Sea otters are perhaps my favorite marine mammal, for several reasons. They’re adorable, endangered, ecologically fascinating, and simultaneously the smallest marine mammal and the largest member of the weasel family (Mustelidae). When I was a kid, I would see the sea otters at the Monterey Bay Aquarium and out on the rocky Northern California coast, especially in Monterey Bay – to the point where I actually got a bit bored of them. Once I got into marine mammal research as a twentysomething, however, I realized that 1) they’re only in the North Pacific, 2) endangered, and 3) sort of a success story. I realized that if I traveled to anywhere in the world other than central California, Kamchatka, Alaska, or British Columbia, I would not be seeing any sea otters. This post is going to go through some older ideas of sea otter evolution, what we know, and what we don’t know, and summarize some of the major findings of a paper I published back in 2017.
A few years back, I slowed down with blog writing. But I have realized one thing, especially with this paper: often just writing a scientific article is not enough. In the four years since my otter paper first appeared online, I’ve read multiple articles by ecologists, molecular biologists, and review articles that have either ignored it or misinterpreted the findings (e.g. Pacific invasion of Enhydra 1-3 million years off of my estimation). I think that my blog historically gets more readers than my actual papers, and many scientists outside my narrow discipline have stumbled across these posts and used them as a bridge to our paleontological literature. So, it is in the hope of clarifying what we do and do not know about the evolution of the sea otter that I am writing this more palatable version of one of my papers (with a much larger background review).
A large raft of northern sea otters; photo by Milo Burcham.
A brief and totally non-comprehensive review of relevant see otter biology
All otters are members of the subfamily Lutrinae (=otters) and are the only major aquatic clade of the weasel family (Mustelidae). Mustelids used to include badgers and skunks, now split out into their own families - but weasels, minks, fishers, martens, and the wolverine are still in the Mustelidae. Mustelids are members of the dog-like carnivores, the Caniformia; canids (dogs) are the first diverging branch, followed by bears (Ursidae), raccoons (Procyonidae), pinnipeds (Pinnipedia), and the red panda (Ailuridae), and eventually the mustelid-skunk-badger clade (pinnipeds occasionally are recognized as sister to bears, or at least traditionally were, and their origins now are somewhat murkier - though clearly not derived from mustelids). That means that there are at least five completely disparate lineages of marine carnivores that independently returned to the ocean: the pinnipeds, the polar bear, the sea otter, and the extinct sea mink, and the south american marine otter (same genus as the North American river otter, Lontra). Historically, there have been two ecomorphs of otters, not necessarily forming monophyletic groupings: the "fish eating" otters, and the "crab eating" otters. The fish eating otters tend to have longer snouts/toothrows and the crab eating otters tend to have shorter snouts with larger posterior crushing teeth (=carnassials of terrestrial carnivores) positioned closer to the jaw joint for greater mechanical advantage (Timm-Davis et al., 2015).
There are three named subspecies of modern sea otters, which seem to be anatomically differentiated from one another Enhydra lutris is quite large for a mustelid (20-40 kg) but still quite a bit smaller than fur seals (60-270 kg in males, depending upon the species). They tend not to stray into deep water and while they are typically considered to be associated with kelp forests, they are not bound to the presence of kelp forests; instead, they seem to be tied to rocky shore environments. All Enhydra have large, blunt, clam-cracking teeth, whereas all other mustelids (even river otters) have quite sharp cusps on their teeth; they have sagittal crests and large temporalis muscles, and thus have a powerful bite. Nevertheless, not all sea otter individuals put this bite force to equal use: some specialize on softer prey items (fish, cephalopods, sea cucumbers), and others specialize on mollusks, crustaceans, and echinoderms. In other words, the seemingly specialized feeding adaptations of Enhydra are not manifested as ecological specialization – in fact, the opposite: it allows individual otters to be specialized on different foods in a diverse rocky shore habitat – a sort of interspecific niche partitioning. This likely allows otters to inhabit geographically restricted rocky shore habitats in a higher population density. Otters starve if the shoreline freezes, which explains why sea otters live no further north than the southern limit of seasonal sea ice.
