March 1, 2012
The Torosaurus Identity Crisis Continues
Triceratops (left) and Torosaurus (right). Art by Nicholas Longrich.
More than 120 years ago, the Yale paleontologist Othniel Charles Marsh described two of the most spectacular horned dinosaurs of all time. The first, named Triceratops in 1889, had three impressive horns jutting out of its face and a solid, curved frill. Two years later, Marsh named Torosaurus, another great, three-horned dinosaur, but with a longer frill perforated by two round holes. Although the two overlapped in space and time, they seemed distinct enough that paleontologists considered them to be separate dinosaur genera. That is, until Museum of the Rockies paleontologists John Scannella and Jack Horner suggested that these two dinosaurs were really one in the same.
Scannella and Horner presented their “Toroceratops” hypothesis at the 2009 Society of Vertebrate Paleontology meeting in Bristol, England, and the following summer their paper came out. Based on skull anatomy, bone microstructure and other lines of evidence, the paleontologists proposed that Marsh’s Torosaurus was really the skeletally mature form of Triceratops. As Triceratops grew, the dinosaur’s frill would have changed size and shape, and those trademark Torosaurus holes would have opened up. An enigmatic fossil named Nedoceratops seemed to show this intermediate anatomy and was cited by Scannella and Horner as a dinosaur caught in the act of changing. Poor reporting on the research sent the public into a tizzy—Triceratops fans wept, wailed and gnashed their teeth at the suggestion that paleontologists were taking away one of their favorite dinosaurs, but only those with an affinity for Torosaurus had anything to fear. Since Triceratops was named first, the name had priority and Torosaurus would therefore be sunk. (No one seemed to care a whit that poor, neglected Nedoceratops would suffer the same fate.)
But should we sink Torosaurus? In the two years since Scannella and Horner’s paper came out, paleontologists have gone back and forth about whether such a radical, late-life transformation in Triceratops was even possible. Early last year, ceratopsian expert Andrew Farke of the Raymond M. Alf Museum of Paleontology criticized the Triceratops transformation hypothesis and pointed out that Nedoceratops did not actually fit neatly into the sequence of changes Scannella and Horner had proposed. Naturally, the Museum of the Rockies paleontologists disagreed, and in a response published in December of 2011, Scannella and Horner reaffirmed the relevance of Nedoceratops to the extreme changes Triceratops might have undergone as it grew up.
Now another set of challengers has appeared. In a paper published last night in PLoS One, Yale University paleontologists Nicholas Longrich and Daniel Field concluded that Triceratops and Torosaurus truly were distinct dinosaurs, after all.
Most of what we know about Triceratops and Torosaurus has been extracted from skulls. Post-cranial skeletons are rare and, in the case of Torosaurus, incompletely known, and so the current argument is centered on how the skulls of these horned dinosaurs changed. In the new study, Longrich and Field coded twenty four different characteristics—relating to bone surface texture, fusion between skull bones, and other features—in a swath of Triceratops and Torosaurus skulls. The paleontologists then used this data to sort the different specimens into growth stages based on their cranial development. If Torosaurus truly represented the mature form of Triceratops, then all the Torosaurus should have come out as adults.
The skulls of Torosaurus YPM 1831 and Triceratops YPM 1822 compared. Image courtesy Nicholas Longrich.
Of the six Torosaurus examined, five fell into a range between young and old adults. But there was one particularly large individual that seemed to be significantly younger. When Andrew Farke issued his critique of the “Toroceratops” hypothesis last year, he noted that a skull designated YPM 1831 was a possible candidate for a young Torosaurus. The paper by Longrich and Field supported this idea—YPM 1831 grouped with the subadult dinosaurs. “It’s a little surprising considering how damn big the skull is—probably about nine feet long—but it’s not fully mature,” Longrich said. “It’s like a teenager,” he noted, “a physically big animal but not all that mature yet.” The development of ornaments on the skull, the fact that some bones are not fused, and a bone texture associated with rapidly growing bone are possible signs that this dinosaur was not yet an adult.
