Guest post: Grazing giants – sauropod feeding

In today’s guest post, John Whitlock takes us through his recent PLoS One paper on the wear facets of sauropod teeth and what this might mean for their feeding habits. I’ve long had a general curiosity about this and was delighted to see renewed interest in this area, not least from someone who takes their sci-comms seriously enough to have signed up for AAB! Take it away please John:

I first got interested in sauropods, and diplodocoids in particular, in my first graduate-level course at the University of Michigan, back in 2003. We were looking at sauropod teeth, and (future advisor) Jeff Wilson was explaining the odd nature nature of the wear facets on Diplodocus’ teeth, with labial (external) wear surfaces on both uppers and lowers, rather than labial and lingual (internal) facets that faced each other when the mouth was closed. I decided that, as a first year graduate student, I should have no problem solving a (then) 117-year-old mystery.

As it turns out, it wasn’t so simple after all. What I did figure out in the course of investigating this, however, was that there was a lot more about diplodocoids that was unusual. The skull itself was something of an oddball among sauropods: long, horse-like faces and broad, square snouts, with all the tiny peg teeth pushed up to the very front of the jaws. It’s the sort of anatomy that could easily make you think that these animals were doing something really unusual to feed, and the hypotheses that have been put forth certainly reflect that: eating aquatic vegetation from riverbeds, plucking clams from riverbeds, scraping algae from rocks, capturing fish, and stripping or combing leaves off of branches. All I saw when I looked at the skulls, though, was a big Mesozoic cow, and I wondered why it wouldn’t have just eaten in a similar fashion: close-cropping of low-growing vegetation is as good a plan as any, after all.

Sauropod skull shapes. From Whitlock, 2011.

Low browsing wasn’t a new idea, but like other hypotheses of browsing behavior, it had never been really rigourously tested. Fortunately, because low browsing was something modern mammals do, there was a wealth of information and techniques that could be adapted to try and test these more “normal” hypotheses of feeding. An obvious choice was to include data based on dental microwear studies, which look at the damage food and other ingesta do to teeth and make assumptions about diet based on that damage, are well established as valid for fossil mammals, and have received a fair bit of attention from dinosaur workers as well (Paul Barrett, Jorge Calvo, Tony Fiorillo, and Vincent Williams, for example). From microwear data, we can get an idea about where an animal was eating, and to a more limited extent what types of food it was eating while there. For example, the closer to the ground you are, the more bits of extraneous grit you’ll find on plants, and these are the objects that tend to cause pitting. If you bite through woody stems or eat a lot of fruits, spores, or seeds, you’ll tend to end up with larger gouges or coarse scratches.

I also wanted to look at snout shape; paleontologists have always recognized the square snouts in diplodocoids as unusual, but it wasn’t until I was invited to work on Nigersaurus that I really realized how important that square shape was to the biology of these animals. In mammals, snout shape is highly correlated with feeding behavior, such that pointed snouts (think deer) belong to animals that feed selectively, and square snouts (think cows) belong to animals that feed less selectively; the idea was that the same pattern would hold true for sauropods.

Based on how mammals score for these features, I laid out a set of hypothetical data I might expect to see in sauropods based on how and where they ate:

-Browsing at ground height: square snouts, more pits than other features, fine scratches

-Browsing at mid-height: either square or round snouts, fewer pits than in ground height browsers

-Browsing in the upper canopy: either square or round snouts, very few pits

-Non-selective browsing: square snouts, subparallel scratches, small wear features

-Selective browsing: round/pointed snouts, cross-scratching, large wear features

I also wanted to know what conditions would result from branch-stripping behavior, and based on apes (who do strip leaves with their teeth), I came up with the following criteria:

-Branch-stripping: fine, subparallel scratches, but very few/no pits.

When all the results came in (skipping over all the stats), they looked like this:

Non-selective, ground-height browsing: Apatosaurus, Diplodocus, Nigersaurus

Mid-height, selective browsing: Dicraeosaurus, Suuwassea, Tornieria

Mid-height/Upper canopy, selective browsing: Brachiosaurus, Camarasaurus

Because they seemed so overtly cow-like to me, I wasn’t surprised to see animals like Diplodocus and Nigersaurus scoring out as ground-height, non-selective browsers, or that none of the microwear from any of the sauropods I looked at suggested branch-stripping; what I didn’t expect was that some diplodocoids, like Dicraeosaurus and Suuwassea, would actually present data suggesting they were intermediate browsers, both in terms of browse height and browse type. When we compare these results to what we know about the plant communities where diplodocoids lived, however, it becomes less surprising.

A Suuwassea tooth

Large diplodocids like Apatosaurus and Diplodocus lived in what is now preserved as the “southern” Morrison Formation, where at least intermittent savanna-type environments existed—lots of open spaces and low, soft ferny browse. Dicraeosaurus and Tornieria, which scored as mid-height selective browsers, lived in what is now Tendaguru: a (relatively) more densely forested environment, with few ferns and lots of woody browse. Suuwassea, another selective browser, hails from what is now the “northern” Morrison, typically characterised as a swampy, woody environment, unlike the savanna to the south. The snout of each animal was shaped in precisely the way you’d expect, based on the environmental conditions in which they lived.

Of course, I’m still not sure how best to explain those wear facets, but there’s always next year.