Friday, September 27, 2013

Q&A III: Nanotyrannus – my turn in the hot seat.

A cast of the Cleveland Skull next to a cast of an adult Tyrannosaurus rex skull in the striking dinosaur gallery of the Royal Ontario Museum.

What follows are my answers to Mickey Mortimer’s extensive questions regarding my post “Nanotyrannus isn’t real, really”. I think they are excellent and require response, especially since it gives me the opportunity to clarify and expand some points regarding my work on the Cleveland skull, the increasingly irrational defenses of the taxon, recently uncovered specimens, and the inaugural article. However, I presently have three projects in the works that have direct bearing on this taxonomic issue. Therefore, I cannot state the conclusions or share data before they are published. You'll have to read the previous post for the appropriate context.
#1: Notably it's only really engaging with the first and worst arguments for Nanotyrannus' validity.
A diagnosis sets out the evidence for the validity of a taxon; for a systematist, it is transparency at its best. In my mind, there is no reason not to accept the diagnosis of Nanotyrannus (or any other taxon) as, prima facie, the best argument for establishing a scientific name. Also, a diagnosis is a hypothesis that can be tested. This is a crucial point, and my hypothesis - the Cleveland skull is a juvenile T. rex - can be falsified if it can be shown that the diagnosis of Nanotyrannus is defensible. What could be easier? Has anyone done this?
The defenders of Nanotyrannus have not even mentioned the diagnosis in their articles.
Perhaps none of them have given the diagnosis any thought, or they assume that it is perfectly fine. Instead, they have just heaped on new characters as if that’ll do the job of rescuing the taxon. In the end, it does not matter if that blind spot is out of agreement, complaisance, avoidance, ignorance, or indifference; we must all agree that the diagnosis is central to the question of the validity of Nanotyrannus, and it has been on borrowed time.
In the case of Nanotyrannus, the diagnosis is important because if the purported type specimen (the Cleveland skull) is not diagnostic, then there is no taxon to which new specimens can be referred. That is why skeletons such as Jane and the ‘dueling tyrannosaur’ are completely irrelevant to this taxonomic issue. Arguably, the diagnosis isn’t low-hanging fruit; in this case, it’s the whole orchard.
#2: Sure Bakker's 25 year old ideas are flawed, but any modern defense of the genus includes a swath of newly recognized characters, and of course Jane.
There are two flaws that I see in Bakker et al. (1988): the article splits tyrannosaurids into many taxa without first assessing ontogeny, including the splitting of Nanotyrannus from Tyrannosaurus rex; I think the last paragraph in their article reveals the philosophical view that motivated the decision. The second flaw was in not conducting a cladistic analysis of the data listed in the caption of their branching diagram. As such, their phylogenetic scheme is a subjective arrangement and not a true test of character congruence.
Aside from those issues, the rest of the article is an excellent and refreshing review of tyrannosaurid cranial osteology. For example, I use their terminology for the basicranium, because no one before or since has done a better job of clarifying the salient landmarks of that region, or in providing a more useful nomenclature for those features.
If by new characters you mean the ones in Larson (2013), then I cannot say anything more than this: I deal with them in my in-progress work on T. rex ontogeny. That article includes a review of all of the literature and ‘evidence’ advanced against the hypothesis in Carr (1999). As you might guess, when the smoke clears, that a juvenile T. rex is still standing.
Again, specimens such as Jane and the ‘dueling tyrannosaurid’ are irrelevant to the issue unless it can be shown that they are different from juvenile T. rex as represented by the Cleveland skull. In that case, we’d have a new taxon or two, but we wouldn’t have Nanotyrannus.
#3. So yes, Bakker was wrong to think the Nanotyrannus type had extensive cranial fusion. But Larson (2013) reports Jane has presacral neurocentral, scapulocoracoid and pelvic fusion. By not addressing these characters (also including glenoid position, subnarial foramen position, quadratojugal pneumaticity, maxillary tooth count, etc.), your analysis loses most of its force.