If YPM 1831 really was a subadult Torosaurus, then it is probable that Triceratops and Torosaurus were distinct dinosaurs. Indeed, if Torosaurus truly was the fully mature form of Triceratops, then we should not find any juvenile or subadult Torosaurus specimens. “[B]oth Torosaurus and Triceratops,” Longrich and Field concluded, “span a range of ontogenetic stages,” and the features which distinguished each dinosaur appear to have developed before full maturity.
But Scannella disagrees. “Nothing in this paper falsifies the synonymy of ‘Torosaurus‘ and Triceratops,” he says. In particular, Scannella notes that the new study relies on comparative anatomical techniques, but does not employ studies of dinosaur bone microstructure which shows how individual skull bones were changing. Scannella explained:
Comparative morphology is useful in examining dinosaur ontogeny, however it shouldn’t be considered in a vacuum. There are other factors which provide a wealth of information on dinosaur growth. For example, by examining histology, the microstructure of the bones, we can actually see how the thick, solid frill of Triceratops expanded, became thinner, and developed the characteristic holes of the ‘Torosaurus‘ morph. You can look at a Triceratops squamosal under a microscope and see how it was transforming. We are also finding that the stratigraphic position of specimens is critical to understanding morphological trends.
Other subtle skull modifications are also in contention, such as how fusion between bones in the skull relates to maturity. Among other features, Longrich and Field looked at the fusion of skull bones to help determine which age bracket particular specimens fell within. “We think that what the fusions are telling you is that growth has slowed,” Longrich explained, “because you can no longer deposit new bone between those bones. This seems to be a fairly reliable indicator of maturity in relatively fast-growing animals like lizards, mammals, and birds.” In the case of Triceratops and Torosaurus, skull fusion seemed to occur in a particular sequence. “First the skull roof is fused, next the hornlets on the frill and cheeks fuse, then the beak and the nose fuse on. It’s a very regular pattern which suggests we can use this as a reliable way of getting at roughly where the animals fit in the developmental series,” Longrich said.
Yet Scannella and Horner have previously argued that the timing and degree of skull bone fusion aren’t as clear. Recently discovered specimens are contributing to the picture of how variable skull fusion might be. “The Museum of the Rockies has collected over a hundred new Triceratops from the Hell Creek Formation of Montana in the last decade,” Scannella said, and these specimens indicate that the details of skull fusion varies between individuals. “We have some huge, fairly mature Triceratops in which much of the skeleton is unfused; and there are also smaller, less mature specimens with many skeletal elements fused,” Scannella explained.
How the skulls of dinosaurs like Triceratops fused is not yet entirely clear, but, according to Andrew Farke, the degree of fusion between skull bones might be reliable for getting a general idea of how old an animal was. “There is little argument that the individual bones of the braincase tend to be unfused in young animals, and fused in old animals,” Farke pointed out, and further explained that “The same goes for the hornlets (epinasals and epijugals) on the face of ceratopsian dinosaurs,” he said, since “young animals tend to have unfused hornlets and old animals have fused hornlets.” Such features are what made the YPM 1831 Torosaurus stand out as a possible subadult to Farke’s eye.
Exactly which dinosaur YPM 1831 represents remains uncertain. The skull is the best candidate so far for a teenage Torosaurus, but this ambiguous specimen alone cannot end the debate. In fact, we have so much left to learn about Triceratops and Torosaurus—particularly about how their post-cranial skeletons changed as they aged—that a great deal of exploration and description remains to be done before this debate can be resolved. And this isn’t the only dinosaur name game in progress. The tiny tyrant “Raptorex” may have been a juvenile Tarbosaurus, the huge Anatotitan likely represents a mature Edmontosaurus, Titanoceratops was probably a big Pentaceratops, and the thick-skulled Dracorex and Stygimoloch might represent early growth stages of Pachycephalosaurus. Some of these changes sting—both Torosaurus and Anatotitan were childhood favorites of mine, and I’d hate to see them go—but, ultimately, these debates will help us better understand how dinosaurs grew up.
