Junk in the trunk: why sauropod dinosaurs did not possess trunks

Before I start: TIANYULONG TIANYULONG TIANYULONG TIANYULONG TIANYULONG.

Ok, moving on...

It is the contention of some that the field of Mesozoic reptile research is plagued with bizarre, nonsensical hypotheses. You may or may not agree with me that skim-feeding giant pterosaurs, wind-surfing sail-crested pterosaurs, dedicated-to-scavenging Tyrannosaurus, and crampon-using dromaeosaurs are all, shall we say, unlikely. In some cases the science has been done to smack these ideas down, in other cases it has not. In my opinion, one of the most illogical hypotheses entertained within recent decades is the 'trunked sauropod' one.

Unlike most tetrapods, the bony nostrils of most sauropods are dorsally located: in diplodocoids they're located right up over the eyes in a region you might term the forehead. In extant mammals, the species with dorsally located bony nostrils have a proboscis, or trunk. Ergo, sauropods might have had trunks.

This idea is familiar to most dinosaur scientists and aficionados because it's often trotted out in popular books as one of those things that 'might be possible, but we'll never know, woooo!'. Gregory Irons produced a short-trunked Dicraeosaurus for Robert Long and Samuel Welles's All New Dinosaurs and Their Friends (Long & Welles 1980) [shown in the middle of the adjacent composite], Bakker illustrated a trunked Diplodocus in The Dinosaur Heresies (Bakker 1986, p. 141) [shown at top of adjacent composite], as did John Sibbick for When Dinosaurs Ruled the Earth (Norman 1985, p. 7) [shown below]. I'm sure there are others.

Walter Coombs and the trunked sauropod movement

The trunked sauropod movement is usually said to have started with Walter Coombs's seminal 1975 paper 'Sauropod habits and habitats': Coombs (1975) looked at numerous lines of evidence and concluded that, while sauropods may have entered the water on occasion, they were not amphibious, but were strongly adapted for terrestriality (though, he did note that 'Calling the entire ... Sauropoda a homogenous group is probably misleading ... as the diversity of sauropod morphology probably reflects diversity in habits and habitat preferences' (p. 29)). Thanks to his 1971 article in Nature, Bakker has gotten most of the credit for initiating the terrestrial sauropod movement (Bakker 1971), but Coombs's far more detailed article was equally important.

Anyway, Coombs (1975) noted that the size, shape and position of the bony nostrils in sauropods 'is similar in some respect to mammals which have, or are thought to have had, either a proboscis or at least a very large nose' (p. 6). He does seem to have concluded that a proboscis of some sort was present in at least some members of the group, though he noted that 'There is certain reluctance to accept a sauropod fitted with a proboscis because no living reptile has anything comparable to an elephantine or tapiroid nose' (p. 6). Coombs also pointed to the general absence of the required facial musculature in reptiles, and noted that this would be a problem for the trunk hypothesis. We'll look at this some more in a minute.

Coombs's proposal was not widely followed - in fact, it wasn't really followed at all. Long & Welles (1980) were interested enough to get that trunked dicraeosaur drawn in their book, but they noted that 'It must be stressed that we will probably never have direct evidence for sauropod trunks, but it is an interesting suggestion, and we wanted to take this opportunity to see just what a sauropod would look like with a trunk!'. Irons's drawing is dated 1975, so was obviously produced soon after the appearance of Coombs (1975).

Martin and Neave's big-lipped diplodocoid

More recently, John Martin has sometimes championed the trunk hypothesis, though to my knowledge he's never published anything on it: in the world of sauropod research, he is better known for arguing that sauropod necks functioned as mostly immobile horizontal beams* (Martin 1987, Martin et al. 1998), and for his work on the systematics and anatomy of Cetiosaurus (Upchurch & Martin 2002, 2003).

* A notion that I and my colleagues totally disagree with, incidentally.

