Gallery and print store

Thursday, 24 December 2015

Dinosaur scales: some thoughts for artists

Turns out that Triceratops horridus had some of the coolest scales of any dinosaur: huge, interlocking tubercles with low bosses and spikes. No other dinosaur has skin like this - at least, not without supporting osteoderms. But what are dinosaur scales actually like, and are we depicting them accurately in our art?

The discovery that many Mesozoic dinosaurs were superfuzzyfilamentouspinyalidocious has been an major influence on contemporary Mesozoic palaeoart. This has affected more than just how we depict the gross appearance of dinosaurian subjects, but also our attitudes to their behaviour, demeanour and place in the Mesozoic world. I've written a fair bit about scientific and artistic attitudes to filamentous dinosaurs and joined choruses arguing that it's important to get these new depictions 'right': we want to see filaments of appropriate morphology, size and distribution in reconstructions of these animals.

In light of this, it's a little peculiar that we have slightly more lax attitudes to how we reconstruct scaly integuments in these animals. We have some truly spectacular skin impressions from scaly dinosaurs which provide a wealth of information about their detailed appearance, and yet many of our reconstructions incorporate little of this data. Instead, we often create 'generically' scaly or wholly speculative integuments. Common issues include rendering of scales of homogenous size and shape across an entire animal, showing little difference in scalation between species, and issues with the size, proportions and shape of individual tubercles. Other times, and most egregiously, some individuals understate just how good the records for scales in certain species are, this seemingly giving license to render a more speculative, but flamboyant body covering. It's not just amateurs making these mistakes and, in the interests of not being a hypocrite, I'll state early on that I'm guilty of some of these issues in my own work.

With this in mind, I want to see out 2015 with a fresh look at four exceptionally interesting samples of dinosaur scales, providing something of a refresher for myself and other about scaly dinosaur integument and food for thought on restoring these animals. The amount of scaly skin we have from dinosaurs means this list could easily comprise 10 or even 20 examples, but for the sake of brevity and detail I'm keeping the count low. The specimens here may be familiar to veterans of dinosaur literature, but I hope to cover them in sufficient detail that much of this information will be new to many readers.

The Carnotaurus holotype skin impressions

Outside of the feathered coelurosaurs, substantial remains of theropod dinosaur skin are pretty rare. There are lots of scraps, many of which are only cautiously referred to Theropoda, but large pieces of skin associated with specific skeletons are very thin on the ground. These circumstances make the extensive scaly skin impressions known from the Late Cretaceous Carnotaurus sasteri type specimen quite special. This specimen is already impressive: described in detail by Bonaparte et al. 1990, it comprises a near complete skeleton missing only parts of the legs and end of the tail. The fact this specimen also preserves a host of skin remains means Carnotaurus is an especially well represented large theropod. Many readers will know the skin remains associated with this specimen makes it quite integral to debates over the ancestral state of dinosaur and theropod skin. As one of the few relatively 'basal' theropods known with decent skin remains, Carnotaurus has quite a bit of sway in discussions about filament development in theropods.

Illustration of the tail base Carnotaurus skin impressions from Bonaparte et al. (1990). The deep grooves in the specimen represent topography of the associated axial skeleton, in this case the haemal arches. Scale bars represent 10 cm.
The skin remains of Carnotaurus are a little patchy, but represent many different parts of the body: the anterior neck, shoulder girdle, mid-torso, and the base of the tail. The skull also bore skin impressions before they were accidentally prepared away. The largest piece of skin covers the tail base, and is figured above. A huge amount of detail can be seen across the various skin pieces. They have a relatively uniform texture, each piece showing a mix of two scale types. The most obvious are the large, 4-5 cm diameter tubercles which protrude slightly from the rest of the skin. Instead of being randomly arranged, these are spaced regularly from each other at roughly 10 cm intervals, separated by large numbers of relatively tiny, 5 mm wide scales. The larger tubercles bear something of a keel, but the smaller structures are quite featureless. Parallel furrows with vertical orientation, perhaps representing creases, are impressed into the mosaic of smaller tubercles, but do not seem to leave an impact on the larger structures. Figures in Bonaparte et al.'s (1990) description suggest that this general skin texture extends right the way around the tail - the reduction in tubercle size and density on the ventral surface commonly seen in artwork is erroneous in this respect.

For artists, the Carnotaurus skin impressions enable us to 'connect the dots' as goes the appearance of this dinosaur's hide. It seems scales were present from skull to tail base, and it doesn't seem much of a stretch to assume most or all of the animal was scaly. There are a few reconstructions of extensively filamentous Carnotaurus out there but, sorry guys, this just doesn't jive with what we know of the skin of this animal. It also seems we shouldn't be drawing Carnotaurus with obvious differences in skin texture across the body - it looks pretty homogenous in the fossils. Also noteworthy is the size of most of the scales. It seems we'd only notice the larger, keeled tubercles and furrows on this animal unless we were standing very close. Those 5 mm tubercles might perhaps register as mottled colouration, but I doubt anyone without superhuman vision could distinguish each scale from afar. Note that Carnotaurus is not unusual in this respect - a lot of dinosaurs had much smaller scales than we show in our illustrations.

The Howe Quarry diplodocids

One of the most striking components of the 1999 Walking with Dinosaurs Diplodocus reconstruction was the tall dermal spines adorning the midline of the animal. These structures were not the idle fantasy of sculptors and artists, but actually based sauropod skin fossils from Howe Quarry, a famous Wyoming Jurassic locality. Described by the late palaeoartist Stephen Czerkas in 1992, these finds are frequently discussed by palaeoartists because sauropod skin impressions are extremely rare. The impressions are associated with incomplete skeletons representing animals from 2-3 to 14 m in length, with some skin pieces being exceptionally large at 25 x 75 cm. Unfortunately, Czerkas (1992) did not identify the remains of these animals. Howe Quarry yields at least one named diplodocid, the recently named Kaatedocus siberi, but it remains to be established if these scaled remains represent the same taxon.

The Howe Quarry diplodocid skin can be described as tessellating hexagonal scales with a rough surface, each about 3 cm across. There is no sign of these scales being divided by differently sized scales to form a pattern like those seen in Carnotaurus. The roughened texture of each scale is formed by small (2-3 mm) tubercles dotted across each large scale. As noted by several authors, this morphology is reminiscent of other examples of sauropod hide and seems common to at least Neosauropoda (e.g. Foster and Hunt-Foster 2011; Upchurch et al. 2015). As a rule, sauropods must've been quite rough to the touch.

Illustrations of the Howe Quarry diplodocid spines from Czerkas (1992). Top row, illustrations of specimens as preserved; bottom, interpretative drawings and reconstructed outlines. Scale bars equal 5 cm.
The truly exceptional part of the Howe Quarry diplodocid skin remains are the 14 subconical structures found dotted amongst the sauropod skeletons (above). Some were isolated, but several of these structures were found in connected rows. Perhaps the most significant of these were associated with a skin impressions wrapped around the tail base of one individual. It's from these remains that we can deduce that they were arranged in a row along back of the animal. This might seem like a minor feat, but - as anyone who's attempted to reconstruct stegosaur or titanosaur osteoderm arrangements might attest - being confident about the arrangement of extraneous pieces of dinosaur integument is nothing to be sniffed at. These cones vary quite a bit in size and shape. The largest, estimated at 18 cm tall when complete, seem to stem from the proximal end of the tail, but those of the distal end are smaller. Some cones are quite tall and straight, others blunter and recurved. The tips of all the cones are flattened laterally, but the bottoms more or less round in cross section. As with hexagonal scales on the body, these spines bear small tubercles across their surface. That these were purely comprised of the dermal tissues, and not osteoderms, is confirmed by the total absence of bone from any of the cones. Quite how far these conical structures extended across their owner's bodies cannot be said from the known remains, nor should we feel confident that we have the full spectrum of size or morphological variation of the spines (Czerkas 1992).

The detail and specificity of the Howe Quarry specimens give artists an atypically good insight into the appearance of these sauropods, and remain significant specimens or this reason. But as cool as this all is, the Howe Quarry skin specimens could be more useful. For instance, it is not clear how large each sauropod individual with associated skin remains was, and it's thus not clear how large those spines or scales were in comparison to each specific animal. The range of body lengths for the Howe Quarry specimens (2-3 -14 m) perhaps indicates that the scales of these animals (3 cm across) might be larger against body size than those of most other dinosaurs, but how visible they might be to observers is really dependent on knowing the sizes of the animals concerned. Likewise, the only published illustrations of these unique, interesting remains are pretty basic: it would be neat to get these specimens figured and described in a lot more detail. Hopefully, these details will be forthcoming soon.

The Sternberg/Osborn Edmontosaurus mummy

You can't discuss scaly dinosaurs without mentioning hadrosaurs. Research into hadrosaur skin is only second to that going into the fuzzballs at the other end of the dinosaur tree, there being so many skin impressions from these dinosaurs that we can gauge variation between species, see pathological skin tissues, and reconstruct virtually complete integuments for some taxa. This relative glut of data has spurned investigation into just why hadrosaur skin crops up so often. The exact cause remains elusive (it's seemingly unrelated to the rocks they occur in, nor their palaeoenvironmental or palaeoclimatic preferences), and it is suspected that there is something intrinsic to their skin anatomy which makes it more preservable (Davies 2012).