My analysis is on the way! In the meantime, I’m going to sound like a broken record here, but I do think it is worthwhile repeating: Jane is irrelevant to the Nanotyrannus issue because (a) it is not the type specimen, (b) the diagnosis of Nanotyrannus is indefensible; therefore, (c) no specimen can be referred to a taxon that does not exist in Nature.
If it can be shown that Jane is a different taxon from the Cleveland skull and adult T. rex, then it becomes the holotype of a new taxon. There is no way that Jane can be referred to Nanotyrannus, because the taxon hasn’t a single diagnostic character. Also, I did propose a hypothesis to account for the difference in tooth count (Carr, 1999).
By the way, I am at the helm of the description of Jane, and we do take care of the purported diagnostic characters listed in recently published articles (e.g., Larson, 2013). As such, my hands are tied on discussing those characters specifically until the article is published.
#4. There's also the problem of using similarity to juvenile stages of other species as an argument. In the 4 stage ontogeny of your 1999 paper, Albertosaurus simply never gets to stage 4. It stops at what is stage 3 in Daspletosaurus. Indeed, in 1996, you used this to argue Daspletosaurus are just adult Albertosaurus. You later changed your mind on this, but couldn't Nanotyrannus be a tyrannosaurine that never grows past stage 1? This makes your huge table moot, just as a huge table of stage 3 characters listed as "Yes!" in Albertosaurus would not mean it was a subadult Daspletosaurus.
The goal of Carr (1999) was to establish a baseline of what the primary growth stages of a  tyrannosaurid looks like. In this case, Albertosaurus libratus was the point of reference given the high number of specimens and completeness of the growth series. It turns out that the growth series can be blocked out into four stages (small stage 1, large stage 1, stage 2, stage 3). In this series, the stepwise transformation of the juvenile features into adult can be clearly seen. I then assumed, based on common ancestry, that this pattern is plesiomorphic for Tyrannosauridae.
Working under that parsimonious assumption, I then reviewed the evidence for pygmy tyrannosaurids, such as Nanotyrannus and Maleevosaurus (it is curious that people are a lot less upset about my hypothesis that Maleevosaurus is a juvenile T. bataar). It turns out that both have the same features as juvenile A. libratus; ergo, they must be juveniles. Since that time, every juvenile of other derived tyrannosauroids (e.g., Bistahieversor, Daspletosaurus, T. bataar) that have come to light have the same constellation of juvenile features as are seen in A. libratus, the Cleveland skull, and "Maleevosaurus". If the Cleveland skull had characters seen in adults of A. libratus, Daspletosaurus, T. bataar, or T. rex, then that would be an indication that it might be a pygmy taxon and my hypothesis would have been different, perhaps validating Nanotyrannus.
As the Cleveland skull and juvenile Daspletosaurus (Currie, 2003) show, tyrannosaurines have the same starting point as A. libratus; this is consistent with the hypothesis that the growth pattern is conserved (i.e., plesiomorphic) in Tyrannosauridae (Carr, 1999). The calibration of growth series is a major part of a manuscript that I have in progress, some of this I have presented at SVP meetings over the past few years.
As such, it turns out that tyrannosaurids went through the same ontogenetic changes in the skull. In the table I present (implicitly) the hypothesis that Daspletosaurus is peramorphic relative to A. libratus, in that the growth trends are carried further in the tyrannosaurine. That is why Daspletosaurus is under the heading “Stage 4”. In the same vein, T. rex is peramorphic relative to Daspletosaurus.
The salient point you make is the suggestion that the Cleveland skull is an adult, but it does not grow past stage 1. I think it is reasonable to expect that an adult of a true ‘pygmy’ tyrannosaurid would have all of the adult characters seen in its closest relatives. The Cleveland skull certainly does not have any adult features - of any derived tyrannosauroid - that I can see.
On a related note, the original manuscript for Carr (1999) did not have the summary table of characters. One of the insightful reviewers required that I put the reams of features mentioned in the text in one, easy-to-follow, table. I like the outcome: the table of growth characters is a useful tool that people can use for a quick assessment of the relative maturity of any new tyrannosaurid specimens.
