One of the quintessential depictions of prehistoric times is that of an ancient, often volcano ridden, landscape full of animals bearing large showy sails of skin stretched over their backs. Sailbacked animals are rather rare in our modern day and age, but back in the Mesozoic and Paleozoic there were sails a plenty.
By far the most popular sailbacked taxa of all time would be the pelycosaurs in the genus Dimetrodon. These were some of the largest predators of the Permian (up to 4.6 meters [15 feet] long in the largest species). Dimetrodon lived alongside other sailbacked pelycosaurs including the genus Edaphosaurus. These were large herbivores (~3.5 m [11.5 ft] in length) that evolved their sails independently from Dimetrodon. The Permian saw many species of sphenacodontids and edaphosaurids, many of which sported these showy sails (Fig. 1. [1–8]).
However sails were hardly a pelycosaur novelty. The contemporaneous temnospondyl Platyhystrix rugosus (Fig. 1 ) also adorned a showy sail.
Fast forward 47 million years into the Triassic and we find the rauisuchians Arizonasaurus babbitti, Lotosaurus adentus, Xilousuchus sapingensis, and Ctenosauriscus koeneni , all bearing showing sails on their backs (Fig. 1 [10–13]). Much like in the Permian, many of these taxa were contemporaneous and, while related, many likely evolved their sails separately from one another.
There are currently no fossils of sailbacked tetrapods in the Jurassic (as far as I know. Feel free to chime in in the comments if you know of some examples). However the Early Cretaceous gave us a preponderance of sailbacked dinosaurs (Fig. 1 [14–19]) including the cinematically famous theropod Spinosaurus aegyptiacus, the contemporaneous hadrosaur Ouranosaurus nigeriensis, the gharial-mimic Suchomimus tenerensis, the potentially dual sailed sauropod Amargasaurus cazaui, as well as the allosauroids Acrocanthosaurus atokensis, and Concavenator corcovatus. Lastly, the discovery announced last year (and just now coming to light in the news) of better remains for the giant ornithomimid Deinocheirus mirificus have revealed that it too may have sported a small sail along its back.
Once again we find a group of related, largely contemporaneous, animals, most of which probably evolved their sails separately.
Such a huge collection of sailbacked animals all living around the same time (and sometimes the same place) has begged for some type of functional explanation. The usual go-to for large, showy surfaces like these or the plates of Stegosaurus has been thermoregulation. The thinking being that blood pumped through a large surface area like this, when exposed to the sun, has the ability to warm up faster than other areas of the body. Conversely when the sail is placed crosswise to a wind stream, or parallel to the orientation of the sun, heat will radiate out into the environment faster than other areas of the body. That most sailbacked dinosaurs were “localized” to equatorial areas, coupled with the large sizes of all the taxa (1-10 tonnes depending in species) has favoured a cooling mechanism function for dinosaur sails. Whereas a heating function has been presumed to be the primary function for sails in Dimetrodon and Edaphosaurus. No real function has been ascribed to the sails in rauisuchians or Platyhystrix, though this is probably due to a lack of knowledge/interest in these groups.
Alternate functions proposed for these sails have included a self-righting mechanism for swimming, sexual signaling and other presumed display functions. In certain cases, namely Spinosaurus aegyptiacus and Ouranosaurus nigeriensis, it has even been argued that the enlarged spines did not support a sail, but rather were supports for a large, fatty hump akin to that of camels or bison (Bailey 1996, 1997).
Given the wealth of hypotheses for potential sail functions it would be beneficial to first understand what extant sailbacked taxa use their sails for. Unfortunately—though unsurprisingly—there are few if any scientific studies on sail use in extant sailbacked animals. This has lead to the apparent assumption that there are no extant vertebrates with sailbacks.
There are, in fact, quite a few sailbacked animals alive today. These include various fish, amphibians and even reptile species. Learning what these taxa use their sails for may offer us a glimpse at what extinct animals were doing with their sails.
