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

Testing Unenlagiinae

Since the publication of Buitreraptor, Rahonavis has emerged as a dromaeosaurid in the Theropod Working Group's matrices. Specifically it groups with Buitreraptor and Unenlagia in a basal subclade named Unenlagiinae. This was recovered in Turner et al.'s (2007) recent version in the Mahakala paper, which added several relevent taxa (Jinfengopteryx, Graciliraptor, Microraptor gui, Jeholornis, Sapeornis). It was also found in Senter's (2007) analysis, which is just the TWG analysis with many added taxa and characters. I added several additional paravians to his matrix (Pedopenna, Neuquenraptor, Jinfengopteryx, IGM 100/1005, Sinusonasus, Urbacodon, Mahakala, Shanag, Archaeornithoides, Graciliraptor, Ornithodesmus, Variraptor, Pyroraptor, "Koreanosaurus", Hulsanpes and Yandangornis), then examined the support for Unenlagiinae and its subclades. The character examination led to the correction of errors in the matrix, especially for paravians.

Unenlagiinae was supported by four unambigous synapomorphies in Senter's (2007) tree.

One is the proximodorsal ischial process, which was miscoded as absent in Sinovenator, Sinornithosaurus, Bambiraptor, Achillobator, Utahraptor and Archaeopteryx. It was miscoded as unknown in Mei and Wellnhoferia. It is however present in all of these taxa. One might argue the processes of some of these taxa aren't prominent, but neither is that of Buitreraptor. Archaeopteryx's process can be quite large (Munich specimen). It was miscoded as unknown in Troodon and Yanornis, but is absent in those taxa.

The second character is obturator process forms "a strongly acute angle in lateral view". This is coded present in Rahonavis (29 degrees), Buitreraptor (18), Microraptor (18), Sinornithosaurus (23) and Falcarius (33), and absent or unknown in all other taxa. Most other paravians do have obturator processes more obtuse than 60 degrees (Sinovenator 82, Sinornithoides 76, Saurornithoides mongoliensis 95, Troodon 91, Deinonychus 76, Velociraptor 149, Adasaurus 65, Bambiraptor ~83). Yet Unenlagia seems to have an angle of ~57 degrees, though it was coded unknown by Senter. Sinusonasus (55; unexamined by Senter), Achillobator (46; coded as plesiomorphic) and Archaeopteryx (24, 28 and 50; coded as plesiomorphic) all have more acute processes than Unenlagia. In fact, the Munich and Thermopolis Archaeopteryx have processes comparable to taxa Senter coded as derived. Here, Archaeopteryx was changed to polymorphic, while Unenlagia, Adasaurus and Troodon were changed from unknown to plesiomorphic.

Senter's third character is pedal phalanx II-2 "with transversely narrow posteroventral keel", as clearly described for Neuquenraptor and Buitreraptor. However, Senter incorrectly codes this as absent in Mei, Sinovenator and Microraptor (present in all, as noted in Makovicky et al., 2005). It is also present in Mahakala.

Senter's final unambiguous synapomorphy is "dorsal margin of postacetabular iliac blade concave", which is caused by a prominent dorsally projecting supratrochanteric process. He only codes this as present in unenlagiines, but it is also quite apparent in Achillobator, Adasaurus and scansoriopterygids. A more subtle concavity is present in Saurornitholestes, Bambiraptor, Sinornithosaurus, Wellnhoferia, the Munich Archaeopteryx specimen (but not the Eichstatt one) and perhaps even Sinovenator. Senter coded all of these as lacking the character, except Adasaurus and Sinornithosaurus which he coded as unknown. This is also a synapomorphy of unenlagiines in Turner et al.'s tree, where it suffers from the same miscoding problems.

Senter also listed three synapomorphies that supported the clade in ACCTRAN mode.

The first is six sacral vertebrae, which was only coded as present in Rahonavis among unenlagiines. Yet Unenlagia has six sacrals too (coded as unknown by Senter), as does Deinonychus (also coded as unknown). The Thermopolis Archaeopteryx seems to have six sacrals, suggesting its coding be changed to polymorphic. This is also present in Mahakala.

The second is a problematic character coding for acromion shape. It has three derived states (rectangular, quarter circle and triangular) and a basal state oddly defined as "does not match any of the following descriptions." The latter is inappropriate, as all states should be definable morphologies. A triangular acromion "with apex pointing away from and subparallel to scapular blade" is supposedly diagnostic of unenlagiines in ACCTRAN, based on its presence in Rahonavis. This is also correctly coded as present in Archaeopteryx and ornithurines (sensu Gauthier), so would support Rahonavis being an avialan (sensu Gauthier) instead of a dromaeosaurid. In any case, I feel this state is due to a topology error, as intermediate taxa like Sinornithosaurus show the acromion process apex to be homologous to the anterodorsal corner which articulates with the coracoid in more basal taxa. What occurs is that the scapular blade angles to point more posteriorly than dorsally, causing the apex to be subparallel to the blade, while the proximal coracoid reduces in width, freeing the acromion apex from contact with it. Unenlagia may have a similar morphology to Rahonavis.

The third is a medially or posteriorly directed hallux, which is coded as present in Rahonavis. It is however absent in that taxon (Middleton, 2003). Scansoriopterygids also seem to lack a twist in their metatarsal I, contra Senter's coding.

One character supported Unenlagiinae in DELTRAN, a preacetabular process markedly longer than the postacetabular process. This was correctly coded as present in Unenlagia, Rahonavis, scansoriopterygids, archaeopterygids and ornithurines (sensu Gauthier). Thus is would also support unenlagiines being avialans. Senter miscoded Deinonychus, Beipiaosaurus, Alxasaurus and Segnosaurus as lacking this state, and Nanshiungosaurus as being unknown.

