Sunday, December 20, 2015

Treading Ambiguously: Theropod Versus Bird Footprints

Hello, Dear Readers!

I'm back in the office full-time for the rest of the year, so as part of our promoting our science mandate, I have the opportunity to blog about some of the nifty cool science I've recently published. Here I'm going to talk about one of our newest papers, entitled Birding By Foot: A Critical Look at the Synapomorphy- and Phenetic-Based Approaches to Trackmaker Identification of Enigmatic Tridactyl Mesozoic Traces. Or, in other words, "How do I tell large bird footprints from small theropod footprints?"

We do a lot of work in the Early Cretaceous in British Columbia. By "a lot", I mean that the majority of our dinosaur and other vertebrate footprint localities are in Early Cretaceous (between about 145 million years old and 100 million years old - check out the Geological Society of America's latest timescale) sedimentary rock.

Exciting things were happening in Bird Land during the Early Cretaceous. More and more bird skeletons are being discovered from the Early Cretaceous every year, but most of these are not in North America. Recently, Wang et al. (2015) published on Archaeornithura, a small ornithuromorph bird capable of flight, from the Lower Cretaceous Huajiying Formation of northeastern China. The specimen is also beautifully feathered...except on the long legs, where feathers are conspicuously absent from the tibiotarsus and tarsometatarsus (lower leg). The equivalents on our bodies would be shin and long bones in our foot, but we mammals have a strange skeleton when compared to birds. Long, unfeathered legs strongly suggests a wading bird/shorebird lifestyle interpretation for Archaeornithura.

This piques my interest. In western Canada there is a pretty decent record of bird footprints from the Early Cretaceous. These footprints are from birds similar in habitat preference, shape, and size to our modern shorebirds and wading birds. When the skeleton of a flying (volant is the term used in scientific lit) Early Cretaceous bird that may have skittered along the shoreline is described, I get all chirpy-happy.

Looking pictures of the feet of Archaeornithura, a footprint from this animal would likely not be identified as anything other than a bird footprint: the foot is small (rough estimate of 15mm from the "heel" of the foot to the tip of the longest toe, digit III), and there is a well-developed hallux, or digit I that faces almost straight back from the foot. So far, so good. China's Early Cretaceous bird footprint record is amazing - and growing - so having a early shorebird skeleton gives us a potential trackmaker for one of these footprint types.

Identifying modern theropod tracks is fairly simple: we only have one group of theropods running and flapping about right now, and those are birds. Even though they are also 100% theropod, they are very specialized theropods, or 100% bird. There are no non-bird theropods visiting our bird feeders.

This isn't the case in the Early Cretaceous. There were non-avian theropods. There were Paraves - all theropods more closely related to birds than to oviraptorosaurs, like Velociraptor and Microraptor. There were Avialae - all theropods more closely related to modern birds (Aves) than dromaeosaurs. There are many many more categories, and then we get to Aves, our modern birds. All of these groups were running around in the Early Cretaceous, leaving footprints. How do we recognize a bird footprint as a footprint that could not have been made by anything other than a bird? It's like trying to find Cinderella (the trackmaker) using the glass slipper (the footprint), except that Cinderella died a long time ago and all we have is the skeleton of her foot to try to fit into the shoe.

So...what would the footprints from these different groups look like? How birdy does a footprint need to be in the Early Cretaceous before we call it a bird footprint? Welcome to my area of research! Since I see a lot of Early Cretaceous footprints, I'm starting to see a few track shapes that are baffling (exciting) this bird nut. Are they bird tracks that just look theropody? Are they theropod tracks that just look birdy? Are they really small plant-eating dinosaur tracks that look birdy and theropody?

NOTE: Birdy and theropody are not technical terms. I've just spent months and months writing "possessing the characteristics of bird footprints" when I really wanted to write "this feature looks birdy", so...yeah. "Birdy" is me rewarding my brain for behaving itself.