Not all sails are created equal
When Bailey (1997) reviewed neural spine morphology in dinosaurs vs. pelycosaurs he noted a distinct difference in spine shape between these two distantly related groups. In the pelycosaurs the elongate neural spines were gracile, subcircular in cross section, and in some taxa, laced with numerous projections coming off the main shaft. These spines appeared to be poor supports for any type of epaxial muscle, and were used as examples of a good sail-supporting spine. In contrast, the spines of S. aegyptiacus and O. nigeriensis were anteroposteriorly expanded with the widest expansion occurring at the apices of the spines. This morphology is very similar to the morphology of extant bison. This lead Baily to argue that the elongate spines of these and other sailbacked dinosaurs, actually housed a hump of either muscle, or fat.
Bailey was right in that the elongated neural spines of pelycosaurs do stand out as very different from the spines of allegedly sailbacked dinosaurs. However his argument becomes less robust upon comparison with other allegedly sailbacked animals. For instance, the rauisuchian Arizonasaurus babbiti also shows the same anterposteriorly widened neural spines with wasting at the base (Nesbitt 2005), as does its relative Ctenosauriscus koeneni (Butler et al. 2011). The temnospondyl Platyhystrix rugosus shows, perhaps, the most extensive degree of anteroposterior neural spine widening of all the sailbacked taxa. Bailey (1997) argued that these widened neural spines served to anchor epaxial musculature and would not work well as sail scaffolding. It is difficult to imagine the rather small P. rugosus needing such extensive musculature to help it scuttle about the Permian landscape. Nesbitt’s (2005) redescription of A. babbiti mentioned extensive grooving along the elongated neural spines. It is pretty likely that these grooves housed adjacent blood vessels. This would seem like a good construction for holding up a sail.
Bailey’s argument also does not agree with extant sailbacked (or tailed) lizards. Bailey (1997) argued that the spines of pelycosaurs were similar to the spines on extant Basiliscus and Hydrosaurus. However, an examination of the spines on these lizards suggests otherwise. Looking at the bottom of figure 2 we can see that the neural spines holding up the sail in Hydrosaurus pustulatus are both elongated and anteroposteriorly expanded, just like the dinosaurs.
Bailey (1997) also pointed out that O. nigeriensis showed a latticework of ossified tendons along its sail , and argued that this was a result of back stiffening, indicating that the spines were aiding in holding the body up, and thus were unlikely to be a sail. However it is important to keep in mind that Ouranosaurus nigeriensis was a iguanodontian. This group, and Ornithischia in general, are known for evolving a latticework of mineralized tendons, thus the presence in O. nigeriensis, of these tendons, could be nothing more than the phylogenetic baggage of its ancestors.
Given the comparisons with extant, related animals. I think the data are more indicative of sails than they are humps for the dinosaur or other archosaur taxa.
However, I don’t mean to discount Bailey’s work entirely. One important aspect of Bailey’s comparison was the revelation of how very strange the spines of pelycosaurs were (more on that below).
As originally brought up earlier, this is a question that has popped up from time to time in the literature. The typical go-to answer has been to say: thermoregulation. The concept being that a sail, with its large surface area, provides an ideal location for blood to spread out and absorb the sun’s heat before traveling back into the body core. Conversely, a sail also provides a means of radiating that heat when placed cross-wise to a wind stream, or dunked in water. The same high-surface area allows for blood cooling just as much as it does for blood heating.
This argument is bolstered by the presence of contemporaneous sailbacked taxa living in similarly hot, arid, environments. For example Spinosaurus aegyptiacus, Ouranosaurus nigeriensis, and Amargasaurus cazaui all lived in the Early Cretaceous of Gondwana. Ctenosauriscus koehni, Arizonasaurus babbiti, Lotosaurus adentus, and Xilousuchus sapingensis lived during the Middle Triassic all around Pangea, and of course there was once a Permian world loaded with sailbacked pelycosaurs (and at least one sailbacked temnospondyl). So the concept that sails were a response to climatic pressures has some merit.
