Guest post: when pachycephalosaurs attack.

Today Eric Snively takes us through hi very recent PLoS One paper on head-butting in pachycephalosaurs. Eric is linked to Larry Witmer’s lab, so it’s no surprise that this is a paper heavy on function and analysis, though he started this work while still based in Canada with Phil Currie’s group. The idea that these dinosaurs may have fought each other by butting heads, or even just smashing into an opponent’s flanks have been around for ages and gone back and forth many times. Here Eric shows how the head of Stegoceras was apparently adapted for hard impacts in a way otherwise only found in heed-butting animals.

For someone who works on tyrannosaur feeding, pachycephalosaurs are an innocuous and probably tasty group of dinosaurs. Pachycephalosaurs were thick-headed, in a good way with expanded bone above the brain and even extra bones above the eyes. Some like adult Pachycephalosaurus, Prenocephale, and Stegoceras had big domes on their heads. These domes raise a red mist of intraspecies conflict, not among pachycephalosaurs (who are dead), but among scientists with different ideas about their development, functional morphology, and phylogenetics. Some of these friendly disagreements are close to resolution, thanks to extensive work in the field, museums, and labs, and some quantitative tests of our intuition.

First, an all-star team led by Ryan Schott and David Evans (University of Toronto) showed that many isolated pachycephalosaur domes are from one paleospecies, Stegoceras validum. Ryan and colleagues found that these domes form a continuum of growth, using morphometrics of external dome geometry and internal vascularity. External dome measurements plot against each other predictably, as expected if the domes had a single ontogenetic trajectory characteristic of one species. Relative amounts of superficial compact bone and deep cancellous (spongy) bone also have a continuous trajectory, as the vascular spongy bone regressed and the compact bone grew in thickness. The authors used micro-CT scanning to examine the 3D internal histology (tissue structure) of the bone.

Stegoceras skull sections.. From Snively & Theodore, 2011.

This was a cool study for me. Many of the authors are old field colleagues from Alberta (I was there when Caleb Brown assembled one of the figured specimens). The micro-CT scans supported the dome growth hypotheses of Mark Goodwin and Jack Horner, which were based on spectacular and arduous histological work. Finally, it corroborated suspicions that struts of cancellous bone were closely-packed enough to absorb energy, if pachycephalosaurs bashed their heads together.

That leads to a second recent work on pachycephalosaurs, which came out the same week and in the same journal (PLoS One) as Ryan’s publication. Jessica Theodor and I tested the hypothesis that Stegoceras could butt heads, a controversial idea advanced by Peter Galton, Hans Sues, Teresa Maryańska, and McNeil Alexander.  Jessica studies the ear evolution of artiodactyls (cloven-hoofed mammals), to unravel how whales evolved from within Artiodactyla. She CT scanned skulls of lots of artiodactyls from the University of Calgary collections. Andrew Farke generously gave us scans of a duiker, a little head-butting antelope (from Dutch for ‘diver,’ because they dive into undergrowth to escape predators). Thanks to Phil Currie, Mike Caldwell, and Larry Witmer, we got great scans of Stegoceras and another pachycephalosaur called Prenocephale.

The most pachycephalosaur-like artiodactyls, with alternating thick layers of compact and cancellous bone, were head-butters like musk oxen, bighorn sheep, and especially the duiker. But who was best at head butting? We constructed finite element (FE) models of the animals’ crania, and subjected them to simulated collision forces. FE analysis divides a continuous structure into lots of discrete elements. With forces, constraints, and material properties, the analysis determines displacement of nodes connecting the elements. From the displacements we get stresses, strains, energy density, and safety factors showing how far a structure is from breaking or permanent deformation. Osteologically, the dome of Stegoceras was best able to survive collisions, and the flat-headed llama (not a head-butter) was the worst. Our force for the giraffe was too high, but bring it down and Stegoceras still wins. Among modern artiodactyls, musk oxen were the best overall, because we incorporated their energy-absorbing keratin pad into our models. With a similar pad, Stegoceras would do even better (Andrew Cox and I showed this in a previous paper).

We went beyond testing whether Stegoceras could butt heads with each other, and addressed whether they did. Usually we require trace evidence, like trackways and tooth marks, to demonstrate behavior in fossil animals. Using tooth marks and local sedimentology, Dave Hone and Mahito Watabe demonstrated scavenging by Tarbosaurus, and Ken Carpenter and John Happ showed that Tyrannosaurus tangled with live prey. What if we lack trace evidence, or clear phylogenetic evidence bracketing extinct animals’ behaviors between those of living relatives?

Jessica and I introduced a probabilistic method of inferring behavior in fossil animals, from morphology and behavior in possible modern analogs. Called correlate disruption, our method builds on a medical diagnostic algorithm (introduced to paleo by Tobin Hieronymus) to see how well morphologies diagnose behavior. If you mis-diagnose an animal with lots of morphological ‘symptoms’ for a behavior, it screws up morphology-behavior linkages by a discrete amount. Stegoceras has a full complement of anatomical features seen in modern head butters, and some improvements of its own. If we say Stegoceras didn’t butt heads, it disrupts morphology-behavior probabilities for head-butting about as badly as if we mis-assign a musk ox or duiker.

In paleontology it’s always good to avoid definitive proclamations, and instead talk about ranges of possibility. With our method and future refinements, we can discuss probabilities of behavior. The method can incorporate continuous biomechanical or morphometric results (like Ryan Schott’s), to complement morphological features. So did pachycephalosaurs butt heads? We don’t yet know, but if I tripped over one’s skull in the woods and fell on a duiker’s, I’d have a quantifiable idea.

Stegoceras, a musk ox and pronghorn. From Snively & Theodor, 2011.