In the table, each column of characters can be thought of as analogous to a series of lines of arrested growth. I have since tested the data for hierarchical structure using parsimony analysis for T. rex (Carr and Williamson, 2004) and A. sarcophagus (Carr, 2010), which was found to be present. A hierarchy from this sort of data is best explained by ontogeny. As such, the table – and the cladistic analyses drawn from it – summarizes the hard-wired sequence of growth changes the members of the clade inherited, and altered, from their recent common ancestor.
#5. A related point is that Larson agrees Nanotyrannus could be the sister taxon of Tyrannosaurus, so your long list of characters they share works well for both hypotheses. We don't need to assume Bakker's odd phylogeny with a basal Nanotyrannus convergent with Tyrannosaurus for the genus to be valid.
Could Nanotyrannus could be a juvenile specimen of the sister taxon of T. rex? For that to be defensible, it has to be shown that the purported new taxon is diagnostic. The weight of evidence shows that it is a juvenile T. rex, so that avenue of thought doesn’t rescue the identity of the Cleveland skull from T. rex.
With regard to Bakker et al.’s (1988) tyrannosaurid phylogeny, I do think that context matters. Their arrangement complements what they say in the text, together showing the deductions that led to their taxonomic decision. It is an excellent article in that the data are absolutely clear.
#6. Further, it implies a taxon can be invalidated merely by finding the original diagnosis wanting. Let's see how that works for Tyrannosaurus, based on Osborn's (1905) original diagnosis.
1. "Carnivorous Dinosaurs attaining very large size." Tarbosaurus got as large as the average Tyrannosaurus, so this is invalid even within Tyrannosauridae.
2. "Humerus believed to be of large size and elongate (Brown)." This was based on an incorrectly identified bone.
3. "No evidence of bony dermal plates (Brown)." This was to distinguish it from Dynamosaurus, whose holotype had incorrectly referred plates.
Well, guess Tyrannosaurus is an invalid taxon we can't refer the Cleveland skull to...
Is there any serious reason to doubt the validity of T. rex as there is for Nanotyrannus? The diagnosis issue has to be taken on a case-by-case basis; there is no reason to think that T. rex is invalid, given the quality of specimens (including the type) and sample size that has been amassed. Certainly, the time has come to bring the diagnoses for T. rex and many other taxa up to date, as I did for Albertosaurus sarcophagus (Carr, 2010). I have this in the works for other tyrannosaurids, including T. rex.
#7. Finally, couldn't the Cleveland skull be a juvenile and a valid genus, just without known adult specimens? That's what you argue for in Alioramus, which is (1) based on two specimens of juvenile tyrannosaurid, (2) has an original diagnosis which relies entirely on characters claimed to be ontogenetic in the case of Nanotyrannus (long snout, high tooth count, compressed teeth, etc.), (3) has derived characters only known in the sympatric Tarbosaurus, so is by your reasoning (4) therefore juvenile Tarbosaurus. Yes, Alioramus has additional suggested apomorphies, but so does Nanotyrannus and they seem to be of equal import (e.g. additional pneumatic features in each). So I'd like to know why you came to divergent conclusions in these similar cases.
I was not at the helm of that project, which accounts for the difference in results (this also pertains to your recent follow-up comment on Alioramus).
#8. Is that really fair to paleontologists in 1988? Almost no theropod specimens were even recognized as juveniles, besides some Gallimimus and Allosaurus individuals. The few known juvenile tyrannosaurids (e.g. the then unnamed Jordan theropod and Shanshanosaurus) were thought to be small basal tyrannosauroids or even dromaeosaurids. Similarly, there was no consensus the smaller Nemegt tyrannosaurids were young bataar. AFAIK, no one had thought to look for immature bone grain in any Mesozoic dinosaur yet, and there was no consensus of (Alberto(Daspleto(Tarbo,Tyranno))) like we have now to use to find nested traits that suggest synonymy. The sample size of Tyrannosaurus was low enough to make the number of specimens a notable media figure, when now we have more than most species of theropod known from more than just their holotype. The argument Nanotyrannus is a juvenile Tyrannosaurus thus depends on a lot of data which didn't exist in 1988 or wasn't organized yet. Rozhdestvensky (1965) was remarkably prescient decades earlier regarding some of these issues, though his same methodology incorrectly found Lufengosaurus to be young Yunnanosaurus, so we might question whether his seemingly correct tyrannosaurid ideas were validly justified or at least partly luck.