However, despite the popularity of this argument, there has been little empirical verification for it. Of the few tests that have been done, the results suggest that sails may not work as well as our intuition would presume. Haack (1986) did perhaps the most thorough test of sails as a means of thermoregulation. Haack applied thermodynamic equations to the estimated surface area of the sails in three species of Dimetrodon with estimated masses of 50 kg – 250 kg. Haack’s results proved to be rather sensitive to the location and type of blood flow that was being pushed through the sail (vessels derived from the skin offered little help in raising or lowering body temperature. Extended vertebral vessels were more effective). Nonetheless the general results suggest that the sail was variably effective in raising core body temperature. Earlier estimates suggested that Dimetrodon grandis (the largest species of Dimetrodon) could have raised its body temperature 6°C in a period of 80 minutes (Bramwell and Felgett 1973). However, under Haack’s model, such a large rise in temperature took nearly 4 hours. This was largely due to Haack allowing for radiative heat loss on the side of the sail that was not facing the sun. This was probably an overly conservative approach as extant reptiles (and other amniotes) are remarkably adept at asymmetric heat flow via vasoconstriction of one side of the body and vasodilation on the other (Cowles 1958). Regardless, Haack’s model routinely showed that the biggest influence on heating and cooling was body size, not sail size. When the reverse was tested—cooling—the sails appeared to be of little use. All in all the large sails of pelycosaurs appeared to only offer a slight advantage for thermoregulation over no sail at all. Haack’s conclusions were that these small advantages were either good enough for what the animals needed, or that we are more likely missing the actual function of these structures.
It’s worth noting that similar arguments have been put forth for the enlarged plates of species in the genus Stegosaurus (Farlow et al. 1976, 2010; Main et al. 2005). Here the presence of extensive vasculature to the plates and their alternating arrangement were found to enhance heat flow, with radiative cooling being the more effective (Farlow et al. 1976). However, it’s important to note that this study was less detailed in its modeling, and did not actually attempt to quantify heat flow through the entire body.
So if sails aren’t all that good at thermoregulation then what are they good for?
To answer this our best option is to turn to the extant realm and look at what modern-day animals are doing with their sails.
The sails of sailfish, marlins and lionfish may provide extra control over body orientation in the 3D realm of the ocean (Fig. 3). Similar functions may be employed by the dorsal sails on newts. However all three taxa are capable of doing something that no sailbacked reptile seems able to do: they can dynamically raise and lower their sails. So if we limit our comparison to reptiles only what do we see?
Basiliscus, Hydrosaurus, and the sailbacked chameleons in the genus Trioceros, do not live in similar environments. Both Basiliscus and Hydrosaurus are tropical forest dwellers, whereas the Trioceros species (T. montium, T. quadricornis, T. cristatus) live in mountainous cloud forests. Despite the difference in habitat both groups do experience a general lack of direct sunlight. Perhaps the sails on these lizards act to increase body temperature during brief intervals of basking?
Perhaps, but then again I remind readers about Haack’s (1986) test of sail efficiency. Sails just don’t seem to be the great radiators / absorbers that intuition would have us believe. Besides there is another major kink in this argument for extant sailbacked reptiles: The sails are not equally distributed among the sexes.
It’s all about sex ba-by
Yes, all four sailbacked groups of lizards show a distinct bias in sail growth. In each case it is the males that grow the showy sails, whereas the females show reduced sails or no sails at all (e.g., Klaver & Bohme 1992). So it appears that extant reptiles have evolved sails as a means of sexual signaling, not thermoregulation. Now, it’s possible that the potential increase in heat absorption that these sails might afford, are used as a way for males to warm up sooner, allowing for better defense of territories. Unfortunately, aside from the difference in morphology between the sexes there has not been any work done on the potential other functions of sails in these species. We don’t know how appealing sails are to the opposite sex in these lizards. We don’t even know about the ontogeny of sail development. All these unknowns make it more difficult to extend interpretations of sail function into the past.
