Of Becklespinax and Valdoraptor

As discussed in previous articles on the 55th SVPCA, I've decided to publish a version of my talk here on the blog. Some time ago I published a collection of theropod images (go here), some of which might have seemed a bit random or unconnected. As we'll see here, those images are indeed connected, sort of, and now all will be revealed...

Ok, so, the talk was titled The large theropods Becklespinax and Valdoraptor from the Lower Cretaceous of England. I might get into the habit of publishing other talks that I give or, at least, I might in those cases where there's a script or pile of notes that can be easily converted into a readable article. In keeping with the rest of my technical research on dinosaurs, my talk concentrated on obscure, poorly known taxa from the Lower Cretaceous Wealden Supergroup of southern England. At some stage (when the other 20-odd manuscripts I'm working on are out of the way) I will get round to publishing a paper on all this stuff, but not today.

The Wealden Supergroup of southern England preserves one of the most representative, most diverse assemblages of Lower Cretaceous dinosaurs in the world. While theropods have been known from the Wealden for as long as this group of dinosaurs has been recognised, it's only recently that good, informative Wealden theropod specimens have been discovered. Most Wealden theropods come from the Barremian Wessex Formation of the Isle of Wight, and this is the case for Eotyrannus, Neovenator and the small taxa Aristosuchus, Calamosaurus and so on (Naish et al. 2001, Naish & Martill 2007). Far less well known are the much older theropods of the Hastings Beds Group [picture above combines reconstructions of Neovenator, Eotyrannus, Thecocoelurus and Calamosaurus (shown roughly to scale) with a map of the Isle of Wight: the black patches show the exposed dinosaur-bearing sediments of the island].

The Wealden Supergroup is divided into three groups: the Hastings Beds Group and Weald Clay Group (both of the Weald sub-basin of the English mainland) and the Wealden Group of the Wessex sub-basin of the Isle of Wight. In being Berriasian and Valanginian in age, the Hastings Beds Group is between 10-20 million years older than the Barremian Wessex Formation (Radley 2004, 2006a, b), and most evidence indicates that these units don't share the same genera, and certainly not the same species. Two Hastings Beds Group theropods were first described in the 1850s: their systematics has been extensively argued over, but nobody's ever gotten round to producing redescriptions. They are Becklespinax altispinax, best known from three tall-spined dorsal vertebrae, and Valdoraptor oweni, known only from an incomplete metatarsus. While both taxa are - obviously - frustratingly fragmentary, I think they still have some useful things to show us about Wealden theropod diversity.

Becklespinax is historically interesting for a couple of reasons. For one, it appears to have been the first dinosaur in which the fossae on the sides of the neural arch were interpreted as pneumatic: this idea came from Richard Owen, who first figured the specimen in 1855 (Owen 1855, Britt 1991). The initial referral of Becklespinax to Megalosaurus meant that the Becklespinax holotype became incorporated into Waterhouse Hawkins' Crystal Palace model of Megalosaurus [shown at top of article]: Owen misidentified the Becklespinax vertebrae as being anterior dorsals, and the height of the Becklespinax neural spines led him to suggest that Megalosaurus must have had a hump over its shoulder region.

The taxonomy of Becklespinax is rather... involved. For a long time the specimen was known as Altispinax, but that's unsatisfactory because there's no way of knowing that the type specimen of that taxon is anything to do with the tall-spined vertebrae. To cut a very long story short, the best solution has been to coin a new generic name for this specimen - that name is Becklespinax (Olshevsky 1991) - but even that's not without problems.

Anyway, the Becklespinax holotype consists of three articulated posterior dorsal vertebrae, probably representing dorsal vertebrae 9 to 11, all of which sport tall neural spines. The centra are taller than they are long, pneumatic foramina on the centra are definitely absent, and the ventral surfaces of the centra are markedly concave. Strut-like laminae surround three neural arch fossae (the infraprezygapophyseal fossa, infradiapophyseal fossa, and infrapostzygapophyseal fossa). There is no spinodiapophysial lamina, as there is in baryonychines. The hyposphene - which can be observed on the third of the vertebrae - is particularly large and notably rectangular. Becklespinax is of course best known for its neural spines, and in the past these have led to suggestions that it is a spinosaurid, or a close relative of Acrocanthosaurus. But there are no shared derived characters that might unite Becklespinax with either spinosaurids or carcharodontosaurids, and these suggestions can't be supported.