Little known is that John worked together with a forensic anatomist, Richard Neave of Manchester University, to produce anatomical models of a Diplodocus with reconstructed soft tissues (Anon. 2000) [one of the models is shown here]. This model doesn't actually have a 'trunk' at all: instead, it has massive, flexible lips, and its nostrils are shown as being placed posterior to the lips, but not fused with them (a trunk or proboscis is best imagined as a fusion of the narial and lip musculature). That nostril position is pretty interesting, as it's similar to the position since advocated by other workers (more on this below). Unfortunately there doesn't seem to be any published information on how Martin and Neave's model was created, nor on the logic behind it: so far as I can tell, it's entirely speculative, providing the sauropod with lip and snout muscles that are not present in any reptile. The model has been featured on TV, though I can't remember where I saw it.

The trunked brachiosaur

Last year ace model maker Bill Munns - perhaps best known for his outstanding life-sized Gigantopithecus model - indulged in a bit of speculative model-making and produced another trunked sauropod [it's shown in the third image down in the composite used at the very top]. Whereas all other depictions have been of diplodocoids, Munns's model depicts the macronarian Gira- - Brachiosaurus. Like Coombs, Munns wondered if the dorsally located bony nostrils of sauropods might indicate the presence of a proboscis, and he went as far as providing the animal with a fairly long, elephant-like, tubular proboscis. Loren Coleman thought all of this was pretty interesting and wrote about Munns's new model at Cryptomundo. Munns has now produced a lengthy discussion of his sauropod models, and of the logic behind them, starting here. Because - I assume - relatively few Cryptomundo readers are going to be up to date with current views on dinosaurs, I thought it worth pointing out that the trunked hypothesis is not supported by present evidence, and a sort of gentle rebuttal to the idea was featured by Loren here.

After the usual maddening delay of several months, I here expand on these points and show just why the trunk hypothesis can be - in my opinion - conclusively rejected. I've always thought that the trunk hypothesis is totally non-sensical, and contradicted by a pile of evidence, and what irks me most about it is that it all rests on the fact that sauropods have dorsally located nostrils. As we'll see, and as is obvious if you look at the proboscides of extant mammals, there's a lot more to having a trunk than having large, dorsally located nostrils. Here, we go through the various anti-trunk points in turn. Some have been discussed before in connection with the hypothesis, others have not.

Skull shape. Mammals with proboscides or inferred proboscides - I'm thinking tapirs, pyrotheres, astrapotheres, the basal whale Makaracetus, some amynodontid rhinos, some oreodonts, Macrauchenia, dik-diks, saigas, desmans, sengis and so on - have narrow snouts. More precisely, the premaxillary and anterior maxillary part of the skull is narrow, typically being about half as wide as the back part of the skull (by which I mean everything from the orbits posteriorly). Many proboscideans differ from the aforementioned mammals in that they have strongly modified, shortened snouts and hypertrophied incisor sockets that obscure the ancestral skull shape, but note that proboscideans with unexpanded incisor sockets (things such as Phiomia, amebelodontids, and even juvenile and tuskless modern elephants) are typically narrow across the premaxillae, compared to the width across the cheeks. Given that the proboscis is used for selective foraging, and hence has to be narrow and prehensile at its tip, it follows that it must typically be the 'extension' of a narrow snout [the adjacent figure, from Paul (1998), shows the mouth widths of various tetrapods compared. Note in particular the broad mouths of (A) Apatosaurus, (B) the Tendaguru brachiosaur, (C) Camarasaurus and (D) Diplodocus].

Sauropods are a whole different kettle of tetrapod: their snouts are broad, sometimes remarkably so. Diplodocoids, popularly imagined by those unfamiliar with these animals to have lightly built, narrow, delicate skulls, had robust, rectangular skulls where the mouth was as broad as, or even broader than, the rest of the skull. Macronarians like Camarasaurus, brachiosaurs and titanosaurs, also had broad snouts. While a few trunked mammals have broad snouts (deinotheres come to mind*), the fact that there aren't really any narrow-snouted sauropods weighs heavily against the trunk hypothesis.