Even if there were no skeletons of small shore- and wading birds from the Early Cretaceous, small bird footprints are relatively easy to identify. In fact, the first bird footprints described from the Early Cretaceous of North America were done without any contemporaneous skeletons to support that identification. They looked birdy enough to identify as fossil bird footprints. If a footprint is small, (under 5 cm long), has widely splayed digits, and has a backwards-facing hallux (digit I), very few people are going to look at that footprint and say it's not a bird.  This is called the phenetic-based method of identifying a potential trackmaker. Basically, if it looks like a bird footprint, and walks like a bird, it's a bird. Lockley et al. (1992) provided this birdy footprint checklist:

1. It looks like a modern bird footprint,
2. Small size,
3. Slender toes,
4. Toe splay (or digit divarication) of the forward-facing toes wider between 110-120 degrees, or more,
5. Backwards facing (or posteriorly directed) hallux,
6. Slender claws, and
7. The claws on the inner and outer toes curve away from the middle toe.

There are a few more, such as track density, associated feeding traces, and associated fauna, but these have little to do with the shape and structure of a bird foot, so I won't be talking about them right now.

The other method for proposing a trackmaker for a footprint is called the synapomorphy-based method. If there is a feature in a footprint that is made by a derived character that is shared by all the members of that group, you can say the potential trackmaker came from that group. Carrano and Wilson (2001) listed the synapomorphies that have a chance of preserving as trace fossils. I refined the list further and stated the ones that were foot-specific:

1. We can say the trackmaker is a theropod if the footprint has toes with claws, and if the axis of the footprint is formed by the middle digit (mesaxonic).

2. We can say the trackmaker is Paraves if the footprint has a reduced inner toe (digit II) that lacks a claw impression.

3. We can say the trackmaker is a bird if the footprint was made on thin substrate and has a backwards-pointing hallux. This is the only synapomorphy-based feature that can be used to identify a footprint as a bird footprint in this method.

A frustratingly short list, no? Yes and no. Synapomorphies, if they impress, will provide an unambiguous "YES! This is most definitely a footprint of Awesomasaurus megarex!" A synapomorphy-based identifucation will also guard against rather, um, let's call them over-excited people, from sticking a sauropod identification on every large hole in the ground, or a tyrannosaurid identification on every large poorly preserved footprint with three toes from the latest Cretaceous Period. A major problem with the synapomorphy-based method is that most synapomorphies are NOT foot-based. Unless a Tyrannosaurus rex goes Hellraiser-wild, strips the skin and flesh off of its skull, and then face-plants into a nice clay-rich silt, most synapomorphies are going to remain unimpressed, and unpreserved, in the trace fossil record.

What do modern bird footprints tell us?
There are issues with both methods, and all of the issues are best demonstrated by looking at modern bird footprints. Birds, in their glorious diversity, come in all shapes, sizes, and habitat preferences. They are a perfect test case for both the phenetic- and synapomorphy-based methods. I looked at the four most common features that are used in the scientific literature to identify bird footprints: the backwards pointed hallux, small size, widely splayed toes, and skinny toes.

Synapomorphies Are Not Always Impressive
Let's look at the posteriorly-oriented digit I, a.k.a. the hallux. This feature is used in both the synapomorphy-based and phenetic-based methods of identifying a trackmaker. Here's an image of the foot of a bird that spends a lot of time on the ground. Festively enough, it's a turkey foot.
Wikimedia photo of turkey feet. The hallux is the lowest backwards-pointing toe. The structure higher up on the metatarsals (or "lower leg") is a spur, common in male gamebirds.
The toe hanging off the back of the foot is digit I. This is the equivalent of our big toe. If a human foot was shaped the same way as a bird's foot, our big toe would be hanging off the sole of our foot, and we'd have to walk on our tip-toes to avoid squishing it. Let that mental image sink in for a minute.