Yes, it is fair because sufficient data was at hand (e.g., Rozhdestvensky, 1965), and there were two growth series (A. libratus, T. bataar) available with which to make comparisons. Bakker et al. (1988) had the opportunity to define tyrannosaurid ontogeny, despite the stumble over suture closure in the Cleveland skull.
#9. It's that old standing on the shoulders of giants cliche, and while you are a major giant here, suggesting Currie, Russell, Molnar, Paul, Carpenter, Bakker, etc., basically every theropod worker of the 80s, was unreasonable seems unfair.
 I did not cast shade over the entire lot; I just just think Bakker et al. (1988) were unreasonable in their approach to tyrannosaurid taxonomy; they split taxa before assessing ontogeny. Aside from that, sincerely, it is an excellent osteological work.
#10. A valid criticism is that Bakker et al. (1988) never cite or mention Molnar's (1980) description of what would become the type of Dinotyrannus, that was referred to lancensis at the time. It was published 8 years earlier in a well read journal, so should have gotten some comment. 
That’s a good point! I have to admit that I was late on the scene as well (Carr and Williamson, 2004).
#11. The supposed unreasonableness of Bakker et al. (1988) also affects your main argument, as some of it depends upon the truth of our current consensus. Accepting Nanotyrannus as a juvenile Tyrannosaurus seems more plausible in a world where Maleevosaurus is just a young Tarbosaurus, Stygivenator is a juvenile Tyrannosaurus, etc.. But if the well-studied Cleveland skull is valid based on newly recognized characters, what does this mean for Maleevosaurus, which has never been studied firsthand since its description to my knowledge? Larson considers Stygivenator a younger Nanotyrannus, for instance, so the former can no longer be used as an example of what juvenile Tyrannosaurus are like.
The bottom line is that the relative maturity of specimens and their taxonomic identity must be considered separately, so a change in the decision regarding the Cleveland skull does not reverse the decision on Maleevosaurus; it doesn’t change the fact that both have the same constellation of juvenile features that are seen in all other derived tyrannosauroids, nor does it change the fact that the Cleveland skull has features that are only seen in T. rex. Williamson and I (2004) showed that ‘Stygivenator’ is referable to T. rex, but Larson (2013) did not engage with the evidence that we presented.
In the same fashion, the Larson (2013) article did not deal with the diagnosis issue in Bakker et al. (1988), or the evidence presented in Carr (1999). Therefore, the taxonomic decisions in that article carry no weight because the validity of Nanotyrannus is uncritically assumed to be valid despite all of the evidence to the contrary.
I trust all of that satisfies your query.

References cited
Bakker, R.T., Williams, M., and Currie, P.J.  1988.  Nanotyrannus, a new genus of pygmy tyrannosaur, from the latest Cretaceous of Montana. Hunteria 1: 1-30.

Carr, T. D.  1999.  Craniofacial ontogeny in Tyrannosauridae (Dinosauria, Theropoda). Journal of Vertebrate Paleontology 19:497-520.
Carr, T. D. 2010. A taxonomic assessment of the type series of Albertosaurus sarcophagus and the identity of Tyrannosauridae (Dinosauria, Coelurosauria) in the Albertosaurus bonebed from the Horseshoe Canyon Formation (Campanian–Maastrichtian, Late Cretaceous). Canadian Journal of Earth Sciences 47:1213-1226.
Carr, T. D. and T. E. Williamson. 2004. Diversity of Late Maastrichtian Tyrannosauridae from western North America. Zoological Journal of the Linnean Society 142:479-523.
Currie, P.J. 2003. Cranial anatomy of tyrannosaurid dinosaurs from the Late Cretaceous of Alberta, Canada. Acta Palaeontologica Polonica 48(2): 191-226.