* And, yes, I'm fully aware of the controversy over deinothere trunk and facial anatomy.

Necks! The idea that sauropods might have had trunks seems particularly bizarre given that these animals had already evolved one of the most extreme and remarkable food-gathering organs in tetrapod history: namely, super-long necks. While it has been suggested that sauropod neck length might have been driven by sexual selection (Senter 2007) - an argument that is itself based on a flawed hypothesis about neck evolution in giraffes - and while some workers have argued that sauropod necks were beam-like, largely immobile, and hence overall useless for anything other than feeding off the ground, it is generally agreed that sauropod necks provided these animals with unparalleled vertical and lateral foraging ranges. Notably, trunked mammals are almost all short-necked. It looks as if trunks are only evolved when the animals concerned (a) have the required musculature and whatnot and (b) are relatively short-necked. So, we see trunks in such animals as tapirs, rhinos, astrapotheres and so on. A few proboscis-bearing mammals ruin this correlation, such as dik-diks. Then there's the litoptern Macrauchenia, traditionally depicted with a proboscis, and known to have had a reasonably long neck.

Nevertheless, the ridiculous, hypertrophied necks of sauropods seem to have played the same role that trunks do in the mammals that have them, and it really seems like overkill for both structures to have been present in tandem. This argument is circumstantial, but I still think there's something in it. Incidentally, I should note here that members of at least one sauropod lineage (Dicraeosauridae) evolved short necks. For the other reasons discussed here, trunks can still be excluded in these animals [Greg Paul illustration above shows high-browsing Morrison barosaurs, camarasaur and apatosaur].

Cranial musculature. One argument that gets consistently trotted out whenever the possibility of sauropod trunks is mentioned concerns the absence of facial muscles in sauropods, and in dinosaurs and reptiles in general. In mammals, a group of muscles ancestrally associated with the upper lip and nose have been co-opted to form a proboscis, including the levator labii, rectus nasi, and caninus (Witmer et al. 1999, Shoshani & Marchant 2001). Unless you're going to propose something way out there and totally, totally novel (like, say, an amuscular proboscis composed entirely of pneumatic cells, somehow controlled by valves), the total absence of these muscles in reptiles means that reptiles lack the basic equipment required to evolve a trunk. In other words, because the extant phylogenetic bracket (EPB) for sauropods (that is, the extant animals that 'bracket' sauropods in the phylogeny - namely, squamates, crocs and birds) shows that these facial muscles (or anything like them) were absent, it would be an unjustified speculation to infer their presence.

The loophole here is that bizarre novelties can arise that violate (if you will) the EPB*. The EPB for ornithischians and sauropodomorphs, for example, shows that cheeks (by which I mean sheets of tissue enclosing the lateral surfaces of the oral cavity) must be assumed absent, yet various lines of evidence show that such structures were present in these animals. Cheeks thus evolved as novelties within these clades (for previous Tet Zoo discussions of cheeks in dinosaurs see Therizinosauroids and Altengerel Perle and Ankylosaur week, day 5: Edmontonia). A similar line of logic could be used for cranial musculature: might sauropods have somehow evolved totally novel musculature that 'allowed' the evolution of a trunk?

* For a previous reaction to my use of the phrase 'violating the EPB' go here.

In short, no. In, err, long... muscles - particularly big, strong muscles like those involved in any hypothetical trunk - leave visible attachment sites, such as crests, scars, or fossae. These sorts of structures are obvious in extant trunked mammals. In the tapir skull shown here (and the one shown previously, in dorsal view), you can see obvious lumps and bumps and deep concavities (termed fossae), all of which are associated with the attachment of musculature. Such structures are also present in proboscideans, and also in extinct mammals inferred to have had a proboscis (Wall 1980). They are entirely absent in the skulls of sauropods. Also worth noting is that some workers have argued that the bony bars around the nostrils of sauropods appear far too weak to have anchored proboscis musculature (Paul 1987).