Each toe (or digit) is a series of smaller bones called phalanges. Each set of phalanges is attached to a metatarsal. Metatarsal I, or the metatarsal for the hallux, isn't a fully-developed bone in most theropods: it's short and attaches somewhere on metatarsal II, like this:

Comparison of theropod (Coelophysis, Deinonychus) and bird (Archaeopteryx, Pigeon) feet, showing where digit I (the short one) attaches to the foot. Digit I of theropods attaches much higher than the digit I for birds. Image created and used with permission by Emily Willoughby. Once you follow this link to the page with the above image, check out the rest of the site. The feathered theropods are glorious!
The big difference between birds and non-avian theropods is where the metatarsal for digit I attaches. In theropods, you can see that digit I is attached higher up on the shaft of metatarsal II. Compare that to the images of Archaeopteryx and the pigeon on the right: metatarsal I attaches low down on metatarsal II - much lower than in the non-avian theropods. With a metatarsal I (and the toe that attaches to it) attached so low on the metatarsal, it's easy to assume that the hallux would always make an impression in a bird footprint. Unfortunately (for us ichnologists) that's not the case.

First, there is a HUGE diversity in hallux length for birds that spend a great deal of time on the ground: this means gamebirds (grouse, turkeys, pheasants), shorebirds (plovers, sandpipers), and long-legged wading birds (cranes, herons, storks). Some of our small shorebirds, like plovers, have a relatively poorly developed digit I: metatarsal I doesn't really make much of an attachment mark on metatarsal II, and the toe for digit I is very short.

Distal metatarsals of the Semipalmated Plover (Charadrius semipalmatus). There really isn't an obvious attachment point for the hallux, which should be somewhere to the left of the rounded end of metatarsal II.

Plover footprints don't often have a digit I impression with their footprints, even if they walk in the best-case sediment. A fine-grained sediment (like silt or clay) can't preserve a hallux that doesn't hit the ground. If a bird in the Cretaceous has a plover-like foot, it's going to leave a footprint without a hallux (check out Paxavipes footprints here). Paxavipes can be identified as a bird footprint by the phenetic method, but since it lacks a hallux in a fine-grained sediment, it cannot be considered a bird footprint using only the synapomorphy method.

No one is going to have much trouble convincing other people that Paxavipes is a bird footprint - after all, it's teeny! But size is a really crappy way to identify a potential trackmaker (remember Carnosauria?) Many birds may be (and were) small, but birds are not exclusively small-bodied. OK, compared to T. rex, all birds, even the ostrich, are going to seem small. However, birds like ostrich, herons, cranes, storks, turkeys are not only large in the bird world, but they are similar in size to small non-avian theropods, as are their footprints. Sandhill Crane footprints range in size between 9 cm and 12 cm in length (Elbroch and Marks, 2001), and the small theropod footprints called Irenichnites are about 10 cm in length.

Second, there's a huge diversity of sediments that birds of all hallux types walk in. Not all sediments are the best at preserving all the features (see webbing versus skin impressions in my previous post here). For birds such as the Great Blue Heron, their digit I is so long and robust that it will likely impress in some fashion in many types of sediment.

One of my Great Blue Heron footprints. This heron was walking in a very water saturated, sloppy silt. Some of the details are great (like the webbing in between toes III and IV), but the hallux (I) is a little sketchy.
However, not all large wading birds have a heron or egret-like long hallux: Sandhill Cranes have a much shorter hallux than do herons and cranes. Also, their footprints don't always leave a digit I impression depending on the consistency of the ground they are walking on (Elbroch and Marks, 2001). And there's no doubting that Sandhill Cranes have a well-developed digit I - check out this fossil tarsometatarsus of a Sandhill Crane from the La Brea Tar Pits.
Sandhill Crane distal tarsometatarsus (LACM G4882, Pleistocene) from the La Brea Tar Pits, with a well-developed hallux (see attachment point circled in red), but a hallux that doesn't always show up in the footprint.
This is where assumptions in science can cause trouble. If you assume that all bird footprints have some sort of hallux impression, and if you assume that any tridactyl footprints from the Cretaceous above 10cm in length definitely cannot be bird if there is no hallux, you might be missing out on interesting paleodiversity data for early birds. The footprint evidence for large wading birds in the Early Cretaceous is growing. There is Limiavipes curriei  from western Canada (McCrea et al. 2014), a similar (yet slightly larger) footprint called Wupus agilis from Chongqing, China (Xing et al. 2014), and footprints from Dinosaur Cove, Victoria, Australia (Martin et al. 2014). All of these footprint types indicate that there was likely more than one crane or heron-sized bird wading along the Early Cretaceous shorelines.