Larson, P. L. 2013.  The case for Nanotyrannus in J. M. Parrish, R. A. Molnar, P. J. Currie., and E. B. Koppelhus (eds.) Tyrannosaurid Paleobiology, University of Indiana Press, Bloomington and Indianapolis, pp. 15-53.
Rozhdestvensky, A. K. 1965. [Growth changes in Asian dinosaurs and some problems of their taxonomy]. Palaeontological Zhurnal 1965:95-100. (Translated from the Russian.)

Sunday, September 15, 2013


A cast of the Cleveland skull in the Hall of Horns and Teeth at the Museum of the Rockies (Bozeman, MT), where it is presented as a juvenile Tyrannosaurus rex, next to a cast of the subadult MOR 1125.
The impending auction of the ‘Dueling Dinosaurs’ has brought Nanotyrannus back into the pop culture spotlight. I repeat here the evidence that I published in support of the hypothesis that the Cleveland skull is a juvenile Tyrannosaurus rex (Carr, 1999). A critical evaluation of the diagnosis for Nanotyrannus lancensis follows this, and I close with evidence documented by Bakker et al. (1988) that supports the specimen’s referral to T. rex.
What was Bakker et al.’s (1988) characterization of Nanotyrannus? They considered the skull to represent a fully adult individual (skull length: 572 cm) of the most primitive (basalmost) tyrannosaurid from the Late Maastrichtian Hell Creek Formation of Montana, which was sympatric with T. rex. They considered Nanotyrannus to be convergent with T. rex upon an identical and extensive suite of adaptations for stereoscopy and striking behavior (i.e., head-neck kinematics).
Isn’t the Cleveland skull from an adult? No – the basis for adulthood by Bakker et al. (1988) was based on the mistaken notion of suture closure in the skull (Carr, 1999). In fact, very few skull sutures actually close during the ontogeny of tyrannosaurids; regardless, every observable suture in the Cleveland skull is open, aside for the internasal and interparietal sutures, which are closed in other tyrannosaurid juveniles (1999).
Also, the surfaces of the bones in the Cleveland have a distinct striated texture that is typically seen in juvenile dinosaurs (Carr, 1999). A similar texture is seen in adults, but its distribution is patchy and the striae are less dense and distinct than the condition seen in small specimens.
Therefore, based on the absence of suture closure (beyond what is usually seen in tyrannosaurid juveniles), the presence of immature bone grain, and its small size, the Cleveland skull is unquestionably a juvenile tyrannosaurid.
Is the Cleveland skull similar in any way to other juvenile tyrannosaurids? Yes – in virtually every detail; this table summarizes the salient features (Carr, 1999):
Albertosaurus libratus juvenile condition
Is it seen in the Cleveland skull?
Premaxilla, narrow, rostral view
Premaxilla, lateral margin, concave, rostral view
Premaxilla, alveolar region, shallow, rostral and lateral views
Maxilla, mediolaterally narrow, lateral and rostral views
Maxilla, first tooth, incisiform, all views
Maxilla, teeth except the first, labiolingually narrow
Maxilla, ventral jugal process, not breached by neurovascular sulcus
Maxilla, lateral surface along rostral end of the antorbital fossa, not strutlike, lateral view
Maxilla, promaxillary fenestra, tall and not recessed, lateral view
Antorbital fenestra, longer than tall, lateral view
Maxilla, maxillary fenestra, size, small, lateral view
Maxilla, maxillary fenestra, position, midway between the rostral margins of the antorbital fossa and fenestra, lateral view
Lacrimal, rostroventral ala, rostral margin, concave, lateral view
Lacrimal, rostroventral ala, contact with jugal, exceeds that of the ventral ramus, lateral view
Jugal, maxillary ramus, dorsoventrally shallow and long, lateral view
Jugal, pneumatic recess, rostrally restricted slit, lateral view
Jugal, postorbital joint surface approaches the orbit floor, lateral view
Jugal, base of the postorbital process, convex, lateral view
Jugal, caudal margin of the postorbital process, convex, lateral view
Jugal, caudal rim of the joint surface for the lacrimal, subvertical, lateral view
Postorbital, laterodorsal margin, vertically oriented, dorsal view
Postorbital, cornual process, absent, lateral view
Postorbital, subocular prong, absent, lateral view
Frontal, lacrimal notch, long and narrow, dorsal view