You could play devil's advocate and say that all of this is negative evidence. Well, ok, but remember that we can only infer structures - particularly super-controversial and counter-intuitive structures like sauropod trunks - when we have strong positive evidence. Or, simply, 'evidence'.

Nostril position and facial vasculature. As mentioned above, the sauropod trunk hypothesis came about because most sauropods have dorsally located bony nostrils. As mentioned earlier, the trunk hypothesis has not been widely adopted among dinosaur workers, but - on those few occasions when it has been adopted - people have imagined the nasal airway to emerge from the bony nostrils and to travel all the way down the trunk such that the fleshy nostrils are at its tip. Far more typical has been the assumption that the fleshy nostrils were actually located within the bony nostrils, and usually in a position somewhere round about the middle or posterior parts of the bony openings.

As Witmer (2001) showed, this is almost certainly incorrect, given that, in extant reptiles (including birds), the fleshy nostril is located anteriorly within the nostril region. And, by 'nostril region', I mean the area that also incorporates the nasal vestibular vascular plexus (NVVP), a complex mass of richly vascularised erectile tissue that surrounds the fleshy nostril. The NVVP leaves behind various osteological signs for its presence (predominantly foramina and canals for blood vessels). In sauropods, Witmer (2001) argued that the osteological correlates of the NVVP are located anterior to the anterior margin of the bony nostril: in other words, the fleshy nostril and its associated soft tissues were located down on the front part of the snout [see figure above, from Witmer (2001)]. This is why newer illustrations of sauropods show them with nostrils that are more anteriorly located than those in older illustrations. Luis Rey has taken to showing distinct fleshy tubes extending from the bony nostrils to the fleshy nostrils. However, like most structures in the heads of extinct archosaurs (I'm thinking bony fenestrae and the margins of bones), it seems more likely that structures such as the narial tubes were flush with the other soft tissues of the head, and didn't 'stand out' like this. Anyway, the osteological evidence for fleshy nostril position in sauropods is flatly incompatible with the trunk hypothesis.

Cranial neurology. Trunks are complicated, prehensile organs, and as such they require sophisticated muscular control. Knoll et al. (2006) noted that elephants have a huge facial nerve (10 mm wide near its emergence from beneath the jaw joint) that unites with a large maxillary branch of the trigeminal nerve. A huge facial nerve also innervates the proboscis in tapirs (Witmer et al. 1999). Evidence for huge cranial nerves is also present in fossil mammals thought to have had proboscides: in Astrapotherium and some amnyodontid rhinos, for example, an enlarged infraorbital foramen (this is the bony opening present on the zygomatic arch, ventral to the orbit) shows that the nerves and blood vessels associated with the snout region were hypertrophied as they are in extant proboscis-bearing mammals (Wall 1980).

Dinosaurs do not have infraorbital foramina, but we can still work out how big their facial nerves were if we have casts of their brains. It is entirely reasonable to assume that, if any other tetrapod were to evolve a proboscis, it also would need specialised, hypertrophied nerves. So: how do sauropods match up? As Knoll et al. (2006) showed, brain casts of Diplodocus and Camarasaurus show that the facial nerve roots were tiny (1 mm and 3 mm wide respectively). This demonstrates that the neurology required for a proboscis is emphatically absent in these sauropods. Because Knoll et al. (2006) were only able to get data from Diplodocus and Camarasaurus, they said that the trunk hypothesis remains viable for other sauropods. However, as argued here, other lines of evidence demonstrate convincingly that trunks were absent across the group [adjacent image, from Knoll et al. (2006), shows (A) endocast of Diplodocus compared with (B) facial nerve distribution in an elephant embryo. While the facial nerve (nerve VII) is tiny in the Diplodocus endocast, it's huge in the elephant].