With these caveats in mind, could it still be possible that all the sailbacked animals of prehistory were rocking sexually signaling sails?
One of the apparent difficulties with the interpretation that these sails were sexually dimorphic signals is the implication that all the sailbacked animals found so far are all one sex (likely male, if the sexual selection of yesteryear followed the same rules as sexual selection today). On the outset this does seem to be preposterous. One would think that of all the specimens that have been fossilized, we should see a more even distribution of males and females in the fossil record.
This argument is valid. There is currently no way of testing for sex in the fossil record. That doesn’t mean we don’t always know the sex of a given fossil (preserved embryos/eggs inside the mother tend to be a dead giveaway), but in general there is no real way to test for sex for a given fossil. All that said, given what we know of the fossilization process, it is unlikely that we are seeing a preservational sex bias in the specimens collected. In fact it is very likely that we do have both sexes preserved, and collected, for many of these species. The key fact to keep in mind is just how few specimens represent the animals in question, as well as how little of that specimen may be preserved. For instance Spinosaurus aegyptiacus was originally known from a single specimen collected in 1915 by Ernst Stromer. To date, despite numerous other trips to Northern Africa over the past 98 years, all other fossils attributed to S. aegyptiacus or similar species (e.g., S. maroccanus) are skull material. Of that most of it is teeth. It is certainly likely that some of those teeth and skulls belonged to males and females. However the singular vertebral series known for Spinosaurus is either a male or a female, and thus gives us a glimpse at the shape of one of the sexes. A similar dearth of material is known for other sailbacked taxa such as Arizonasaurus babbitti, known from two specimens, only one of which preserved any vertebrae (Nesbitt 2003). Ctenosauriscus koehni is known from a single individual (Butler et al. 2011), Ouranosaurus nigeriensis is known from four specimens, only two of which preserve more than a handful of bones (Taquet 1976), Amargasaurus cazaui is also known from a single specimen (Salgado and Bonaparte 1991), and so on.
It’s worth noting that the spinosaurid, Baryonyx walkeri, does not preserve any type of sail (Charig and Milner 1997). If one operates under the assumption that these sails were working as sexual signals it is possible that the B. walkeri holotype was the female of the species. At one point, the gharial-mimic, Suchomimus tenerensis was thought to be a different species of Baryonyx (Hutt and Newbery 2004). However these interpretations were presented as a talk and have yet to surface as an actual technical article. Nonetheless S. tenerensis and B. walkeri still remain each other’s closest relatives. The close relationship of these two taxa may also indirectly illustrate some of the differences that might have existed between the sexes of spinosaurs. The Brazilian spinosaurid Irritator challengeri (Sues et al. 2002), may also represent the female of the species. However the only postcranial material for it is referred and consists of some sacrals and proximal caudal vertebrae (Bittencourt and Kellner 2004). These regions of the back may not have supported a sail anyway (e.g., Basiliscus plumifrons males show a dip in sail size in the sacral and proximal caudal region).
To date the only species that have provided enough specimens to say much of one thing or another about dimorphism and sails are the pelycosaurs. Unfortunately despite the wealth of specimens available for testing this, not much has been done regarding sexual selection. Romer and Price (1940) may be the last time that this topic had been addressed. Spines appear to have been present in both sexes, with size being the biggest difference (Romer and Price 1940). So the one group that has enough specimens to (somewhat) test this hypothesis, appears to invalidate it. This doesn’t bode well for the sexual dimorphism hypothesis. However the fact that pelycosaurs are quite a ways removed from sailbacked archosaurs, especially in comparison to lizards, this phylogenetic distance could mean that pelycosaurs were following a different set of rules.
As it currently stands there is just not enough data to say one way or the other.
Sexual selection in prehistory
Sexual selection is not an easy topic to broach in paleontology. As mentioned earlier, extinct animals leave very little evidence of their sex. When it comes to looking for evidence of dimorphism, be it morphological or size based, we need to look at multiple examples of the same species.