Covering the Spread: Details on Digit Divarication

What about the wide splayed digits? Ichnologists take a measurement called digit divarication. This is the angle made by the outer toes, digits II and IV, and it's measured like this:
The arc between digits II and IV is the total divarication. Modified from Xing et al. (2014).
Birds, in general, are supposed to have a total divarication that is larger than 110 degrees, with theropod footprints having a splay that is generally below 90 degrees. This doesn't always work. Here's a small sample of data from modern bird footprints I've collected:
  • Great Blue Heron: 97.7 degrees (88 - 110, N = 10)
  • Canada Goose: 82.6 degrees (73 - 95, N = 15)
  • Spotted Sandpiper: 106 degrees (86 - 133, N = 34)
  • Solitary Sandpiper: 111 degrees (90 - 130, N = 20)
These prints each come from one individual, so this is a sample of the natural variation in toe splay that one trackmaker can exhibit. In general, the larger birds have a slightly lower digit divarication than the smaller birds. Not all bird footprints are going to have a high digit splay.

The same goes for theropod footprints: Magnoavipes was once thought to be a very large bird footprint - we're talking footprints that are over 20 cm long (Lee 1997) - because the splay of the toes was above 90 degrees. Several studies since (Lockley et al. 2001; Matsukawa et al. 2014; Xing et al. 2014) have demonstrated the likely trackmaker was a long-legged theropod, like an ornithomimid (Matsukawa et al. 2014). Just as birds can have narrow footprints, some theropod footprints can have widely spread toes. That's natural variation. As for size, digit splay can't be the only criterion used to say "Aha! Bird footprint!"

The Skinny on Digit Thickness

Remember before when I said that size is a crummy feature to use when identifying a trackmaker? Size and the thickness of the toes of a trackmaker are, in general, closely related. Check out how thick the toes are on this emu (a large bird) footprint).

Wikimedia image of an emu footprint. There's also skin impressions!
Small birds can also have relatively thick digits.

Figure from my thesis comparing digit thickness in Cretaceous bird footprints. A, Tatarornipes (Lockley et al. 2012); B, Koreanaornis dodsoni (Xing et al., 2011); C, Morguornipes (Xing et al. 2011); D, Aquatilavipes swiboldae (Currie, 1981). There is a great deal of variation in the thickness of the toes.
Some of this variation in toe thickness may be due to soft-tissues on the feet of the trackmakers. Some of this variation might be related to the water content in the ground the birds were walking on, and how those footprints preserved. Wet, gloopy sediments will collapse in on themselves after the toes leave, making a footprint with unnaturally skinny toes. Again, using just one feature, like skinny toes, is not good enough to identify a footprint as having been made by a bird. Observing shorebirds in their natural habitats is a great way to become familiar with how one bird can leave a variety of very different looking footprints.

How Do We Identify Large Bird Footprints (Or Small Theropod Footprints) Without Stepping In It?

The only way to make sure we make the most accurate identification we can is by using ALL the information present. This means that we can't just rely on one measurement or observation to make the identification of an avian trackmaker. We also can't ignore all of the other information present in favor of one measurement or observation - that's called cherry-picking, and that's a big data interpretation no-no. Birders have a term, gestalt, that roughly translates to "the whole being greater than the sum of its parts" (Sibley 2002). While this may sound quite unscientific, it's a good reminder to those of us looking at bird footprints that we need to look at all of the variables at play when making our identifications. In a way, gestalt tracking combines both the phenetic and the synapomorphy methods, but also takes into account all of the reasons why those key features might not be there, or might look different than expected. Gestalt tracking can make us a little more comfortable with the ambiguities inherent in ichnology.

We also have to make sure that we make use of the living laboratories that we have outside of our museums and computer labs. This is why I spend some time every summer slogging through muck and mire to watch birds make footprints in their natural habitats - I like to see all the variables that go into the bird taking a step to the final shape of the footprint. This type of fieldwork is crucial to understanding how birds leave the traces that they do, and will give us a better understanding of how fossil birds did what they did.