The amount of data we have for hadrosaur skin is really impressive. Here, in grey, you can see the skin impressions known for several hadrosaurid taxa: A, Brachylophosaurus canadensis; B, Edmontosaurus annectens; C, Gryposaurus notabilis; D, Maiasaura peeblesorum; E, Saurolophus angustirostris; F, Saurolophus osborni; G, Corythosaurus casuarius; H, Lambeosaurus lambei; I, Lambeosaurus magnicristatus; J, Parasaurolophus walkeri. From Bell (2014).
Even among hadrosaurids, Edmontosaurus annectens stands out as having particularly exemplar skin remains. Collectively, we have skin impressions from virtually its entire body (above). One of the most spectacular Edmontosaurus fossils with scaly remains has to be the "Trachodon mummy", discovered by George Sternberg (Charles Sternberg's son) in 1908 and described by Henry Fairfield Osborn in 1912. Osborn lavished attention on the integument of this near complete, fully articulated specimen, of which skin impressions covered the posterior jaws, neck, shoulders, chest, belly and forelimb. This specimen also revealed the presence of a low frill along at least the posterior part of the neck. Osborn's work on this animal stands out as a landmark document on extinct reptile integument, and interested parties really should download this article from the American Museum of Natural History here (NB. this is a 75 Mb download, it coming bundled with historic descriptions of the skulls of Tyrannosaurus and Allosaurus, whatever they are).

Pectoral (lower) and manual (upper) skin remains from the "Trachodon mummy" specimen. Notice the scales extending onto the unguals - these animals did not have nails or claws on their hands. From Osborn (1912).

Osborn's description revealed details of dinosaur skin which were, at that time, poorly known from other animals. He remarked on how thin the skin layer was and the remarkably small size of the scaly tubercles covering the body (1-5 mm). The fineness of the skin resulted in perhaps a third of it being accidentally destroyed during collection - 'dinosaur mummies' were an unknown quantity before this specimen, and collectors had no idea such data was at risk when skeletons were being uncovered. Edmontosaurus skin was a mosaic of larger and smaller tubercles, but their size variation is more continuous the obviously bimodal configurations of other species. The smaller (1-3 mm) tubercles were rounded structures located between larger (5-10 mm) hexagonal ones. Osborn called these 'pavement scales', and noted that they occurred in small (5-10 cm wide) clusters in some areas, such as the neck, inner surface of the arm and belly, but covered entire other parts of the body, such as the side of the chest, lateral surface of the arm and above the hips. The largest pavement scales, about 10 mm wide, occur on the lateral surface of the arm and tail. Both large and small scales occur on the frill (below). Folds, creases and smaller tubercles seem to correspond with intervertebral spaces, likely reflecting where these tissues flexed and creased with neck movement. The actual height of the frill is unknown from this specimen, the free margin being damaged during collection.

Osborn's illustration of the frill of Edmontosaurus. From Osborn (1912).
We could go on as there's so much detail on this specimen, but you're better off just checking out Osborn's description. He certainly provided lots of interesting details for artists: a visual summary of the distribution of larger and smaller scales in a cartoon hadrosaur (below), comments on his collaboration with Charles Knight to produce a 'trachodont' reconstruction in line with his new information on hadrosaur skin (also below), and even speculation on how pigmentation may pertain to the scale pattern. Of further interest is Osborn's comparison of the skin of Edmontosaurus with other hadrosaurs, this noting that the scales of his mummy specimen were a lot smaller than those of other, closely related animals. Other differences in hadrosaur skin texture has become even more apparent in subsequent years.

Left, Osborn's illustration of Edmontosaurus outlining the distribution of large scale clusters, with their size much enhanced for visibility; right, Charles Knight's iconic 1912 painting of the same taxon, an artwork produced in collaboration with Osborn and data from the "Trachodon mummy". From Osborn (1912) and The World of Charles R. Knight.

So, other than the obvious take-home - that we know a heck of a lot about the skin of Edmontosaurus -are there any obvious pointers for artists here? As noted for Carnotaurus above, it's doubtful that we'd be able to define individual scales or the patchy distribution of pavement scales on this large bodied (12-13 m long) species unless we were right next to it. Secondly, of all dinosaurs, surely this is one species to consider off limits to extensive filamentation. I suppose you could argue that filaments filled the few parts of this animal's hide left unrepresented in the fossil record, but that fuzz is going to look like weeds growing through a pavement if you're paying attention to where we know scales were. I also think it's worth paying attention to what Osborn meant by 'frill' along the back of this species: it does not appear to be a narrow, fibrous structure as commonly depicted, but a scaly continuation of adjacent dermal tissues.

The (unpublished) Triceratops superscales

I've saved what I consider to be one of the most interesting and impressive set of scale impressions for last, even though they are represented by specimens which have only currently received only very superficial publication through online news articles. These specimens belong to one of the most familiar and famous dinosaurs of all, the ceratopsid Triceratops horridus, and yet they demonstrate a scale topography completely unlike that of any other dinosaur. Their discovery is a particularly fun curve-ball because we have skin samples from a number of other ceratopsians, none of which are particularly like those now known for Triceratops. I'm reminded about earlier discussions of 'one skin fitting all': it seems ancient dinosaurs really could be just as varied in skin morphology as modern animals.

Huge patch of Triceratops skin, preserved as an internal mould - look at the size of the individual scales! Borrowed from the Rapid City Journal.

These extensive skin impressions were associated with one of the most complete Triceratops specimens ever found, a Wyoming individual known as 'Lane'. This specimen, including its skin, is now on display in the Houston Museum of Natural Science. Without a full description it's a little difficult to give much in the way of specifics about the skin, but published photographs reveal a network of very large (I'm estimating 50-60 mm wide based on the adjacent images) hexagonal tubercles dividing larger tubercles (perhaps c. 100 mm) with central, conical projections. These large scales are sometimes described being as 'nipple-like', for obvious reasons. Divisions between these tightly interlocked scales are marked, and we might have been able to distinguish individual scales on these animals from some distance away. The function of the larger tubercles with their prominences has been the source of much speculation in art - do these structures represent bosses and low spikes, or tubular supports for large, coarse filaments? I must admit to considering the latter unlikely as neither hair or scales in modern animals grow through scales, but instead around them. I'm happy to be wrong on this, though, and both interpretations could be easily tested by looking for apertures at the tip of each prominence. Hopefully these specimens will get a full write up soon, which might provide such details.

Detail of the large tubercles adorning the outside of Triceratops. Also borrowed from the Rapid City Journal.
Lane's skin impressions suggest that the scales of Triceratops were characteristically coarser, certainly a lot larger and perhaps more sculpted than those of most other dinosaurs. Their overall appearance is very different to the hadrosaur and theropod skin mentioned here, contrasts markedly from the scales known from other ceratopsians, and is rather unexpectedly most similar to the scales of sauropods. It's difficult not to intuitively equate Triceratops skin with that rhinos and armadillos: there's something almost armour-like about those heavy scales and low, projecting bosses. Perhaps this chimes with the unusually solid, reinforced cranial frill we find in this species - was Triceratops something of a horned dinosaur tank? I reckon there's a lot of fun to be had with depicting this animal as looking particularly tough and grizzled, with big skin creases and heavy folds - such a depiction can be seen at the top of this article. It's perhaps worth noting that the actual appearance of Triceratops is not a million miles off the Charles Knight's famous painting of 'Agathaumas' (probably = Triceratops) with its speculative heavy scaling.

Summary time

I hope what's becoming clear here is that we can obtain quite a lot of information from dinosaur skin impressions, and that they show scaly dinosaur species have their own characteristic integuments in the same way that filamentous ones do. There really doesn't seem to be a 'standard' type of dinosaur scale, and even closely related species show some significant variation between them. We have to conclude that those of us hoping to restore these animals accurately really need to pay close attention to these data, considering variation in tubercle size, texture and distribution. I particularly emphasise this for artists who draw every scale: if that's the route you're taking, make sure you're drawing them correctly! Moreover, the specimens outlined here are good reasons to be inventive when skin impressions are lacking. It seems most relatively extensive skin impressions of scaly dinosaurs reveal things like spines, keeled scales, armour-like structures, frilled projections and so on. Mesozoic dinosaur skin must've been as interesting as that of modern reptiles, and we might expect many species to have elaborate structures of some kind.

And that's it for 2015

OK folks, we're done here for this year, but there's plenty more to come in 2016. Weird archosauromorphs, stem mammals, some retropalaeoart and the publication of Recreating an Age of Reptiles will be covered early on. Huge thanks to everyone who's been reading and supporting this blog throughout 2015 - I hope you've enjoyed what I considered to be one of my best blogging years so far. All the best to you all for the festive period, and see you all in 2016!