Skull of a sea otter, illustrated by N.N. Kondakov.
A famous college statistics problem involves the asymmetry of the sagittal crest in sea otters, as otters express handedness in chewing, and with regards to their forepaws, Kenyon (1969) concluded that sea otters were typically right-handed. Roest (1993) is likely the origin of the statistics exercise, and investigates the frequency of various aspects of asymmetry in the skulls of Enhydra. Growth studies show that full biting ability is delayed during growth, and that males develop structures associated with powerful bite force more quickly than females, likely to gain a competitive advantage against other males (Law et al., 2017). Sea otters bite at an unusually wide gape angle in order to use their posteriormost enlarged teeth - upper p4 and the lower m1 - to crush hard prey with the greatest mechanical advantage (just like a lever-style nutcracker). They have unusually wide zygomatic arches to provide a longer distance for contraction of the masseter muscles (Timm-Davis et al., 2017) along with unusually high and long sagittal crests and nuchal crests. The bite of Enhydra lutris is quite a bit different from the only other major marine durophagous carnivore, the extinct "oyster bear" Kolponomos. It hasn't really been the focus of any particular study, but I've observed that the coronoid process (which the temporalis muscle attaches to - the major jaw closing muscle) is comically big and dorsally very elongate with respect to many terrestrial carnivores - including other otters. Enhydra has a loose mandibular symphysis (jaw joint at the chin), and so there is a bit of flexibility between the mandibles; Kolponomos instead has a fused mandibular symphysis and high mandibular stiffness - so its bite is more like a sabertooth cat than a sea otter, despite sharing the enormous rounded molars. Sea otters instead have low mandibular stiffness, allowing for a crushing bite on one side, a bit more like a hyena. Kolponomos probably did not have tactile paws like a sea otter, and therefore relied on its canines and front of the jaws to pry mollusks off of rocks, prior to crushing (Tseng et al., 2016).
Sea otters are famous for their tool use – often keeping a favored anvil stone in a famous armpit ‘pouch’ – rather, there’s a bit of loose skin that keeps the hammer stone from falling out if the otter holds it in place with its forelimb. Sea otters are often credited with carrying around ‘favorite’ anvil stones, and though I’ve seen this claim repeated endlessly across the web, I’ve never seen it credited to an actual biological observation in nature. It’s not mentioned by Kenyon (1969) nor is it mentioned in any more recent papers on tool use. If anyone knows of the origin of this claim, please share it! The one documented case of repeated use of the same rock I know of are stationary boulders in Elkhorn Slough where otters smash bivalves against, leaving little deposits of mussel shells with consistent fractures (a fascinating “archaeological” study of otter anvil stones by Haslam et al. 2019). Otters apparently do not use hammer stones; they use rocks, or occasionally bivalve shells, on their stomach and slam the bivalve downwards into the rock. I’ve witnessed it many times but always while fossil collecting on the cliffs in Santa Cruz, so I’ve been within earshot but never close enough to see what was being slammed into what.
Sea otters with an anvil stone and a stationary anvil boulder, from Haslam et al. 2019.
The tiny, mitten-like forepaw of the sea otter, from Strobel et al. 2018.
We know thanks to some careful genetic work (Ralls et al., 2017) that 1) not all otters use tools, but those that do are not closely related to one another; and 2) tool use is taught by mothers to their pups. This distributed tool use behavior suggests that the common ancestor of all sea otters, which likely lived during the last glacial maximum, was capable of tool use. Possibly correlated with tool use are the hilariously tiny forelimbs and mitten-like paws which frequently lack individual divisions between the fingers. For some reason, sea otters ended up reducing rather than lengthening their fingers, likely for heat retention (see below). Unfortunately, there is no obvious expression in brain endocast topography that distinguishes tool using Enhydra from the clawless otters, which use their paws extensively for prey manipulation.
Skeleton of a sea otter showing the disproportionate fore and hind limbs, from twitter user @seaottersavvy
A sea otter drying out its hindfeet while rafting. Photo by Shetzers Photography.