Buckley, L. G., R. T. McCrea, and M. G. Lockley. 2015 . Birding by foot: a critical look at the synapomorphy- and phenetic-based approaches to trackmaker identification of enigmatic tridactyl Mesozoic traces. Ichnos 22(3-4):192-207

Carrano, M.T., and J. A. Wilson. 2001. Taxon distributions and the tetrapod track record. Paleobiology 27(3):564–582.

Currie, P. J. 1981. Bird footprints from the Gething Formation (Aptian, Lower Cretaceous) of northeastern British Columbia. Journal of Vertebrate Paleontology 1(3–4):257–264.

Elbroch, M., and E. Marks. 2001. Bird tracks and signs: a guide to North American species. Stackpole Books, Mechanicsburg, Pennsylvania, 456 pp.

Lee, Y.-N. 1997. Bird and dinosaur footprints in the Woodbine Formation (Cenomanian), Texas. Cretaceous Research 18:849–864.

Lockley, M. G., J. Li, M. Matsukawa, and R. Li. 2012. A new avian ichnotaxon from the Cretaceous of Nei Mongol, China. Cretaceous Research 34:84–93.

Lockley, M. G., J. L. Wright, and M. Matsukawa 2001. A new look at Magnoavipes and so-called “big bird” tracks from Dinosaur Ridge (Cretaceous, Colorado). Mountain Geologist 38:137–146.

Lockley, M. G., S.-Y. Yang, M. Matsukawa, F. Fleming, and S.-K. Lim. 1992. The track record of Mesozoic birds: evidence and implications. Philosophical Transactions of the Royal Society B 336:113–134.

Martin, A. J., P. Vickers-Rich, T. H. Rich, and M. Hall. 2014. Oldest known avian footprints from Australia: Eumeralla Formation (Albian), Dinosaur Cove, Victoria. Palaeontology 57(1):7-19.

Matsukawa, M., K. Hayashi, K. Korai, C. Peiji, Z. Haichun, and M. G. Lockley. 2014. First report of the ichnogenus Magnoavipes from China: new discovery from Lower Cretaceous inter-mountain basin of Shangzhou, Shaanxi Province, central China. Cretaceous Research 47:131–139.

McCrea, R.T., L. G. Buckley, A. G. Plint, P. J. Currie, J. W. Haggart, C. W. Helm, and S. G. Pemberton. 2014. A review of vertebrate track-bearing formations from the Mesozoic and earliest Cenozoic of western Canada with a description of a new theropod ichnospecies and reassignment of an avian ichnogenus. New Mexico Museum of Natural History and Science Bulletin 62:5–93.

McCrea, R. T., L. G. Buckley, A. G. Plint, M. G. Lockley, N. A. Matthews, T. A. Noble, L. Xing, and J. R. Krawetz. 2015. Vertebrate ichnites from the Boulder Creek Formation (Lower Cretaceous: middle to ?upper Albian) of northeastern British Columbia, with a description of a new avian ichnotaxon, Paxavipes babcockensis, ichnogen. et, isp. nov. Cretaceous Research 55:1–18.

Sibley, D. A. 2008 Sibley's Birding Basics. Knopf Doubleday Publishing Group, 168p.

Xing, L., L. G. Buckley, R. T. McCrea, M. G. Lockley, J. Zhang, L. PiƱuela, H. Klein, and F. Wang. 2015. Reanalysis of Wupus agilis (Early Cretaceous) of Chongqing, China as a large avian trace: differentiating between large bird and small theropod tracks. PLoS ONE 10(5): e0124039. doi:10.1371/journal.pone.0124039

Xing, L.-D., J. D. Harris, C. K. Jia, Z. J. Luo, S. N. Wang, and J. F. An. 2011. Early Cretaceous Bird-dominated and Dinosaur Footprint Assemblages from the Northwestern Margin of the Junggar Basin, Xinjiang, China. Palaeoworld 20:308–321.