  • Bonaparte, J. F., Novas, F. E., & Coria, R. A. (1990). Carnotaurus sastrei Bonaparte, the horned, lightly built carnosaur from the Middle Cretaceous of Patagonia. Contributions in Science. Natural History Museum of Los Angeles County, 416, 1-42.
  • Czerkas, S. A. (1992). Discovery of dermal spines reveals a new look for sauropod dinosaurs. Geology, 20(12), 1068-1070.
  • Davis, M. (2012). Census of dinosaur skin reveals lithology may not be the most important factor in increased preservation of hadrosaurid skin. Acta Palaeontologica Polonica, 59(3), 601-605.
  • Osborn, H. F. (1912). Integument of the iguanodont dinosaur Trachodon. Memoirs of the American Museum of Natural History v. 1

Friday, 11 December 2015

The lifestyle of Tanystropheus, part 2: coastal fisher or first-day-on-the-job aquatic predator?

The new Tanystropheus cf. longobardicus skeletal reconstruction I presented in my last post. What the dickens did this crazy animal do? That's what we're discussing today.
What sort of animal was the Triassic, long-necked Eurasian protorosaur Tanystropheus? As we discovered in the last post, the lifestyle of Tanystropheus remains controversial over a century after it was first discovered. There is near universal agreement that it ate swimming prey such as fish and squid, but opinion is divided over whether it was obligated to aquatic, swimming lifestyles because of the burden of its long neck, or whether it was a water margin specialist that plundered small prey from shorelines. Previously, we discussed a core argument for the aquatic hypothesis, that the Tanystropheus neck would over-balance the animal. Calculations presented in the last post suggested that the mass distribution of Tanystropheus is not as weird as we might think, and certainly less so than than that of another group of long necked reptiles we are confident lived out of water, the azhdarchid pterosaurs. Based on this very basic test, I expressed some scepticism about the neck being simply too heavy to permit a terrestrial existence.

In this second discussion, I want to look at some finer aspects of Tanystropheus anatomy and palaeontology, how they've been interpreted, and what they might mean for its lifestyle. There are several areas which are relevant here: what we know of Tanystropheus diet, the palaeoenvironmental context of Tanystropheus fossils, aspects of tail and limb anatomy, and of course, the functionality of its neck. There's a lot to get through here, so let's not waste any more time on preamble.

Fossil record

An obvious line of inquiry about ancient animal habits is the palaeoenvironmental bias of its fossil remains, and the fossil organisms it is found with. We mentioned last time that Tanystropheus was a wide-ranging taxon, occurring across Europe, Israel and China in locations representing coasts and shallow waters around the ancient Tethys ocean. About half of Tanystropheus fossils come from shallow marine settings, the rest being derived from more coastal environments: river and estuarine environments, lagoons, intertidal settings and so forth (for a brief overview, check out the Fossilworks entry on this animal: there's a few localities missing, and the 'terrestrial' occurrence of Tanystropheus there is erroneous, but it gives a flavour of its depositional context). We often find marine fish and seagoing reptiles in the same beds as Tanystropheus, but it also occurs alongside terrestrial or freshwater species such as temnospondyls, terrestrial reptiles, stem mammals and plants in a number of locations. The link of Tanystropheus to these faunas seems complex: in one locality, Tanystropheus fossils only occur in horizons containing a mix of highly terrestrial and highly marine reptiles, without many 'intermediate' semi-aquatic species (Renesto 2005). Because Tanystropheus was likely not adapted for a truly seagoing lifestyle, this has been argued as evidence of it being part of a terrestrial community rather than a marine one (Renesto 2005).

Collectively, it seems difficult to argue a strong terrestrial or marine bias in this record. Tanystropheus seems to have lived in or around aquatic environments, maybe with a bias to those under marine influences, but it does not seem a stranger to brackish or freshwater settings either. There is perhaps something of a skewed association with marine animals, but it occurs with enough 'terrestrial' forms to keep the idea of a coastal fishing lifestyle buoyant. It would be interesting to put some actual numbers on this and see how commonly associated with terrestrial influences Tanystropheus is, or whether a couple of sites are skewing our perception of data. Maybe that's a job for another blog post - until then, we probably need to look at other sources of information for clearer lifestyle indications.

Gut content

The idea that Tanystropheus ate swimming prey is verified by the association of digested fish remains and cephalopod hooks in the gut regions of articulated specimens (Wild 1973; Li 2007). The latter is sometimes considered smoking gun evidence for the swimming Tanystropheus lifestyle hypothesis, it being reasoned that cephalopods are exclusively marine animals, mostly found far out to sea, and unlikely to be eaten from land (e.g. Nosotti 2007).

A number of heron species, including the globally distributed black-crowned night heron (Nycticorax nycticorax), are known squid-eaters. Image from Wikimedia (CC ), by Kuribo.
The fossilised squiddy gut content of some Tanystropheus specimens certainly matches the idea of a marine-influenced lifestyle, but several non-marine, and sometimes non-aquatic, birds and mammals challenge the idea that it had to be a swimming animal to have obtained them. Examples include night herons (Hall and Cress 2008) and several types of mustelid (e.g. Hartwick 1983; Beja 1991). Exactly how night herons obtain squid is not documented in detail, but photographs of two other heron species demonstrate squid can be apprehended without venturing out to sea, or even into deep water. As might be expected, cephalopods also frequently wash up on beaches (sometimes still alive, and in huge numbers) allowing animals such as bears and wolves to also access cephalopod meat. Humans are also adept predators of squid in coastal settings. Shore-based squid angling is reportedly a growing hobby around the world (and apparently requires only very basic fishing equipment) and we routinely collect cephalopods from intertidal environments for use as bait or cooking ingredients (Denny and Gains 2007). Contrary to expectations, accessing cephalopod prey from shore environments appears quite possible for a number of differently adapted species. It seems premature to rule out a coastal fishing lifestyle for Tanystropheus just because it sometimes ate squid-like animals.


One of the most famous and complete Tanystropheus longobardicus specimens known, MSNM BES SC 1018. This illustration is from Nosotti's huge (2007) monograph.
With the fossil record and gut content providing slightly ambiguous insight into Tanystropheus habits, its functional anatomy is probably going to be a deciding card here. A lot has been said about the functional morphology of Tanystropheus, and there is a lack of consensus on many issues. For instance, its neck flexion has been described as almost 'swan-like' (Wild 1973); broom handle-stiff (Tschanz 1988), or somewhere inbetween (Renesto 2005). Its tail has been considered lousy for aquatic propulsion by some (Wild 1973; Renesto 2005) but well suited for the job by others (Tschanz 1988; Nosotti 2007). Clearly, some of these ideas must be erroneous, them being too polarised for all contributing parties to be correct. Such confused functional interpretations are not without precedent: Darren Naish and I noted a similar situation with azhdarchid pterosaurs in our 2008 paper: maybe this is simply what happens when we try to understand weird fossil species.

The main points of contention about Tanystropheus functional anatomy concern its tail, limbs and neck. We might link these attributes to two principle functions: locomotion and foraging. Let's start with the former. Proponents of the aquatic Tanystropheus hypothesis suggest the tail was the likely propulsive organ, it being considered that the limbs are too long and gracile to function as effective paddles (Tschanz 1988; Nosotti 2007), even if the foot might have some aquatic adaptations (below; Kuhn-Schnyder 1959; Wild 1973). Near 'horizontal' articulations between the posterior trunk and tail vertebrae appear to have permitted this part of the body to undulate laterally, permitting a crocodile-like sculling approach to swimming.

Soft-tissue preservation around the tail of Tanystropheus cf. longobardicus specimen MCSN 4451. We're looking at the underside of the tail in the left of the image here - note the width of the soft-tissue (the big grey mass). The verts on the right are shown in left lateral view. From Renesto (2005).
A fly in the ointment here is the gross tail anatomy of Tanystropheus. Rather than being long, and comprised of the robust, tall vertebrae expected of a sculling aquatic reptile, its tail is slender, relatively short and actually broader than tall - hardly an ideal sculling organ (Renesto 2005). This fact has been noted by proponents of the swimming lifestyle hypothesis, and it has been proposed that the tail sported some sort of fin to modify it into a swimming organ (Nosotti 2007). Well, maybe, but this idea is entirely without support from fossil data. Readers may recall that marine reptile workers have been quite ingenious in their ability to detect fins and flukes from osteological correlates, none of which are obvious in the tail of Tanystropheus. Moreover, preserved soft-tissues from the anterior Tanystropheus tail region (above) show no signs of fins but instead a broad tail base poorly suited to aquatic propulsion (Renesto 2005). Also worth mentioning is recent work on the relationship between vertebral articulation and swimming capability in crocodyliforms. They can reflect sculling behaviour, but articulations like those seen in Tanystropheus can also be linked to preventing trunk collapse during non-aquatic locomotion (Molnar et al. 2014). We could go on, but I think the point has been made that arguments for the Tanystropheus tail being a swimming organ are, at best, not without complication, and, at worst, uncompelling.