Bonebeds can be especially helpful for this. The bonebeds of Coelophysis bauri were used to determine the presence of a robust and gracile morph between individuals. These were one of the first cases of apparent sexual dimorphism in dinosaurs (Colbert 1989). Similar comparative work has found evidence of potential dimorphism in a variety of other dinosaur species (Chapman et al. 1997).
Of course these interpretations were not without some controversy. In 2011 Kevin Padian and Jack Horner came out with a counter hypothesis for all the bizarre structures seen in dinosaurs. They argued that, rather than being sexually dimorphic structures, they represented elaborate species recognition devices (Padian and Horner 2011a). The authors rightly point out that many of the examples used for sexual dimorphism in dinosaurs, are far too small to be tested statistically, and thus could go either way in regards to sexual dimorphism. Another problem, more recently coming to light, is that most dinosaur fossils are not of adult animals, but of juveniles and subadults. Thus it is hard to say exactly what structures are secondary sexual characteristics, and what are just ontogenetic effects. This is especially problematic for any study that purports to see changes in robusticity or overall size. Padian and Horner argued that species recognition is the more parsimonious explanation for the varied visual signals seen across Dinosauria.
This argument was and still is considered controversial (Knell and Sampson 2011, Padian and Horner 2011b). Rob Knell and colleagues put forth an alternative hypothesis. They argued that the bizarre structures in dinosaurs may have been sexually selected, but that the selection was not for dimorphism. Rather, both sexes were selecting for structures in the other (Knell et al. 2013a). These two hypotheses elicited an extensiveprofessional back and forth in the literature that is going on till this day (Knell and Sampson 2011, Padian and Horner 2011b, Knell et al. 2013a, b, Padian and Horner 2013, Hone and Naish 2013). Despite the hefty citation dropping there I’m do not intend to delve too much into that argument. Sufficed to say both parties make good arguments and both parties are hampered by the same problem: lack of statistical support. As is often the case in biology, if something can do one thing or another, it’s likely that it did a mix of both. Dinosauria was certainly a diverse enough group that a one size fits all approach to answering questions about their bizarre structures is likely to be wrong.
All that said, a common theme to all of these papers is that the structures in question functioned as display. Bringing us back to our prehistoric sailbacked animals it is quite likely that these showy structures were doing just that: putting on a show. As with all the other dinosaurs in which function for strange structures has been proposed, we lack enough specimens to say one way or another regarding the function of these structures. Even if they were for show that still doesn’t preclude them from doing other duties. If this leaves the reader feeling a bit unsatisfied I don’t blame you. Knowing how little we know it is all too easy to just throw one’s hands up and declare:
“Well, what’s the point in trying?”
Though this lack of data is frustrating it is puzzles like this that are what drive science forward.
Sails, or high backs?
This post has focused a lot on the large spinous processes of various reptiles (and one amphibian) and their likely role as sail supports. We have also delved a bit into the alternate hypothesis that these spines held up a large “hump” of muscle or fat. However what if it was a mix of both? What if these spines produced a high back, rather than a sail?
Astute (critical?) readers may have noticed that I placed Acrocanthosaurus atokensis in the sailback category, despite its purported sail being modest in comparison to most of the other theropods listed on that image (Paul  referred to it as a finback). This view of A. atokensis as a sailbacked critter is not a wholly supported one. Many other authors seemed more content to stuff a wodge of axial musculature along those elongate spinous processes (e.g., Stovall and Langston Jr. 1950, Bailey 1997). Such extensive musculature would have stiffened the back considerably, giving Acrocanthosaurus atokensis a rather rigid appearance. Similarly, the theropods Becklespinax altispinax and Metriacanthosaurus parkeri, as well as the ornithopod Hypacrosaurus altispinus, have all been considered to have had muscular, high backs, rather than low “fins” or sails. Stovall and Langston Jr (1950), and Bailey (1997) argued for high backs in comparison to extant mammals, but one need not have to go all the way down the tree of life to basal amniotes before coming across potentially good modern analogues. Extant Reptilia has its fair share of high-backed taxa too.
