New ichthyosaur family tree by Ji et al. 2015

A recent paper on ichthyosaur systematics
(Ji et al. 2015, Fig. 1) adds newly discovered taxa and the tree is getting nice and big.

Unfortunately,
at the base of their cladogram Ji et al. place a distinctly different proximal outgroup for ichthyosaurs than what was recovered in the large reptile tree (subset shown in Fig. 1, click to enlarge). They appear to be guessing. Apparently they are not sure how ichthyosaurs are related to other reptiles.

Here 
proximal outgroup taxa for ichthyosaurs include Wumengosaurus, Thaisaurus and Xinminosaurus (in ascending order) not Thadeosaurus. These large reptile tree taxa demonstrate a gradual accumulation of basal ichthyosaur traits. The Ji et al taxa, Hovasaurus, Claudiosaurus and Thadeosaurus do not. In the large reptile tree these three are basal younginiformes, related, yes, but much more distantly related to ichthyosaurs.

So, as an experiment, 
we’ll delete the large reptile tree proximal outgroup taxa in order to match more closely the Ji et al taxon list. What is recovered now?

1. Hovasaurus, Claudiosaurus and Thadeosaurus now nest together in an outgroup clade.
2. Xinminosaurus, Grippia and (Shastasaurus alexandrae + Utatsusaurus + (Shastasaurus  pacificus + Hupehsuchus) now form an unresolved clade at the base of the Ichythyosauria.
3. Then Chaohusaurus nests at the base of the rest of the Ichthyosauria with the same topology as the subset of the large reptile tree.

A few differences between the two topologies without deletions…
Note the morphological mismatches in the Ji et al. topology not found in the large reptile tree.

1. In the large reptile tree Chaohusaurus nests between two similar taxa, Parvinatator and Besanosaurus. In the Ji et al. tree Chaohusaurus nests between the mismatched and odd Hupehsuchus and a clade of basal ichthyosaurs as the basalmost ichthyosaur, even though it has a derived ichthyosaur shape and traits.
2. In the large reptile tree the derived, but still Triassic, Cymbospondylus petrinus nests between its contemporary, Mixosaurus and several other giant serpentine ichthyosaurs. All have a depressed cranium with a central ridge. The unrelated flat-headed C. buchseri nests elsewhere with similar deep-bodied, high-crested Shonisaurus popularis. By contrast, in the Ji et al. tree C. piscosus (= petrinus) and C. buchseri nest together with the very primitive, very small, Xinminosaurus, which does not have such a depressed cranium with a central crest.
3. Ji et al. have a clade of Shastasauridae that includes only shastasaurs. In the large reptile tree, that clade also includes the odd little hupehsuchids and demonstrates how these little toothless enigmas evolved from larger forbearers. Ji et al. provided several skull reconstructions. Perhaps a few more would help to resolve the distinct topologies.

Those are the major issues.
The rest can be swept up later. I’d like to see the authors either expand their own taxon list or work off the large reptile tree to confidently establish a series of outgroup taxa for the Ichthyosauria that actually demonstrate a gradual accumulation of character traits, instead of doing what they did. Then we might have closer correspondence in tree topology. And we’re going to have to figure out Cymbospondylus… is it derived? or primitive?

References
Ji C, Jiang D-Y,  Motani R, Rieppel O, Hao -C & Sun Z-Y 2015. Phylogeny of the Ichthyopterygia incorporating recent discoveries from South China, Journal of Vertebrate Paleontology, DOI: 10.1080/02724634.2015.1025956