Ceratosaurus, saber-toothed theropod?
Ceratosaurus - Saber-toothed Theropod?
A discussion on Oswald and Curtice 2026
Figure 1. Ceratosaurus - one of the most underrated and coolest theropods!
Ceratosaurus is one of the most distinct and underrated dinosaurs of all time, at least in my opinion. Let's look at how it is cooler than any other theropod (see Fig. 1 for visuals):
- Giant nose horn for which the "nose lizard" was named. That thing is seriously
- Oversized head
- A row of osteoderms (armor) down its back
- Vicious, long, thin teeth
We have four reasonably complete skulls, including a juvenile (Fig. 2, now sold).
Figure 2. Juvenile Ceratosaurus original bone. Note the elongate "saber" teeth in this small beast
Figure 3. Ceratosaurus nasicornis holotype (USNM 4735)
Figure 4. Ceratosaurus magnicornis holotype (MWC 1, from Fruita Paleontological Area). I consider it C. nasicornis)
Figure 5. Ceratosaurus dentisulcatus holotype (UMNH 5728, from CLDQ). I consider it C. nasicornis)
Figure 6. Ceratosaurus with me for scale (much more of this skull is known)
You can tell I enjoy Ceratosaurus and have always taken a moment to see the material when visiting museums where its bones are curated (Fig. 3-5, the pics I took were on film that I should scan someday). I have long wanted to be part of describing the Brigham Young University material (Fig. 6), and remember Jim Madsen showing me the CLDQ elements and Dr. Brooks Britt the Fruita Paleontological Area material. Mix in the fact that I looked at it at the Smithsonian, and I've pretty much touched 'em all!
When Taylor Oswald, a paleontologist friend out of Utah with whom I've published several papers (seriema feet and their 'killer claw' use, and Early Cretaceous theropod teeth from Utah that represent new genera), shared with me his thoughts on how
Ceratosaurus may have used its wildly awesome maw to quickly dispatch smaller prey to avoid kleptoparasitism and risk of injury I was delighted. I thought twas a cool idea, but how would he test it? He proposed we conduct a morphometric analysis of Late Jurassic Morrison Formation theropod teeth and see how they match up quantitatively.
In our "Here Be Dragons" (co-authored with the amazing Colin Boisvert and the awesome Dr. Domenic D'Amore) work on Early Cretaceous Cedar Mountain teeth, Taylor and team showed me how theropod tooth workers have standardized tooth measurements (Smith et al. 2005, Hendrickx et al. 2015, Noto et al. 2022). There is a prescribed methodology they all adhere to. This is fantastic in science because we know how they conducted their research and, if we emulate their methods, we can compare our teeth to those they measured, thereby increasing our sample size.
We discussed saber use and accepted the assumption that sabers were, in most cases, used to deliver a "quick kill" rather than solely for big-game hunting. Clearly, they could be used to kill larger animals if desired, but that wasn't their primary function in our and many others' opinion (Van Valkenburgh and Hertel, 1993; Anton et al., 2004; Salesa et al., 2005; Van Valkenburgh, 2008; Salesa et al., 2010). I must admit I was surprised to see the mammalian literature shift to what I call "speed to feed" (= a quick kill) from what I grew up with as a child in the 1970s/80s, "big teeth for big prey" (Gonyea, 1976; Emerson and Radinsky, 1980; Akersten, 1985; Turner and Antón, 1997). "Speed to feed" results in energy conservation, reduced injury risk, and less kleptoparasitism because bully predators can't easily steal what is already in a stomach.
A secondary goal of the project was to test the suggestion by Bakker and Bir (2004) that Ceratosaurus was a semiaquatic predator. This paper's citation is often DM'd or posted on Fossil Crates' socials by folks telling me that Ceratosaurus ate mostly fish. I've never bought into it as the postcranial evidence, especially the tail portion, didn't carry any weight based upon my sauropod research. Taylor suggested we could do morphometrics on spinosaur teeth alongside all of the Morrison Formation theropod teeth we could get our hands on, borrow from previous papers, and see what the computer spat out.