The veiled chameleon (Chamaeleo calyptratus), a native of the Middle East, is a high-backed species. Looking at the spinous processes of its vertebrae (figure 4) one might be tempted to place a small sail or fin along its back. However one look at the living animal makes it clear that these processes appear to serve as a means of extending the height of the back without becoming a full-fledged sail. I know of no anatomical data on the axial musculature of chameleons, so it’s difficult to say if the increased height of the neural spines are there to help stiffen the back or perhaps support the large casque on the head of these lizards (it’s large, but light). Chameleons are pretty stiff critters to begin with. They don’t do much lateral bending, so it’s certainly possible that these elongate spinous process are aiding in this function. The other thing that these processes are doing is increasing the size of the animal from the side. This can be beneficial if the lizard is looking to scare away predators, but may be even more beneficial if it wants to intimidate rivals. The colour-changing ability of C. calpytratus also means that this increased real estate provides for a larger visual signal that can be observed from further away. Perhaps, then, the increased spines are still serving a function in display. A function that might hint at how sails evolved in the first place (compare figure 4 with figure 5).
Whether the spines on these high-spined—but not necessarily sailed—dinosaurs were for increased axial musculature, fat storage, neither or both, it would seem that all of these animals would be making themselves appear much larger from the side. It’s hard to imagine how such structures could not serve as ways of communication, even if that wasn’t their primary function.
Sails: What’s the point?
That brings us back to our initial question: what was the function of all these sails?
The short answer is: we don’t know. We probably never will know. However we can still make reasonable inferences of potential functions for these sails. These function can be tested biomechanically and via comparison with extant analogues. So far, according to extant relatives (lizards, in this case), it looks like these were largely display structures. They may even have been sexually dimorphic display structures. Regardless of what their primary functions were, signaling would almost have to have been one of their jobs. It’s hard to imagine evolving a billboard like the 2 meter sail on Spinosaurus aegyptiacus, and not use it to signal something to another animal.
It’s worth noting that our only extant amniotes with sails today are also close relatives of archosaurs. It’s also worth noting that lizards are a largely quiet, but visually active, group of animals. Were these extinct reptiles following a similar routine? Given what we are able to infer about extinct archosaur sound production we are pretty sure that these animals could make sounds, but it may not have been a large part of their daily repertoire (Senter 2009). So perhaps sails and high backs—much like frills, crests, spikes, and plates—were just another means for a largely quiet group of reptiles to communicate with one another.
Pelycosaurs: The oddballs
So dinosaurs, rauisuchians, the temnospondyl Platyhystrix rugosus, and extant sailbacked and high backed lizards seem to all follow a very similar pattern in their sail formation. This could have something to due with parallel evolution in the reptiles, and coincidental convergent evolution for the temnospondyl. It could also hint at a similar function for these structures. What about the pelycosaurs?
As Bailey (1997) discovered when he initially compared the spinous processes in these groups, the sails of pelycosaurs seemed to be built differently from all other sailbacked tetrapods. Romer and Price (1940) found evidence of fractured, and later healed, spines in some Dimetrodon species. This was strongly suggestive that these fairly fragile spines were supported by a web of skin or cartilage. They also had a series of grooves and sometimes spikes, which made them quite different from the more paddle-shaped spines of other sailbacked tetrapods. Were pelycosaurs using their sails for display too?
Probably. As with the dinosaurs it’s hard to imagine that such showy structures did not play a communication role in the lives of these animals. However pelycosaur sails, unlike the sails of all the other taxa we have looked at, show the most evidence for a separate function aside from simply being a billboard. The accessory spines in edaphosaurids suggest increased support for the web of skin/cartilage in the sail as well as increased surface area for blood vessels to permeate. Despite the results of Haack’s (1986) study, the sails of pelycosaurs still hold the most promise as thermoregulatory structures.