Turning our attention to the limbs, I mentioned in the last post that I was surprised how 'leggy' Tanystropheus was when restored as walking rather than, as we're used to seeing it, squatting. The limb proportions and girdle sizes of Tanystropheus compare well with non-aquatic protorosaurs such as Macrocnemus and Langobardisaurus (e.g. Renesto 2005; Nosotti 2007) and, as alluded to above, it is immediately clear that these limbs are not flippers. Not only are they too long and gracile for effective use as hydrofoils, but their long bones are hollow - unexpected features of an aquatic animal. Another protorosaur - Dinocephalosaurus - gives an insight into how these reptiles could modify their limbs into efficient flippers (below), and, without going into detail, they're nothing like the limbs of Tanystropheus (see Renesto 2005 for a long discussion of this). Tanystropheus limb joints are mostly robust and well-defined (but see below), and its hands and feet are strongly built and compactly structured. Some differences between hand and foot proportions can be seen: the hands are short, the feet rather long, and the latter characterised by a peculiarly long first bone in the fifth toe. The limb girdles are well developed, looking proportionally comparable (speaking from pure eyeballing here, not precise measurements) to those of large monitor lizards and crocs. I find the shoulder blade of particular interest, as it is rather large and broad, subequal in proportions to the coracoid (the lower portion of the shoulder girdle). This contrasts with many aquatic animals, which tend to maximise the size of the coracoids while reducing the scapula.

Variations in protorosaur limb anatomy, demonstrated by the aquatic Dinocepahlosaurus (A-B) and Tanystropheus (C-D). Note how both the arm (A) and leg (B) of Dinocephalosaurus are short and wide compared to their equivalents in Tanystropheus (forelimb = C, hindlimb = D), making them much more effective flippers. You can also see the reduced mineralisation in the Tanystropheus wrist here. From Renesto (2005).
I have to agree that Tanystropheus limbs were probably unchallenged by non-aquatic habits (Renesto 2005) and, if this were any other species, I don't think we'd be disputing the fact that its limbs were likely capable of terrestrial locomotion. That said, there are undeniably some hints that Tanystropheus was not always walking on land. Several authors have noted that the wrist and ankle bones of Tanystropheus are not as well ossified as those of other protorosaurs (e.g. Rieppel 1989; Nosotti 2007), and some have suggested that the pelvic bones may also be somewhat less defined (Rieppel 1989). Moreover, the elongation of the fifth toe is atypical for a purely terrestrial reptile, but common among aquatic creatures (see Kuhn-Schnyder 1959 for a good illustration of this point). Proposals that this made the foot somewhat paddle-like, or supported Tanystropheus on soft, saturated substrates do not seem unreasonable. These are fairly minor modifications to the skeleton when viewed overall however: the reduced ossification in the wrist, ankle and pelvis is pretty minor - especially when we consider how cartilage-filled the joints of many giant terrestrial archosauromorphs can be (Holliday et al. 2010) - and the reconfiguration of foot bones do not override the otherwise elongate, gracile structure of the hindlimb. My overall interpretation of these limbs broadly agrees with that proposed by Renesto (2005): a bauplan suited to terrestrial locomotion with some aquatic leanings, rather than sustained aquatic propulsion.

Finally, we come to the neck. I've saved discussion of this for last because I consider much of its anatomy so significant to Tanystropheus habits that discussing it earlier might have rendered other points a bit superfluous. We make a lot of noise about how strange the neck of this animal is, but Tanystropheus neck anatomy frequently converges with those of other long necked reptiles - pterosaurs and sauropods - and even some long-necked mammals. That doesn't necessarily make it less weird - it's definitely still an 'extreme' biological structure - but does help us put its neck anatomy in perspective with other animals, as well as highlighting significant adaptive differences to neck elongation in aquatic and non-aquatic species.

As with pterosaurs and sauropods, Tanystropheus went to great lengths to lighten its neck. Firstly, its neck is comprised of relatively few (13), slender vertebrae rather than dozens of short ones (see Rieppel et al. 2010 for discussion of cervical vertebra counts in this animal). This is about half as many as some other protorosaurs had (Reippel et al. 2008), and a far cry from the vertebral counts of some dinosaurs (including birds). A low vertebral count reduces the number of heavy joints and muscle attachments in any part of the axial column, so this is a good basis to having a lightweight neck. More weight was lost through hollowing the bony core of each vertebra, a condition Tanystropheus took so far as to need bony struts supporting the interior cavities of each vertebra. Note that there is no evidence that these bones were pneumatised, seemingly lacking openings through which airsacs could penetrate the bone walls. However, simply removing bone - one of the densest, heaviest materials in our bodies - would still throw out a lot of weight. The neck was likely lightly muscled, the mid-series vertebrae being long tubes with highly reduced processes for muscle anchorage (below). In many respects, the vertebral bodies are similar to those of azhdarchid pterosaurs. The role of these tubular, slender mid-series neck vertebrae is confusing at first, but they make a bit more sense once we realise that most terrestrial animals control their necks via musculature anchoring to the top and base of the neck. This was likely true for Tanystropheus and azhdarchids because anterior and posteriormost neck vertebrae are the most complex parts of the neck skeleton, presumably reflecting attachment of more muscles in these regions. We might therefore assume their necks worked in a broadly similar to those of modern animals, weird as they are.

Three dimensionally preserved mid-series Tanystropheus vertebra described by Dalla Vecchia (2005).
The seemingly lessened set of neck muscles on the Tanystropheus neck would likely limit neck performance (i.e. the size of prey that could be lifted into the air) but, again, would facilitate weight reduction. Strong, restricting joints between the majority of the neck bones and bundles of elongate cervical ribs aided reduction of musculature further, passively resisting inter-vertebral movements which otherwise require muscle action or thick ligaments to control. Elongation of cervical ribs provides another bonus for mass reduction, this trait being linked to shifting muscles down the neck in sauropods and thus lightening the cranial end (Taylor and Wedel 2013). With passive support structures in place, muscles operating around the neck base may have been able to support and move the neck quite easily. Indeed, areas where neck elevator muscles (such as levator scapulae and the trapezius) anchor on the shoulder blade are unusually broad and well developed in Tanystropheus compared to other protorosaurs, and certainly a lot larger than those of long-necked aquatic animals (Araújo and Correia 2015). These are useful muscles to emphasise if you're looking to economise neck mass, being able to both lift and turn the neck by simply varying the symmetry of their activation. We also see a good set of short, robust cervical ribs and broad coracoids at the base of the neck, anchoring muscles related to strong downward neck motion (unless Tanystropheus differed from all other tetrapods). As Mark Robinson preemptively commented on my last post, this is starting to sound a lot like a mechanical crane: a lightweight, strong beam operated by long muscles and ligaments (cables and pulleys in our analogy) from a powerful, mobile base. Quite how much motion was possible at the neck base is debated, but the fact that a number of articulated Tanystropheus specimens are preserved with distinctly elevated neck bases suggests it was more flexible than the rest of the neck, and perhaps capable of a large range of motion (Renesto 2005). This, of course, has implications for balance: if the neck could be drawn up as in fossil specimens the centre of mass would be quite far back in the body (see the last post for more on Tanystropheus mass distribution).

To me, this is all sounding quite sauropod- and azhdarchid-like: an economically constructed neck capable of somewhat limited, but sufficient motion to procure food in terrestrial habits, albeit food that doesn't put up too much of a fight. By contrast, the Tanystropheus neck compares poorly to those of long necked aquatic animals. For one, we expect a large number of short vertebrae in long-necked aquatic species, this permitting greater numbers of muscles working on the neck skeleton. Aquatic animal neck bones are frequently expanded to enlarge the size of muscles attaching to them, these being required to move any long appendage through water. This makes for a heavy neck, but perennial aquatic support renders this a moot issue. Indeed, weight is often a commodity in water rather than a problem, it providing ballast against air-filled lungs or positively buoyant tissues - it's widely known that swimming tetrapods often have entirely solid bones to increase their mass further. The neck of Tanystropheus doesn't really match any of these features. While the number of neck bones is somewhat increased compared to other protorosaurs, the aquatic Dinocephalosaurus has almost twice as many more - 25 - in a neck of similar proportions. Tanystropheus neck length is mainly achieved by stretching each vertebra tremendously, the addition of another three vertebrae perhaps merely being a secondary measure to boost neck length overall (birds and sauropods do the same thing - adding more neck vertebrae is not strictly an aquatic adaptation). Reduction of neck mass in Tanystropheus neck (and limb) bones is also at odds with expectations for an aquatic animal, the hollow cores, stiffened joints and posterior displacement of musculature being unnecessary and even disadvantageous in an aquatic setting. It's actually hard to imagine the neck of Tanystropheus being pulled through water efficiently at all, the reduced muscle profile and long vertebrae being quite problematic and under-powered for this task. It certainly does not seem well suited to chasing and grabbing fast moving aquatic prey such as squid and fish. To me, Tanystropheus neck anatomy just seems to make a lot more sense out of water and, given how much emphasis Tanystropheus put on its neck tissues, I think this is a pivotal consideration when attempting to understanding its lifestyle.

Summary time: a twist in the tale?

Let's sum up these three lines of discussion. The fossil record of Tanystropheus suggests we could find it in a variety of aquatic settings - we might average these out to say it was a denizen of coastal and nearshore environments. It clearly had a taste for seafood, although we need to be careful not to over-state what this might mean about its lifestyle. Anatomically, it seems its propulsor apparatus is best suited to non-aquatic settings and that strange neck finds overwhelmingly superior comparison to terrestrial tetrapods than it does aquatic ones. I therefore have to agree with pretty much everything said about Tanystreopheus anatomy reflecting a 'coastal fishing' habit rather than a strictly aquatic one (e.g. Renesto 2005). I actually really struggle to see how this animal can be argued as a swimming predator, and note that even proponents of this lifestyle acknowledge that Tanystropheus must have been a sluggish, ineffectual aquatic animal, limited to ambushing prey from darkness (e.g. Nosotti 2007). This brings us to a twist to our Tanystropheus story: acknowledging some big issues with the Tanystropheus swimming hypothesis, Nosotti (2007) proposed that it was a newcomer to the aquatic realm, still carrying a lot of anatomical baggage from terrestrial ancestors. It doesn't look much like an aquatic animal because, in evolutionary terms, it's Tanystropheus first day on the job and it's still learning the adaptive ropes for being a successful marine predator.

