I'll break into non-theropods for a while, as I've been concentrating on them lately. I must warn you all that my ornithischian knowledge and references are much less extensive, so I rely more on the original description and less on my own interpretations. Thanks to Jaime Headden for his suggestions.
Liaoningosaurus Xu, Wang and You 2001
L. paradoxus Xu, Wang and You 2001
Etymology- "paradoxical lizard from Liaoning"
Barremian, Early Cretaceous
Yixian Formation, Liaoning, China
Holotype- (IVPP V12560) (330 mm, juvenile) premaxilla, maxilla, mandibles, quadrate, pterygoid?, basioccipital?, basisphenoid?, paroccipital process?, teeth, several cervical vertebrae, three dorsal vertebrae, twenty-three dorsal ribs (to 54 mm), four sacral vertebrae (6 mm), two sacral ribs, about twenty caudal vertebrae, chevrons, scapulae (32 mm), coracoids, sternal plate, humeri (27 mm), radii (19 mm), ulnae (21 mm), metacarpus (5 mm), manual phalanges, manual unguals, ilium (51 mm), proximal pubis, ischia (30 mm), femora (28 mm), tibiae (28 mm), fibulae (25 mm), metatarsus (13 mm), pedal phalanges, pedal unguals, cervical scutes?, five cervicopectoral spines (5-8 mm), three cervicopectoral plates (7-11 mm), posteroventral plate (67 mm)
Diagnosis- sternum trapezoidal with slender, curved and tapered posterolateral process; enlarged armor plate covering pelvic area ventrally.
Description- The holotype is a juvenile as shown by the unfused cranial and neurocentral sutures. It is 330 mm as preserved, although few vertebrae are present, so this is probably not completely accurate.
Several cranial elements are preserved, but the small figure, lack of labels, crushing in ventral view and my poor knowledge of ankylosaur anatomy make them difficult to identify. The skull is longer than wide. It has a toothed premaxilla, and a maxilla with an antorbital fossa and ten teeth. Portions of the posterior skull are present, probably including pterygoid, quadrate, basisphenoid, basioccipital and paroccipital process. The premaxillary teeth lack cingula and have five cusps. The maxillary and dentary teeth have cingula and low rounded crowns with eight cusps, with fluting extending to the cingula. The mandible probably possesses an external mandibular fenestra.
Several cervical vertebrae seem to be preserved, but they are mixed with the pectoral armor and details can not be seen. Similarily, three dorsals and four sacrals are preserved. Eleven almost complete dorsal ribs are present on the left side, while twelve ribs (some partial) are seen on the right side. At least two long sacral ribs contact the preacetabular process. About twenty caudal vertebrae are preserved including fragments of the proximal section of the tail and articulated sequences from the mid and distal sections. The proximal caudals have transverse processes over twice as long as their neural spines and centra equal in width and length. The latter character is apparently otherwise only known in Gargoyleosaurus. Distal caudals have elongate zygopophyses and chevrons that contact each other. However, there are no ossified hypaxial tendons and no distal club.
The scapulae are bowed anteriorly are exapanded slightly distally. There is a pseudoacromial process, but as the scapulae are visible in medial view, its shape and orientation cannot be determined. Xu et al. suggest that it may be obliquely directed, but Shamosaurus' non-obliquely directed process looks identical in medial view. The acromion itself is poorly developed and the scapular blade narrow, reminiscent of Texasetes. The coracoid is a tall parallelagram and the coracoid glenoid is about half the size of the scapular glenoid. The sternal plates differ from Silvisaurus, Sauropelta and Euoplocephalus at least in having a short tapered posterolateral process. They are more like nodosaurids than ankylosaurids in not having greatly expanded trangular medial ends.
The humerus is plesiomorphically slender with a short deltopectoral crest. The radius and ulna are more slender than most ankylosaurs, the latter having a small olecranon process. The manus has a phalangeal formula of 2-3-3-2-0, metacarpals I and V are subequal to III in width. The unguals are claw-like, being longer than wide and laterally grooved.
