Saturday, March 13, 2010

A Scientific Patina to an Unscientific Endeavor

I hunt for fossils along the western shoreline of the Chesapeake Bay, with the Calvert Cliffs towering over me as I walk the narrow beaches at their base.

So, I was pretty eager to read a recent article in the Journal of Vertebrate Paleontology by Christy C. Visaggi and Stephen J. Godfrey (marine biologist and paleontologist, respectively) in which they presented a quantitative analysis of the distribution of Miocene fossil shark teeth by genus and species along the Calvert Cliffs. [Variation in Composition and Abundance of Miocene Shark Teeth From Calvert Cliffs, Maryland, Journal of Vertebrate Paleontology, January 2010]

To be honest, after reading it, my excitement was generated less by the interesting results of their study, and more by a key aspect of the methodology they used for sampling teeth from the cliffs. That provided me with one of those moments when my avocational interest in fossils might actually intersect with scientific analysis, and when I have a chance to make a fool of myself.

Let me set the scene. The Calvert Cliffs, stretching for more than 30 miles down the coast, are composed of sediments deposited periodically during a period of roughly 10 million years, from about 18 million years ago to about 8 mya during the Miocene Epoch. Erosion from the cliffs exposes a multitude of marine fossils. Some of these fossils are found in situ in the cliff face (digging is illegal in the cliffs on public land and legal on private land only with permission), but most are caught in the waves and then captured by collectors as “float” along the shore. The fossils from the Calvert Cliffs reflect the breathtakingly rich marine fauna that characterized this arm of the Atlantic Ocean that, during the Miocene (and earlier), ebbed and flowed here in the so-called Salisbury Embayment.

Three different formations make up the cliffs – the Calvert Formation deposited during the Early and Middle Miocene, the Choptank Formation deposited during the Middle Miocene, and the St. Mary’s Formation deposited in the Late Miocene. Significantly, these formations are tilted, sinking at the rate of about 11 ft/mile as one moves from north to south. As a result, the earliest (oldest) formation, the Calvert, is exposed only at the northern end of the western shoreline because it drops out of sight (and out of the reach of the forces of erosion) as one treks south along the cliffs, until only the St. Mary’s Formation is exposed at the southern end of the cliffs. For reasons that will be clear, my collecting is focused on the northern end, that is, on fossils eroding out of the Calvert Formation. The pictures above are of the Calvert Formation, though it is still beyond me to identify the specific layers shown.

Scientists Clench Their Teeth

Visaggi and Godfrey examined teeth secured in two independent ways – 1,866 teeth currently residing in museum collections gathered originally in situ from the various layers of the cliffs by collectors (presumably, professional scientists and others), and 24,409 teeth gathered along the shore as float material.

It’s this latter source of teeth where the story gets particularly interesting for me. These teeth were collected by Calvert Cliffs residents, volunteers with the Calvert Marine Museum (Solomons, Maryland), and members of the CMM fossil club, and were originally donated to the CMM’s Discovery Room where children search through sand to discover fossil shark teeth. During a three-year period, 40 donations with good provenance data were diverted to the analysis of shark teeth distribution along the cliffs.

The authors divided the aggregate float sample of over 24,000 teeth into ten unequal groups based on where along the cliff shoreline they were collected. For example, the group from the beach at the northernmost end was designated CC10 and contained 1,716 teeth, while the southernmost group was CC1 with 614 teeth. Each one of the ten groups was analyzed separately, as was the total float sample.

Each of these sources of teeth – the in situ collection and the float collection – comes with decided biases. Among the various biases discussed by the authors were these two:

Payoff bias
– Collectors of float material have a natural tendency to head north to hunt for teeth because that’s where, by reputation, the payoff is the biggest. In their analysis, the authors emphasized the proportional distribution of teeth for different sections of the cliffs in an effort to mitigate against that potential bias. They suggested that this fossil wealth in the north reflected a deep, open marine paleoenvironment highly supportive of sharks that prevailed when the Calvert Formation was being laid down.

Size and Prize bias – Most collectors are likely to be biased against small teeth through a lack of interest in them or an inability to spot them. Teeth that are large, rare, or unusual are likely to be the focus of collectors’ energy. Interestingly, the authors conclude these sought-after teeth were likely to be overrepresented among the in situ teeth (adding these to a museum’s collection would be a priority), but underrepresented among the float teeth (the typical collector of these teeth would keep the best finds). The authors suggested that for the large, rare, or unusual teeth, the actual proportional representation would fall between those of the in situ and the float samples.