References:
Longrich, N., & Field, D. (2012). Torosaurus Is Not Triceratops: Ontogeny in Chasmosaurine Ceratopsids as a Case Study in Dinosaur Taxonomy PLoS ONE, 7 (2) DOI: 10.1371/journal.pone.0032623
Sign up for our free email newsletter and receive the best stories from Smithsonian.com each week.
19 Comments »
RSS feed for comments on this post. TrackBack URI
Excellent as usual Brian, thanks!!
I haven’t read the latest paper yet, but it seems so far that potential sexual dimorphism has been overlooked. Would be interesting to see that discussed wrt cranial ornamentation
Can’t they be different species in the same genus, Triceratops? Why does it have to be an ontogenetic explanation?
It seems like, when Andrew Farke and now Longrich & Field have published criticisms to “Toroceratops” based on testable variables, Horns and/or Scannella come back and say “those variables are too variable.” That seems like moving the goalposts to me.
Skull fusion? Oh, sorry, too variable. Number of epoccipitals? Nope, that’s too variable too. What about development of the parietal fenestrae? Ooh, that’s REALLY variable. You have to be able to TEST this hypothesis from multiple angles, not just bone microstructure. And Longrich & Field make a good point: nobody’s presented a good transitional skull yet with expected superficial morphology. With hundreds of Triceratops skulls, I’d expect one of them to have parietal fenestrae.
Could it be that not all Triceratops underwent the change into the Torosaurus morph?
What I think was going on was that Triceratops formed harems- dominated by a single, dominant male triceratops. When a male Triceratops achieved alpha status by physically ousting another harem leader hormonal cues would set in motion the changes that resulted in the Torosaurus morph. This alpha male would then have a visual signal to other Triceratops that he was the leader of the harem. Males would then fight him and females would mate with him. But it would only be after dominance of a harem was achieved that the torosaurs visual morph would arise.
For me this scenario accounts for all the incongruities- the rareness of Torosaurs, the prevalence of classic triceratops morph, differential transition into the torosaurus morph at differing ages- it all fits in this scenario. Remember not all males will become dominant harem leaders and the window of dominance was probably very short-lived.
I don’t know how this hypothesis could be tested but for me it fits the evidence.
Terrific post, lovely drawing.
interesting perspective Zach.
And where are the southern specimens in all this? According to their online database, the Simthsonian have a couple partial skulls from Utah’s North Horn formation. Where do they fit into all this?
“and the thick-skulled Dracorex and Stygimoloch might represent early growth stages of Pachycephalosaurus.”
Probably not, given the existence of young Stygimoloch specimens (MPM 7111 & MPM 8111) that are the same size as the near-adult Dracorex, yet obviously recognizable as Stygimoloch ( http://www.childrensmuseum.org/themuseum/dinosphere/draco_rex/dracorex_hogwartsia.pdf ). Besides that, Horner & Scannella didn’t even study the original Dracorex fossils (See the Saulsbury quote).
Quoting Saulsbury ( http://boards.420chan.org/dino/res/16237.php ): “They state that they utilized computed tomography to reveal internal skull development of Dracorex and Stygimoloch in order to reveal ontogenetically (developmentally) derived features. Unfortunately, they did not scan the Dracorex skull or even examine the original fossils. These were donated to a public institution and are available for study. There are, of course, no internal skull features to the resin cast they studied. In fact, the Dracorex skull has been CT scanned twice, both in and out of matrix, revealing fused sutures, an adult feature. Similarly, the jugal and postorbital bones are fused, which is evident upon examining the fossil.”
