Unravelling the History of Seals

Allodesmus, a desmatophocid

Trying to determine the largest patterns in evolution can be a daunting task. Here, we often want to look at large numbers of species, comparing the living ones and filling in the gaps with fossils that are often incomplete, ambiguous, or that simply haven't been discovered yet. As a result, there are several big transitions in mammalian evolutionary history that we'd like to get a better look at. Bats are a significant case in point; their small fragile skeletons don't preserve well if we want to see more than teeth, and how they developed their forelimbs into wings remains obscure.

With some groups, however, we do have sufficient fossil evidence that we can look at a whole group of animals and get some idea, not just of how it originated, or where it fits in the larger mammalian family tree, but what ups and downs it has faced over the course of its existence. This can tell us what alterations in climate or geography drove changes within the group and how and when particularly evolutionary innovations developed.

The pinnipeds are one group where we do have enough fossils to build up a picture on this scale. These are aquatic mammals with flippers instead of feet that still come ashore in order to breed, and they have long been divided into three families. The largest, in terms of the number of species, is the seal family which, in turn, has three subfamilies: one with the harbour seal and its relatives, another including the elephant, monk, and Antarctic seals, and a third known only from the Middle Miocene about 15 million years ago. (For more information on this family, I did a review of all living species a few years back). The other main family is best known for the sea lions but also includes the "fur seals", which are anatomically much the same.

Studies taking a systematic look at the broad sweep of pinniped evolutionary history date to at least the 1950s. Then, the conclusion was that pinnipeds are a natural group of animals, first appearing in the Arctic Ocean and subsequently splitting into the two families when a land bridge over the Bering Straits cut off those in the Pacific from those elsewhere. This is the opposite of what we normally hear of land bridges doing - allowing animals from different continents to mingle and interact - but when we're talking about sea creatures, they're a barrier, not a bridge. 

By the 1970s, a new theory had come along, suggesting that the similarities between the families were the result of parallel evolution, not a common ancestor. Under this view of things, seals evolved in the North Atlantic from relatives of the mustelids and sea lions in the North Pacific from bear-like animals. This, however, was comprehensively scuppered by genetic and molecular studies in the 21st century; it's not parallel evolution. We now know that the two families diverged from one another roughly around the dawn of the Miocene 23 million years ago long, long after, their common ancestor had diverged from that of their closest living relatives, the weasel/skunk/raccoon group.

In the last decade or so new fossil discoveries, many from previously underexplored fossil deposits in the South Pacific, have added to our knowledge. A recent study took a detailed look at the current state of play, including not just fossils from the living families but numerous fossils that are placed further out on the relevant branch of the larger mammalian family tree.

The first thing to do was, naturally, to clarify how different pinnipeds related to one another, using  known genetic data to provide a backbone to the tree, and filling in the rest as best we can based on the physical features of the fossils. In total, they looked at all 34 living species of pinniped, the two species that went extinct during the 20th century, 76 named fossil species, and a further 17 fossils that appear to belong to distinct species, but have yet to be formally named.

The resulting tree looks, on a large scale, not too different from the less complete versions we already had. I can't put the full 129-point tree here, and it would require too much explanation if I did, but I can include a greatly simplified overview. Here, the seal family is shown in blue, and the broader sea lion clade in green. I've had to cut out many of the minor branches to make it fit here, but even so, it's obvious that two branches descend from the last common ancestor of all living species without fitting in either of those two main families.

One of those is, unsurprisingly, the walrus family. This has only one living species today, although there were many more in the past - this particular study looked at no fewer than 23. There is no surprise here about its placement; it has been known for a long time that walruses are more closely related to sea lions than to true seals, and they were placed in a superfamily together as early as 1821. (Indeed, the 1950s study I mentioned above actually considered the two together). Molecular evidence suggests that they split from the sea lions about 20 million years ago in the Early Miocene, long after their common ancestor had split from the true seals. 

The other branch consists of a fourth family of pinniped, distinct from seals, sea lions, and walruses, and that is now extinct. These are the desmatophocids, the first large-bodied pinnipeds, living in the North Pacific during the first half of the Miocene. They are well-known from fossils, first having been described in 1906. Their position in the family tree has been a matter of debate for decades. A common view today is that they were related to sea lions and walruses, but you'll note that this latest one disputes that, placing them closer to the true seals.

