O Crest-less One

My brief forays into ornithischians is usually relegated to those heterodont taxa that are basal ornithopods and their ancestors (but not descendant thyreophorans, ceratopsians or pachycephalosaurs) because of the wonderful arrangement and oddity of their teeth. Once you get past them and into ankylosaurs, stegosaurs, ceratopsids, and iguanodontoids, things get a little more boring when it comes to variation. These taxa become essentially homodont and even isodont (their teeth remain roughly the same size mesiodistally). While understanding tooth variation includes understanding its restriction to the morphology of elongated and supernumerary tooth rows, as in hadrosaurs and ceratopsians (acquired convergently), I am more interested in understanding how they got there.

This foray is into a new taxon of hadrosaur, Acristavus gagslarsoni, and how it doesn’t tell me anything about the central thesis of this blog. One will recall, though, that one of my very first posts was on a hadrosaur, Glishades ericksoni.

Described by Terry Gates, Jack Horner, Rebecca Hanna and C. Riley Nelson, Acristavus gagslarsoni is based on two specimens representing a single cohesive morphology. The holotype is MOR 1155 from the base of the Two Medicine Formation, Teton County, Montana, USA; and a referred specimen is UMNHVP 16607 from the Middle Mudstone Member of Wawheap Formation (Grand Staircase-Escalante National Monument), Kane County, Utah. It is primarily known only from these two specimens, which represent partial skulls and portions of the postcranium (the referred specimen includes only a cervical vertebra).

Due to preservation, little is said of the dentition, which is a pity. The skull lacks preservation of the full mandible, including the predentary, while the cranium lacks portions of the premaxilla. Features said to be diagnostic of Acristavus gagslarsoni (named for the collector and not some stab at Gary Larson, he who made many gags) are extensive and separated into autapomorphic features and an apomorphic suite: 1. a developed dorsolateral “enlarged” orbital margin of the anterior process of the postorbital, 2. posterior extension of the basioccipital from the margin of the foramen magnum (elongation of the occipital condyle), 3. a laterally “squared” frontonasal suture, 4. “swollen” interdigitating suture between prefrontal and frontal, 5. lateral surface of postorbital with deep “depression,” a sulcus posterior to the anterior and orbital-margin “swelling,” 6. alar process (squamosal, exoccipital, paroccipital flange, supporting various muscle groups and enclosing the “otic notch”), 7. expanded basipterygoid ridge, 8. supratemporal bar dorsally “flat” and deep, 9. absence of nasal ornamentation, 10. lachrymal wide in posterior view,  11. anterodorsal nasal process above external naris does not extend anterior to the narial margin, and 12. a dentary with a “flat” ventral gular margin. These selective features are unique in combination to the new taxon.

Skulls of "brachylophosaurin" hadrosaurs. Left, outlines of skulls of Acristavus gagslarsoni (based on type and referred specimen), B, Maiasaura peeblesorum, TCMI 2008.89.2, C, Brachylophosaurus canadensis, MOR 794. At right, skulls of D, MOR 1154 in 1, left lateral and 2, dorsal views; and E, UMNHVP 16607 in dorsal view, with 2 representing a line drawing of 1. All images modified from Gates et al., 2011. Numbers in A are apomorphies refered to in the text, and black arrows in B and C refer to the upturned frontal of Prieto-Márquez (2010). Arrows in D and E refer to feature "3" in the text, and also to the posterior buttress of the nasal. This is explained in the text.

However, bizarrely, the most conspicuous things about the paper are two things missing from it: the first is the selection in phylogenetic analysis, and the second is an understanding of the ontogeny of the specimen.

1. Phylogeny

Gates et al. (2011) provide a cladistic analysis sampling 15 taxa and 116 characters, modified from Horner et al. (2004) (which had 105 characters). Below and to the left (with two taxa removed due to the “scope” of the analysis and the ingroup focus), this phylogeny produces a relationship between Brachylophosaurus canadensis and Maiasaura peeblesorum to the exclusion of other hadrosaurids at the base of the tree, whereas Acristavus gagslarsoni is placed sister to Maiasaura peeblesorum. Rather peculiarly, Gates et al. had the option of also using the analysis of Prieto-Márquez (2010), which had 53 taxa and 286 characters (350% the number of taxa and 240% the number of characters), which is shown below and to the right.