The ilium is preserved in ventral view and has an extremely elongate preacetabular process oriented horizontally, like ankylosaurs. The preacetabular process is not as laterally deflected as stegosaurs, nodosaurids or ankylosaurids, being closer to Mymoorapelta or Minmi (~25 degrees). The acetabulum is closed, and the postacetabular process is greatly reduced and faces laterally. The latter is a plesiomorphy retained in Scelidosaurus and Mymoorapelta among ankylosaurs. The pubis is larger than most ankylosaurs, being comparable to Scelidosaurus and Minmi, though probably somewhat larger than the latter. The postpubic process extends at least halfway down the ischial shaft. The ischium is typically ankylosaurian- slightly bowed, tapering distally and having a convex acetabular edge.
The femur is fairly robust and has a crest-like (non-pendent) fourth trochantor located proximally. Proximally, the anterior and greater trochantors are unfused, as in Minmi, Cryptosaurus and Hoplitosaurus. The tibia is notable for having subequal proximal and distal expansions (anterior view) and being subequal to the femur in length. The fibula has a more gradual proximal expansion and greater distal expansion than Sauropelta. Both astragalus and calcaneum are preserved, the latter is said to be larger than most ornithischians, though it seems smaller than Scelidosaurus. The metatarsi are more elongate than ankylosaurs other than Scelidosaurus, being almost 50% of tibial length. This contrasts with other ankylosaurs, which have metatarsotibial ratios of 20% or so. The pedal phalangeal formula is 0-3-4-5-0. Metatarsals I and V are reduced and splint-like and the unguals are elongate and claw-like. Metatarsals II and IV are shorter than III and the proximal phalanges are longest in digits II and III.
Armor is amazingly sparse. No osteoderms are reported to be present on the skull or dentary. Armor is only preserved in the cervical and pectoral area. At least three pairs of flattened, slightly recurved spines are present in the cervical area. Also, two pairs of oval scutes were present, the more posterior larger and positioned over the pectoral girdle. Small ossicles and scutes may be present in the cervical area as well. Strangely, the back, hips and tail are devoid of armor. Half of a large plate is preserved ventral to the pelvis. It is thickest medially and covered with hexagonal and rhombic tubercles about .5 mm wide. In life, it would have started midway down the dorsal column, ended at the posterior ilial edge and extended transversely across the whole belly. Impressions indicate another plate was present anteriorly that extended to the pectoral girdle.
We must keep in mind that the holotype is a juvenile, so ontogeny may have affected various characters used in ankylosaurian phylogenetic analyses. Comparison to juvenile Pinacosaurus suggests that the elongate tibia and metatarsus are not juvenile characters in ankylosaurs.
The authors conducted a phylogenetic analysis taken directly from Sereno's (1999) thyreophoran analysis excluding the stegosaur characters. They added Liaoningosaurus and condensed the stegosaurs into a single OTU. The final analysis includes 14 taxa and 95 characters. The result is 24 MPT's of 123 steps with a large polytomy within Ankylosauria due to Liaoningosaurus jumping around from nodosaurid to derived ankylosaurid.
Examination shows many problems with the matrix however. For instance, stegosaurs are coded as having a sinuous dentary tooth row, when it is actually curved in one direction. They are also coded as having only five premaxillary teeth, but Huayangosaurus has seven. The character "scapular blade parallel-sided" really needs work, as it is coded as present in all eurypods, but Stegosaurus, Liaoningosaurus and especially Polacanthoides have distally expanded blades. The character "preacetabular blade laterally deflected 45 degrees" is incorrectly coded as present in Liaoningosaurus and Minmi. Scelidosaurus has an unexpanded distal ischium similar to Liaoningosaurus and Minmi, so is coded incorrectly. I don't think that stegosaurs or ankylosaurs have especially spreading metatarsi compared to Liaoningosaurus or Scelidosaurus. Stegosaurs have an occiput that's wider than tall, cingula on their cheek teeth and oval premaxillary palates, so are coded incorrectly. The character "Snout low: maximum preorbital depth less (0), or more (1), than twice maximum preorbital length." makes no sense, as no thyreophore has that short of snout. The dentary has a sinuous ventral edge in stegosaurs, just as in ankylosaurs (and Scutellosaurus slightly), so this character is coded incorrectly. Stegosaurs have a proximal scapular blade much less than 75% the acromial width, contra Xu et al.. I don't see any lateral component to stegosaur glenoids, so I don't understand why they're coded as having glenoids oriented "posteroventrally and laterally". Liaoningosaurus is coded as lacking a fused skull, but it's a juvenile, so this is expected in ankylosaurs. It should be coded as unknown. Pinacosaurus is coded as having state 2 for "accessory dermal ossifications on posterior skull roof" when only states 0 and 1 are defined. It should be 1 of course, but I don't know how the other ankylosaurs were coded, as they are only known from adult material lacking the sutures to show the presence of "tabulars", "postfrontals" and such. Two characters are repeats of each other. "Distal chevrons T-shaped and contact each other" is just like "distal chevrons contact each other". What's even odder is that Mimmi, Hylaeosaurus and "other nodosaurines" are coded as having the first, but lacking the second. This is impossible and makes me trust the matrix even less than before. Indeed, I found all the above errors while trying to validate the codings of stegosaurs, which are very well studied and illustrated. I'm sure going through the codings of the other taxa would show similar problems. I lack the references and expertise to check many codings, so even if I went through the whole thing as best I could, I would not trust it to be accurate. Because of this, I have not attempted to correct and rerun the matrix. I recommend the papers phylogenetic results not be trusted and must look to other sources to evaluate Liaoningosaurus' relationships.