Scientific Guise

I appreciate the cleverness in amassing over 24,000 teeth from donations to the museum to create an impressive sample of teeth from the cliffs. Despite the clear biases and limitations of relying on float material, the sheer number of teeth suggests some confidence in what can be gleaned from this sample. Still, there might some interesting idiosyncrasies of float collectors that influence what is gathered.

[Later edit: I dropped a paragraph in which I speculated about other biases of float collectors and the number of individuals who might have been involved. It was misdirected because I failed to realize that 40 donations might have involved lots of people contributing to each donation, not a maximum of 40 contributing the whole sample (which raised the question of just a few collectors influencing the whole).]

So, what’s a rank amateur to do? Explore the idiosyncrasies, of course. I decided to devote my next trip to the Calvert Cliffs to answering this question:

How does the distribution of genus and species among the teeth I might collect on a single day compare to the distribution of the float sample reported by Visaggi and Godfrey?

A foolish venture, I recognize. One that was patently unscientific – a single sample taken by one collector on one day for, it turned out, three and a half hours (my back could only take so much stooping and bending on the beach) certainly doesn’t carry any weight. Any divergence from the paper’s results could be easily explained away. But, my motives were pure, I think, and it gave the trip some structure and I could delude myself that it added a very thin scientific patina to the effort. Plus, it was a damned sight more fun than going back through the teeth I had already collected from the Calvert Formation to see what their distribution looked like.

Let’s start with a few of my credentials to engage in this folly. I am a collector of float material and I reflect the payoff bias in spades since my collecting from the Calvert Cliffs is concentrated in an area at the northern end where the float samples CC10 (1,716 teeth) and, possibly, CC9 (7,523 teeth) came from.

It’s interesting to consider the possible influence of the size and prize bias on my collecting this day. I clearly have this bias – it’s hard not to. But, since I wasn’t donating my finds, it wouldn’t matter for the results. I wouldn’t be skimming off the “good stuff.” As for the tiny stuff, I decided to gather up every tooth that I spotted, regardless of size. This resolution presumably didn’t reflect one held by the collectors of the float material in the Visaggi and Godfrey study. Also, to be honest, I couldn’t follow through on this for the entire three and a half hours I was collecting. For roughly the last hour, I just couldn’t bear to stoop for yet another speck washing up around my boots.

What the Scientists Found

First, I think it’s appropriate to review the general results that Visaggi and Godfrey present in their paper because they are broadly relevant to my little exercise.

1) Overall, along the Calvert Cliffs, teeth come predominantly from gray or requiem sharks (Carcharhinus spp.), sand tiger sharks (Carcharias spp.), tiger sharks (Galeocerdo spp.), snaggletooth sharks (Hemipristis serra), and mako sharks (Isurus spp.). (“spp.” is the plural abbreviation for species – in other words, there are multiple species within the specific genus.)

2) The vast majority, about 90%, of all teeth, in situ or float, are found in the generally northern end of this expanse of shoreline, the section associated with the Calvert Formation.

3) The further south one goes, the relative distribution of teeth from these five genera of sharks changes – grays and sand tigers diminish in their representation among the teeth being found, while teeth from snaggletooth sharks and makos increase proportionately.

Relevant Study Samples Meet Mine

More relevant than their overall findings for my project were the results from the float samples Visaggi and Godfrey analyzed that came from the stretch of beach where I collect. This is principally the sample they identified as CC10, though it’s unclear precisely where the boundary between CC9 and CC10 falls. I suspect my wanderings took me from CC10 across that boundary, hence I not only compared my sample to the CC10 results, but also to a second aggregate sample I created from data in the article for CC9 added to that for CC10.

My sample consisted of the 157 teeth I collected on one cold, windy day in March, from about 2 hours before low tide to about 1 ½ hours after. Lots of wave action, and steady erosion from the cliffs. Little feeling in my fingers. Altogether, a typical winter day at the bay.


The table and graphic below present the relative distribution among the genera accounting for the largest proportion of teeth found. Any genus accounting for at least 1% of the total of any specific sample – mine, CC10, or CC9+10 – is included. These eight genera include the five listed above and the following: cow sharks (Notorynchus primigenius), lemon sharks (Negaprion eurybathrodon), and hammerhead sharks (Sphyrna laevissima). These eight account for between 87% (my teeth) and 92% (CC10) of the teeth in these samples.

Genus/SpeciesMy TeethCC10 TeethCC9+10 Teeth

Notorynchus primigenius0%4.6%1.2%
Carcharias spp.7.6%10.1%8.8%
Isurus spp.3.2%4.5%2.5%
Hemipristis serra7.0%8.0%9.4%
Carcharhinus spp.45.2%56.9%50.7%
Galeocerdo spp.17.2%7.0%17.6%
Negaprion eurybathrodon3.8%0.1%0.1%
Sphyrna laevissima2.6%0.6%0.9%

Some nice examples of teeth in these predominant categories are shown below. This first one contains some teeth from grays.