Frontal, paired bones are longer than wide, dorsal view
Frontal, dorsotemporal fossa, shallow, dorsal view
Frontal, dorsotemporal fossa, indistinct rostral margin
Parietal, nuchal crest, delicate in form
Parietal, nuchal crest, low in height
Ectopterygoid, jugal ramus, not inflated, ventral view
Ectopterygoid, muscle scar on jugal ramus, caudolateral in position
Supraoccipital, dorsal process, narrow, caudal view
Basioccipital, occipital condyle, shape of an inverted triangle, caudal view
Basioccipital, occipital condyle, form of caudoventral surface, flattened, caudal view
Basioccipital, basituberal web, flat ventral surface, ventral view
Basioccipital, basituberal web, ventral surface is dorsally arched, caudal view
Basioccipital, subcondylar recess, deeply excavated, caudal view
Basisphenoid, pneumatic foramina are small, caudoventral view
Basisphenoid, pneumatic foramina are situated ventrally within the basisphenoid recess, caudoventral view
Basisphenoid, oval scar, smooth and lateroventrally oriented, caudoventral view
Dentary, dorsoventrally shallow, lateral and medial views
Dentary, mediolaterally narrow, ventral view
Surangular, dorsoventrally shallow, lateral view
Surangular, surangular shelf, horizontally oriented, lateral view
Prearticular, dorsal margin is restricted caudally, medial view
Prearticular, caudal ramus is dorsoventrally shallow, medial view
Prearticular, rostral ramus is dorsoventrally shallow and straplike, medial view
Prearticular, rostral ramus, caudodorsal surface is smooth, medial view

Therefore, every bone in the Cleveland skull is identical to the condition seen in small juvenile tyrannosaurids, not adults (Carr, 1999). Therefore, we can add the details of the bone structure to the stack of evidence regarding its immaturity.
What about the diagnosis of Nanotyrannus lancensis? Doesn’t it adequately define the new genus and species? This is the heart of the matter because the validity of a taxon rests on its diagnosis. A diagnosis is a list of the anatomical features that set a new taxon apart from its closest relatives; presumably the diagnosis of Nanotyrannus includes indisputably unique features. For Nanotyrannus, let’s consider each character (Bakker et al., 1988) one at a time:
1) “muzzle width greatly constricted to only one fourth the width of the temporal region” (Bakker et al., 1988:3)—This feature is not unique; a narrow snout and wide temporal region is also seen in T. rex. Also, the snout is not as narrow in the Cleveland skull as the authors claim, a condition that has  been exaggerated by mediolateral crushing. This can be seen in the plate of the specimen in dorsal view (1988:28), where  fractures extend across the right jugal and maxilla, and the base of the snout extends rostrally at an unnaturally sharp angle.
2) “extraordinarily wide basicranium between basitubera and basipterygoid processes” (Bakker et al., 1988:3)—This is also seen in adult T. rex (Bakker et al., 1988).
3) “two large pneumatic foramina, placed one behind the other, near the midline of the basisphenoid” (Bakker et al., 1988:3)—This condition is the result of damage (Carr, 1999).
4) “Differs from Daspletosaurus, Gorgosaurus and Alioramus and agrees with Tyrannosaurus, Albertosaurus and Tarbosaurus spp. in having the derived features of a marked expansion of the width across the temple, relative to the skull length…”(Bakker et al., 1988:3)These are not a diagnostic characters since they are seen in other taxa; there is no reason to think that it didn’t evolve once in the common ancestor of Tyrannosauridae and later became lost in several species.
5) “…and basitubera that are displaced forward towards the basipterygoid processes” (Bakker et al., 1988:3)—The same issue applies here as in (4).
6) “Differs from all other tyrannosaurids, except Alioramus, in retaining the primitive characters of a long, low snout…”(Bakker et al., 1988:3)—A long, low snout is typical of juvenile tyrannosaurids, and it has recently been shown (Brusatte et al., 2012) that the extremely long and shallow depth of the snout seen in Alioramus is derived beyond the norm in tyrannosaurids; i.e., the Cleveland skull does not have the condition seen in Alioramus, its snout depth is normal for a tyrannosaurid of its size.