My preferred lifestyle interpretation for Tanystropheus: a Triassic croc-o-heron which snatched prey from shorelines and promontories around coastal waterways. Note the animals perched on rocks out to sea - I have no problem with this animal swimming per se (as noted above, there is reason to think it was somewhat aquatically adapted), I just don't think it lived in water.
Personally, I find this sort of argumentation weak. It implies Tanystropheus is somehow exempt from relationships between morphology and function well established in other animals, and seems like an excuse to dismiss contrary evidence more than it does a robust hypothesis. It elevates the proposal of aquatic Tanystropheus to a foregone truth of its palaeobiology and structures other lines of evidence around that, which I cannot see as a positive approach to these sort of palaeontological investigations. To my mind, Tanystropheus taphonomy, gut content and functional anatomy are fully consistent with it being a Triassic variant on a heron, an animal which struck at swimming prey while supported on land or in bodies of shallow water. Its smattering of minor aquatic adaptations might have been useful to cross small bodies of water, support itself on wet, soft substrates and access better fishing sites. However, the morphological onus seems to be on movement unsupported by deep water, and it might be assumed these formed a minority of adaptive pressures on Tanystropheus anatomy. Although it is difficult to think of a perfect modern analogue for this, we might find comparable functionality and behaviours in a variety of birds, crocodylians and lizards.

OK, time to call it a day with Tanystropheus for now, although we're not done with weird Triassic taxa yet. I've definitely caught their bug, and I'm sure we'll be spending time with several more of these fascinating oddballs in the near future. Before then, the last post I have planned this year returns us to familiar dinosaur territory, featuring an especially obscure species none of you will be familiar with. I can barely remember what it's called... Threecerasaurus? Trihornedabottoms? Dang - I'm sure I'll remember by next time.

This overly-long article and its artwork are made possible by Patreon

Regular readers will know that this blog and artwork is sponsored by patrons who pledge support at my Patreon page. For as little as $1 a month you can help keep this blog going and, as a reward, you get to see a bunch of exclusive content such as prints, a discount at my art store, and bonus posts not seen anywhere else. Articles posted here also typically get some 'bonus content'. For this post, I'll be discussing the scientific and palaeoartistic reasoning behind the two Tanystropheus paintings seen accompanying my two articles on this animal. As always, I'm very grateful to everyone who signs up!


  • Araújo, R., & Correia, F. (2015). Plesiosaur pectoral myology. Palaeontologia Electronica, 18(1), 1-32.
  • Beja, P.R. (1991). Diet of otters (Lutra lutra) in closely associated freshwater, brackish and marine habitats in south-west Portugal. Journal of Zoology (London), 225: pp. 141-152
  • Dalla Vecchia, F. M. (2005). Resti di Tanystropheus, saurotterigie e “rauisuchi”(Reptilia) nel Triassico Medio della Val Aupa (Moggio Udinese, Udine). Gortania, 27, 25-48.
  • Denny, M. W., & Gaines, S. D. (2007). Encyclopedia of tidepools and rocky shores (No. 1). Univ of California Press.
  • Hall, C. S., & Kress, S. W. (2008). Diet of nestling Black-crowned Night-herons in a mixed species colony: implications for tern conservation. The Wilson Journal of Ornithology, 120(3), 637-640.
  • Hartwick, B. (1983). Octopus dofleini. In Cephalopod Life Cycles, Vol. I: Species Accounts, ed. P.R. Boyle, pp. 277-293. Academic Press, London
  • Holliday, C. M., Ridgely, R. C., Sedlmayr, J. C., & Witmer, L. M. (2010). Cartilaginous epiphyses in extant archosaurs and their implications for reconstructing limb function in dinosaurs. PLoS One, 5(9), e13120.
  • Kuhn-Schnyder, E. (1959). Hand und Fuss von Tanystropheus longobardicus (Bassani). Paläontologisches Institut der Universität Zürich. 921-941.
  • Li, C. (2007). A juvenile Tanystropheus sp.(Protorosauria, Tanystropheidae) from the Middle Triassic of Guizhou, China. Vertebrata PalAsiatica, 45(1), 41.
  • Molnar, J. L., Pierce, S. E., & Hutchinson, J. R. (2014). An experimental and morphometric test of the relationship between vertebral morphology and joint stiffness in Nile crocodiles (Crocodylus niloticus). The Journal of experimental biology, 217(5), 758-768.
  • Nosotti, S. (2007). Tanystropheus longobardicus (Reptilia, Protorosauria): Re-interpretations of the Anatomy Based on New Specimens from the Middle Triassic of Besano (Lombardy, Northern Italy). Società Italiana di Scienze Naturali e Museo Civico di Storia Naturale.
  • Renesto, S. (2005). A new specimen of Tanystropheus (Reptilia, Protorosauria) from the Middle Triassic of Switzerland and the ecology of the genus. Rivista Italiana di Paleontologia e Stratigrafia, 111(3), 377-394.
  • Rieppel, O., Li, C., & Fraser, N. C. (2008). The skeletal anatomy of the Triassic protorosaur Dinocephalosaurus orientalis Li, from the Middle Triassic of Guizhou Province, southern China. Journal of Vertebrate Paleontology, 28(1), 95-110.
  • Rieppel, O., Jiang, D. Y., Fraser, N. C., Hao, W. C., Motani, R., Sun, Y. L., & Sun, Z. Y. (2010). Tanystropheus cf. T. longobardicus from the early Late Triassic of Guizhou Province, southwestern China. Journal of Vertebrate Paleontology, 30(4), 1082-1089.
  • Taylor, M. P., & Wedel, M. J. (2013). Why sauropods had long necks; and why giraffes have short necks. PeerJ, 1, e36.
  • Tschanz, K. A. R. L. (1988). Allometry and heterochrony in the growth of the neck of Triassic prolacertiform reptiles. Palaeontology, 31(4), 997-1011.
  • Wild, R., (1973). Tanystropheus longobardicus Bassani: neue Ergebnisse. In: Kuhn-Schnyder, E., & Peyer, B. (eds). Die Triasfauna der Tessiner Kalkalpen, XXIII. Schweizerische Paläontologische Gesellschaft 95, 1-162.
  • Witton, M. P., & Naish, D. (2008). A reappraisal of azhdarchid pterosaur functional morphology and paleoecology. PLoS One, 3(5), e2271.

Friday, 13 November 2015

The lifestyle of Tanystropheus, part 1: was that neck too heavy for use on land?

Two Tanystropheus longobardicus tussle in Triassic Europe. There's a distinct lack of water supporting their necks in this scene, and some might suggest this makes such behaviour impossible for these animals. But does it? Read on...
One of the most famous non-dinosaurian denizens of the Mesozoic is Tanystropheus, a spectacularly long-necked reptile which lived across Europe and Asia in Middle-Late Triassic times. We've known about this 5-6 m long animal since fragmentary fossils were pulled from Italian Triassic rocks in 1855, and now regard it as a particularly large and anatomically extreme member of the Protorosauria. This is a Permian-Triassic group of archosauromorphs (all reptiles more closely related to crocodylians and birds that lizards) that spawned numerous aberrant taxa, such as drepanosaurs, Sharovipteryx and Dinocephalosaurus. Within Protorosauria, Tanystropheus can be considered a tanystropheid, closely related to similar, but shorter-necked and smaller-bodied species such as Tanytrachelos and Langobardisaurus. Tanystropheus longobardicus is by far the best known Tanystropheus representative, and the one we always think of when discussing this animal, but something like five Tanystropheus species have been named over the years. It is currently uncertain how many of these should be considered valid and, of those, which ones truly represent Tanystropheus and not some other type of protorosaur. There are hints that longobardicus might be the sole representative of this genus, but work on this is ongoing.

We know a lot about Tanystropheus because it's fossils are not uncommon, and many of them are complete or every nearly so. Its remains occur in Alpine Europe, the Middle East and China and we can conclude that, weird as it seems, the Tanystropheus bauplan and life strategy was a successful one. But exactly what that strategy was remains a bone of contention for palaeontologists. Summarised simply, opinion is divided over whether Tanystropheus was confined to aquatic habits, at least above a certain age and body size, or else capable of living terrestrially as a shore-patrolling, 'animated fishing rod'. Unsurprisingly, the principle source of this contention is its neck anatomy: clearly long and relatively stiff in life, was it so heavy that would over-balance the animal if not supported in water? Or was the neck anatomy not as heavy as commonly supposed and really no great hindrance to life on land? Other aspects of Tanystropheus form have also influenced this debate, including limb structure and tail anatomy, but it seems fair to say these discussions persist because experts disagree about the significance of that crazy neck. Renesto (2005) and Nosotti (2007) provide recent overviews and contributions to this long-running controversy.