I said, "Sign me up! What do you need from me?" He replied, "Machairodontinae teeth." To which I grinned as I happen to have access to 'em. We made a plan, then executed. Happily, our paper was published as part of the glorious Morrison Formation volume “New Developments in the Paleontology and Geology of the Upper Jurassic Morrison Formation,” edited by a truly epic team of Morrison specialists: John R. Foster, Kelli C. Trujillo, ReBecca K. Hunt-Foster, and Spencer G. Lucas. We gave it an appropriately academic title: A Morrison “Saber-tooth”? – Comparison of Ceratosaurus Dentition to other theropods and Machairodontinae and its implications for Ceratosaurus predatory ecology. You can download it here. I'll provide some level of detail below, but nothing beats the primary literature.
Quick Results (aka TLDR ;-))
Saber-tooth Similarities?
Semiaquatic predator results
We went in assuming spinosaurs are piscivorous to a large degree. We used CBR to compare Ceratosaurus teeth with spinosaur teeth and found that Ceratosaurus teeth are vastly different in 'morphospace' from spinosaur teeth. "No kidding, look at the pics!" I can hear you thinking, and you are not wrong. However, we wanted to make the comparison quantifiable and testable. So we mathematically showed they are quite different. We don't see a world where Ceratosaurus is an, ahem, aquatic pursuit predator. Nor do we think they were snapping at fish on the regular; their snouts are nowhere near as elongate as those of spinosaurs and other fish-heavy reptiles. However, we'll never know for sure, but we eagerly await other opinions to wade in via testable hypotheses!
Brief(ish) Methodology of the Paper
The paper uses the following measurement abbreviations. I highly recommend that if you aren't familiar with tooth work and you really want to dig into our study, then make a 'cheat sheet' (I know I did!). In hindsight, we probably should have included one in this paper, along with the exact parameters we used to measure. The three acronyms below serve as the basis for our comparison. The first two are derived from comparing additional measurements. Yep, lots of caliper work on this project!
- Crown Base Ratio (CBR) = Crown Base width (mediolateral width of the base of the tooth crown) / Crown Base Length (ant-post length of the base of the tooth crown)
- Crown Height Ratio (CHR) = Crown Height (distance from the crown base to apex) / CBL
- Crown Angle (CA) = angle representing the displacement of the crown apex from the center of the base, roughly correlates with crown curvature.
We used 18 Ceratosaurus, 14 Allosaurus, 6 Marshosaurus, 4 Ornitholestes, and 2 Torvosaurus maxillary teeth, plus 8 Smilodon, 5 Homotherium, and 4 Megantereon canines (I'm going to call them sabers from here on out, canines for felids just sounds... odd.) in our study.
![]()
Figure 7. Crown Height Ratios (CHR) and Crown Base Ratios (CBR) Comparisons (Modified from Fig. 3A in Oswald and Curtice, 2026)
Figure 7's scatterplot is best read as follows. On the X-axis, crown base ratio (CBR), the farther to the left a taxon is, the more bladelike (labiolingually compressed) its teeth. The lower down on the Y-axis the crown height ratio (CHR) of a taxon is, the shorter its tooth crown. We used the average of CHR and CBR to account for variation among different folks who measured the teeth; this allowed us to plot a single number for each taxon (though we also provide each individual tooth measurement in our Tables 2 and 3).
If you've ever held a Smilodon saber, you know how thin it feels. Amazingly, these three saber-tooth felid taxa all plot out as "middle-of-the-road" in robustness while being exceptionally long. I say amazingly because the teeth of Ceratosaurus are exceptionally thin, and Marshosaurus even more so. A caveat on the Torvosaurus teeth, we only had two, of guestimated position, versus the 18 for Ceratosaurus of confident position. When the Cincinnati Torvosaurus publication comes out, we'll hasten to include it and see what known-good positioning does to its location on the chart.
With the above Torvosaurus caveat considered, that gives Ceratosaurus the highest average crown height ratio (CHR) among theropods. Keep in mind, this ratio was derived from 18 teeth; if you focus only on the longest teeth in the middle of the maxilla, these teeth surpass those of Homotherium (as noted on the chart).
![]()
Figure 8. Crown Angles (CA) and Crown Base Ratios (CBR) Comparisons (Modified from Fig. 3B in Oswald and Curtice, 2026)
Figure 8 is a scatterplot comparing crown angles (CA) to crown base ratios (CBR). The X-axis is best read as taxa on the left possess 'gracile' teeth, while those on the right have proportionately stouter teeth. The higher up on the Y-axis, the crown angle (CA) of a taxon, the straighter its teeth. The crown angle roughly corresponds to tooth curvature, and plotting CA to CBR has been found, among carnivorans, to be the most significant factor in functional optimization within the machairodonts, with Smilodon occupying an optimal zone of extreme morphology.