Tooth wear. A reasonable amount of work on sauropod tooth wear has been produced. Little known outside of the field of dinosaur research, it seems, is that sauropod teeth are typically heavily worn, with large wear facets produced by abrasion from food (Barrett & Upchurch 1994, 1995, Calvo 1994, Christiansen 2000, Upchurch & Barrett 2000, Sereno & Wilson 2005). The steeply inclined wear facets that are usually present seem to have been a direct result of biting vegetation. Furthermore, inclined wear facets on the labial surfaces (= outside surfaces) of diplodocoid upper and lower jaw teeth show that these dinosaurs were grabbing foliage in the mouth, and then either pulling the head and neck sharply upwards (thereby producing inclined wear facets on the labial surfaces of the lower jaw teeth), or downwards (thereby producing facets on the labial surfaces of the upper jaw teeth). Upchurch & Barrett (2000) referred to this style of feeding [reconstructed below, from Barrett & Upchurch (1994)] as unilateral branch stripping. It is, again, flatly incompatible with trunk presence, because a diplodocoid that grabs branches in its mouth and then pulls its head sharply downwards is going to find that branch banging into and/or injuring its proboscis.

It's obvious that the extensive wear on sauropod teeth - which indicates raking or slicing - demonstrates extensive use of the teeth as feeding tools. Again, this seems incompatible with the presence of a trunk, as if sauropods had proboscides, wouldn't they be grabbing foliage with the trunk tip, and then passing it into the mouth, over the teeth? It's certainly difficult to imagine how they would be creating heavy wear on the labial sides of their teeth. Note that some titanosaurs possessed guillotine-like jaw edges that they surely used as pseudo-beaks for slicing off vegetation. Again, this is contradictory with the need for a trunk given that the whole point of a trunk is that it, rather than the jaws, is used to pluck vegetation.

Final thoughts

So, there we have it. I argue that the broad muzzles and super-long necks of sauropods are incompatible with trunk presence. The lack of appropriate facial musculature, the absence of muscle attachment sites, and the presence of small facial nerves all show that sauropods did not, and could not have had, trunks. Furthermore, the data we have on soft tissue nostril position, and on tooth wear, is also completely incompatible with the presence of a trunk. Given that, as mentioned a few times in this article, the trunk hypothesis has NOT been widely adopted by dinosaur workers - in fact it is very much a minority fringe opinion, rarely taken seriously - this rather lengthy appraisal might be akin to using a sledgehammer to open a peanut (or whatever the phrase is). Nevertheless, I felt it needed doing, in part because a detailed appraisal such as this, involving numerous lines of evidence, hasn't been produced before.

I must say that, goddammit, this all took far longer than planned. As I neared completion of this article it occurred to me that I might turn it into a paper at some point. True, it's all been published before, but then this was, after all, a review.

If you've come here because you think sauropods are neat, and if you don't already know about it, may I draw your attention to SV-POW!, a blog run by myself and two colleagues. It is devoted entirely to sauropod vertebrae. And for previous Tet Zoo sauropod posts see...

• Bones of the Krayt dragon
• The hands of sauropods: horseshoes, spiky columns, stumps and banana shapes
• The horror that is LOLSAUROPODS
• The world's most amazing sauropod
• Gigantoraptor, Eocursor and... baby Toni

Incidentally, astrapotheres were mentioned once or twice in this article. You must therefore visit...

• Dude, where's my astrapothere?
• Snorki the giant's friends and relatives

Refs - -

Anon. 2000. Just when you thought it was safe...... The Dinosaur Society Quarterly Magazine 3 (5), 8-9.

Bakker, R. T. 1971. Ecology of the brontosaurs. Nature 229, 172-174.

- . 1986. The Dinosaur Heresies. New Theories Unlocking the Mystery of Dinosaurs and their Extinction. William Morrow, New York.

Barrett, P. M. & Upchurch, P. 1994. Feeding mechanisms of Diplodocus. Gaia 10, 195-203.