Turner et al. supported Unenlagiinae based on short anterior dorsal transverse processes, which they failed to code as present in Rahonavis, Archaeopteryx, Sapeornis and Confuciusornis. The presence in basal birds (as correctly coded by Senter) makes this more likely a paravian synapomorphy lost in Microraptor and derived dromaeosaurids.

They also support it based on a reduced but not absent supracetabular crest, but miscode Archaeopteryx as having an absent crest, when it fact it has the condition present in unenlagiines. Senter miscodes Unenlagia and Rahonavis as lacking the crest, in addition to miscoding Archaeopteryx this way.

Similarly, Turner et al. support Unenlagiinae based on their vertical pubis, while Archaeopteryx has a nearly identical angle but is coded as opisthopubic. Senter corrects the coding for Archaeopteryx, but codes Sinovenator, Troodon, Achillobator, Velociraptor, Adasaurus and Microraptor as being less opisthopubic than they are. He also neglects to code Buitreraptor, which is mesopubic.

Unenlagia and Buitreraptor were further united to the exclusion of Rahonavis in Senter's tree based on one character, the supratrochanteric process. Yet this is present in basically all paravians except seemingly Sapeornis, as well as Erliansaurus, Neimongosaurus, Chirostenotes and Citipati. Senter only codes it present in Unenlagia, Buitreraptor, Confuciusornis and Mei.

Unenlagia and Rahonavis were united to the exclusion of Buitreraptor in Turner et al.'s tree based on five characters.

The first is all presacral vertebrae pleurocoelous. Velociraptor was miscoded as having this condition, and new evidence suggests Utahraptor lacks it (coded as unknown).

The second is six sacral vertebrae, which Turner et al. correctly coded as present in Unenlagia (unlike Senter), but still coded Deinonychus as unknown when in fact it has six as well. They also coded Adasaurus as unknown (it has five), Archaeopteryx as having five (it can have six too), Jeholornis as having five (it has six), Sapeornis as having eight (it has seven), and Apsaravis as having seven (it has ten). The miscoding of almost all birds for such a simple character is disturbing.

The third is an elongate preacetabular process, which Turner et al. miscode Deinonychus, Alxasaurus, Segnosaurus and Apsaravis as lacking.

The fourth is a posteriorly extensive m. cuppedicus fossa. Turner et al. and Senter each treat this character differently. Turner et al. code it as present only in Velociraptor, Unenlagia and Rahonavis. Senter codes it as present in all oviraptorosaurs and paravians. The reality is in between these extremes. In addition to those three taxa, the condition is also present in Caudipteryx, Microvenator, Sinovenator, Achillobator, Saurornitholestes, Bambiraptor and Scansoriopteryx, as in Senter's matrix, contra Turner et al.'s. However, it is absent in IGM 100/42, Khaan, Ingenia, Adasaurus, Microraptor and Archaeopteryx, as in Turner et al.'s matrix, contra Senter's. Deinonychus has been illustrated with both conditions, so agrees with both matrices.

The final character is an enlarged proximodorsal ischial process. This was only coded in taxa which had such processes, so Microraptor (which lacks one) should be coded unknown instead of plesiomorphic. Utahraptor and Achillobator should have been coded plesiomorphic instead of unknown, as they have small processes. Unenlagia has a much smaller process than Rahonavis, Sapeornis or Confuciusornis, so should be coded plesiomorphic in my opinion. Archaeopteryx can have a large process (Munich specimen), or a small one (Eichstatt specimen) so should be polymorphic instead of plesiomorphic. Jeholornis should be coded as apomorphic instead of unknown.

After making the corrections noted above, a new tree was found. In the revised tree, Unenlagia clades with Rahonavis based on having all presacral vertebrae pleurocoelous, an elongate preacetabular process, and a small but present supracetabular crest. Additionally, they clade with Mahakala based on having six sacral vertebrae.

Buitreraptor is closer to derived dromaeosaurids than unenlagiines based on a sinusoidal frontal margin of the supratemporal fossa (absent in Mahakala), a large paraquadrate foramen with associated tab on quadrate (unknown in unenlagiines), teeth unconstricted basally (unknown in unenlagiines), an elongate radius (absent in Mahakala; paralleled in Rahonavis), a longitudinal ischial ridge (absent in Rahonavis and Unenlagia), and a proximally placed hallux (unknown in unenlagiines).

The tree below follows the format of that on my website, where taxa which are incertae sedis within a clade are shown connected to that node by an asterisk. Variraptor (based only on the holotype dorsal and sacrum) could be a dromaeosaurid or an oviraptorid, so is not shown here.

|  `--+--+--Pedopenna
|     |  `--Yandangornis
|     `--+--Patagonykus
|        `--+--Mononykus
|           `--Shuvuuia
  |  |--Epidendrosaurus
  |  `--+--+--Archaeopteryx
  |     |  `--Wellnhoferia
  |     `--+--Jeholornis
  |        `--+--Sapeornis
  |           `--+--Confuciusornis
  |              `--+--Protopteryx
  |                 `--Yanornis
  `--+*-IGM 100/1005
     |  `--+--Neuquenraptor
     |     `--+--Mei
     |        `--+*-Urbacodon
     |           |*-IGM 100/44
     |           |--Byronosaurus
     |           |--Sinusonasus
     |           `--+--Sinornithoides
     |              `--+--Troodon
     |                 `--+--Saurornithoides mongoliensis
     |                    `--Saurornithoides junior
        |  `--+--Unenlagia
        |     `--Rahonavis
                 |  `--+--Graciliraptor
                 |     `--+--Sinornithosaurus
                 |        `--+--NGMC 91
                 |           `--Microraptor

Mickey Mortimer