Reading the graph, Torvosaurus and Homotherium have the straightest teeth in our study set (note their Y-axis placement), while Marshosaurus and Allosaurus had the most curved teeth. Ceratosaurus possessed mostly straight, gracile teeth.
Morrison Formation Theropods Studied Here
The theropods in this study coexisted, with Dry Mesa Quarry providing bones for all 5 theropod taxa (Allosaurus, Ceratosaurus, Marshosaurus, Ornitholestes, Torvosaurus) in one locality. This diversity suggests that these taxa co-existed to some extent. I say to some extent because Dry Mesa Quarry might represent an extreme drought, which might have brought critters together that otherwise wouldn't interact much, if at all. Cleveland-Lloyd Dinosaur Quarry may have 4 of the 5 present, and other quarries have 3+.
Zimbabwe's Hwange National Park, where I have spent some time, has lions, leopards, cheetahs, caracals, and servals hunting within the park. This is only possible because they, for the most part, hunt in different niches. Kleptoparasitism poses a risk to smaller felids. A lion may come across a cheetah's kill and decide the kill now belongs to the king of the jungle.
We found 9 papers discussing possible Allosaurus feeding behaviors. Some focused on the mechanics, such as open-mouth slashing or forelimb-assisted bite-and-hold asphyxiation, while others considered the kinds of feeding, such as cannibalism, cooperative group hunting, scavenging, and agonistic aggregations of unaligned individuals (Komodo dragon-style).
Allosaurus appears to have snacked upon large terrestrial prey, based on tooth marks on many a sauropod limb bone. Whether they caused its demise or simply opportunistically fed upon animals that died from natural events like floods, lightning (I'm convinced that was the top killer of sauropods :-)), old age, starvation, disease, and the like, or were actively taken down is impossible to know. Putative Allosaurus toothmarks on large sauropods, medium-sized stegosaurs, and small ornithopods do suggest a generalized diet.
Theropod researchers have taken a stab (ahem) at what the Ceratosaurus hyper-elongate teeth might have been used for. The long teeth were suggested to serve a communication/recognition function, or to indicate that it was a semiaquatic predator. The differences among theropod skulls, teeth, and arms suggested distinct feeding strategies, providing evidence for niche partitioning. One author proposed that Ceratosaurus and Torvosaurus were large-prey specialists.
We agreed the theropods figured out some sort of niche partitioning schema, and being the only dinosaur in the Morrison Formation with such glorious blades, teeth that would likely have jutted out were lips present (I'm still stuck on the fact that arguably the most important part of that argument hinges upon N = 1 of a 'tyrannosaurid' tooth) we could see them possibly serving a recognition function.
Semiaquatic Ceratosaurus?
Figure 9. Bakker and Bir (2004) suggest that Ceratosaurus routinely swam.
Taylor wanted to address whether Ceratosaurus was a "semiaquatic predator," as proposed by Bakker and Bir in 2004 (Fig. 9). Assuming spinosaurs were fish specialists, we could compare their teeth quantitatively to see how they differed. The same analysis could then be applied to other Morrison Formation theropods to explore similarities and differences in tooth shape and to use any discoveries to examine potential functional impacts.
We used CBR to describe "conicalness." Conical teeth possessing reduced to absent serrations and a ~circular crown base (a CBR close to 1) are associated with piscivory. This build appears to be excellent for stabbing and holding slippery fish.
T. rex has a high CBR, approaching 1, but its teeth are strongly serrated. Allosaurus premaxillary teeth CBR also approach 1, but they are serrated and D-shaped rather than round. With these caveats in mind, demonstrating that we'd need to use our eyes as well as math, we decided CBR would work well enough for our study. A bonus for us was in getting to see the differences between the heterodonty of Allosaurus and Ceratosaurus, noting if their teeth differed in thickness in the differing mouth regions.