Phylogenies of Hadrosauridae. A, after Gates et al., 2011, based on the analysis of Horner et al., 2004. B, after Prieto-Márquez, 2010.

Gates et al. choose not to use an anlysis they regard as “the most comprehensive to date” on what seems irrelevant grounds:

Despite utilizing more morphologic characters, as well as incorporating a morphometric approach that had never been utilized in hadrosaurid phylogenetics, the phylogeny presented by Prieto-Marquez[sic] (2010a) unites Maiasaura and Brachylophosaurus in a clade on the basis of one character, whereas the analysis presented here posits three characters supporting a clade consisting of Maiasaura and Acristavus. We suggest that the Prieto-Márquez (2010a) result is not substantial enough to discount the phylogenetic hypothesis presented in our analysis and, therefore, a more critical examination of the characteristics presented in both phylogenies will be required before the exact placement of Acristavus can achieve consensus.

[Gates et al., 2011, pg.809; “Prieto-Márquez (2010a)” refers to Prieto-Márquez (2010), and is included in references below.]

It seems odd to produce an analysis and then claim that running it in Prieto-Márquez’ analysis will have to wait for more concensus. The taxon was run in the analysis, and may merely have needed to be recoded to accord with the updated morphologic analysis presented in the paper. This may have corrected the issue of any possible error. Or, say, run both analyses and produce a level of certainty where they agree (a three-taxon Brachylophosaurini). It is not unheard of, and in some analysts’ minds desirable, to run taxa in multiple analyses in order to find concensus when presenting said taxon. In many cases, this reveals more about taxon-character agreement than it does the inverse, and can only increase information. Instead, they reject the analysis in part due to is use of Bayesian analysis (when the analysis can be run without it, and was: figs. 7 and 8 in Prieto-Márquez, 2010); the topology excluding Bayesian analysis does not differ in this specific point of the tree, and concretely supports the position of Acristavus gagslarsoni with Brachylophosaurus canadensis and Maiasaura peeblesorum, although basal to them.

This aspect of the phylogeny, the specific placement of Acristavus gagslarsoni and a few odd taxa (such as Shantungosaurus giganteus), appears inconsequential. A single feature supports either placement, although Gates et al. conflate two additional characters.

Gates et al. regard the unique feature of a broad postorbital with a squared anteromedial corner (character 48, state 1 in Gates et al., 2011) as indicative of Acristavus gagslarsoni + Maiasaura peeblesorum, while also regarding two additional characters supporting this placement to the exclusion of Brachylophosaurus canadensis, the anterior process of nasal terminates prior to anterior margin of external nares (31,0) and the ventral margin of maxillary process of jugal straight to slightly curved (53,0). However, these optimize instead as apomorphies of Brachylophosaurus canadensis to the exclusion of the Acristavus gagslarsoni + Maiasaura peeblesorum clade, and thus are not necessarily useful in rejecting the topology of Prieto-Márquez (2010).

On the other hand, the latter analysis uses the support of ten characters to affirm the monophyly of Acristavus gagslarsoni + (Brachylophosaurus canadensis + Maiasaura peeblesorum): The angle between horizontal and proximal slope of dentary diastema less than 113° (character 35, state 1 in Prieto-Márquez, 2010); the angle between maxillary tooth row and rostral “premaxillary” process between 25-39° (87,1); the slope of the ectopterygoid shelf of the maxilla between 4-10° (98,2); the ventral margin of maxillary process of jugal pointed and about as deep as long (105,1); the caudoventral “horn” of jugal greatly expanded relative to the immediately caudal dorsoventral depth of the quadratojugal process by over 155% (110,2); the ventral margin of quadratojugal process of jugal relatively anteroposteriorly long and with a “well-pronounced” concavity (112,1); the frontal forming a very narrow, rather than broad or absent, portion of the orbital rim (143,1);  the alar process of the basisphenoid well-developed and large; a caudal constriction of basisphenoid very narrow, nearly twice as broad at the spheno-occipital tubercles (155,1); the skull being more than 25% broader across the postorbitals than at the squamosals (195,0).

Meanwhile, a single character supports Brachylophosaurus canadensis + Maiasaura peeblesorum: The prefrontal has a wing-like flaring dorsolaterally around the margin of the orbit (126,1).