I've been wanting to cover Tanystropheus lifestyle here for some time now, and I've ended up with sufficient material to spread discussion over two posts. In the next article I'll be discussing nuances of arguments for aquatic and terrestrial habits, but, first, I want to satisfy some personal curiosity over how Tanystropheus was constructed. Specifically, I'm interested in the mass distribution of this animal: is it really _that_ front heavy? There are plenty of terrestrial animals with very long necks - sauropods, giraffes, some pterosaurs - and we don't worry about them toppling over. Moreover, although neck mass is frequently mentioned as critical to understanding Tanystropheus lifestyle, to my knowledge, there isn't any information available on its body volumes or mass (if I'm wrong, please tell me below). I thought I'd see what I could find out about this myself using the GDI (Graphic Double Integration) method of volumetric mass estimation, a quick and easy way to get a sense of mass and body fractions of fossil animals. It basically involves chopping up drawings of animals to determine volumes of body segments, then multiplying these by a suitable density - check out this excellent SV:POW! summary for a full lowdown.

Tanystropheus longobardicus as reconstructed by Rupert Wild in 1973. Image borrowed from Palaeos.
GDI methods require a clear layout of animal form to measure and divide into segments. There is no shortage of life restorations of Tanystropheus out there, and plenty of photographs of near complete specimens, but objective, modern portrayals of its anatomy are hard to come by. Rupert Wild's skeletal reconstruction from the 1970s (above) seems to remain a common frame of reference, and a David Peters reconstruction is sometimes used as an alternative. Neither really seemed suited for my purposes here, the crouched poses obscuring anatomical details, some specifics of vertebral count being inaccurate to modern interpretations, and the latter being produced with techniques of questionable reliability. I decided to try my hand at producing a new skeletal reconstruction based on the large, near complete Tanystropheus skeleton described in detail by Rieppel et al. (2010): GMPKU-P-1527:

Tanystropheus cf. longobardicus specimen GMPKU-P-1527, as depicted by Rieppel et al. (2010).
I wanted a large animal because the Tanystropheus neck seems to increase in length disproportionately to body size (Tschanz 1988). I want to give this animal the best chance of falling over, so it makes sense to use the largest neck possible. GMPKU-P-1527 is articulated and includes most of the neck, missing only the relatively short anterior 3.5 vertebrae (of 13), the skull, and the end of the tail. I reconstructed these missing parts using smaller Tanystropheus specimens (from Nosotti 2007) and Wild's widely-used 'adult' skull reconstruction. These came together to form a skeleton measuring 3.5 m as reconstructed, and likely over 4 m if the vertebral column were completely straightened. This is not as large as we think this animal could get, but is c. 70% of maximum size, and the minimal amount of proportional inference and cross scaling means we should be looking at a fairly authentic image of Tanystropheus form. The results are below.

Tanystropheus cf. longobardicus skeletal reconstruction, almost entirely based on GMPKU-P-1527. See text and illustration below for details on which bits are borrowed from other specimens.
The length of the limbs here is quite striking. Note that they aren't depicted in a true sprawling pose, because foreshortening would impact measurements for the mass calculation, but I depicted a crouched pose which hopefully conveys something of a low, sprawling gait. I also followed Nosotti's (2007) suggestion of digitigrady, which boosts the standing height a bit. Despite the low pose, I immediately get a different vibe from this image to that of Wild's classic, sitting reconstruction. Simply putting the animal on its feet gives the impression of the limbs and body being more proportionate to the neck. The arc of the neck follows that preserved in GMPKU-P-1527 quite closely, a pose also occurring in several other articulated Tanystropheus specimens. As depicted, I don't think it conflicts with recent interpretations of Tanystropheus neck arthrology (e.g. Renesto 2005). The body outline should be non-controversial, pretty much following the outline of the skeleton and hitting major muscle landmarks.

Time to chop this guy and up see what it's made of. Ideally, we'd want full orthographic views for a GDI mass estimate, but I've not had time to produce a multi-view skeletal. This means we're going to have to make predictions of body width. For the neck, body and tail, I decided to calculate width as 2/3 of body segment height, this being indicated by the proportions of Tanystropheus neck and tail verts, and the fact the dorsal ribs straighten out as they approach the lateral margins of the body. The 2/3 figure is a little arbitrary and arm-wavy, but seems more precise than assuming a circular cross section across the entire body. Other elements - the head and limbs - were modelled as having circular cross sections, however. You can see how I chopped the reconstruction up below: note that this uses an earlier, differently posed version of the skeletal shown above, and that the limbs are somewhat straighter. The bone sizes are no different, however, so influences on mass estimation should be negliable.

GDI mass estimation on Tanystropheus cf. longobardicus. Grey portions of the skeleton show which parts were modelled on other specimens. Numbers in parentheses give mass fractions for each body component, and the grey shapes indicate the cross-sectional shape used for that part of the body.
The entire animal shakes out to 26.7 L, and using a middle-of-the-road reptile density of 0.85 kg/L, the animal masses 22.9 kg. Of more interest to us is the mass percentages of each component, which are indicated in parentheses in the illustration above. You can see that the neck and head together are a hair away from 20% of the body mass, despite accounting for something like half the length of the animal. Virtually all of this 20% represents neck, of course, the head being less than 1% of the overall mass. As is usual for tetrapods, the trunk volume dominates all, being 50% of overall bulk despite only just exceeding 50 cm long in a 3.5 m long animal.

What do these figures actually mean? 20% doesn't seem like that much in the grand scheme of things, it being balanced by the other 80% of the body. These are certainly not values which make me think this animal perpetually toppled over unless it was in water. But can we be more precise here - how does this neck fraction stand up to other long-necked animals? For brachiosaurid sauropods, Mike Taylor (2009) suggested the neck accounted for 14% of the body mass, while Don Henderson (2010) suggested 8% for the same animals, noting that this was the largest neck mass fraction in his dataset of 10 volumetric sauropod mass predictions. Mitchell et al. (2013) did not report exact head and neck mass fractions for a large set of giraffes, but eyeballing their data suggests male giraffe necks and heads account for around 14% of body mass, with females slightly less than that. These are all significantly lower than the 20% I've estimated for Tanystropheus, implying that my gut feeling might be wrong: maybe it did have quite a heavy neck and, perhaps, was at greater risk of overbalancing.

However, it strikes me that giraffes and sauropods are not particularly good analogues for Tanystropheus, because their anatomy is built around a fundamentally different set of demands: processing of plant material. Herbivores need large guts to get the most from their nutrient-poor diet, equating to proportionally larger trunk volumes. Anyone who's played with volumetric mass estimations will know that even slight adjustments to trunk proportions can have a big impact on absolute mass and tissue fractions because they represent the biggest components of most animal bodies. We therefore can't ignore the requirement for herbivore torsos to be large when comparing them to non-herbivores like Tanystropheus. Our problem here is that finding a long-necked terrestrial carnivore to compare with Tanystropheus is challenging. Such body plans have been rare throughout geological time and are entirely unrepresented nowadays. We're not entirely licked, though: following the laws of monster movie science, any challenge involving a poorly understood, freakish creature is best solved with another poorly understood and freakish animal: in this case a long-necked azhdarchid pterosaur. Azhdarchid palaeoecology has a history of contention and controversy, but no-one believes that they were aquatic animals, or herbivorous, or at risk of toppling forward without environmental aid. This is despite azhdarchids bearing neck/trunk proportions similar to those of Tanystropheus, as well as much larger heads. We just assume they could carry their heads and necks one way or another, because all indications are that they were not adapted for an aquatic existence.

Taking a GDI approach to the Zhejiangopterus linhaiensis skeletal I produced earlier this year, I attempted to gather some data on azhdarchid body volumes and masses. As before, I estimated widths rather than producing full orthogonal views. The head and neck were assumed to be half as wide as tall, with the neck widths not permitted to exceed those of the skull. All other elements are treated as having circular cross sections. Azhdarchid torsos, forelimbs and necks are all highly pneumatised, so I gave these low tissue densities of 0.7 kg/L (about the lowest density recorded for modern birds), while the legs were given a more typical density of 0.85 kg/L. The breakdown and results:

Zhejiangopterus linhaiensis gets the GDI treatment. As above, numbers in parentheses indicate mass fractions, and the grey shapes indicate cross sectional area used in the calculations. Skeletal based on data in Cai and Wei (1994).
First things first, I was happy to see the animal come out at 7.9 kg - that's in line with most post-2000 interpretations of pterosaur mass, and seems about right for an animal with a 2.5-3 m wingspan. That makes me think the constituent volumes and masses are probably in a sensible ball park. In terms of body component masses, the torso - famously small in derived pterodactyloids like azhdarchids - provides less than 25%, the neck is just over 25%, and the head and paired forelimbs are 22% each. The legs account for just under 6%, and the tail might as well not exist. This puts - wow - almost 50% of the mass in front of the shoulders in this reconstruction. But even accepting that I've been generous with neck tissue in my reconstruction (following a reptilian, rather than avian pattern of neck musculature), and that some cross sectional shapes used here could be refined, it seems unlikely we could slim the head and neck tissues down to the mass fractions seen in long-necked herbivores. Even if my estimates are out by a factor of 2, the neck and head will still account for more mass than the same components in Tanystropheus. This finding makes the neck tissue fraction of the Tanystropheus model look a lot less aberrant, as well as verifying the suspicion that lifestyles, and not just anatomy, are important factors when comparing animal bauplans.