- . & Upchurch, P. 1995. Sauropod feeding mechanisms: their bearing on palaeoecology. In Sun, A. & Wang, Y. (eds) Sixth Symposium on Mesozoic Terrestrial Ecosystems and Biota, Short Papers. China Ocean Press (Beijing), pp. 107-110.

Calvo, J. O. 1994. Jaw mechanics in sauropod dinosaurs. Gaia 10, 183-193.

Christiansen, P. 2000. Feeding mechanisms of the sauropod dinosaurs Brachiosaurus, Camarasaurus, Diplodocus, and Dicraeosaurus. Historical Biology 14, 137-152.

Coombs, W. P. 1975. Sauropod habits and habitats. Palaeogeography, Palaeoclimatology, Palaeoecology 17, 1-33.

Knoll, F., Galton, P. M. & López-Antoñanzas, R. 2006. Paleoneurological evidence against a proboscis in the sauropod dinosaur Diplodocus. Geobios 39, 215-221.

Long, R. A. & Welles, S. P. 1980. All New Dinosaurs and Their Friends. Bellerophon Books, Santa Barbara.

Martin, J. 1987. Mobility and feeding of Cetiosaurus (saurischia, sauropoda [sic]) - why the long neck? In Currie, P. J. & Koster, E. H.(eds) Fourth Symposium on Mesozoic Terrestrial Ecosystems, Short Papers. Boxtree Books (Drumheller, Alberta), pp. 154-159.

- ., Martin-Rolland, V. & Frey, E. 1998. Not cranes or masts, but beams: the biomechanics of sauropod necks. Oryctos 1, 113-120.

Norman, D. B. 1985. When Dinosaurs Ruled the Earth. Marshall Cavandish, London.

Paul, G. S. 1987. The science and art of restoring the life appearance of dinosaurs and their relatives - a rigorous how-to guide. In Czerkas, S. J. & Olson, E. C. (eds) Dinosaurs Past and Present Vol. II. Natural History Museum of Los Angeles County/University of Washington Press (Seattle and London), pp. 4-49.

- . 1998. Terramegathermy and Cope's Rule in the land of titans. Modern Geology 23, 179-217.

Senter, P. 2007. Necks for sex: sexual selection as an explanation for sauropod dinosaur neck elongation. Journal of Zoology 271, 45-53.

Sereno, P. C. & Wilson, J. A. 2005. Structure and evolution of a sauropod tooth battery. In Wilson, J. A. & Curry-Rogers, K. (eds) The Sauropods: Evolution and Paleobiology. University of California Press, Berkeley, pp. 157-177.

Shoshani, J. & Marchant, G. H. 2001. Hyoid apparatus: a little known complex of bones and its "contribution" to proboscidean evolution. In The World of Elephants - International Congress, Rome 2001, pp. 668-675.

Upchurch, P. & Barrett, P. M. 2000. The evolution of sauropod feeding mechanisms. In Sues, H.-D. (ed) Evolution of herbivory in terrestrial vertebrates: perspectives from the fossil record. Cambridge University Press (Cambridge), pp. 79-122.

- . & Martin, J. 2002. The Rutland Cetiosaurus: the anatomy and relationships of a Middle Jurassic British sauropod dinosaur. Palaeontology 45, 1049-1074.

- . & Martin, J. 2003. The anatomy and taxonomy of Cetiosaurus (Saurischia, Sauropoda) from the Middle Jurassic of England. Journal of Vertebrate Paleontology 23, 208-231.

Wall, W. P. 1980. Cranial evidence for a proboscis in Carducodon and a review of snout structure in the family Amynodontidae (Perissodactyla, Rhinocerotoidea). Journal of Paleontology 54, 968-977.

Witmer, L. M. 2001. Nostril position in dinosaurs and other vertebrates and its significance for nasal function. Science 293, 850-853.

- ., Sampson, S. D. & Solounias, N. 1999. The proboscis of tapirs (Mammalia: Perissodactyla): a case study in novel narial anatomy. Journal of Zoology 249, 249-267.