Figure 10. "Conicality" comparison (modified from Figure 2 of Oswald and Curtice 2026 with images from Madsen 1976, Madsen & Welles 2000, Stromer 1915)
The large red squircle (Fig. 10) shows the range of spinosaur tooth crown base ratios (CBR). Because spinosaurs lack heterodonty, they appear only once on the chart. The small red circle on the left is an outlier dentary tooth of Ceratosaurus. Note that both the dentary and maxillary teeth of Ceratosaurus are quite compressed labiolingually. Ceratosaurus heterodonty is a function of the more robust and differently shaped premaxillary teeth.
Allosaurus had higher CBR than Ceratosaurus for nearly all 3 regions, but aligned with spinosaurs quite a bit (though twas more varied). Allosaurus maxillary teeth have a lower CBR than their premaxilla and dentary CBRs, but they are still largely higher than the maxillary teeth of Ceratosaurus and have a wider range, too. Additionally, the CBR of Ceratosaurus maxillary teeth does not overlap with the range of its premaxillary teeth, unlike Allosaurus.
Allosaurus and Ceratosaurus differ in dentary CBR ranges as well. Allosaurus has more robust crowns on the lower jaw, similar to its premaxillary teeth, while Ceratosaurus has thin, bladelike teeth in the dentary similar to its maxillary teeth. These statistical observations match visual examination of the teeth.
Our ANOVA and post hoc analyses showed the maxilla and dentary of Ceratosaurus were statistically similar to one another, but significantly different from all Allosaurus teeth and, especially, spinosaurs.
Statistics in hand, we rejected the idea that Ceratosaurus was a semiaquatic predator that hunted aquatic prey via swimming. Could they eat a fish? Sure! But the big blades differ significantly from the conical teeth of piscivores, even the Ceratosaurus premaxillary teeth, which were more robust than the maxilla and dentary teeth, were significantly different from those of Spinosaurus.
Figure 11. Genyodectes (modified from Woodward, 1901)
What Might They Have Been Doing?
Perhaps the thicker Ceratosaurus premaxillary teeth were functionally analogous to the machairodont incisors, used to pick flesh from bones or tear apart carcasses, while the blade-like maxillary and dentary teeth inflicted deep wounds on terrestrial prey?
Maybe the Ceratosaurus maxillary teeth functioned like hyper-optimized machairodont-like sabers? They are proportionally long, curved, and especially thin. Ceratosaurus somewhat mirrors machairodonts in the extreme degree of dental hypertrophy, bearing the longest teeth of any dinosaur proportionally (with Genyodectes being longer still (Fig. 11)), suggesting a degree of ecological similarity between Ceratosaurus and machairodonts.
However, saber-tooths have a pair of elongate canines near the front of the mouth, while Ceratosaurus has a mouthful of saber-like maxillary and dentary teeth positioned farther back. Thus, different killing mechanisms must be at play, with the Ceratosaurus teeth being less precise. However, the similarity of the teeth themselves warrants attention, and we went on to hypothesize on the ecology of Ceratosaurus based on these similarities. As a reminder, we are aware that tooth shape similarity does not automatically imply identical or even closely similar ways of life. However, teeth are a reasonable proxy for diet and behavior interpretation in our opinion, especially since we will never truly know (short of inventing a time machine). The next best option is to use testable models like this.
When predicting animal behavior, most leap to "adult versus adult," forgetting that before one is an adult there is hatching, the early years, tween and teenagers, then adult. Adult Ceratosaurus likely competed with adult Allosaurus and Torvosaurus, and juvenile and subadult Ceratosaurus with Marshosaurus, smaller theropods like Ornitholestes, plus juveniles of the larger genera.
Based on what we know today, Torvosaurus and Allosaurus were heavier than Ceratosaurus and thus could have engaged in kleptoparasitism against Ceratosaurus. Thus, hypertrophied, blade-like teeth may have evolved akin to functional analogues of machairodont teeth... to kill prey quickly. In our scenario, Ceratosaurus used its teeth to stab or slice deeply, likely from ambush. A small enough animal would rapidly bleed out if bitten with a dozen "knives." A mouth bristling with such hardware could kill prey quickly without prolonged struggle, save energy, reduce the risk of injury, and give the predator more time to eat.
Allosaurus teeth appear more generalized. They are shorter and thicker, especially when compared to those of Ceratosaurus. A more generalized diet means less ecological pressure for specializations that might limit prey options.
