Sea otter (Enhydra lutris) mussel pounding

Just north of Monterey, California, there’s an enormous underwater canyon. From a depth of over 2km, this incised channel and its branches rise steadily from the Pacific Ocean bed, narrowing eastward from a broad sedimentary fan into a thin, crooked finger that almost touches the modern shoreline of Monterey Bay. That finger is pointed directly at the small town of Moss Landing, and its surrounding low-lying plain.

The cold offshore water brings nutrients to the Bay, driven by complex changes in the winds (and not, as originally assumed, by the funnelling shape of the canyon itself). A food chain that begins with tiny plankton creates a linked series of seaborne buffets, via herring and krill up to seabirds and even Earth’s largest animal, the blue whale.

And bobbing alongside all those creatures is the sea otter. Inured against cold water by the densest fur of any mammal—a mixed blessing given the human predilection for wearing other animals’ coats—sea otters gather in groups to sleep, feed, clean themselves, and not much else. In a little side pocket of shallow water just outside Moss Landing, known as Bennett Slough, that feeding involves a lot of mussels. And that’s where we find ourselves today.

A cracking mess

It’s also where I found myself in 2015, standing on a small bridge over the Bennett Slough Culverts with sea otter experts Tim Tinker and Jessica Fujii, and ethologist Natalie Uomini. That’s no coincidence - today’s post is about our team’s research at the site, which we published in 2019.

We reported on how the local sea otters use rocks to break open mussels. In fact, they collect and break so many mussel shells that they create piles that could be mistaken for archaeological shell middens made by past people. In this post I’ll describe how we did the work, focusing on the archaeological side of things because that’s my specialty. However, up front I want to emphasise that none of it would have been possible without the long years of dedicated sea otter tracking and recording by Tim and Jess, along with co-authors Katherine Ralls, Sarah Espinosa, and Karl Mayer (of UC Santa Cruz, Monterey Bay Aquarium and the Smithsonian) that gave our work invaluable context. Oh, and I’ll also be sharing various sea otter photos along the way. Because they’re undeniably cute.

Case in point: here’s an adult female otter at the site, munching on mussels with her puffy baby (its fur so full of air that it can’t help but float) safely nearby:

Sea otters are one of the few non-primate mammals known to habitually use stone tools. Not all sea otter groups do so—then again, most chimpanzee groups don’t use stone tools either—but those that do follow a similar pattern.

First, they pick up a rock during a foraging dive, along with some prey that needs smashing. Back at the surface they lay on their backs, with the rock on their chest. Sea otters do just about everything on their backs, so this is a stable position for both them and the rock. The animal then holds the prey, for example a hard-shelled mussel, in both paws, and repeatedly strikes it against the stone.

It’s a very linear action, first raising their arms towards their face, and then quickly slamming the food down onto the anvil. Repeat as needed. They don’t have to be looking at the anvil as they do this, and in fact the sheer effort means that they sometimes have their head fully underwater, stretched back, as they strike.

This is a typical pounding action that I recorded at Bennett Slough. Lifting on the left, striking on the right:

Sea otters can find and use anvils anywhere that there are loose rocks on the seabed (or slough-bed). They are also happy to smash things against rocks that they’re not physically holding, including stones either above or just underneath the waterline. Our team named this behaviour emergent anvil use, to distinguish it from the chest anvils that have been the focus of most sea otter tool use discussion.

The emergent part refers to the fact that the shoreline rocks targeted by the otters are periodically submerged and then exposed by the shifting tides. In order to use these fixed anvils, the otters either raise the front of their body out of the water and strike downwards onto the rock, or (for anvils just at or below the waterline) they can approach them swimming on their side. In both cases the linear, two-pawed striking action is very similar to that seen with chest anvils. Have a look at the photo at the top of this post for a typical emergent anvil pounding position.

Bennett Slough is actually a perfect spot for studying emergent anvil use. The small bridge that crosses the Slough is made of six large corrugated iron pipes that connect the two sides of the crossing, with a roadway over the top. This particular bridge was built in 1991, following the collapse of a previous structure in the October 1989 Loma Prieta earthquake (the same one that caused significant destruction in San Francisco, Oakland, Santa Cruz and other nearby areas).

To orient you spatially, here’s Monterey Bay, with its deep canyon. Bennett Slough is right where the coast splits at the tip of the canyon, midway between Santa Cruz to the north, and the Monterey peninsula to the south:

The Bennett Slough pipes channel the tides into and out of the surrounding wetlands. The nutrient flow and solid substrate makes it an excellent site for mussel growth, which thus makes it a great sea otter picnic spot. Plus, when the culvert was re-built, large boulders were placed between and around the pipes to support the road. Those boulders have a lot of angular surfaces that the otters can exploit for their cracking.

On our first visit to the site, back in 2015, Natalie and I didn’t know about emergent anvil use there; after all, it wasn’t a common topic of scientific study. But we saw a puzzling amount of broken shells around some of the boulders. And when the tide goes out, massive shell deposits can be seen underwater on the north side of the culverts, thickest at the rocky bridge but extending for metres out into the slough:

In the photo on the left, you can see the edges of some emergent rocks (the leftmost one has natural green seaweed on it), and thousands of broken mussel shells just under the water. On the right, you can see the end of a pipe, and get a better idea of the extent of the shell mounds. In the lower centre of the right photo, the muddy bed of the slough is visible where outflow from the pipe has scoured away the shells.

