Learning to Identify Fossil Species

I think the most intimidating thing that happened to me when I started my Ph.D. work was being presented with drawers of fossil teeth and being instructed to identify them to species.

How do you even begin?

I still struggle with this, twenty years on. But now I have tools to get past the initial steep, seemingly insurmountable, learning curve.

Today, I have a bunch of fish scales. I don’t know what fish they come from, but I’d like to. I don’t know if you can even identify species from fish scales. But what I’ve got is a bunch of fish scales and I need to know more than just ‘fish.’

Let’s see. My main research focuses on mammals. Mammals don’t have scales. So it looks like I’m starting from scratch.

To make matters more challenging, when I started, I was in the field, separated by days from reference materials including the internet. But that’s OK. I can do without.

Step 1 – Group the fossils

Take a subset and come up with groups of similar-looking fossils. You might not know names and terms yet, but you can tell what’s similar and what’s not.

For my fish scales, I quickly came up with three groups: Gar, Toothed, and Pumpkin Seed.

I’d seen gar fish scales before. They’re little, black, and diamond-shaped. The black is because of the fossilization process as much as anything. There were lots of gar.

The pumpkin seed ones look a lot like seeds, maybe more like sunflower seeds, but they have a little keel at one end.

Toothed scales have a lot of surface texture that seems to focus into a toothed edge.

Step 2 – Hit the books.

Now that I had them grouped, I could think about what features might be important for distinguishing different species. Now it was time to develop the vocabulary.

First, I learned there are four different types of fish scale:

Placoid – found in sharks

Ganoid – found in gars, sturgeons, and bowfins. Often have peg-in-socket joints that connect each scale to the one in front of it.

Cycloid – a subdivision of leptoid scales with a smooth edge found in salmon and carp

Ctenoid – a subdivision of leptoid scales with a toothed edge, found in most other ray-finned fishes.

What did I learn from this? Well the obvious gar scales and the pumpkin seed scales are all ganoid scales. The ‘keel’ of the pumpkin seed is just the bottom of the scale, whereas the scales I already identified as gar show only the top surface.

What about the toothed scales? It seems like they may be ctenoid scales, but I need to do more research.

Step 3 – Find out what animals might have been in the area at that time.

More research, now into the technical literature. Usually, I already know something about a site before I start trying to do identifications. Most often, an approximate age is already known. There are also frequently localities of similar age that have been described elsewhere. I read about those fossils. I look at pictures.

The first paper I looked at had a photo of a scale that looked essentially identical to the ‘toothed’ scales. I dug in. This scale belongs to a member of the gar family.

So it appears that what I have are gar, gar, and gar. It’s possible I have more than one species of gar (or gar relative) but it seems that everything I have is gar.

I suppose also it could be gar, sturgeon, and bowfin. So, time for more research.

Step 4 – Focus the research down.

I have learned that bowfins were at least around during the time that the gars were living in the same area. We have some skeletal evidence for that. Sadly, none of the scales we have really look much like bowfin scales, so it’s probable that they’re still gar.

I’m not entirely satisfied with this result, so I’ll continue doing research. But you can see the basic process.

You start with what you can figure out without a lot of prior knowledge, and then jump into the technical literature.

For my dissertation research, when I had 2000 isolated mammal teeth to start with, I first separated them into piles of ‘complete’ versus ‘broken.’ Then I counted cusps and separated them into molars and not-molars, and also upper and lower. Then, with some guidance from the literature, I was able to divide the teeth into families, like plesiadapidae versus condylarthra.

It’s an iterative process. For my dissertation, I narrowed the 2000 teeth to 800 that I could identify, and I probably looked at each of those 800 teeth a dozen times before I was satisfied with the identifications. It took me two years. That’s about right.

For the fish scales, it took me about a day. It’s not as complex of a system as mammal teeth, and I only had three morphologies (and 30 specimens) to work with.

But the process was essentially the same: Examine what you’ve got. Learn what might have been there. Compare. And keep going until you’re convinced of your identifications.

No doubt other paleontologists have different approaches. How do you begin such a daunting task?