The hindlimbs of sea otters are quite long, and the hindfeet are developed into rudimentary flippers much like pinnipeds, with the outermost digits being very elongated, and extensive webbing along the entire length of the digit. Pinniped hindflippers have elongated outer (5) and innermost (1) digits with digit 3 being the shortest, whereas in Enhydra, digit 5 is longest and digit 1 is shortest; the hindflipper of Enhydra looks startlingly like the foreflipper of a phocid seal. All other otters instead have five toes with incomplete webbing that look nearly identical to the forepaw. Sea otters swim by dorsoventral undulation of these hindflippers as well as the elongate, muscular tail – but unlike other lutrines, most of the propulsive force comes from the feet. Opposite from the hindlimbs, the forepaws are highly reduced relative to the ancestral condition, and are not used in locomotion; the result are proportions not unlike many bipedal dinosaurs. Enhydra is quadrupedal, and the extreme disparity between forelimb and hindlimb length makes for an extremely strange and unsettling posture while walking around on land. Their reduced agility on land is probably a major contributing factor for haulout behavior by the southern sea otter, which appears to haul out less often than the northern sea otter in Alaska. This may also be a result of the extreme population bottleneck, and perhaps there is a greater fear of humans in the southern population.
Sea otters running on land; photo credit: seaottersavvy.org
Sea otters have no blubber, unlike other marine carnivores like pinnipeds. Pinnipeds have either blubber and no fur (sea lions, earless seals, walrus) or a combination (fur seals). Fur seals have some of the thickest fur in the animal kingdom – but fur is inadequate for diving at depth. Now, I just read a Scientific American article with a Q&A about fur and the first answer is “there’s no difference between fur and hair”, which from the perspective of homology may be true (both are cylindrical keratin growths from skin) but in terms of morphology and function is absolutely bullshit. Mammalian carnivores have basically two types of insulation: guard hairs, which are the longer, wider, and stiffer hairs you see externally on your dog or cat, and underfur, which is the really fine stuff that is often kind of wiggly. As I’m typing, I see a couple strands of my cat’s underfur right now. This is the stuff that you need to brush your dog or cat to remove. Hair does nothing for insulation underwater, but fur traps air bubbles. To help with this, it has a different scaly pattern under microscopy. This is only effective at shallow depths because once you go below the water pressure pushes the air bubbles out (Liwanag et al., 2012). So fur only works for shallow-diving, early marine invaders of aquatic lineages – once marine carnivores start to dive below 100 meters, they need to evolve blubber. Hair was lost independently within sea lions and earless seals, and more than likely a third time within walruses (and, if we ever had a way to test it, I’d wager it was lost a fourth time in the desmatophocids, based on body mass). Sea otters are shallow divers with a very tiny body mass – the smallest marine mammal – and rely completely on fur for insulation in the absence of blubber. The earliest pinnipeds, scarcely larger than Enhydra, very likely primitively had only fur and no blubber.
Map of modern sea otter populations, with original caption, from Kenyon (1969).
Sea otters nowadays have three very geographically disparate (and somewhat anatomically disparate; Wilson et al., 1991) subspecies populations: the western sea otter (Enhydra lutris lutris), today living in the Komandorskiye/Commander Islands and Kuril Islands south of Kamchatka, the northern sea otter (Enhydra lutris kenyoni) living throughout the Aleutians and east along the southern Alaska coast to Prince William Sound, and the southern sea otter (Enhydra lutris nereis), living only in central California from Santa Cruz County down to Ventura/LA counties and the channel islands. Historically, they lived continuously from Hokkaido and Sakhalin in the western North Pacific, eastward throughout the Aleutians and Alaskan panhandle as well as the Pribilof Islands, British Columbia, Washington, Oregon, California, and down to Baja California, at least as far south as Morro Hermoso on the Pacific side. The Spanish wrote that otters used to be seen frequently inside the San Francisco Bay as there used to be more kelp beds – now buried by the extensive silt deposits that filled in the bay and expanded the wetlands that make the area so famous for its natural history. These wetlands and buried rocky outcroppings mean that there is no longer very much kelp in the bay, and so there is little reason for otters to venture within. They also rarely venture further north than Point Ano Nuevo – though there are stray individuals occasionally as far north as Eureka; I saw one in 2017 at Pescadero State Beach (southern San Mateo county) and I lost my shit I was so excited.
