New paper: Constructing and testing hypotheses of dinosaur foot motion

Today sees the online (in press) publication of work I’ve been showing off since early in my Marie Curie Post-doc, way back in 2012!

Falkingham, P.L., Turner, M.L. and Gatesy, S.M. (2020), Constructing and testing hypotheses of dinosaur foot motions from fossil tracks using digitization and simulation. Palaeontology. doi:10.1111/pala.12502

The paper’s open access, so click the link above and have a read.

We used photogrammetry to digitize a really impressive 4-peice track from the Beneski Museum of Natural History, Amherst. The track was in four pieces, and in different cabinets, having at some point in the past been separated. While they all have the same, original, specimen number carved into them, the each have more modern numbers which are non-sequential. It was an exciting time when we realised they fit together (I think it was Steve that realised this first). It’s not surprising they were seperated – the tracks on each surface look fundamentally different, from short toes and a long metatarsus impression at the ‘top’, to three parallel scratches lower down:

With the digital models aligned in Maya, we were able to locate analogous points on each surface, toe tips, hypex, and so on. With those positions marked in 3D space, we could constrain the size of the foot. Having done that, we created a simplified foot from cylinders (one cylinder for each toe, one for the metatarsals), and animated the foot such that it hit, as best as possible, each landmark:

We termed this a ‘hypothesis of motion’.

Hypotheses are there to be tested though, so to test our hypothesis we ran the motion through a particle simulation, using LIGGGHTS running on ARCHER.

The idea being that if sediment motion resulted in a very different looking track, our hypothesis would be rejected, but if the simulation matched the fossil, then our hypothesis would be supported.

Hurrah! Whilst this doesn’t prove our foot motions are correct, it does support our hypothesis.

Our simulated track then gives us a wealth of information about how the track formed, and what it looks like volumetrically, as we’ve seen before in Guineafowl tracks (Falkingham and Gatesy 2014).

I’ve been sitting on the bulk of this work for years, so it’s nice to see it out finally. Ahead lies some really interesting work exploring how accurate we can get, or need to be. What effect does adding joints to the toes have? What about widening the metatarsal cylinder, or adding claws, or scales? All in due course…