Let's bring all this together. While the sums outlined here are provisional, back-of-the-envelope-type stuff, I find them sufficient to at least make me sceptical of claims that Tanystropheus has a terrestrially-untenable mass distribution. At least one group of non-aquatic Mesozoic carnivores seem more front heavy, and a basic model of Tanystropheus mass distribution does not raise major alarm bells about relative neck and head weight. I could be convinced otherwise, and obviously there's a lot more than could - and should - be done to investigate this issue, but I currently don't see neck mass as a significant barrier to terrestrial habits. This exercise has also brought home the fact that we might not know much about the adaptive and structural significance of extremely long necks in carnivorous animals, and that we should be careful comparing them to other long-necked creatures. Perhaps our unfamiliarity with this extinct bauplan, along with our generally poor intuitive sense of animal mass and tissue fractions (see this discussion and comment field at SV:POW!), means we should be extra cautious about gut-feeling interpretations of such creatures. I guess the bottom line is that running numbers to test our intuitions is an essential part of understanding unfamiliar animal types, especially if we're suggesting those assumptions are significant for extinct animal behaviour and lifestyle.

There'll be more on Tanystropheus in the next post, where the plan is to review recent arguments for and against different lifestyles in this animal. In the mean time, I'm very curious to know what others make of the ideas presented here. Would you interpret these results differently? Would you have reconstructed Tanystropheus in a different way? The comment field is open...

Tanystropheus was brought to you by Patreon, a weekend of downtime and the letter 'S'

I've really enjoyed putting this research-led post together, and would like to do more of this sort of stuff in future. Obviously they take a little longer than 'routine' posts but, thankfully, this time investment is possible because of my patrons, individuals who support this site via my Patreon page. From $1 a month you can be a patron too: this gets you access to cool, exclusive content and rewards, and helps me make my art and writing more detailed and interesting. Thanks to all those who've signed on already - your contributions are really appreciated.


  • Cai, Z., and Wei, F. (1994). "On a new pterosaur (Zhejiangopterus linhaiensis gen. et sp. nov.) from Upper Cretaceous in Linhai, Zhejiang, China." Vertebrata Palasiatica, 32: 181-194.
  • Henderson, D. M. (2004). Tipsy punters: sauropod dinosaur pneumaticity, buoyancy and aquatic habits. Proceedings of the Royal Society of London B: Biological Sciences, 271(Suppl 4), S180-S183.
  • Mitchell, G., Roberts, D., Sittert, S., & Skinner, J. D. (2013). Growth patterns and masses of the heads and necks of male and female giraffes. Journal of Zoology, 290(1), 49-57.
  • Nosotti, S. (2007). Tanystropheus Longobardicus (Reptilia, Protorosauria): Re-interpretations of the Anatomy Based on New Specimens from the Middle Triassic of Besano (Lombardy, Northern Italy). Società Italiana di Scienze Naturali e Museo Civico di Storia Naturale.
  • Renesto, S. I. L. V. I. O. (2005). A new specimen of Tanystropheus (Reptilia, Protorosauria) from the Middle Triassic of Switzerland and the ecology of the genus. Rivista Italiana di Paleontologia e Stratigrafia, 111(3), 377-394.
  • Rieppel, O., Jiang, D. Y., Fraser, N. C., Hao, W. C., Motani, R., Sun, Y. L., & Sun, Z. Y. (2010). Tanystropheus cf. T. longobardicus from the early Late Triassic of Guizhou Province, southwestern China. Journal of Vertebrate Paleontology, 30(4), 1082-1089.
  • Taylor, M. P. (2009). A re-evaluation of Brachiosaurus altithorax Riggs 1903 (Dinosauria, Sauropoda) and its generic separation from Giraffatitan brancai (Janensch 1914). Journal of vertebrate Paleontology, 29(3), 787-806.
  • Tschanz, K. A. R. L. (1988). Allometry and heterochrony in the growth of the neck of Triassic prolacertiform reptiles. Palaeontology, 31(4), 997-1011.

Thursday, 22 October 2015

The Spinosaurus saga continues

A year after the 'Spinosaurus reboot' as a small-legged, early whale-mimicking aquatic quadruped, experts remain divided over fundamental aspects of Spinosaurus palaeobiology. This depiction shows Spinosaurus aegyptiacus as generally imagined prior to 2014.

The long, tragic and occasionally controversial research history of the giant, enigmatic theropod Spinosaurus aegyptiacus will be familiar to many readers of this blog*. First named and described in the early 20th century by Ernst Stromer from remains found in Late Cretaceous strata of Egypt, our principle Spinosaurus material fell victim to Allied bombing raids in World War II and was completely destroyed. Stromer's detailed illustrations and descriptions are all that remains of this material, and these have formed a variably interpreted foundation of all subsequent Spinosaurus research. For much of the 20th century the life appearance of Spinosaurus remained mysterious. Depicted as a nondescript sailed giant theropod early on, discovery of well represented spinosaurids like Baryonyx and Suchomimus, as well as fragments of new Spinosaurus material, permitted more confident interpretations of Spinosaurus size and form as we approached the new millennium. By the 2010s, Spinosaurus was recognised as a gigantic, derived and perhaps semi-aquatic spinosaurid, adapted for feeding on large aquatic prey (above). Much of this interpretation relied on new Spinosaurus remains from multiple locations in northern Africa, including the famous Moroccan Kem Kem Beds, an expanse of Late Cretaceous rocks roughly contemporaneous with those Egyptian deposits yielding the original, destroyed Spinosaurus remains.

*For succinct overviews of Spinosaurus research prior to 2014, check out posts at Tetrapod Zoology and Laelaps.

Famously, last year saw Spinosaurus reinvented again, this time as a quadrupedal, knuckle-walking, long-bodied, tiny-legged dinosaurian take on a crocodile or early whale (below). The authors of this widely publicised study, Nizar Ibrahim and colleagues (2014), synthesised existing and new data on African spinosaurids to create this reconstruction, synonymising several taxa into S. aegyptiacus and presenting new Spinosaurus remains obtained from the Kem Kem beds. The most significant of these was a set of associated vertebrae, pelvic and hindlimb remains which were proposed as a neotype specimen for Spinosaurus (a specimen to hold the Spinosaurus name now that the original material is lost to science). That this neotype represents Spinosaurus was bolstered by it bearing similar hindlimb and vertebral proportions to 'Spinosaurus B', a collection of Egyptian spinosaurid specimens described by Stromer, considered referable to Sp. aegyptiacus by Ibrahim and colleagues. Spinosaurus B is also now lost, also being destroyed in WWII. The Ibrahim et al. study provided a lot of new data on Spinosaurus and has helped cement the concept of it being a semi-aquatic animal, but several aspects of the paper didn't meet the warmest reception from a number of academics. Specific issues were scaling of the skeletal components, how sensible it was to lump so much north African spinosaurid material into one species, and uncertainty about the provenance of the neotype specimen. Some of these concerns were diffused by the authors, but we await a promised monograph for answers to all the questions raised by their first paper. In the mean time, the 2014 Spinosaurus interpretation remains a debated topic among those interested in dinosaur palaeontology.

The Ibrahim et al. (2014) take on Spinosaurus aegyptiacus. Different colours represent different specimens: red is the neotype; brown is the original Spinosaurus material; yellow is referred, isolated Spinosaurus remains; green bones are borrowed from other spinosaurids, and blue bones are crafted to fit the skeleton based on neighbouring elements. Image borrowed from

One year later...

This week, the Spinosaurus tale has taken another twist with publication of a mammoth (open access) paper penned by a team of European spinosaurid experts, led by Serjoscha Evers. Evers et al. have reappraised the affinities of Moroccan specimens seemingly related to Spinosaurus: Sigilmassasaurus brevicollis and Spinosaurus maroccanus. These animals, known only from vertebrae, were subsumed into Sp. aegyptiacus by Ibrahim et al. (2014) as part of their trans-African Spinosaurus concept, and that decision is a core focus of the Evers et al. paper. Their work contains extensive commentary on the detailed anatomy of Moroccan spinosaur material and what it might mean for recent interpretations of Spinosaurus form and lifestyle. Given the wide interest in Spinosaurus and the 2014 reconstruction, I thought it might be of interest to summarise some of what they outline here.

Firstly, taxonomic revisions proposed by Evers et al. present a very different picture of what fossils we can identify as belonging to Spinosaurus. Their work on Si. brevicollis and Sp. maroccanus suggests these species are probably one and the same (the latter being sunk into the former), and that Sigilmassasaurus should be considered distinct from Sp. aegyptiacus. They go on to suggest that other Kem Kem vertebrae hint at a second spinosaurid species in the Kem Kem fauna, and outline several reasons why the Ibrahim et al. 'neotype' specimen cannot be referred to Spinosaurus. For one, the neotype is anatomically quite different from Stromer's Egyptian 'Spinosaurus B' specimen. Ibrahim et al. considered Spinosaurus B as representing Sp. aegyptiacus, but Evers and colleagues argue that Spinosaurus B is anatomically more similar to Sigilmassasaurus than Spinosaurus. Spinosaurus B therefore might have no use for linking any specimens specifically to Sp. aegyptiacus, including that all-important neotype.