Other studies of ankylosaur phylogeny have focused on the skull, which is largely undescribed and poorly illustrated in Liaoningosaurus. However, Kirland (1998) published an analysis with 17 taxa and 46 characters, including fifteen from the postcranium. This matrix looks more accurate than Xu et al.'s (though I see a few problems, ex. Gastonia is the only taxon coded as having an obliquely directed pseudacromial process) and adding Liaoningosaurus results in 3 MPT's of 80 steps. These trees have the following topology-
I like this phylogeny more than Sereno's because it matches the most extensive phylogenetic analysis yet (Carpenter, in press) in many details and does not assume monophyly of certain clades ("other nodosaurines", "other ankylosaurines") a priori. In addition, I see many characters suggesting Liaoningosaurus is more basal than nodosaurids, polacanthids and akylosaurids. These include- antorbital fossa present; external mandibular fenestra probably present; claw-like manual and pedal unguals; preacetabular process laterally deflected only 25 degrees; postacetabular process faces laterally; pubis not extremely reduced; unfused anterior and greater trochantors; tibia subequal to femur in length; tibial distal width subequal to proximal width in anterior view; metatarsi almost 50% of tibial length. Minmi also has claw-like manual unguals, a slightly deflected preacetabular process, slightly reduced pubis and unfused femoral trochantors. It is more derived in lacking an antorbital fossa and having a ventrally facing postacetabular process. The presence of ventral armor is the only character I can find that would unite the two, so placing Minmi as more derived than Liaoningosaurus is most parsimonious. Ankylosaurid-like characters include- extremely elongate distal caudal prezygopophyses; indistinct femoral head. Others suggested by Xu et al. are plesiomorphic (non-enlarged radial condyle of humerus), also seen in Minmi (ischial acetabular margin convex) or erronous (chevrons interlock; straight medial sternal margin; deltopectoral crest and transverse axis through distal condyles in same plane). The nodosaurid-like characters cited are plesiomorphic (maxillary and dentary teeth with cingulum; slender humerus; deltopectoral crest extends less than of half shaft length; proximally located fourth trochantor), unknown in polacanthids and more basal taxa (sub-trapezoidal sternum), also present in polacanthids and ankylosaurids (preacetabular process faces ventrally) or undefendable (obliquely directed pseudacromial process). The two ankylosaurid-like characters are best seen as convergences when the many primitive characters are taken into account. Still, despite its basal placement, Liaoningosaurus is more derived than Scelidosaurus based on- strong maxillary/dentary cingula; elongate proximal caudal transverse processes; elongate distal caudal zygopophyses; distal chevrons contact each other; pseudacromial process on scapula; horizontallly oriented preacetabular process; non-pendent fourth trochantor. Only the probable presence of an external mandibular fenestra would suggest otherwise. In conclusion, I suggest placing Liaoningosaurus as a basal ankylosaur, more derived than Scelidosaurus, but probably less than Minmi.
Anyone who wants the pdf, contact me offlist. Before I venture back into the Theropoda, I'll take a swing at Jeholosaurus.