Cartorhynchus: rebuilding the small ichthyosaur mimic

Mistakes were made here
earlier while reconstructing Cartorhynchus, the basal sauropterygian ichthyosaur-mimic. Those mistakes are corrected here (Figs. 1, 2) and already updated in earlier posts. All of these repairs further cement the relationship of Cartorhynchus to its sister, Sclerocormus  (Fig. 3) and its ancestral sister, Qianxisaurus (Fig. 4), taxa nesting near the base of the Eosauropterygia, not the Ichthyopterygia in the large reptile tree (LRT, 1401 taxa).

Cartorhynchus lenticarpus (Motani et al. 2014; Early Triassic) was originally considered a strange basal ichthyosauriform and a suction feeder. Here it nests with Sclerocormus and Qianxisaurus as a basal eosauropterygian representing a new clade of ichthyosaur-mimics with a very early appearance of flipper-like limbs. Neotony played a part in the appearance of a short rostrum, large eyes, short neck, poorly ossified phalanges and small size. The supratemporal was large here, and the splenial can be seen in lateral view, though just barely. These are also results of neotony as most sauropterygians lack them. The outgroup taxon, Pachypleurosaurus, fuses the large supratemporal and squamosal

Ichthyosaurs have the following traits by convergence.
Ichthyosaurs have robust scleral rings (eyeball bones) while most eosauropterygians do not. Distinct from most eosauropterygians and like ichthyosaurs, Qianxisaurus has small supratemporals and gracile scleral rings. The splenials are not visible in the present exposure. Like Cartorhynchus, the digits of Qianxisaurus are not well developed. 

The premaxilla
of Cartorhynchus was tiny, ideal for nipping small food items. Small teeth were present, contra the original interpretation. The naris and orbit were quite large, distinct from all candidate sister taxa. The in situ maxilla was taphonomically flipped, so Cartorhynchus actually had a ventrally convex maxilla. The heavy ribs and flat bottom make Cartorhynchus look like a bottom feeder. The elevated dorsal neural spines suggest a dorsal fin (Fig. 2). The small scapula and coracoid suggest a weak, passive pectoral flipper. A long tail was probably present, as in its larger sister, Sclerocormus (Fig. 3). This would have been the primary propulsive organ. Note the lack of large flippers in Sclerocormus. Qianxisaurus is the last known common ancestor.

Qianxisaurus chajiangensis (Cheng et al. 2012; Fig. 4) is a Middle Triassic basal eosauropterygian based on a virtually complete articulated skeleton with all digits poorly ossified. Cheng et al. 2012 nested Qianxisaurus as derived from a sister to Wumengosaurus, a taxon that nested closer to thalattosaurs, ichthyosaurs and mesosaurs in the LRT. The Cheng et al. (2012) study had some odd nestings including Kuehneosauridae (rib gliders) a little too close to turtles and thalattosaurs. The LRT widely separated these taxa, as befitting their utterly distinct morphologies.

The LRT
nests Qianxisaurus between Pachypleurosaurus and Lariosaurus. Pistosaurus nests as an outgroup in the Cheng et al. (2012) tree, but closer to Simosaurus in the LRT.

The upper temporal fenestrae of Qianxisaurus
were smaller than in Pachypleurosaurus. The supratemporal formed the posterior rim without fusion to the squamosal. In Qianxisaurus the retroarticular process of the mandible was smaller than in related taxa. The teeth of Qixiansaurus were unusual with a slightly constricted cylinder and short conical crown.

Despite the small size of the ilium
at least four sacrals were present.

The Cartorhynchus-like pectoral girdle of Qianxisaurus.
The scapula (green Fig. 5) had a slim strap-like morphology. The clavicles were broader laterally, meeting medially in an arch shape, as in Cartorhynchus (Fig. 1).

Mimic taxa
appear occasionally in the LRT, which, so far, has been able to lump and split mimics by testing them against all available candidate sisters. Motari et al. 2014, for all his experience and expertise in ichthyosaurs, failed to add basal eosauropterygians, like Qianxisaurus, to their taxon lists and so was not able to consider this possibility. Better not to assume things, but to let the software perform an unbiased analysis starting with a wide gamut of taxa like the LRT.

Correcting mistakes
is part of the scientific process, whether they be internal or external. 

References
Cheng YN, Wu XC, Sato T and Shan HY 2012. A new eosauropterygian (Diapsida, Sauropterygia) from the Triassic of China. Journal of Vertebrate Paleontology. 32 (6): 1335. doi:10.1080/02724634.2012.695983
Jiang D-Y, Motani R, Huang J-D, Tintori A, Hu Y-C, Rieppel O, Fraser NC, Ji C, Kelley NP, Fu W-L and Zhang R 2016. A large aberrant stem ichthyosauriform indicating early rise and demise of ichthyosauromorphs in the wake of the end-Permian extinction. Nature Scientific Reports online here.
Motani R et al. 2014. A basal ichthyosauriform with a short snout from the Lower Triassic of China. Nature doi:10.1038/nature13866

wiki/Qianxisaurus
wiki/Cartorhynchus
wiki/Sclerocormus