7) “…and maxillary teeth that are strongly compressed side-to-side” (Bakker et al., 1988:3)—Narrow teeth are typical of juvenile tyrannosaurids, and the Cleveland skull does not have unusually narrow teeth in comparison with them (Carr, 1999).
8) “Differs from all other tyrannosaurids in lacking strong ridges and striae along the dorsal surface of the nasal” (Bakker et al., 1988:3)—It turns out that this is not true; a smaller and less mature juvenile T. rex (LACM 28471) has smoother nasals than are seen in the Cleveland skull, indicating that the nasals coarsened with maturity in T. rex (Carr and Williamson, 2004). In other tyrannosaurids, such as Albertosaurus, the nasals are coarse relatively earlier in growth (Carr, 1999).
9) “Differs from Tyrannosaurus in retaining the primitive character of a smooth depressed area for muscle origin on the anterior-ventral corner of the lachrimal” (Bakker et al., 1988:3)—This is the antobital fossa, a pneumatic surface and not an origin for muscle, which is seen in adult T. rex, except that this region is swollen by the internal inflation of the bone by the antorbital air sac system (Carr, 1999). This is a simple growth change that is seen in all tyrannosaurids (Carr, 1999).
Why then is the Cleveland skull referable to T. rex?
To answer this we don’t need to go any further than Bakker et al. (1988), who provided a long list of characters that are only seen in T. rex and the Cleveland skull:
1) “In Tyrannosaurus the basituberal width is enormously expanded, reaching 19% of the skull length…In Nanotyrannus the basituberal web is fully 20% of skull length…”
In other tyrannosaurids the ratio is much lower,  12%-15% (Bakker et al., 1988: 9).
2) “in Albertosaurus, Tyrannosaurus and Nanotyrannus [the main tuberous process of the basal tuber] is reduced in bulk” In other tyrannosaurids, the process is a “swollen development” (Bakker et al., 1988: 9).
Given the modern understanding of tyrannosaurid phylogeny, the reduced condition would have almost certainly have evolved twice, once in Albertosaurus, and again in Tyrannosaurus.
3) “In Nanotyrannus the ventral floor of the basisphenoid is nearly flat between the basituberal and basisphenoid webs, a condition seen in Tyrannosaurus” (Bakker et al., 1988: 12).
In other tyrannosaurids the “ceiling of the central basisphenoidal cavity is very tall and the cavity is very deep top-to-bottom” (1988:12). Although the authors are inconsistent in using the terms “ceiling” and “floor”, they are discussing the same feature, namely, the basisphenoid recess.
4) “The occipital condyle…in Nanotyrannus…faces…strongly downward…agreeing with the condition in Tyrannosaurus” (Bakker et al., 1988: 12).
In other tyrannosaurids the condyle presumably faces more caudally than downward.
5) “Moreover, the articular surface on the condyle in Nanotyrannus has been displaced downwards and around onto the ventral surface of the [occipital condyle] neck…again agreeing with [the condition seen in] Tyrannosaurus. Clearly the head was flexed far more sharply downwards in Nanotyrannus and Tyrannosaurus than in more primitive [i.e., all other] tyrannosaurids” (Bakker et al., 1988: 12).
In other tyrannosaurids the articular surface presumably does not extend ventrally onto the neck of the occipital condyle.
6) “…in Tyrannosaurus and Nanotyrannus, the basitubera are displaced forward and project less strongly below the floor of the basicranium” (Bakker et al., 1988:13).
In other tyrannosaurids “the basitubera are located far aft, beneath the occipital condyle, and project strongly downward” (1988:13).
7) “In Nanotyrannus the width of the…entire temporal region is greatly expanded relative to the snout…the breadth between the jaw joints…is 57% of the skull length…[In Tyrannosaurus] the quadratojugal width is 62% of skull length” (Bakker et al., 1988:14).
In other tyrannosaurids this ratio is much lower, ranging from ~33%-42% (1988:14).