Then again we could be looking at this all wrong. Maybe the sails in pelycosaurs were simple showy structures. Rather than a thermoregulatory function, perhaps the extensive vasculature to the sail in these taxa were for pattern generation instead. Their sails may have been more dynamically colourful. Theoretically pelycosaurs still retained the four colour cones that are plesiomorphic to the eyes of all amniotes (mammals later lost two and primates re-evolved a third). As such, pelycosaurs would be more sensitive to nuances of colour. Full on cephalopod quality colour change was probably unlikely, but something akin to what chameleons do, could have been possible. Consider it food for thought. The next time someone feels an inkling to draw a pelycosaur they might want to consider being a bit more extravagant with the sail colours.
Sails in summary
So as it currently stands we can’t really say one way or the other what sails were doing. Early hypotheses about thermoregulation do not seem to agree well with either modeling experiments or observations of extant sailbacked animals. The most likely function for these sails appears to be some kind of signaling, be it sexual, within, or across species. Though the most likely explanation it is also the hardest one to test as it requires a sample size that is currently not available to us in the fossil record.
Though this review of sails across prehistory has not come to any definitive conclusions on sail use I would hope it can at least lay the groundwork for where we have been and where we could one day go. At the very least it would be nice to get a handle on how the sails of extant sailbacked animals are used by their owners. That one step alone could shed light on what advantages sails have to offer.
Until then, these structures will remain rare anatomical novelties.
Novelties that nature sought fit to evolve over and over again.
Bailey, J.B. 1997. Neural Spine Elongation in Dinosaurs: Sailbacks or Buffalo-Backs? JVP. Vol. 71(6):1124–1146.
Bailey, J.B. 1996. Sails or Humps? Inquiry into the Biological Significance of Elongated Neural Spines Among the Dinosauria. GSA Vol. 28(6):28A.
Bittencourt, J.DS., Kellner, A.W.A. 2004. On a Sequence of Sacrocaudal Theropod Dinosaur Vertebrae from the Lower Cretaceous Santana Formation, Northeastern Brazil. Arq. do Museu. Nac. Rio de Janeiro. Vol. 62(3):309–320.
Bramwell, C.D., Fellgett, P.B. 1973. Thermal Regulation in Sail Lizards. Nature. Vol. 242:203–205.
Butler, R.J. Brusatte, S.L. Reich, M. Nesbitt, S.J. Schoch, R.R., Hornung, J.J. 2011. The Sail-Backed Reptile Ctenosauriscus from the Latest Early Triassic of Germany and the Timing and Biogeography of the Early Archosaur Radiation. PLoS ONE. Vol. 6(10):e25693.
Chapman, R.E. Weishampel, D.B., Hunt, G., Rasskin-Gutman, D. 1997. Sexual Dimorphism in Dinosaurs. Dinofest. Int. Proc. p. 83–93.
Charig, A.J. Milner, A.C. 1997. Baryonyx walkeri, A Fish-Eating Dinosaur from the Wealden of Surrey. Bull. Nat. Hist. Mus. Lond. (Geol) Vol. 53(1):11–70.
Coblert, E H. 1989. The Triassic Dinosaur Coelophysis. Mus. N. AZ. Bull. Vol. 57: 1–160.
Cowles, R.B. 1958. Possible Origin of Dermal Temperature Regulation. Evolution. Vol. 12(3):347–357.
Farlow, J.O., Hayashi, S., Tattersall, G.J. 2010. Internal Vascularity of the Dermal Plates of Stegosaurus (Ornithischia, Thyreophora). Swiss. J. Geosci. Vol. 103(2):173–185.
Farlow, J.O., Thompson, C.V., Rosner, D.E. 1976. Plates of the Dinosaur Stegosaurus: Forced Convection Heat Loss Fins? Science, New Series. Vol. 192(4244):1123–1125.
Haack, S.C. 1986. A Thermal Model of the Sailback Pelycosaur. Paleobiology. Vol. 12(4):450–458.
