Mesozoic dinosaurs of several lineages famously possessed horns, frills, bony bosses, crests, frills, blah blah blah – you’ve heard all this a million times before. Pterosaurs were flamboyant creatures too. Why did these animals possess these so-called exaggerated structures? Together with Dave Hone, I’ve just published my latest missive on this issue (Hone & Naish 2013).

As discussed in previous Tet Zoo articles (see links below), I’m one of several researchers who thinks that exaggerated structures evolved primarily under sexual selection pressure: that is, the evolution of horns, frills, crests and so on was driven by the use of such structures in advertising genetic fitness. In wooing mates and intimidating rivals. There are several reasons for favouring this hypothesis: the structures concerned are anatomically similar to those of modern animals that use their structures in sexual selection; the structures exhibit growth rates suggestive of a sexual role; they are frequently costly in terms of growth and maintenance and hence very different from the zero-cost signals associated with other roles, and so on (Knell et al. 2012).

Sexual selection (the set of evolutionary phenomena that are relevant to the enhancing of mating success and fertilisation) is typically imagined to be distinct from natural selection (the set of evolutionary phenomena related to predator avoidance, feeding and foraging, surviving the rigours of climate, and so on). Indeed, the two phenomena frequently operate in contradiction. Palaeontologists have tended to dislike the idea that sexual selection might explain the evolution of exaggerated structures, and have instead proposed that crests, frills, sails and so on evolved within the context of natural selection, most typically as heat-shedding or heat-absorbing structures, or as organs with special aerodynamic or mechanical roles.

An idea that’s been popular among palaeontologists is that exaggerated structures served a special role in species recognition: that is, that they functioned as visual badges allowing members of a given species to recognise conspecifics and differentiate them from heterospecifics. The dinosaur literature basically frames this hypothesis as an alternative to the sexual selection one, at the same time downplaying the sexual selection hypothesis as problematic and without good support.

One major, crippling problem with the species recognition hypothesis is that crests and other exaggerated structures have yet to be shown to play a role in species recognition in any of the extant animals that have them (Knell & Sampson 2011). Yes, you read that right: feel free to read it again and allocate it to memory. Species recognition does not appear to be an important function for the horns, frills, antlers and so on of chameleons, hornbills, cassowaries, rhinos, deer, bovids, rhino beetles and whatever other extant animals you might consider as possible analogues for ornamented dinosaurs! Instead, the animals use those structures in (drumroll) sexual selection, which is partly (though not entirely) why I and my colleagues prefer the sexual selection hypothesis over others (Hone et al. 2012, Knell et al. 2012, 2013, Hone & Naish 2013). [In montage above, anole by Steven G. Johnson; cassowary by safaris.]

Our latest paper is devoted to a discussion of the species recognition hypothesis and, specifically, why we think it’s problematic and should be discarded. We’ve noted that dinosaur workers have increasingly taken to mentioning species recognition whenever they discuss exaggerated structures (see list of citations in Hone & Naish 2013), so now is a good time to try and set the record straight.

Interested parties will also know of the recent exchange of views that occurred in Trends in Ecology and Ecology between our group and Kevin Padian and Jack Horner (Knell et al. 2012, 2013, Padian & Horner 2013). This debate mostly concerns what we mean by ‘sexual selection’. I’ve already said above what I think it means; Padian & Horner (2013) argue that sexual selection should be applied only to the phenomenon that encourages the evolution of sexually dimorphic traits used in attracting mates or repelling rivals. They lean heavily on the idea that Darwin “invented sexual selection” (Padian & Horner 2013, p. 1): seeing as this is science, not dogma, we disagree with the notion that Darwin’s interpretation is sacred and that it should be adhered to above all else. In any case, Darwin did not (so far as we’ve been able to work out) state anywhere that he regarded sexual dimorphism as key to sexual selection. On the contrary, he even referred to the idea that sexual selection might well exist in cases where dimorphism is absent (Knell et al. 2013, p. 1).

And we absolutely reject Padian & Horner’s (2013) argument that sexual dimorphism is essential for the recognition of sexual selection: there’s unambiguous evidence from the living world that sexual selection is at play even when dimorphism is absent (Hone et al. 2012, Knell et al. 2012, 2013). [Adjacent photo by Marek Szczepanek.]