This article forgot to state that in Longrich and Field’s paper, epoccipital splitting was not yet demonstrated to have indeed occurred in the epiparietals. Just a little leeway, Scanella and Horner’s team stated that Vagaceratops irvinensis (Chasmosaurus irvinensis in their usage)is the old adult form of Kosmoceratops richardsoni. Granting that they are correct in their assumption, following their logic in their 2011 PLoS One paper that chasmosaurines split their epoccipitals as they mature, then we should expect to see that Vagaceratops have much more epiparietals than Kosmoceratops. The catch? The two have the SAME number of epiparietals (6 vs 6). That burns a gaping hole in the epoccipital splitting hypothesis and reaffirms the previous studies of other scientists that epoccipital configuations (those are the numbers, positions, orientations, and morphologies of epoccipitals) are unique for each and every species/genera of ceratopids regardless of their phase of ontogeny. Another point: I find the statement in Longrich and Field’s paper that the rate and extent of bone remodeling to be variable between and even within the same element in the same individual as very intriguing because that . Add to that that the paper re-emphasized (even though passingly) that the best location to test the relative maturity of the animals being studied is the appendicular region, particularly the limb bones since their morphology is highly conservative across phylogeny and ontogeny. As I clamor before, every line of examination should be used(morphological, histological, biogeographical, and molecular analyses).
@Zach Miller
It would be great to have a clearly transitional skull (I assume Nedoceratops doesn’t count), or on the other hand, a clearly juvenile (not “subadult” whatever that means) skull with fenestrae.
But clearly there is some degree of massive morphological change happening in these animals. Check out figure 8 from the paper. These Torosaurus skulls are nearly as different from one another as the bottom one is to Triceratops. It doesn’t take a great leap to imagine the bottom skull’s fill being slightly shorter, slightly more saddle-shaped, and with slightly smaller or with no fenestrae.
http://www.plosone.org/article/fetchObject.action?uri=info%3Adoi%2F10.1371%2Fjournal.pone.0032623.g008&representation=PNG_M
Slot YPM 1831 in between YPM 1831 (the mature Torosaurus) and YPM 1822 (the Triceratops pictured above) and YPM 1831 becomes a pretty convincing transitional form. The only kink here is that YPM 1822 is an old adult based on cranial fusion. But as the paper shows there are other Triceratops skulls of the same morphology that are not as fused. So if toro and trike are distinct, this could be complicated by the fact you’d expect a subadult Toro younger than YPM 1831 to be nearly indistinguishable from a mature Triceratops.
@Herman
Scannella was not an author on the Pachycephalosaurus synonomy paper, so it is inaccurate and unfair to level any criticism (especially incorrect criticism) at him.
@Hikaru
The suggestion that Kosmoceratops MAY be a juvenile Chasmosaurus irvinensis was mine; Scannella and Horner were coauthors on the poster but the hypothesis is something I proposed. I do not believe anyone has suggested that all chasmosaurines increased the number of epiparietals though ontogeny. However, numbers of frill epiossifications vary between Triceratops specimens (sometimes even between left and right sides of the same individual, and are higher in larger specimens), even if you remove Torosaurus specimens from the dataset. Further, ontogenetic increase in epiossifications has been suggested before (on the squamosal) by authors other than Horner/Scannella/Goodwin (see Scannella & Horner papers for refs). The process by which this occurs is currently unknown, but we do need to think up hypotheses that we can test.
The authors ignored AMNH 5116 which is as good as a candidate for intermediate between Triceratops and Torosaurus as can be. (that or a younger Torosaurus if they’re distinct from one another) Maybe not ignored but without access since The Museum of the Rockies has the largest collection of Triceratops’ skulls. It’s entirely reasonable all counter-hypothesis of Scannella & Horner’s are going to lack in a way or another due the smaller sample they’re testing. Individualism cannot be ruled out with an insufficient sample.
“Scannella was not an author on the Pachycephalosaurus synonomy paper, so it is inaccurate and unfair to level any criticism (especially incorrect criticism) at him.”
