Sinovenator changii (Xu et al. 2002, Fig. 1) was originally correctly identified as a troodontid. That’s the good news.
Turner, Makovicky and Norell (2012) nested Sinovenator with Mei (a scansoriopterygid bird in the large reptile tree of 654 taxa) and Xixiasaurus (a long-rostrum troodontid), both far from Avialae (birds). By contrast, the large reptile tree (Fig. 2) nests Sinovenator in the troodontid grade (not a distinct clade) close to two other short-faced troodontids, Xiaotingia and Eosinopteryx, which are proximal to Archaeopteryx and birds (Fig. 2).
Turner et al. 2012 reported,
“The discovery of Sinovenator and the examination of character distributions along the maniraptoran lineage indicate that principal structural modifications toward avians were acquired in the early stages of maniraptoran evolution.”
in Figure 2, Sinovenator represents not an early stage of maniraptoran evolution, but a later stage, which, in turn, is an early stage in avian evolution. That was a taxonomic flip-flop. See below for details.*
Turner et al. 2012 concluded: “The morphological gap between the par avian clades has blurred to the point that basal dromaeosaurids, troodontids, and avialans are nearly indistinguishable from one another, and in life these animals would appear extremely similar. However, important morphological divisions exist that allow us to understand the basic relationship between these three clades.
“Taxon sampling within Paraves is the most exhaustive to date, but the phylogenetic hypotheses discussed herein will certainly not be the last word on par avian or coelurosaurian relationships. Indeed, changes and additions to similar data matrices are already yielding interesting results within the various clades of Coelurosauria. Moreover, the potential for new discovery that will modify these results is a given.”
Therein we may be seeing part of the problem with the Turner et al. 2012.
The large reptile tree was built taxon by taxon over several years and theropods were only seriously considered in the last year. I started with a few theropod taxa that were essentially complete and later added taxa that were also complete. In this way the correct tree was built from the beginning with incrementally increasing resolution. Afterwards less complete taxa were added or avoided. As everyone knows, incomplete sisters lead to loss of resolution, which is a party killer.
If Turner et al. had started with a dozen or two well-known taxa,
then added taxa incrementally, perhaps there would have been more lumping and splitting from the get go*.
every new taxon was scored without bias. I did not know where the lesser known taxa should nest. As theropod experts, Turner et al. sort of knew how their tree would be built.
Moreover part 2
Turner et al 2012 did not test individual Archaeopteryx specimens as unique taxa. Thus they were not able to identify Mei as a scansoriopterygid bird. They also did not test the basal paravaian, Tanycolagreus, but they did test the basal paravian (not a tyrannosauroid), Eotyrannus. They did not test Sinocalliopteryx, Suchomimus, Guanlong or Xiongguanlong, which are not paravians, but would have attracted Dilong and Proceratosaurus, causing their deletion from the inclusion set. Turner et al. erroneously nested Juravenator and Sinosauropteryx with Compsognathus. They nest elsewhere in the large reptile tree (Fig. 2).
In phylogenetic analysis it is easy to flip flop taxonomic lineages. As an example, in the large reptile tree the proximal basal taxa to the Solnhofen birds (Archaeopteryx, etc.) are the troodontids, beginning with the primitive dromaeosaurid-like Sinornithoides and ending with the derived bird-like Anchiornis, Sinovenator, Xiaotingia and Eosinopteryx. The Turner et al. study flipped this order by nesting the bird-like Xiaotingia and Anchiornis as basal troodontids while nesting the dromaeosaurid-like Sauronithoides as the most derived troodontid taxon.
Turner et al. did not include Aurornis, which was published a year later. So, that’s not their fault. The long-snouted troodontid Aurornis would have attracted even longer snouted Buitreraptor away from the dromaeosaurids (Fig. 2). (See above for the Turner et al. prediction of this sort of thing happening after their publication.)
(and I’m just now learning this) there were basal forms with medium length skulls like Sinoornithoides, and (so far) two derived clades. One tended toward ever longer skulls, like Buitreraptor. The other tended toward shorter, more bird-like skulls, like Sinovenator. The Turner et al. study flip-flopped and otherwise confused this.
Taxon inclusion does solve problems. But I can see where too many incomplete taxa can cause problems. Flip-flopping taxonomic lineages is a problem one has to avoid. It is also important to reconstruct all taxa, which is something Turner et al. did not do, or at least did not publish.
*The Turner et al study reminds me of the mechanical pterosaur ornithopter video in which the experts added fur, teeth and eyeballs to their flying model before affirming that the wings worked! They didn’t work, unfortunately for everyone in the video and watching the video.
Turner AH, Makovicky PJ Norell MA 2012. A Review of Dromaeosaurid Systematics and Paravian Phylogeny. Bulletin of the American Museum of Natural History 371:1–206. doi:10.1206/748.1. Online here.
Xu X, Norell MA, Xiao-lin W, Makovicky PJ, Xiao-chun W 2002. A basal troodontid from the Early Cretaceous of China. Nature 415: 780–784. doi:10.1038/415780a. PMID 11845206.