Moving on, which arguments have been put forward in support of the species recognition hypothesis, and how do we respond to them?

‘Directionality’ and ‘randomness’ in phylogeny

One argument is that the exaggerated structures present in dinosaurs seemingly evolved in a “relatively random” pattern, and it’s this random pattern that’s supposedly better indicative of species recognition (Padian & Horner 2011). After all, say Padian & Horner (2011), the point of the exaggerated structures within this context is that they serve to distinguish members of a species from its closest relative(s), and hence any anatomical variation that performed that job was evolutionarily advantageous.

However, our argument (Hone & Naish 2013) is that phylogenies are often ambiguous as goes the ‘directionality’ of exaggerated structures, so much so that you can say whatever you want about randomness. Knell & Sampson (2011) used a similar argument. As an example, there’s little (or no) doubt that the facial wattles, iridescent neck feathers, long tail feathers and so on of gamebirds have major roles in sexual selection. Yet, when mapped onto phylogenies, these structures seem to appear and disappear at random, with numerous independent losses and gains (e.g., Bonilla et al. 2010). In other words, the claim that randomly arranged exaggerated structures are more in keeping with a species recognition role than a sexual selection one is erroneous.

And, in keeping with what I just said about directionality, things are complicated when we look at Mesozoic dinosaurs anyway, since there are always trends in the evolution of their exaggerated structures that fit with the idea that exaggerated structures become more elaborate (and hence arguably ‘sexier’) over time. We use centrosaurine horned dinosaurs as an example (Hone & Naish 2013). Witness that, within this clade, we see a loss of brow horns and their replacement by a cornified supraorbital pad, a shortening of the nasal horn and its conversion to a gigantic nasal boss, an elaboration of big scales anterior to the nasal boss, and so on (Hieronymus et al. 2009).

In short, there’s no clear support from the ‘directionality’ or diversity of exaggerated structures for the species recognition hypothesis; their diversity and evolution is better consistent with roles in sexual selection.

Friends and relatives: the argument from sympatry

A cornerstone assumption of the species recognition hypothesis is that the exaggerated structures of the animals concerned are most prominent in sympatric species (that is, in species that lived alongside one another) (Padian & Horner 2011). Seen the other way round, the argument here is that a diverse array of exaggerated structures in sympatric species indicates the role of such structures in species recognition among those respective animals. We think that several observations make this argument moot: that is, the presence, evolution and elaboration of exaggerated structures fits with the concept of sexual selection, not with species recognition.

Firstly, living animals show that you don’t need exaggerated structures in order to differentiate conspecifics from heterospecifics. Consider that there are numerous extremely similar, sympatric frog, mammal, lizard and bird species that successfully do this without exaggerated structures (they use other cues, like body size, pigmentation, scent, vocalisations and so on). Indeed, among Mesozoic dinosaurs, there are sympatric species that seem to have distinguished one another just fine without the use of exaggerated structures: look at the contemporaneous iguanodontians, for example, that inhabited various Early Cretaceous ecosystems in Eurasia and North America (Hone & Naish 2013). In short, the whole idea that exaggerated structures are needed for species recognition is highly suspect – these elaborate organs are more in keeping with sexual selection.

Secondly, what about those dinosaur taxa bearing exaggerated structures that occur on their own, without close relatives? Examples: the Asian spinosaurid Ichthyovenator, the Chinese stegosaur Wuerhosaurus/Stegosaurus homheni. While the presence of unknown, as-yet-undiscovered related species remains possible, the existence of these singletons would seem to show that species recognition is not the mechanism that’s driving the existence of exaggerated structures in these cases.

Thirdly, sympatric dinosaurs species are not always obviously distinct in term of the morphology of their exaggerated structures, as they should be if species recognition was the key function of those structures. We use the apparently contemporaneous ceratopsians Protoceratops hellenikorhinus, Bagaceratops and Magnirostris as examples: these animals are not obviously and reliably distinct in frill and nasal boss anatomy “leaving us to wonder how the animals themselves might identify conspecifics if externally visible morphology was their only guide” (Hone & Naish 2013, p. 5). Other examples come from the plates and spines of contemporaneous stegosaurs, the cranial crests of oviraptorosaurs, and the cranial bosses and hornlets of tyrannosaurs.