Attack pattern

Any shells or other debris left over from a sea otter meal—bits of crab, etc.—are dropped directly down into the water. The otters have an efficient cleaning mechanism to help with this, where they perform a tight body roll that releases the mess that builds up on their chest. They can even do these rolls while securely holding any remaining food, or an anvil rock, then go back to munching.

Their behaviour means that shells dropped out in the more open waters of the slough aren’t visible, and are subject to the currents. But when the otters use the emergent boulders at the edge of the culverts to crack a mussel, they usually don’t go far to eat. The shell mounds leading away from the boulders are therefore the result not of random currents, but of incessant mussel crunching at the site. In theory, we could date the use of the site by excavating through the shell layers and radiocarbon dating the shells, although at this particular site we expect dates only back to the early 1990s, when the bridge was re-built.

A video would probably help: here’s one that National Geographic put together on our research. You’ll see the otters cracking mussels at the culvert site on rocks both above and just below the water. You’ll also see how the otters use their strong paws to push the orange mussel meat out of the shell—and straight into their mouth—once it’s cracked. And keep an eye out for those cleaning rolls…

Despite how it may look, this project wasn’t just an excuse to spend a few weeks hanging out in the California sun, watching these dextrous mammals do their thing (unfortunately). In order to help us find other sea otter anvil sites in future, even if the otters are no longer there thanks to over-hunting, Natalie and I mapped the patterns of damage we found on the rocks at the site.

We mapped the precise location of every rock over 30cm, all 421 of them, and then recorded the presence and intensity of any use-damage. We wanted to be sure that we could distinguish random rubbing on the rocks (from people walking on them, or vegetation rubbing against them) from the crushing damage done by the otters.

We found that the sea otter damage was quite distinctive. It doesn’t smooth or polish the rocks, instead it leaves brighter sections as a result of the underlying fractured quartz. This example, with a crab for scale, shows one of the more heavily otter-used boulders at the site:

Just like in that photo, we found that points and ridges of the boulders were significantly more damaged than the flat surfaces, and that damage was concentrated on the upper side of the rocks that faced the water. Together, our results showed that otters were responsible for the visible damage at the site, and that the archaeological evidence of their mussel pounding—targeting ridges and raised surfaces with downwards strikes—matched our live observations.

As a summary, this figure from our paper maps the intensity of damage on each rock, with darker shading meaning more damage. It shows that the otters had favourite or preferred stones for their activity, but we’d need a longer study to work out why this is (in the figure, BSC stands for Bennett Slough Culverts):

This result gives us confidence that we can recognise similar patterns elsewhere in the sea otter former range, which extended from Mexico, around the northern Pacific rim to Japan, and correctly attribute it to otter activity. If you live somewhere in that range and have some unexplained boulder damage on your rocky coastline, please get in touch!

A striking bias

It’s a promising start, but there’s quite a lot we still don’t know about sea otter anvil use. For example, how exactly do they select the rocks, either as chest or emergent anvils? When do they just bite open prey, rather than adding the time and effort of finding a pounding surface? We only saw female otters using the emergent anvils at Bennett Slough, but we don’t know if that’s a genuine sex bias in the behaviour, or a result of some unknown factor in the times or places we were observing.

And there’s another mystery we need to clear up. Mussels are bivalves, which means they have two shells or valves held together by a ligament near the joining hinge. In a random sample of broken shells collected from the base of a large anvil at the site, some of the shells were incomplete, fractured halves. Others were complete halves, with a small piece of the adjoining shell still attached.

This pattern shows that the sea otters prefer to pound on one side of a shell, rather than directly on the hinge or edge. But more puzzling is the fact that all the intact shells were ‘left’ valves, and all the fractured ones were right valves. That is, it seems as though the sea otters were all holding the shells in roughly the same position as they struck them against the stone.

Could this be evidence for a laterality or ‘handedness’ bias in the way that the sea otters hold their prey? Or is it just more comfortable to hold the mussels that way? Does it help crack them open, or is it a functionally neutral trait specific to the Bennett Slough sea otter group? Humans have a right handed bias, which affects the way we use some tools (although most striking tools—axes, baseball bats, hammers, etc.—aren’t made in right and left handed versions). If sea otters have a similar strong bias, it would be interesting to know why.

Again, if you’re out and about in sea otter mussel-cracking territory, have a look and see what you find. This figure from our paper shows you what we found, and gives you an idea of what to look for:

Archaeottergy

In my work over the past decade I’ve tried various ways to bring archaeology into questions of animal tool use. Sea otters offer a lot of potential for this work: they use durable stone tools, they repeatedly use the same locations (building up debris), and they have group differences in tool use that could help us track behavioural change over time. All those criteria are the same ones that make ancient hominin archaeology possible. Admittedly, there’s more water involved, but coastlines and sea levels shift, and if there’s been uplift then some sea otter sites could well be inland now.

With this study, otters have now joined their primate cousins—including long-tailed macaques in Thailand, bearded capuchins in Brazil, and western chimpanzees in the Ivory Coast—as targets of archaeological investigation. This is just the beginning of this work, though, and we can expect more species to join the list as we go along. Animals deserve to be recognised in archaeology not just as food, or pets, or threats, but as complex beings in their own right, with a discoverable behavioural record.

But, I admit some of them are also quite photogenic.

Sources: Haslam, M. et al. (2019) Wild sea otter mussel pounding leaves archaeological traces. Scientific Reports 9:4417.

Sea otter image credits: Michael Haslam & Haslam et al. (2019) || Map image credit: Science Magazine; https://science.sciencemag.org/content/358/6361/317.4 ||