What's really quite strange about these neural spines is the marked difference in height between the first spine, and the second and third spines. The first spine differs from the other two in that its apex is irregular, rather than parallel to the long axis of the centrum: it almost looks as if the upper third of the spine has been snapped off. The second spine's apex appears to have 'over-grown' the apex of the first. Some authors have eliminated this 'over-growth' when figuring the specimen, but it was accurately depicted by Owen in 1855. At their dorsal apices, the second and third neural spines are mediolaterally thick, exceeding 55 mm in thickness. This is an unusual character, not seen in other tall-spined theropods. In his 1884 History of British Fossil Reptiles, Owen described the spine tips as rugose and expanded compared to the smooth basal portions of the spines, and also described the first and second spine tips as ankylosed. They aren't truly ankylosed - but there is in fact some fusion of the tips of the second and third spines.

On the basis of these observations - the over-growth of the first spine by the second, the rugose, expanded tips to the spines and so on - Molnar (2001) suggested that the Becklespinax holotype might be pathological. We can't be confident about this from these characters alone though: rugose neural spine tips are present in various large tetanuran specimens where there's no indication of pathology. However, the differing heights of the neural spines, and the 'over-growth' of the first spine, are perhaps more difficult to explain. If the first spine really was broken in life, then - given that these are posterior dorsal vertebrae - does this mean that this individual sustained some sort of injury to an otherwise continuous dorsal sail? An alternative explanation is that this difference in neural spine height is natural: if we look at other tall-spined vertebrates we can see a similar disparity in neural spine height in that part of the vertebral column where the neural spines start to increase in height. But if this really was the condition in life, then Becklespinax must have been a rather peculiar theropod: it would have meant that the sail must have been restricted to the posterior dorsal and sacral region (and perhaps the caudal region as well). This is not impossible of course, but intuitively it looks outlandish. Ultimately the only way to confirm or refute this is to either section, or CT-scan, the spines. Pathological neural spines of other dinosaurs have an amorphous, structureless internal bone texture when viewed in section, and it should be possible to see this if Becklespinax is ever CT-scanned.

The affinities of Becklespinax are difficult to determine: many of the features seen in the specimen are widely distributed in tetanurans, and it doesn't exhibit any derived characters that allow it to be placed within any specific clade. Indeed there are relatively few characters in the posterior dorsal vertebrae that allow the members of the different tetanuran clades to be reliably distinguished, and unfortunately the only ones present in Becklespinax are all ambiguous. The lack of ventral keels on the centra might suggest that Becklespinax is an allosauroid, given that most spinosauroids possess ventral keels. The neural spines on the posterior dorsal vertebrae of allosauroids have been regarded by some authors as being anteriorly inclined, but this is often quite subtle and subjective, and, while it might be present in Becklespinax, I wouldn't say that you could be confident about it.

Several other tetanurans exhibit a similar pattern of neural arch laminae to Becklespinax, including Condorraptor, Piatnitzkysaurus and Sinraptor, and Sinraptor in particular exhibits a superficially similar morphology: this could mean that Becklespinax is a sinraptorid, but there are no uniquely shared characters that might demonstrate this. For now, pending further discoveries, Becklespinax remains an indeterminate tetanuran of unknown affinities. Its combination of characters, and uniquely expanded neural spines tips, mean that it is a valid, diagnosable taxon.

While Becklespinax is only definitely known from dorsal vertebrae*, Valdoraptor is presently only known from a partial left metatarsus [shown in adjacent image, from Owen 1856]. The Valdoraptor holotype is actually quite small - 215 mm long in preserved length, with an estimated total length of about 240 mm. The detailed morphology of the specimen has partly been overlooked because Owen's (1856) plate doesn't depict it accurately: not only that, but, to confuse matters, Owen's figure of the specimen is actually a mirror-image of the real thing.

* A partial allosauroid tibia from the Hastings Beds Group has been suggested to be referable to Becklespinax (Naish 2003), but this referral is only plausible if Becklespinax is an allosauroid and, as discussed here, it might not be.

Metatarsal II in Valdoraptor is quite distinctive: it's strongly compressed mediolaterally, and exhibits a prominent ridge along its dorsomedial surface. It's unique - there aren't any other theropods that share such a strongly compressed metatarsal II - and consequently it's diagnostic. The distal end of metatarsal II diverges quite strongly from metatarsal III, creating a sub-oval interosseous space. This isn't typical for a theropod: the distal ends of these metatarsals are normally in quite close contact. Contrary to claims that this specimen is non-diagnostic then, it does in fact seem to possess at least two potential autapomorphies [key features of the specimen depicted in adjacent image].

There is one additional element that can be referred to this taxon: that's an isolated right metatarsal II (BMNH R2661) from the Hastings Beds Group (it appears to have come from the Valanginian Grinstead Clay Formation). It exhibits exactly the same unusual morphology as does metatarsal II of the holotype.