This next picture shows some tiger shark teeth. These are all from the contortus species. (In this little exercise, I’ve followed Visaggi and Godfrey in grouping these specific teeth under the Galeocerdo genus.)

A few sand tiger shark teeth.

Finally, a trio of snaggletooth teeth, all three of these are uppers and would be pointing down in the shark’s mouth – two have serious root damage and the third is quite pretty.


So, what was learned? Well, though I doubt this will offer much comfort to Visaggi and Godfrey, my sample actually closely tracked the distribution of these genera in both CC10 and CC9+10. Gray shark teeth (Carcharhinus spp.) dominated in all three samples. The better match overall with my sample was with the combined float CC9+10.

Anomalies? Sure. I found no cow shark teeth (Notorynchus primigenius) on this day, and my handful of lemon shark teeth (Negaprion eurybathrodon) and hammerhead shark teeth (Sphyrna laevissima) represented a much larger percentage of my sample than in either of the others.

Knowing that I was collecting for distributional analysis made a difference. As noted, I set out to act as a vacuum for much of the day, as a result I chased down all stray teeth (at least for awhile), painfully bending frozen fingers to pick yet another tooth from the waves, from under a downed tree, or from a chink in a clayey block of fallen cliff face. This may account for my luck with lemon and hammerhead shark teeth which are smallish teeth and, on another day, ones I might otherwise have skipped or missed. That doesn’t explain why my sample was missing other little teeth that did show up in the CC10 and CC9 samples, though still at marginal rates.

While I was on the beach, I was continuously comparing this day to the myriad others I’d spent gathering teeth in this same area. For much of the time, I bemoaned how bereft of the really good stuff – read the “large, rare, and unusual” – the beach was. They had taken an inopportune leave of absence. Then, I had one of those moments when I felt I’d stepped into some sort of parallel fantasy universe, actually, the fantasy universe of fossil hunters. In this universe, the fossil gods deigned to reward me. Almost everywhere I looked, there were wonderful teeth, one stuck in the sand, black glistening crown pointing to the cliff top, another two tumbling together in the surf, another one nestling next to the prize I had just grabbed with a sharp intake of my breath. It didn’t last long, but it was grand while it did, and all the while, I was sure it made my sample on this day totally irrelevant. But, not so. The good stuff either fit into expected categories or was outweighed by the other teeth gathered earlier. Besides, who was I kidding, my sample, despite any scientific gloss I tried to give it, was probably irrelevant from the outset


Moral of the story – one day’s collection does not a valid sample make, though it’s not bad.


  1. Fantastic post! I love the citizen science approach. . .

  2. Glad you liked it. I had fun putting it together.

  3. I just found this and really enjoyed your thoughts. I am a science teacher and reformed paleontologist, and have visited the cabins at Calvert Cliffs several times- and have never found a shark tooth. So, I was amazed that you could find 150 in a single day. Can you explain your process? I am taking some kids there in November and I would love to help them bring back a few. Thanks.

    1. Ken:

      Thanks for the comment. I have to admit it was a particularly good day (even though chosen at random). I went out on the one year anniversary of this hunt to do it again and found very few teeth. Fossil gods give and take away.

      You mention the "cabins at Calvert Cliffs" -- I assume you're speaking of the cottages at Matoaka. I don't know their status since one of the owners died in the past couple of years. Also, the one time I collected from there, I had very little luck. For public access points, you're pretty much limited to the Calvert Cliffs State Park (long hike to an unproductive beach), Flag Ponds Nature Park (short hike to an unproductive beach), and Brownies Beach, just below the town of Chesapeake Beach at the northern end of the Calvert Cliffs. You are likely to find something at any of these beaches, but Brownies Beach, or Bayfront as it is officially named, is probably your best bet to find teeth.

      Here's the address to the Maryland Geological Survey's discussion of several of these sites:

      If you're going with younger children, I'd recommend Brownies and just inspecting the wet sand at the edge of the beach there. Keep an eye on what the small waves might churn up. Little teeth will show up around you if you pay attention.


    2. Tony-

      Thanks. I may try Brownies.

      Is it worth trying to build/bring some sort of sieve?

      Are there also inverts to find at Brownies? Last time I was at Matoaka, a big chunk had come down that was full of Isognom and lots of other great stuff.

      Thanks again,


    3. A sieve can be useful. Unless you're getting serious, a kitchen colander with a small mesh works fine. Any of the sites mentioned by the Maryland Geological Survey will produce some shells.


Nature Blog Network