The tooth replacement of Ceratosaurus has been suggested to be slightly lower than that of Allosaurus. Perhaps Ceratosaurus preferred prey less capable of sufficient resistance to break teeth? Ceratosaurus might target small (though still substantial) prey to use its teeth to maximum effect. Long maxillary and dentary teeth are better suited for stabbing than slicing, while short and more robust premaxillary teeth can better dismember a carcass or strip meat from bones, like felids do today. The differing tooth morphology of Allosaurus and Ceratosaurus might reflect actual niche partitioning, with different feeding strategies
Proportionately, Ceratosaurus has an oversized head, especially when compared to Allosaurus. Maybe it rapidly 'Ceratoed' large pieces of a carcass, perhaps flipping prey into the air for a swallowing whole event. We observed the seriema do this over and over. We thought the birds would use their dromaeosaurid-like "killing claw" quite often. The reality was that it was rarely deployed. Usually, the bird would stomp on prey, grasp it with its hooked beak, then flip the prey in the air and swallow it entirely.
However, maybe it did use the replaceable sabers to kill larger prey, like Smilodon populator may have done? The challenge here is that big felids have big forelimbs, but Ceratosaurus' forelimbs are small. Thus, the jaws of Ceratosaurus would have borne the brunt of the stress of immobilizing and killing, increasing the risk of teeth being broken.
We propose that Ceratosaurus wasn't as well adapted to hunting adult sauropods or stegosaurs, which are poor targets for delicate teeth specialized for killing prey quickly. This might have limited competition with Torvosaurus and Allosaurus (but if Allosaurus used an asphyxiating bite and targeted ornithopods, it might overlap with Ceratosaurus more).
We suspect Ceratosaurus might have targeted small ornithopods like Nanosaurus, medium-sized Dryosaurus, and even larger ones like Camptosaurus. Juvenile sauropods are on everything's menu. One wild thought is if Ceratosaurus was actually a sauropod specialist, ambushing sauropods whilst drinking water, one well-placed chomp could fatally wound a sauropod. I saw this in Switzerland, tanystropheid after tanystropheid skull and half a dozen cervicals articulated, the last cervical showing bite marks, the body consumed. The challenge would be in rapidly eating a sauropod before kleptoparasitic-capable theropods like large Allosaurus and Torvosaurus arrived. As I'm typing I'll continue the imagination-land thoughts, perhaps the Ceratosaurus neck bite would eliminate the sauropod's ability to make sound, thereby slowing, if not outright preventing, curious competitors from arriving.
Ceratosaurus adults, at 21'+ and over 2,000 lbs., needed substantial prey, not an endless diet of tiny vertebrates. Ceratosaurus has shorter legs than Allosaurus, suggesting it was slower and perhaps relied on ambush to dine. Juvenile Ceratosaurus might have liked Nanosaurus, but likely not adults.
A Note on Machairodonts! (aka Saber-toothed felids)
Machairodont teeth sure look like Ceratosaurus outwardly; they even have "false denticles," giving those big blades a serrated edge! That surprised the heck out of me. We compared Smilodon, Homotherium, and Megantereon teeth with Ceratosaurus teeth and pondered if they were quite similar; perhaps it's a result of shared ecological pressures.
The higher CBR in machairodont teeth of similar CHR to Ceratosaurus might be due to differences in mammalian dinosaurian dental biology. For example, dinosaurs replace teeth often; mammals only once, so they need greater insurance against breaking their teeth. Theropods, with "always new" teeth, have fewer constraints on tooth thickness. Break some and in a few months, new ones will be up and biting. The high frequency of broken teeth in machairodonts may indicate such vulnerability, which shows that, despite having less laterally compressed teeth than Ceratosaurus, tooth breakage was relatively common compared to short-toothed taxa like coeval Aenocyon dirus.
The End... for now!
Congratulations, you've made it to the end! I hope you enjoyed this blog. I did, as I was able to take more liberties than one can in a peer-reviewed paper, for one :-). It was neat revisiting the paper, as in academia, like in movies, you do the work, then move to the next project, only to learn months later that the paper is out. Speaking of the paper, if you are interested in this topic, please give it a read. You can find it, and many more, on my research collections page here. As always, feel free to make a comment; just know it won't be automatically posted (it runs on a Shopify business platform backend, not actual blog software). You can always shoot me an email at Brian@FossilCrates.com, too. That goes to me wherever in the world I am at.
Thank you kindly,
BC
Time to write: 10 hours (!) [wow time flew!]