Display features in the fossil record

It’s been more than a while coming but here’s an actual normal blogpost for the blog that’s not just PR for one of my own papers or projects (don’t worry, more of that coming sooner or later). This one has been prompted by some repeated comments I’ve seen in recent months about the hypothesis of various features being used for display by academics discussing dinosaurs in particular, but other extinct animals too.

The argument basically runs ‘you say it’s for display only because you don’t know what it is’ and usually followed with ‘like when archaeologists say it’s for ceremonial purposes when they don’t know what it’s for’. I can’t speak for my fellow professionals studying human culture, but I can very much speak for the assessment of display features having written perhaps more on this than anyone else when it comes to dinosaurs and pterosaurs at least.

First off, yeah, some researchers are very much guilty of this. One recent paper did argue something was for ‘display’ and that was the last word on the subject. That is, there was no actual evidence or discussion of the implications and how it might function or have evolved or what it was a good signal etc. and that’s clearly suboptimal at best. And it’s hardly new, it’s a classic old argument for lots of things on dinosaurs that’s been about for a century at this point and so people arguing for display without data isn’t some recent phenomenon. However, for plenty of cases we either do have decent scientific evidence or it’s fairly trivial to make a reasonable argument and that comes from our understanding of sexual selection in particular and signaling structures in general. So here’s a breakdown of the kind of lines of evidence and reasoning that can support display as a function.

1. It has no clear mechanical function. Not every bit of anatomy is functional in presenting a positive advantage to an animal, and some can be optimised for multiple things, or are used only very occasionally, or, yes, can be cryptic and we don’t know what they are for. But in general, selection is very good at getting rid of things that are costly and not useful (see how quickly flightless birds reduce their wings for example) and things argued to be for display are often large and heavy and are unlikely to survive many round of selection.

2. Diversity of form between species. There’s a reason the claws, fingers, ulna, humeri, spine and even ribs of moles, golden moles, marsupial moles, pangolins, aardvarks, armadillos and anteaters look very similar and that’s convergent evolution based on strong selection for a clear mechanical function. Animals, especially closely related ones, doing the same things in the same ways will almost inevitably end up with very similar anatomy. There’s a reason the wings of birds all look similar (flight), but the variety seen in their tails or head wattles etc. (display) are so varied. There’s probably only one or two optimum mechanical shapes and repeatedly deviating from that, especially in close relatives is a display hallmark. There’s also a general suggestion (though I think untested) that these tend to evolve rapidly as well compared to more classic functional traits.

3. Diversity of form within species. Moose all look alike but their antlers can be very different to one another and there’s usually far more variability of display features between individuals than other anatomical features, and that’s before the possibility of things like dimorphism (though an absence of dimorphism is not an argument against a signalling function for various reasons).

4. Rapid growth late in ontogeny. Sexually selected and display structures grow when the animal is at, or close to, sexual maturity and are very small or non-existent before then. So if there’s any indications of the growth rate from having multiple animals at different ages or sizes this can really help.

5. Structures are costly. A related point to 1, but the idea of ‘honest’ signals means that these features should be expensive to grow or maintain and have some kind of disadvantage for bearing them. And that also means they tend to be big and obvious (though with various trade-offs often at play limiting size – things can’t grow forever).

6. Analogy. While few features are clearly analogous to those seen in living clades (though of course some like fossil deer have lots of living and well-studied relatives) it is possible to draw analogies for some. The elongate tail streamers of microraptorines and various Cretaceous birds are obviously similar to those of numerous extant birds which have been shown to be signaling structures, so it’s reasonable to infer that similarly shaped ones in related animals with similar ecologies and behaviours and doing similar things.

Not everything fits these moulds perfectly. Features can be multi-functional like elephant tusks where they are under sexual selection but also are used to fight off predators, strip bark from trees and other things and probably are under selection to optimise multiple activities. And of course functions can change over evolutionary history with, for example, horns potentially shifting from an initial display feature to an anti-predator function or combining the two. Thus what the original function of a feature may have been and what selection pressures drove it to its current condition are not necessarily the same thing (though I suspect often are).

Take something like pterosaur head crests which have repeatedly been suggested to have some kind of steering function. We’d expect there to be only one or two optimised versions of this given the complexities of flight and the extreme similarity of pterosaur wings to each other, but instead we see enormous varieties of crests, they vary between and within species and both grow in size and change shape during ontogeny and are apparently small or absent in young juveniles. Despite the suggestion that this has a mechanical advantage, it’s not clear how it would work and one might expect if head crests were so useful they would have appeared in birds and bats too at some point, and it’s not like pterosaurs are short of flight control surfaces. Plus of course, for such light and flying animals, these would have been heavy features and therefore presumably costly.

So it’s fairly easy to make a case for these as display features even if we can’t do a detailed analysis of their flight mechanics or look at the detailed ontogeny and variation of many (any?) species to the degree we would like. In short, yes, palaeontologists need to be much better at explaining how and why they are arguing for display as a feature and simply saying ‘it’s big and odd’ while kinda hinting at a couple of these points, really isn’t good enough. But on the other hand, a lot of the things argued to be display features (ankylosaur armour, ceratopsian frills, hadrosaur crests, tyrannosaur hornlets, spinosaur sails etc.) fit most or all of these categories and even if in-depth analyses aren’t possible, it’s certainly a reasonable starting hypothesis that they are there for display.

So the often knee-jerk response of ‘ugh, you just say it’s display without evidence’ belies a real lack of understanding of the ways we can make reasonable inferences about these features and the simple fact that big and weird structures almost by default will match these lines of evidence (when say a big tooth or long leg or extra toe will not) should not argue against these as a starting point for discussion. Display features are rampant in large tetrapods at least and it should be no surprise that highly vision-oriented animals like dinosaurs and pterosaurs would have gone down various display routes. Yes, we need better arguments and testing, but I’m more than confident that many of these features will ultimately be shown to have had display as a major part of their functionality.