A well-known graph showing the number of sea otter pelts (y axis in thousands - e.g. 1,000 to 8,500) sold in London (and only London) from the late 19th century to the early 20th century. Source: Wikimedia commons.
Though hunted for thousands of years by indigenous peoples along the Pacific shores, sea otters were by westerners in the 1730s, leading to the first hunting expeditions by Russian fur traders in the Kuril Islands. Russian hunting extended ever further eastward, and the trade in sea otters decimated the populations in the Pacific Northwest so badly that kelp forests have vanished, a ghost of their former distribution – and severely limiting the carrying capacity for otter populations. Fort Ross, built in 1812 about a three hours’ drive north of San Francisco in Sonoma County, was built by the Russians to support otter hunting in northern California. Fort Ross marks the furthest outpost of Russian Imperialist expansion during the 19th century – a high water mark – and it was abandoned in the 1840s. By this time the sea otter population in California had also been decimated. Southern sea otters were eventually considered extinct in the late 19th century, and the species as a whole were commercially extinct throughout their range by the 1850s. In the early 20th century, there were a handful of sightings kept secret by scientists, and thanks to the WPA construction of Highway 1 in the late 30s, otters were formally rediscovered to be very much alive in 1938. However, there was an estimated 50 individuals in this population (alternatively, I’ve read a raft of 50 individuals spotted by a couple testing a new telescope in 1938, presumably part of a larger but unquantified population). They have since rebounded to around 3,000 – though this is only a fraction of their pre-industrial population estimate of about 16,000.
Molecular phylogeny derived cladogram of lutrine relationships, from Koepfli et al. (2008).
Molecular phylogenetic work has focused on the genetic bottleneck of southern sea otters as well as the placement of Enhydra within the Mustelidae – it is typically recovered as the sister taxon to the old world clawless otters (Amblonyx + Aonyx). Molecular clock dating from a paper in the 90s suggests a rather broad (and predictable) range of dates for the common ancestor of all three modern subspecies, sometime about 20,000-200,000 years ago (Cronin et al., 1996). Clearly, more work is needed to refine this molecular divergence date. Morphometric studies (Wilson et al., 1991) seem to support the recognition of three subspecies (regardless of how one feels of the use of subspecies as taxonomic entities).
Next up: the fossil record of sea otters, and my desperate attempts to go out and find more.
Beichman et al. 2019. https://academic.oup.com/mbe/article-abstract/36/12/2631/5481008?redirectedFrom=fulltext
Haslam et al. 2019. https://www.nature.com/articles/s41598-019-39902-y
Kenyon, 1969. https://www.fwspubs.org/doi/abs/10.3996/nafa.68.0001
Larson et al., 2012. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0032205
Law et al. 2017. https://academic.oup.com/biolinnean/article/121/2/420/3045262
Liwanag et al. 2012. https://academic.oup.com/biolinnean/article/106/4/926/2452408
Liwanag et al. 2012. https://academic.oup.com/biolinnean/article/106/4/926/2452408
Ralls et al. 2017. https://royalsocietypublishing.org/doi/full/10.1098/rsbl.2016.0880
Sharpe, 1939. The discovery of the “extinct” sea otter. https://seaotters.org/pdfs/extinct.pdf
Strobel et al. 2018. https://jeb.biologists.org/content/221/18/jeb181347
Timm-Davis et al. 2015. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0143236
Timm-Davis et al. 2017. https://jeb.biologists.org/content/220/24/4703
Tseng et al. 2016. https://royalsocietypublishing.org/doi/full/10.1098/rspb.2016.0044
Wilson et al. 1991. https://digitalcommons.unl.edu/cgi/viewcontent.cgi?article=1090&context=usdeptcommercepub