study of their matrix, small as it is, when I can rather code into Holtz?s
updated GAIA matrix, which I will do shortly. Rauhut and Xu coded some features
using multistate characters for the minimum guesses allowable given the
material, a method I find unique and perhaps informative; they run the taxa as
if they were both state A and state B, the median between which is the likely
result of the character, were it complete. This is true of the way they coded
the height of the astragalar ascending process in *Xinjiangovenator* (states 1
and 2 of possible 0, 1 and 2 states: shorter than height of astragalar body, as
tall as astragalar body and twice as high as astragalar body or taller), yet
the bone is incomplete; while the process can be said to be taller than state
1, it cannot be determined in relation to state 2, so the authors coded it as
?(12).? I find this confusing in the matrix since it assumes what is
essentially an unknown. The bone is neither state 1 nor state 2, and while it
can be the latter, it can _not_ be the former, so treating it as potentially as
state 1 is confusing. In some cases, guesses were made about complete
preservation in the femur and tibia of *Tugulusaurus*, and these were reduced
to ??? because of preservation. The results supported a sister-group
relationship between *Bagaraatan* and *Xinjiangovenator* based on the reduction
of the incision between the femoral cotyles of the tibia and the greater length
of the lateral cotyle caudally than the medial cotyle, as in Rauhut and Xu

  I reran the matrix in PAUP* 4.0b10. The new analysis produced 10 MPTs (most
parsimonious trees) with 249 steps, a CI of 0.6104, a RI of 0.706, a RC of
0.431, and a HI of 0.3986. The strict consensus tree is identical to the one
produced using Rauhut and Xu?s original codings, so the changes in tree
statistics are due to the reduced multistated characters and increase in ???
codings. However, when I bootstrapped the tree (50% majority-rule concensus,
100 replicates) in PAUP*, it produced the following tree:

                  |   `--Therizinosauroidea
                  |   `--Xinjiangovenator
  Note that coelurosaurs more advanced than ?Compsognathidae? (*Compsognathus*,
*Sinosauropteryx*, *Aristosuchus* and *Mirischia*) collapse except for Paraves
(dromaeosaurids, troodontids, and birds), the oviraptorosaur-segnosaur clade,
and again, the *Bagaraatan* + *Xinjiangovenator* clade. Support at 75%
majority-rule retains a basal position for *Tugulusaurus* within Coelurosauria,
for Paraves, and for the sister-group relationship of ?Compsognathidae? and the
other coelurosaurs.


Jaime A. Headden

  Little steps are often the hardest to take.  We are too used to making leaps 
in the face of adversity, that a simple skip is so hard to do.  We should all 
learn to walk soft, walk small, see the world around us rather than zoom by it.

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

Do you Yahoo!? 
Yahoo! Small Business - Try our new resources site!