My bad. I said Scannella, but I meant Goodwin. In any case, Horner & Goodwin admitted to using a cast instead of studying the original fossils in their 2009 paper, so my criticism is applicable to them.
@ Denver:
Regarding the number of epoccipitals, I’m wondering if chasmosaurines in general greatly increased their numbers of epoccipitals through ontogeny. Since it has been established that centrosaurines more or less retained the same numbers and positions throughout ontogeny, if the same pattern also holds true for chasmosaurines, it would make the hypothesis that epoccipitals continuously incresed in numbers as the animals age far less likely. Also, I have read your a recent post of yours in the DML. I’m just wondering if someone is also looking for soft tissues preserved inside the limb bones of Triceratops and Torosaurus so that they could be tested for the presence of original biomolecules(obviously I’m more biased towards molecular techniques). Obviously, if the ontogeny is to be well established for the taxa being studied, confirming the differences observed in morphology with the differences and similarities in their biomolecular profiles would be far more effective (Wuster and Broadley, 2007). Because in the case contradictions arise between histological analyses and molecular examinations, the latter type of test is established as giving more reliable results (Eszlinger,N., Krohn,K., Hauptmann,S., Dralle,H., et al., 2008; Itakura, E., Huang, R.R., Wen, D.R., Cochran, A.J., 2011). Hence, before we could establish “who turned into what”, it would be more appropriate to establish first “who is who”. Also, the consensus biomolecular profile of a species doesn’t change with ontogeny (Chapmann, Tobin, and Hood, 1981); only the amounts of the products of the genes switched on and/or off by ontogeny change, but not how the amino acids in those expressed gene products are sequenced.
References:
Chapmann,B.S., Tobin,A.J., and Hood,L.E. (1981). Complete amino acid sequence of the major embryonic β-like globin in chickens. Journal of Biological Chemistry, 256(11), pp. 5524-5531.
Eszlinger,N., Krohn,K., Hauptmann,S., Dralle,H., et al. (2008). Perspectives for improved and more accurate classification of thyroid epithelial tumors. J Clin Endocrinol Metab. September 2008, 93(9):3286–3294.
Itakura, E., Huang, R.R., Wen, D.R., Cochran, A.J. (2011). “Stealth” melanoma cells in histology-negative sentinel lymph nodes. Am J Surg Pathol., 35(11): 1657-65.
Wuster, W. and Broadley, D.G. (2007). Get an eyeful of this: a new species of giant spitting cobra from eastern and north-eastern Africa (Squamata: Serpentes: Elapidae: Naja). Zootaxa 1532, 151-168.
@Hikaru
>Since it has been established that centrosaurines more or less retained the same numbers and positions throughout ontogeny, if the same pattern also holds true for chasmosaurines…
It doesn’t, and it has been known and published on for over 100 years. Farke (2011) suggested that intraspecific variation in chasmosaurine epiossification number was individual variation; but various other authors (both before and since Farke’s paper) noted an increase from lengthening / expansion of the frill, and suggested increase through ontogeny (See Scannella & Horner 2011). There’s no consensus as to why/how this occurs, but we do see it. We are collecting more specimens with tight stratigraphic data that help figure out what might be going on, but in many ways this is an iterative process; every new specimen we find gives us a little bit more info that helps us (re)interpret previously collected specimens. Some ideas have moved on considerably since the original Triceratops growth paper of Horner and Goodwin, or the synonomy paper of Scannella and Horner.
As far as biomolecules go, their taxonomic potential is probably still very far from being realized in fossils (although the potential is certainly there). We do take samples of Hell Creek bones (of various taxa) for Mary Schweitzer’s research, but I don’t have anything more to do with that process beyond collection so I couldn’t tell you where it is going.