It is not “fair” to regard a single character in association with two unoptimized characters which are effectively uninformative apomorphies of the outlying taxa versus another single character and a host of contradicting characters of the broader clade as supportive of not including the latter analysis as superior. That the authors regard said analysis as more comprehensive but then reject its use is puzzling. Neither analysis necessarily conflicts with one another over a point of the paper, the presence of a monophyletic clade (Brachylophosaurini, clade 13 in Prieto-Márquez, 2010) including all three taxa. They they may not be perfectly arranged in both seems irrelevant to analysis choice.

2. Ontogeny

Gates et al. (2011) describe the braincase and cranial roof as being essentially unfused, where only the exoccipital and opisthotic (posterior braincase bones) are fused, whereas the sutures between the remaining bones and the bones of the cranial roof are unfused. These are features that indicate non-adult status, and suggest, although do not prove, that the material at hand had time to grow. I am curious why the authors chose not only to leave out the ontogenetic assessment of the taxon but, given Jack Horner as a coauthor, left out a histological analysis. Nonetheless, the potential growth of the taxon into a  more developed form is curious because of its placement in the Two Medicine Formation, albeit at the base, from whence Maiasaura peeblesorum derives. Similarly, Brachylophosaurus canadensis is known from the adjacent Judith River Formation, which is a lateral extension of the Two Medicine Formation, and may be said to be contemporaneous with Maiasaura peeblesorum.

Thus these two taxa, supported as sister taxa as far back as Horner (1983), preserve a unique feature that has consistently united them: the character suite relating to their nasal crest. This crest is a broad rounded tabular extension of the nasals and projected caudally over the frontals; they form a deep scarf-join with the frontal, which is uniquely sculptured, and itself forms two longitudinal troughs that terminate at the end of the nasal as a rounded “scoop”-like depression. The frontal is cranially expanded dorsally and forms a broad “buttress” beneath and caudal to the crest; in Brachylophosaurus canadensis, this crest is slanted (Prieto-Márquez, 2005), but in Maiasaura peeblesorum the crest abuts the frontal portion of the crest (the buttress) and this forces the crest more vertically. The crest extends caudally over the frontal and forms a “paddle” in the former, while such a structure is absent in the latter. In Acristavus gagslarsoni, the nasals lack this dorsal or caudal extension, but the contact with the frontal still forms a deep scarf join and there is a distinct buttressing of the anterior frontal (arrows in D and E, above).

Lambeosaur ontogeny by specimens. Top row, Hypacrosaurus stebingeri; middle row, Hypacrosaurus altispinus; bottom row, Lambeosaurus lambei. After Brink et al., 2011.

Ontogenetically, the infratemporal fenestra, which has a narrow lenticular aspect (“kidney”-shaped) in adult hadrosaurs, is rounded and more ovate in subadults and juveniles. This is due to the rotation of the ascending, postorbital process of the jugal becoming more slanted caudally and horizontally and longitudinal growth of the posterior, quadratojugal process of the jugal relative to its length. Thus, the orbit progressively shifts over the ventral extent of the fenestra in adults. This feature is notably more rounded in juvenile Maiasaura peeblesorum skulls, as it is in virtually all hadrosaurid skulls, and progressively becomes dorsoventrally elongated into subadult ontogeny, until is bends under the orbit in the adult stages. In Acristavus gagslarsoni, the jugal’s postorbital process is not bent caudally and articulation with the cranium (see above at A) produces a very slight lenticular shape. The skull is not likely to be fully adult, and I would infer it to be subadult.


This is one of the reasons I am curious the material, which includes a femur, common source for histological analysis, was not sampled for the work to assess age. This is even more important because Jack Horner has been fairly vocal about the need to assess age as a constraint on the validity or utility of taxa. Here we have a Judith River Group hadrosaur only a million years older than its apparent closest relatives, and which is represented by potentially subadult material. It may belong to one of them. The use of juveniles has been shown to have an adverse effect on phylogeny, but this is not really problematic as the specimen is likely a near-adult subadult. The material, however, appears to be unique, and likely does represent a new taxon due to the development of several features of the cranium, regardless of its age (incrassation of the postorbital especially).