Hone, D.W.E., Naish, D. 2013. The ‘Species Recognition Hypothesis’ Does Not Explain the Presence and Evolution of Exaggerated Structures in Non-Avialan Dinosaurs. J. Zool. Vol. 290(3):172–180.
Hutt, S. Newbery, P. 2004. A New Look at Baryonyx walkeri (Charig and Milner 1986) Based Upon a Recent Fossil Find from the Wealden. SVPCA abstract.
Klaver, C., Bohme, W. 1992. The Species of the Chamaeleo cristatus Group from Cameroon and Adjacent Countries, West Africa. Bonn. Zool. Beitr. Vol. 43(3):433–476.
Knell, R.J., Naish, D., Tomkins, J.L., Hone, D.W.E. 2013a. Sexual Selection in Prehistoric Animals: Detection and Implications. TREE Vol. 28(1):38–47.
Knell, R.J., Naish, D., Tomkins, J.L., Hone, D.W.E. 2013b. Is Sexual Selection Defined by Dimorphism Alone? A Reply to Padian and Horner. TREE Vol. 28(5):250–251.
Knell, R.J., Sampson, S. 2011. Bizarre Structures in Dinosaurs: Species Recognition or Sexual Selection? A Response to Padian and Horner. J. Zool. Vol. 283(1):18–22.
Main, R.P., de Ricqles, A., Horner, J.R. 2005. The Evolution and Function of Thyreophoran Dinosaur Scutes: Implications for Plate Function in Stegosaurs. Paleobiology. Vol. 31(2):291–314.
Nesbitt, S.J. 2003. Arizonasaurus and its Implications for Archosaur Divergence. Proc. R. Soc. Lond. B (suppl.) Vol. 270:S234–S237.
Padian, K., Horner, J.R. 2013. Misconceptions of Sexual Selection and Species Recognition: A Response to Knell et al. and to Mendelson and Shaw. TREE. Vol. 28(5):249–250.
Padian, K., Horner, J.R. 2011a. The Evolution of ‘Bizarre Structures’ in Dinosaurs: Biomechanics, Sexual Selection, Social Selection or Species Recognition? J. Zool. Vol. 283(1):3–17.
Padian, K., Horner, J.R. 2011b. The Definition of Sexual Selection and its Implications for Dinosaurian biology. J. Zool. Vol. 283(1):23–27.
Romer, A.S., Price, L.I. 1940. Review of the Pelycosauria. Geological Society of America Special Paper 28.
Salgado, L., Bonaparte, J.F. 1991. A New Dicraeosaurid Sauropod, Amargasaurus cazaui Gen. Et. Sp. Nov., from the La Amarga Formation, Neocomian of Neuquen Province, Argentina. Ameghiniana
Vol. 28(3-4): 333–346.
Senter, P. 2009. Voices of the Past: A Review of Paleozoic and Mesozoic Animal Sounds. Historical Biology. Vol. 20(4):255–287.
Stovall, J.W., Langston Jr., W. 1950. Acrocanthosaurus atokensis, A New Genus and Species of Lower Cretaceous Theropoda from Oklahoma. Am. Mid. Nat. Vol. 43(3):696–728.
Stromer, E. 1915. Ergebnisse der Forschungsreisen Prof. E. Stromers in den Wüsten Ägyptens. II. Wirbeltier-Reste der Bahariye-Stufe (unterstes Cenoman). 3. Das Original des Theropodes Spinosaurus aegyptiacus nov. Gen., nov. Spec. Abhandlungen der Königlich Bayerischen Akademie der Wissenschaften, Mathematisch-Physikalische Klasse, München Vol. 28:1–28.
Sues, H-D., Frey, E., Martill, D.M., Scott, D.M. 2002. Irritator challengeri, A Spinosaurid (Dinosauria: Theropoda) from the Lower Cretaceous of Brazil. JVP. Vol. 22(3):535–547.
Taquet, P. 1976. Geologie et paleontologie du gisement de Gadoufaoua (Aptien du Niger). Cahiers de Paleontologie, Centre National de la Recherche Scientifique Paris, p. 1-191.