However, the real purpose of this is finding new detail in the large body of data. One thing we can do, for example, is look at the rate at which new species appear and then go extinct. This shows that the number of known species of pinniped has increased over time, even after the initial burst in which they first appeared, especially among the sea lions. It's possible, of course, that is a reflection of the fossils that we have - fewer from the more distant past, for example - but it's interesting to note that it seems relatively steady, with no major setbacks where pinnipeds temporarily declined.

The exception is among the walruses, where new species pop up at an unusually high rate starting around 12 million years ago, but all them die out almost immediately to be replaced by some new variant. This could have initially been due to them replacing the then-vanishing desmatophocids, and may have continued for so long because the sea level was dropping at the time, making it more likely that some communities ended up being cut off from others. 

It may be significant that an unusually large number of new cetacean species turn up at around the same time. On the other hand, those didn't tend to immediately go extinct again, which is why we have so many dolphin species today. Around 5 million years ago, however, while new walrus species continued to pop up rapidly, the extinction rate became even higher, so that the number of walrus species at any given time started declining - leaving us, eventually, with just the one. This, the authors argue, may have been due to the relatively specialised requirements of walruses, which allowed them to be out-competed by sea lions and fur seals - which, being comparatively deep-water animals, were less bothered by changing coastlines.

We can also, of course, look at where the various fossil species were found and how old they are to see how the group spread across the world. Here, we can use the family tree to try to fill in gaps; once we know that two particular species are closely related, then we can infer where their mutual ancestor is likely to have lived, and where the mutual ancestor of that animal and the next nearest group lived, and so on back down the tree.

The oldest pinniped fossils, including Enaliarctos, which lived 28 million years ago, before the existing families separated, all hail from the North Pacific. So it's clear that this is where the group first appeared. The ancestors of the sea lions and walruses remained there for a long time, with the latter only heading to the South Pacific around 6 million years ago, and using that as a route to colonise the South Atlantic and the Indian Ocean in even more recent times. The last surviving walrus species, on the other hand, was eventually forced north into the icy waters of the Arctic.

The story for the true seals, however, is very different. Their origin has been dated to 23 million years ago and if their closest relatives really are the lost desmatophocids, it must have taken place in the North Pacific - the only place that the latter have ever been found. But the oldest fossils identifiable as seals are 22 million years old and hail from the North Atlantic.

What seems to have happened here is that the ancestors of the seals left their original home almost immediately, heading eastwards into the Atlantic via the wide waterway that still existed between North and South America at the time. From there they later reached the Mediterranean and Arctic Oceans, and, via the South Atlantic, the Antarctic and then the South Pacific. (There are no true seals in the South Atlantic today, but intermediate fossils of the right age have been found there). By a circuitous route that probably took them over 20 million years to complete, they even got back to their origin point in the North Pacific, in the form of the northern elephant seal, whose ancestors swam up from the south where the southern elephant seal still lives.

This new analysis, however, adds a wrinkle to the above story, which we could already infer from simply looking at the dates of existing fossils. This concerns the initial diversification within the seal family, previously assumed to have taken place in the North Atlantic after their departure from the Pacific. This is due to the presence of some early seal fossils, including all of the extinct devinophocine subfamily, not in the North Atlantic but the Mediterranean. That they reached such a place so early on makes it entirely possible that they originated in the Paratethys, a substantial sea that once existed northeast of today's Mediterranean, and of which the Black and Caspian Seas are the last remnants. 

The study estimates that the very first pinnipeds lived around 29 million years, in the North Pacific. That's slightly later than most other estimates, but not dramatically so. What it can't do is tell us where the land-dwelling ancestors of those first pinnipeds lived. That's because we have two or three possible candidates for the closest fossil relatives of that animal - the almost-but-not-quite-yet pinnipeds of the Late Oligocene. Pujila lived in what is now the Canadian Arctic, which makes sense if pinnipeds originally entered the water in the North Pacific, as you would expect. But Potamotherium and Amphicticeps have both been proposed as close relatives of the first pinnipeds... and one of those lived in Europe, and the other in Mongolia. 

The real problem there is the blurred and muddy history of the carnivorans that lived so early on, before most of the current families had come into existence. Perhaps somebody needs to sort that one out... but that's an even more daunting task.

[Photo by "Momotarou2012" from Wikimedia Commons. Cladogram adapted from Park et al. 2024.]