@ Denver:
Do we have enough juvenile (nestlings, puppies, and late juveniles) and adolescent (early and late subadults) remains for the other chasmosaurine taxa aside from Triceratops, Pentaceratops, and Chasmosaurus mariscalensis? Because if there is a dearth of material concerning those stages of other chasmosaurine taxa, then we can not say for certain that they generally did not follow the same trend as what had been established in centrosaurines (besides, it has been established long ago that, just like chasmosaurines, centrosaurines also elongate their neck crests as they mature).
However, I give very high praises to your efforts of collecting specimens in order to test them for the presence of those creatures’ original biomolecules. At least I know I’m not the only one who stands up to Dr. Horner’s advocacy of using biomolecular examinations to build a better picture of the species richness of the world before it was hit by that asteroid 65 million BC. Nevertheless, I do lament that only a few teams are willing to do such a daunting task (AFAIK, except for Dr. Scnweitzer’s team, only Dr. Manning’s team and Prof. Lindgren’s team are performing biomolecular examinations of fossil taxa that have the potential to be used for phylogeny). I surely hope that someday, some scientists would discover some workable DNA sequences of those animals’ genes, not just amino acid sequences of some proteins.
Just to note – there is still no good evidence for an increase in epiossifications through ontogeny in chasmosaurines. The Tyrrell has a specimen of presumed Chasmosaurus that is a *tiny* squamosal with just as many bumps for episquamosals as on adult-sized specimens (discussed in my 2011 paper). The evidence for increase in Chasmosaurus is based on a presumed mixed-species sample (and the Agujaceratops squamosals are far less complete than most people realize!). However, I will defer to evidence from the new Triceratops sample *if* adults & juveniles from the same level consistently show an increase through ontogeny.
We’re always open to alternative explanations of epiossification number, and some previous hypotheses presented concerning aspects of Triceratops morphology may be wrong. However, individual variation is a possibility to be considered only once you exclude stratigraphic (first) and ontogenetic variation. We can test strat and ontogeny, but individual variation (and sexual dimorphism) are much harder to test (if at all). As with most things Triceratops, there are stratigraphic patterns which we need to weed out of the dataset in order to address Trike-Toro synonomy: something that Longrich & field did not do: they average 2 million years of Triceratops into one dataset with no differentiation. Ceratopsid morphologies are observably different over timescales of ~300-400ky (Sampson et al., 2010 [ceratopsian indiana volume]; Fowler, 2006 [svp abstract]), so we would expect to see variation over the timescles represented in the Lance / Hell Creek. this is a critical omission from their analysis.
Also, Andy Farke (who is much too modest for such self-promotion) just wrote a short note on the Longrich & Field paper stating that the cluster analysis does not provide adequate support for any particular interpretation (Note, Andy still considers Trike-Toro as separate taxa).
http://www.plosone.org/annotation/listThread.action?inReplyTo=info%3Adoi%2F10.1371%2Fannotation%2F4be4a8ad-f2d2-4978-b283-f0569665ed01&root=info%3Adoi%2F10.1371%2Fannotation%2F4be4a8ad-f2d2-4978-b283-f0569665ed01
It is my understanding that other workers also doubt the cluster results based on misapplication of the method. Maybe there will be more on this subject in the near future.
@ Denver & Andy:
Good day, Drs. Fowler and Farke. I just want to know if scientists like you could someday reconstruct the biomolecular profiles of ceratopids from preserved viable soft tissues(even if only one step at a time), would stratigraphy also apply to specimens that will be subjected to those kinds of examinations?
Very interesting read. I wonder however if any of these paleontologists have considered sexual dimorphism in triceratops/toroceratops? In many reptile and bird species, the skeletal structures, weight distribution, colors, etc are dissimilar between males and females. If you look at a male green iguana, they are much more robust and impressive looking than the female. The same is true of the basilisk lizard, who’s males sport three sets of “fins” on their back, while the females is barely noticeable.
It could be that toroceratops is simply the male version of triceratops… wonder if this has considered?