How does Valdoraptor compare to other theropods? It's been argued that the cross-sectional shape of the metatarsals - particularly that of metatarsal III - demonstrates a tetanuran identity. In fact, this isn't really clear, but on the basis of its width : length ratio, the metatarsus is most comparable to those of carnosaurs and short-footed coelurosaurs like some oviraptorosaurs and dromaeosaurs. Again, the specimen lacks derived characters that would allow it to be identified as a member of any specific tetanuran clade. The cross-sectional shape of metatarsal IV - it's laterally convex and mediolaterally quite wide - differs from the narrower, differently shaped metatarsal IV of allosauroids. Tom Holtz and colleagues recently suggested that Valdoraptor might prove to be synonymous with the Wessex Formation taxa Neovenator or Eotyrannus (Holtz et al. 2004). We can now reject these suggestions with some confidence, but, as for the true identity of Valdoraptor, it remains enigmatic (much like Becklespinax!).

It appears, then, that there were some unusual theropods in the Hastings Beds Group, and - perhaps not surprisingly given that they're known from such poor material - we can't yet allocate these taxa to any specific tetanuran clade. And this in itself is interesting because, while the better-known Wealden theropods can be confidently identified as spinosaurids, allosauroids or coelurosaurs, there are some additional taxa that can't yet be classified with such certainty. Whether these animals are unique to the Hastings Beds Group, or occur throughout the Wealden Supergroup is a good question: we can say that - in the Wessex Formation - we have tantalising glimpses of unusual Wealden theropods that don't seem to match any of the better-known taxa [the adjacent image shows the metatarsals of Valdoraptor, Eotyrannus and Neovenator, depicted very roughly to scale].

So, to conclude... (1) Hastings Beds Group theropods - far older than those of the Wessex Formation - remain poorly known and represented only by fragmentary remains; (2) Becklespinax doesn't exhibit any derived characters that allow it to be confidently identified as belonging to any specific tetanuran clade, though the strongest similarity is with sinraptorid allosauroids; (3) Becklespinax is a valid taxon as there are no similar-aged taxa with the same neural arch lamina and neural spine morphology; (4) Valdoraptor also lacks characters that allow it to be confidently identified as belonging to any specific tetanuran clade; (5) Valdoraptor is diagnostic and is a valid taxon

And that's that. Several people helped with the research behind the talk: Sandra Chapman, Steve Hutt and Cathy Waller provided access to specimens, and Roger Benson, Dave Martill, George Olshevsky and Mike P. Taylor are thanked for discussion and assistance.

Oh, and last night I got to stroke a real live pipistrelle (I don't know if it was a 45 or a 55 pipistrelle, and neither did its carers). More on that another time. And thank you Vicki for dragging us into the pub. Again.

Refs - -

Britt, B. B. 1991. Pneumatic postcranial bones in dinosaurs and other archosaurs. Ph.D. dissertation, University of Calgary (Calgary).

Holtz, T. R., Molnar, R. E. & Currie, P. J. 2004. Basal Tetanurae. In Weishampel, D. B., Dodson, P. & Osmólska, H. (eds) The Dinosauria, Second Edition. University of California Press (Berkeley), pp. 71-110.

Molnar, R. E. 2001. Theropod paleopathology: a literature survey. In Tanke, D. H. & Carpenter, K. (eds) Mesozoic Vertebrate Life. Indiana University Press (Bloomington & Indianapolis), pp. 337-363.

Naish, D. 2003. A definitive allosauroid (Dinosauria; Theropoda) from the Lower Cretaceous of East Sussex. Proceedings of the Geologists' Association 114, 319-326.

- ., Hutt, S. & Martill, D. M. 2001. Saurischian dinosaurs 2: Theropods. In Martill, D. M. & Naish, D. (eds) Dinosaurs of the Isle of Wight. The Palaeontological Association (London), pp. 242-309.

- . & Martill, D. M. 2007. Dinosaurs of Great Britain and the role of the Geological Society of London in their discovery: basal Dinosauria and Saurischia. Journal of the Geological Society, London 164, 493-510.

Olshevsky, G. 1991. A Revision of the Parainfraclass Archosauria Cope, 1869, Excluding the Advanced Crocodylia. Publications Requiring Research, San Diego.

Owen, R. 1855. Monograph on the fossil Reptilia of the Wealden and Purbeck formations. Part II. Dinosauria (Iguanodon). (Wealden). Palaeontographical Society Monographs 8, 1-54.

- . 1856. Monograph on the fossil Reptilia of the Wealden Formation. Part IV. Palaeontographical Society Monographs 10, 1-26.

Radley. J. 2004. Demystifying the Wealden of the Weald (Lower Cretaceous, south-east England). OUGS Journal 25 (1), 6-16.

- . 2006a. A Wealden guide I: the Weald sub-basin. Geology Today 22 (3), 109-118.

- . 2006b. A Wealden guide II: the Wessex sub-basin. Geology Today 22 (5), 187-193.