Furthermore, a species recognition role for exaggerated structures indicates that species would differ with respect to the form of a single structure. Instead, however, we see numerous different signals that would be redundant in this context. “In ceratopsians, for example, we see elaboration in nasal horn, brow horn, jugal boss and frill midline and frill edge morphology (in addition to differences in body size, proportions, and integumentary anatomy!)” (Hone & Naish 2013, p. 5). Living animals verify that these multiple signals are better in keeping with the sexual selection hypothesis where, basically, the showier you are, the better (look at the adjacent Mandarin duck Aix galericulata).

According to the species recognition hypothesis, then, these dinosaurs have evolved (frequently) costly and almost wholly redundant structures in order to distinguish one another when they would have been better off evolving ‘zero-cost’ signals for such purposes (Knell & Sampson 2011, Hone & Naish 2013).

Fourthly, let’s consider how exaggerated structures evolved in the first place. According to the species recognition hypothesis, they’re ‘badges’ that sort species x from species y. Ok, but ‘badges’ of this sort would only be needed if speciation was occurring sympatrically (that is, a single population sorted itself into two). Needless to say, if the structures evolved in allopatric species, they’re redundant within the context of the species recognition hypothesis and can’t be used as evidence in its support.

So, we have to think in terms of sympatric speciation when considering the species recognition hypothesis. The only model we can imagine here (Hone & Naish 2013) is that members of the ancestral population evolved incipient versions of the respective exaggerated structures; other members of said population then decided to mate preferentially with those ornamented individuals… eventually, speciation occurred. However, if we’re talking about individuals choosing ornamented individuals over unornamented ones, the ornaments obviously evolved within the context of sexual selection and ‘species recognition’ cannot be considered integral to their origins.

No sexual dimorphism, so no sexual selection. Wait… what?

Another argument used to support the species recognition hypothesis is that an absence of sexual dimorphism in dinosaurs negates the sexual selection hypothesis (Hieronymus et al. 2009). This ‘no sexual dimorphism, thus no sexual selection’ thing keeps on coming up at the moment and it’s utterly erroneous.

Firstly, it’s often impossible to be sure that sexual dimorphism truly is absent in fossil animals: the sample sizes we’re working with are frequently just not good enough. Secondly, even in those cases when we can be confident that sexual dimorphism was absent, the possible existence of mutual sexual selection (the phenomenon whereby individuals of both sexes are ornamented) means that sexual selection is still at play.

Like many of the phenomena discussed here, mutual sexual selection is well established and uncontroversial among biologists: it’s fairly widespread in birds and hence our proposal that it might also have been widespread in other dinosaurs (Hone et al. 2012) is not only intriguingly but theoretically reasonable.

Ontogenetic morphing and species recognition: wait a minute… aren’t those concepts incompatible?

Jack Horner is one of the primary proponents of the species recognition hypothesis in dinosaurs. As is well known, Horner also advocates the hypothesis that certain dinosaurs (most notably chasmosaurine ceratopsians and pachycephalosaurs) underwent profound anatomical changes during ontogeny. So – to take the most famous example – short-frilled Triceratops underwent a major change in morphology at some stage during adulthood, morphing into long-frilled Torosaurus. Note at this point the evidence (from medullary bone) showing that dinosaurs of many (or most or all) lineages were capable of breeding before reaching full skeletal maturity (Lee & Werning 2008).

If the model of ontogenetic morphing is correct (I agree with others* that it may well not be), we therefore have to imagine dinosaur populations where an anatomically diverse array of individuals are available as mating partners. Within this model, a courting Triceratops thus has short-frilled animals as potential partners as well as long-frilled animals, as well as all manner of intermediates. In other words, it’s faced with several transforming morphs, the more disparate of which are more different from one another than some are from the members of other species.

* Farke (2011), Longrich & Field (2012).

In short, the ontogenetic morphing hypothesis is in direct contradiction with the species recognition hypothesis since the latter requires that members of a species are enough alike that we can obviously identify the anatomical features that allowed them to distinguish conspecifics from heterospecifics. By the way, the notion of late ontogenetic development of a given anatomical structure is not incompatible with a role for such structures in sexual selection (Knell et al. 2012).