8) “As a consequence of the temporal breadth/muzzle breadth ratio in Nanotyrannus and Tyrannosaurus, right and left eyes have a wide overlap of visual fields. Stereoscopy was probably present over a range of 30 degrees or more” (Bakker et al., 1988:15).
In other tyrannosaurids this overlap is not seen.
9) “the orbits face upwards as well as forwards in Tyrannosaurus…In Nanotyrannus, the upward orientation of the orbits is less marked” (Bakker et al., 1988:16).
Although the authors are making a contrast here, it is worth noting that both taxa have orbital fenestrae with an upward component, a condition that is not seen in other tyrannosaurids.
10) “The paroccipital processes of Nanotyrannus agree with those of Tyrannosaurus in being oriented nearly directly outwards and having much less of a backward orientation than that seen in gorgosaurs and daspletosaurs” (Bakker et al., 1988:16).
Arguably, the authors made a better case for referring the Cleveland skull to T. rex than I did (cf. Currie, 2003).
What did Bakker et al. (1988) make of these striking similarities between the Cleveland skull and T. rex? How could they think they had a new genus and species?
Here it is, in their own words:
“Whatever its evolutionary origins, Nanotyrannus seems to have evolved its stereoscopy independently of Tyrannosaurus. The low, long snout and primitive teeth of Nanotyrannus seem to preclude an ancestry from an animal as advanced as Gorgosaurus. The two stereoscopic genera, Nanotyrannus and Tyrannosaurus, are the only well known large predators from the Lancian Age, and it is noteworthy that both have achieved the highest degree of potential stereoscopy known among large theropods…Clearly natural selection was favoring greater precision of depth perception…Thus Nanotyrannus seems to have been shaped by an inexorable law of the evolution between predator and prey, a law that operated to produce ever greater finesse in the choreography of attack behavior” (Bakker et al., 1988:26).
The authors considered this suite of characters to have evolved by convergent evolution in Nanotyrannus and Tyrannosaurus, evidently at the same point in geological time and in the same geographic location. Clearly, that is not a parsimonious hypothesis.
Here’s my train of thought on the matter:
1) The Cleveland skull is of a juvenile tyrannosaurid.
2) The diagnosis of Nanotyrannus lancensis fails under scrutiny: the characters reflect the specimen’s relative maturity, damage to the skull, and characters shared with other tyrannosaurids. Consequently, N. lancensis is an invalid taxon; it entirely lacks defining features.
3) The Cleveland skull shares many specializations that are seen nowhere else except in adult T. rex, the species with which it happens to be sympatric.
4) Therefore, the Cleveland skull is a juvenile T. rex.
5) Ergo, the ‘dueling tyrannosaurid’ cannot belong to a taxon that does not exist.
So what species is the dueling tyrannosaurid?
It isn’t in a legitimate repository, so we can't know this at present. We’ll almost certainly never know if the auction goes ahead and the specimen is sold to a private individual – unless, as a colleague recently suggested to me, the people of the US want to intervene and rescue it for science under a successful argument for eminent domain.

References Cited
Bakker RT, Williams M, Currie PJ. 1988. Nanotyrannus, a new genus of pygmy tyrannosaur from the latest Cretaceous of Montana. Hunteria 1: 1-30.

Brusatte, S. L., T. D. Carr, and M. A. Norell. 2012. The osteology of Alioramus, a gracile and long-snouted tyrannosaurid (Dinosauria: Theropoda) from the Late Cretaceous of Mongolia. Bulletin of the American Museum of Natural History 366:1-197.

Carr, T. D. and T. E. Williamson. 2004. Diversity of Late Maastrichtian Tyrannosauridae from western North America. Zoological Journal of the Linnean Society 142:479-523.

Carr, T. D.  1999.  Craniofacial Ontogeny in Tyrannosauridae (Dinosauria, Coelurosauria). Journal of Vertebrate Paleontology 19:497-520.

Currie, P. J. 2003. Cranial anatomy of tyrannosaurid dinosaurs from the Late Cretaceous of Alberta, Canada. Acta Palaeontologica Polonica 48: 191-226.