The impressive scale and thoroughness of Tschopp et al.’s study has understandably meant that any study challenging their results has a high bar to clear in terms of effort and rigor, so it’s not surprising nobody has done so yet. However, I think that a closer examination of their results reveals that the case for considering
Brontosaurus a valid genus is not nearly as well-supported as it seems at first glance.
Most significantly, Tschopp et al. do not consistently recover either
Apatosaurus or
Brontosaurus (as they define them) as monophyletic (even ignoring the position of the problematic
Amphicoelias altus, which renders their
Brontosaurus non-monophyletic in both analyses). Their equal weights topology recovers a clade of
A. ajax and
A. louisae as sister to a clade uniting
B. excelsus,
B. parvus, and
B. yahnahpin, but their implied weights topology found
A. louisae to be outside a clade uniting the other four species, with
A. ajax as sister to
B. parvus. Not only do they fail to consistently recover a topology consistent with their proposed taxonomy, the topology in which both genera are polyphyletic is arguably ever so slightly the more credible of the two topologies: it is more consistent with the only previous phylogenetic analysis of relationships within
Apatosaurus sensu lato (Upchurch et al. 2004), it aligns with the general inclination of 20th-century studies to consider
A. excelsus and
A. louisae as clearly distinct but
A. ajax and
A. excelsus difficult to distinguish (e.g. Riggs 1903, Gilmore 1936, McIntosh 1990), and Tschopp et al. themselves suggest in their methods section that implied weighting may provide more accurate results for immature specimens (and thus the position recovered for the immature holotype of
A. ajax may be more accurate under IW than EW). So while
Apatosaurus sensu lato is unquestionably monophyletic,
Apatosaurus sensu Tschopp et al. and
Brontosaurus sensu Tschopp et al. are both potentially polyphyletic—and I think that alone is enough to seriously call into question their proposed taxonomy.
The basis for Tschopp et al.’s proposal to consider
Apatosaurus and
Brontosaurus as separate genera is their quantitative approach in which they define genera by counting the number of character changes separating clades. They decided on the threshold of 13 steps based on the widely-accepted congeneric species pairs of
A. ajax and
A. louisae (separated by 12 steps) and
Diplodocus carnegii and
D. hallorum (separated by 11 steps; note that
Diplodocus hallorum sensu Tschopp et al. is largely equivalent to most 20th-century usage of
Diplodocus longus). In their implied weighting analysis (which, you will remember, recovered both genera as polyphyletic)
A. louisae was separated from the
excelsus+parvus+yahnahpin+ajax clade by 14 steps, and
A. ajax was separated from
A. parvus by 14 steps, so the threshold for genus separation was just barely cleared (albeit without supporting
A. ajax and
A. louisae as congeneric). In their equal weighting results, which recovered a topology more consistent with their proposed taxonomy,
Apatosaurus and
Brontosaurus were separated by only 11 steps, below their threshold for recognizing separate genera and less than the distance between
A. ajax and
A. louisae. It should also be asked how appropriate the 13-step threshold is in the first place. The diplodocine genera
Barosaurus,
Diplodocus,
Galeamopus, and
Kaatedocus are separated from one another by 20–21 steps (the more fragmentary conditions of
Tornieria and
Leinkupal make it harder to evaluate the distances between more basal diplodocine genera). So perhaps we could conclude that 11–12 steps are typical for species within a genus and 20–21 steps are typical for closely related genera, so Tschopp et al. could have reasonably set their threshold for recognizing separate genera anywhere between 13 and 19 steps. Therefore, the 13-step threshold they selected is actually setting the bar for recognizing
Brontosaurus as a distinct genus almost as low as possible, and
Brontosaurus cannot even consistently reach that threshold. So even if you buy into their approach to quantitative taxonomy (and many don’t, including one of the reviewers of the paper), their approach would seem to provide only equivocal support, at best, for separating the two genera, and it is significant that it is only by choosing apomorphy counts from one analysis and the phylogenetic topology from the other are they able to support
Brontosaurus as a valid genus.
Additionally, their principal coordinate analysis and pairwise dissimilarity metrics show that
A. ajax and
A. excelsus are the two most similar apatosaurine species to one another. In fact, the pairwise dissimilarity between the two species (0.18) is lower than the pairwise dissimilarity within
A. parvus (0.23). Given the phylogenetic instability of
A. ajax, I don’t think we can rule out the possibility that it is sister to
A. excelsus, or possibly even synonymous with it.
Moreover, the proposed diagnoses of
Apatosaurus and
Brontosaurus are inadequate to reliably distinguish the two genera. The horizontal lamina in the SDF reported as an autapomorphy of
Apatosaurus appears to be more or less identical to the horizontal lamina in the SDF reported as an autapomorphy of
B. parvus, and the SDF is not adequately preserved in
A. excelsus to evaluate this character, so it seems more likely that this lamina is shared among most or all apatosaurines (I certainly have yet to see an adequately-preserved apatosaurine neck that lacks it). Although Tschopp et al. diagnose
Apatosaurus as having a straight scapular blade, the scapular blade of the holotype of
A. ajax appears to be slightly curved ventrally, and damage to the distal end of the blade suggests the scapula could have shown stronger curvature when complete. The bases of the posterior dorsal neural spines are not adequately preserved in
A. excelsus to determine whether they exhibit the condition reported as diagnostic of
Brontosaurus. The rounded expansion of the scapular blade reported as diagnostic of
Brontosaurus appears to be ontogenetically variable (it is absent in the holotype of
A. parvus), so its absence in the immature holotype of
A. ajax is not necessarily taxonomically informative. I am not convinced that the difference in the proportions of the astragalus is taxonomically informative, either. Most genera with multiple specimens listed in their supplementary data exhibit specimens on both sides of the state boundary, and the differences between
A. excelsus and
A. louisae are comparable to the differences between left and right sides of the same individual in CM 94. Furthermore, the astragalus is unknown in
A. ajax, and the optimization of this character is unclear because it is polymorphic in both Dicraeosauridae and Diplodocinae. I have so far not been able to determine what difference is meant by the presence or absence of a roughened lateral aspect of the cervical PRDL; I don’t know whether this means I’ve overlooked something or that the character is not well-defined. That leaves a single character—the presence or absence of a shallow fossa on the scapula posterior to the acromial ridge—that I can’t dispute, but I wouldn’t stake a genus on a single subtle character like that.
So to sum up, Apatosaurus and Brontosaurus, as defined by Tschopp et al., are difficult to distinguish morphologically and may not be reciprocally monophyletic. The metrics Tschopp et al. used to justify separating the two genera only provide equivocal support, at best. I do not see the value in replacing one easily recognizeable monophyletic genus with two difficult-to-distinguish, potentially polyphyletic genera. Of all the important work that Tschopp et al. did, I think it’s unfortunate that this is the conclusion that their paper is best known for, as it is, in my opinion, one of the weakest parts of the paper.