Species recognition bad; sexual selection good

Mesozoic dinosaurs (and pterosaurs, and other fossil animals) were surely capable of distinguishing conspecifics and heterospecifics. The point is that they clearly didn’t ‘need’ to evolve exaggerated structures – horns, frills, bony plates, dorsal sails and so on – in order to do this, and all of the arguments put forward in defence of this idea either don’t support it, or are equivocal and just as likely associated with sexual selection. There’s no obvious support from phylogenetic ‘randomness’ that exaggerated structures evolved within the context of species recognition, a workable model of species recognition that disavows a prominent role for sexual selection has yet to be proposed, claims that subtle or absent sexual dimorphism disproves the existence of sexual selection are false, and the ontogenetic morphing thought to occur in dinosaurs is inconsistent with the species recognition hypothesis. Furthermore, living animals show that exaggerated structures are more to do with sexual selection than species recognition.

As is so often the case with the biology of extinct animals, we know so little that a possible role for species recognition in the evolution, persistence and elaboration of exaggerated structures cannot be ruled out entirely.

However, our primary argument is that there’s “no good evidence that might support this hypothesis and it should not currently be considered viable” (Hone & Naish 2013, p. 7). As I and my co-authors have argued in previous papers (Hone et al. 2012, Knell et al. 2012, 2013), the exaggerated structures of dinosaurs (and pterosaurs) are better explained by a role in sexual selection (and sexual selection should not be used in the peculiar and restrictive definition recommended by Padian & Horner (2013)).

I somehow doubt that we’ll get some members of the community to give up on their idea that the term ‘sexual selection’ means something far more restrictive that the understanding of the term favoured by myself, my colleagues and, I think, the vast majority of biologists who work on this subject. But I do hope that we’ve presented a serious challenge to the species recognition hypothesis. At the very least, dinosaur workers should think very carefully before they claim – as they so often have – that it represents the best explanation for the evolution of exaggerated structures.

For previous Tet Zoo articles on sexual selection and related aspects of Mesozoic archosaur behaviour and biology, see…

Refs - -

Bonilla, A. J., Braun, E. L. & Kimball, R. T. 2010. Comparative molecular evolution and phylogenetic utility of 3’-UTRs and introns in Galliforms [sic]. Molecular Phylogenetics and Evolution 56, 536-542.

Farke, A. A. 2011. Anatomy and taxonomic status of the chasmosaurine ceratopsid Nedoceratops hatcheri from the Upper Cretaceous Lance Formation of Wyoming, U.S.A. PLoS ONE 6, e16196.

Hieronymus, T. L., Witmer, L. M., Tanke, D. H. & Currie, P. J. 2009. The facial integument of centrosaurine ceratopsids: morphological and histological correlates of novel skin structures. Anatomical Record 292, 1370-1396.

Hone, D. W. E. & Naish, D. 2013. The ‘species recognition hypothesis’ does not explain the presence and evolution of exaggerated structures in non-avialan dinosaurs. Journal of Zoology doi:10.1111/jzo.12035

- ., Naish, D. & Cuthill, I. C. 2012. Does mutual sexual selection explain the evolution of head crests in pterosaurs and dinosaurs? Lethaia 45, 139-156.

Knell, R., Naish, D., Tomkins, J. L. & Hone, D. W. E. 2012. Sexual selection in prehistoric animals: detection and implications. Trends in Ecology and Evolution 28, 38-47.

- ., Naish, D., Tomkins, J. L. & Hone, D. W. E. 2013. Is sexual selection defined by dimorphism alone? A reply to Padian and Horner. Trends in Ecology and Evolution

- . & Sampson, S. 2011. Bizarre structures in dinosaurs: species recognition or sexual selection? A response to Padian and Horner. Journal of Zoology 283, 18-22.

Lee, A. H. & Werning, S. 2008. Sexual maturity in growing dinosaurs does not fit reptilian growth models. Proceedings of the National Academy of Sciences 105, 582-587.

Longrich, N. R. & Field, D. J. 2012. Torosaurus is not Triceratops: ontogeny in chasmosaurine ceratopsids as a case study in dinosaur taxonomy. PLoS ONE 7, e32623

Padian, K. & Horner, J. 2011. The evolution of ‘bizarre structures’ in dinosaurs: biomechanics, sexual selection, social selection, or species recognition? Journal of Zoology 283, 3-17.

- . & Horner, J. R. 2013. Misconceptions of sexual selection and species recognition: a response to Knell et al. and to Mendelson and Shaw. Trends in Ecology and Evolution 28,