In addition to these morphological objections, Evers et al, also raise palaeobiogeographic issues with the 'neotype' referral. Evidence for Egyptian dinosaur species being present in Morocco is scant at best, most data indicating little mixing of eastern and western African dinosaur species during the Late Cretaceous. It would be unusual, then, to find the Egyptian species Sp. aegyptiacus in Morocco. Palaeobiogeography is not a deal clincher for taxonomy of course - careful examination of the neotype and genuine Spinosaurus remains will be the deciding factor here - but it is another stick in the mud for the neotype proposal. Although the exact identity of the 'neotype' specimen is left in the air by Evers et al. - ongoing descriptive work on the specimen needs to be completed to truly assess this - they reject the proposal of the Kem Kem specimen as a Sp. aegyptiacus neotype, and leave Spinosaurus characterised by features in Stromer's illustrations. This is obviously quite a shake up of the suggestions made last year: Spinosaurus 2014 might be a mix of at least two named species, incorporate material of under-appreciated taxonomic importance, and substantial, newly published material might have little, if anything, to do with Spinosaurus.

The proposed Spinosaurus neotype. Image borrowed from Andrea Cau's excellent Theropoda blog.

Moving on, Evers et al. also raise concerns about interpretations of Spinosaurus in context of Kem Kem fossil collecting practises. Museum exhibitions and PR exercises suggest that the Kem Kem yields complete skeletons of dinosaurs and other fossil vertebrates, but the reality is quite the opposite. Kem Kem vertebrates are typically preserved as isolated, often broken bones in multitaxic bone beds (that is, bone beds comprising many species). Associated skeletons of single individuals do occur, but they're relatively rare and rely on precise collecting documentation to prove their authenticity. Unfortunately, historic and recent records of Spinosaurus occurrences and excavation are often poor. We might chalk a lack of historic documentation to the practises and technological limitations of bygone times, but recent issues are caused primarily by the commercial value of Kem Kem fossils. The greater majority of Kem Kem fossils, including dinosaurs, are collected without extensive documentation and then sold by private dealers. Even if localities are recorded, ambiguity often surrounds association of fossil material prior to excavation. Several alleged associated Spinosaurus specimens are meant to have come from single localities, but being from the same place is really only half the battle if they stemmed from multitaxic assemblages. Concordant size of bones might suggest genuine association, but this is not always certain either: Evers et al. report practises where collectors sort loose material from disparate locations into type and size categories before sale - nefarious individuals making fossil skeletons more substantial with unassociated elements is a real problem the world over. It's sad but true that the monetary value associated with substantial vertebrate fossils makes ascertaining their authenticity crucial for subsequent credible interpretation.

Unfortunately, Evers et al. report these factors as affecting virtually all associated Spinosaurus material, including the 'neotype' and the other specimen key to the 2014 reconstruction, Spinosaurus B. In the case of the latter, all we have to go on to establish association are Stromer's notes, which are not quite as detailed as we might like. For the neotype, we know some of the specimen was directly collected in the field, and that other bits were purchased from dealers by two academic institutions over a two year period - exact documentation of this remains to be presented (hopefully it will in the 'neotype' monograph). Without strict certainty over how many individuals these specimens might represent, Evers et al. suggest some of the odd proportions in recent Spinosaurus reconstructions may reflect the marrying of mismatched bones to one another. That's not a certainty, of course, but it's also something which shouldn't be casually ignored.

Collectively, Evers et al. use these points to provide an alternative take on Spinosaurus to that presented in 2014. Ibrahim et al. argued that their new material helped simplify and integrate different interpretations of African spinosaurid material, but Evers et al. argue the opposite: they emphasise how poorly known Spinosaurus and kin are, and how interpreting fossils of north African spinosaurids is getting increasingly complex. Spinosaurus fossils remain very fragmentary to the point where most cannot be directly compared, they seem to hint at, but don't really crystalise, an apparent high species diversity, and are often of uncertain association or exact origin. At face value, that doesn't leave us with a lot to be confident about, although we'll have to see how this more despondent view goes down with other spinosaurid researchers. More complete and well documented discoveries will soon help smooth out bumps in our knowledge, but it seems likely that a lot of work and discussion remains to sort out what is really going on with north African, Late Cretaceous spinosaurids.

What does this mean for 'the Spinosaurus reboot'?

That's not quite the end of our discussion, however. It might be assumed that the points outlined above sound the death knell for the strangely proportioned 2014 Spinosaurus reconstruction, and that we should go back to our traditional interpretation of this animal. That might not be quite right, for two reasons. Firstly, given how distinctive many 'Spinosaurus' remains now seem to be, it's actually questionable what specimens should be considered Sp. aegyptiacus at all, other than the first specimen described by Stromer. A lot of referred isolated Spinosaurus specimens have been incorporated into our 'traditional' reconstructions in recent years, and we might need to think hard about their role in our interpretations of this animal. What we've become typically used to thinking of as Spinosaurus may not entirely be Spinosaurus!

Secondly, while some aspects of the 2014 interpretation of Spinosaurus have clearly been challenged by the Evers et al. paper, not all proportional aspects of the recent Spinosaurus reinvention are obviously erroneous. Last year, Ibrahim et al. noted that both Spinosaurus B and the 'neotype' have reduced hindlimbs with respect to their associated vertebrae, and used this fact as support for the diminutive legs in their reconstruction. Although arguing that there is no longer evidence for short hindlimbs in Spinosaurus itself, Evers et al. don't completely dismiss the notion of some African spinosaurids being short legged. The hindlimb proportions of those specimens is very similar despite the vagaries surrounding fossilisation and exhumation of ancient animal remains, maybe more similar than you'd expect from chance alone. If it is coincidence, it's certainly a startling one.

Stromer's 'Spinosaurus B' material: proportionally similar to the 'neotype' specimen, but does that tell us anything about spinosaurid proportions? Another image borrowed from Theropoda.
However, Evers et al. also attach some important caveats to this point. Stromer's notes clearly state that he did not consider the 'Spinosaurus B' material to represent one individual, and his testimony is the closest thing we have to a report on the excavation of the material. He specifically comments on the hindlimb being too small and slender to match the vertebrae, and thus interpreted them as representing a second individual. Other workers have agreed that this material must represent multiple animals or even several types of dinosaur (discussions about the possibly chimeric nature of Stromer's spinosaur specimens are not new - e.g. Rauhut 2003; Novas et al. 2005). Interpretation of the Spinosaurus B material as representing one animal is thus against some current thought and, of course, Stromer's original declaration. While the 'neotype' specimen might make a case for Stromer being mistaken, we really need to know more about the collection history to ascertain that. We're left with an intriguing set of measurements hinting at the reduced hindlimbs proposed by Ibrahim et al., but little in the way of objective information to explain their significance. The discovery of new specimens is needed to establish whether some spinosaurids were really short-legged, or if confusion of specimen inventories just made it look that way. In short, and no-doubt to the disdain of people who lose sleep about 'what science has done' to one of their favourite theropods, there's still something to play for with these short-legged spinosaurids.

So that's the latest chapter of research in Spinosaurus, then: I don't doubt that it's going to cause a lot of discussion in popular and academic circles. My personal take-home is that we seem to know less about Spinosaurus than might have been recently suggested, or at least that some issues need to be ironed out before we can develop a clear picture of what Spinosaurus is, and what sort of lifestyle it led. I don't know that any recent proposals about this animal have been shot down entirely yet, although clear gauntlets have been established for some of the more extreme ideas suggested in the last few years. It's going to be very interesting to see how others interpret these latest developments in the ongoing Spinosaurus saga, and where our understanding of this animal moves to next.

Mighty Spinosaurus has nothing on the power of Patreon

Production of the artwork and words you see here is funded by my patrons, individuals who support this site via my Patreon page. From $1 a month you can be a patron too: this gets you access to cool, exclusive content and rewards, and helps me make my art and writing more detailed and interesting. Thanks to all those who've signed on already - your contributions are really appreciated.


  • Evers, S. W., Rauhut, O. W. M, Milner, A. C, McFeeters, B, & Allain R. (2015) A reappraisal of the morphology and systematic position of the theropod dinosaur Sigilmassasaurus from the “middle” Cretaceous of Morocco. PeerJ 3:e1323
  • Ibrahim, N., Sereno, P. C., Dal Sasso, C., Maganuco, S., Fabbri, M., Martill, D. M., & Zouhri, S., Myhrvold, N. and Iurino, D. A. (2014). Semiaquatic adaptations in a giant predatory dinosaur. Science, 345(6204), 1613-1616.
  • Novas, F., Dalla Vecchia, F., & Pais, D. (2005). Theropod pedal unguals from the Late Cretaceous (Cenomanian) of Morocco, Africa. Revista del Museo Argentino de Ciencias Naturales nueva serie, 7(2), 167-175.
  • Rauhut, O. W. M. (2003). Special Papers in Palaeontology, The Interrelationships and Evolution of Basal Theropod Dinosaurs (No. 69). Blackwell Publishing.