Phylogenetically, then, Acristavus gagslarsoni may be either a crestless taxon nested at the base of a clade producing crested forms (Prieto-Márquez’s hypothesis) or nested within them with a crested form at the base (Gates et al.’s hypothesis). Of these, based on the multitude of data and the far more comprehensive work, the former appears to be more consistent, and this is where the phylogenetic hypotheses differ where it matters: if the new taxon lies at the base of the group, it may arise from within a basal lineage of crestless forms, producing convergent crested taxa among “saurolophin” and “brachylophosaurin” hadrosaurines; however, if the new taxon lies amongst crested forms, it has merely lost the crest that is basal to it, and the development of the crests in “saurolophin” and “brachylophosaurin” hadrosaurines occurred once. Perhaps this is why the analysis choice was made, although that is empty speculation.

An analysis of the differences between character selection may be affordable in the future, but here the major difference in phylogenies would be revealed: why choose an analysis with less than a third of the taxa and less than half the characters?

Brink, K. S., Zelenitsky D. K., Evans, D. C., Therrien, F. & Horner, J. R. 2011. A sub-adult skull of Hypacrosaurus stebingeri (Ornithischia: Lambeosaurinae): Anatomy and comparison. Historical Biology 23(1):63-72.
Gates, T. A., Horner, J. R., Hanna, R. R. & Nelson, C. R. 2011. New unadorned hadrosaurine hadrosaurid (Dinosauria, Ornithopoda) from the Campanian of North America. Journal of Vertebrate Paleontology 31(4):798-811.
Horner, J. H. 1983. Cranial osteology and morphology of the type specimen of Maiasaura peeblesorum (Ornithischia: Hadrosauridae), with discussion of its phylogenetic position. Journal of Vertebrate Paleontology 3(1):29-38.
Horner, J.R. & Makela, R. 1979. Nest of juveniles provides evidence of family structure among dinosaurs. Nature 282:296–298.
Horner, J. R., Weishampel, D. B. & Forster, c. A. 2004. Hadrosauridae. pp.438–463 in Weishampel, Dodson & Osmólska (eds.) The Dinosauria (second edition). University of California Press (Berkeley, California, USA).
Prieto-Márquez, A. 2005. New information on the cranium of Brachylophosaurus canadensis (Dinosauria, Hadrosauridae), with revision of its phylogenetic position. Journal of Vertebrate Paleontology 25:144–156.
Prieto-Márquez, A. 2010. Global phylogeny of Hadrosauridae (Dinosauria: Ornithopoda) using parsimony and Bayesian methods. Zoological Journal of the Linnean Society 159:435–502.
Sternberg, C. M. 1954. Classification of North American duck-billed dinosaurs. Journal of Paleontology 28:382–383.

This entry was posted in Paleontology, Science Reporting, Systematics and tagged , . Bookmark the permalink.

4 Responses to O Crest-less One

  1. Excellent post. I agree with your assessment on both ontogeny and phylogeny, and think it’s pretty clear the authors used Horner et al.’s analysis because Horner’s a coauthor. Which if true would be a pretty terrible example of subjectivity. I still say my favorite part of the Acristavus paper is the worried expression on the Shantungosaurus in their cladogram. ;)

  2. My biggest pet peeve about this paper is the exclusion of Lambeosaurinae from the published cladogram because the paper doesn’t focus on lambeosaurs. What? Isn’t knowing the scope of Saurolophinae important? Is Lophorothon a saurolophine? non-saurolophid hadrosaurid? non-hadrosaurid hadrosauroid? Same for Telmatosaurus…. You have to find the supplemental file and run the analysis yourself!

    But as for the tree they chose….. I think the location of Hypacrosaurus stebingeri in Prieto-Marquez’ cladogram may indicate why they chose not to use it more than any discussion of brachylophosaurin character states.

    • Well, note that both Lambeosaurus and Hypacrosaurus are not monophyletic in regards to total included taxa (laticaudus near Velafrons coahuilensis, and Hypacrosaurus altispinus sister to that group, which is in turn sister to all other “disk-crested” lambeosaurines). I wonder if it is better to be safer not discussing a position you are not focusing on or willing to test than it is to focus on your taxon of choice. They do provide the full complement of the analysis, but do not discus this in the paper. Sadly, their referenced analysis is 6 years old at this time, while Albert’s is so much younger and fresher.

  3. Pingback: Hadrosauridae /Eendesnavel-dinos « Tsjok's blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s