The First Whales to Use Sonar?

Xenorophus

Other than their obvious physical adaptations, one of the most familiar features of dolphins and whales is their ability to use ultrasound to echolocate. All dolphins, porpoises, and toothed whales (collectively called odontocetes) can echolocate and, while some of the fine details do vary between, say, sperm whales and some of the smaller species, the basic mechanism is much the same. Ultrasound would not have been useful to the large, ground-dwelling ancestors of whales in the same way that it is for bats, so it must have evolved after they entered the water. But how soon after?

We can put some limits on this. At the younger end, since all odontocetes echolocate, it's unlikely to have evolved any later than their last common ancestor, which is estimated to have lived around 34 million years ago. On the other hand, it's notable that the toothless, baleen, whales do not use ultrasound; in fact, they are specialised in the exact opposite direction to produce sounds well below, not above, the range of human hearing. This suggests that they evolved along different lines, and that the origins of ultrasonic echolocation lie somewhere after the two split. 

Whether an animal uses ultrasound is, as one might expect, not all that easy to tell when all you have to go on is a skeleton, and often one that's incomplete or badly damaged at that. Many of the adaptations required to produce ultrasound lie in the soft tissues... but not, as it happens, all of them. Crucially, such animals need air gaps around the ear to stop sound being transmitted from the skull and confusing directional hearing, and a concave shape to the bones under the forehead, leaving space for the melon that transmits the outgoing signal.

If those features are both present in a fossil, it's a fair bet that the animal in question could use ultrasound to echolocate. Thus, for instance, we can be confident that the very earliest whales, while they may have had directional underwater hearing, did not yet have ultrasound. It's something odontocetes evolved, not something that baleen whales lost.

In 1923, Remington Kellogg published a description of a "dolphin" skull recovered from a deposit in South Carolina that we now know to have been around 28 million years old. He named it Xenorophus, for reasons that he didn't bother to explain ("xeno" means "strange", but what, if anything, "rophus" was intended to mean is unclear). Although it was obvious that this was an early odontocete, quite where it fit in relation to other whales was unclear and it ended up being shifted around quite a bit. It wasn't until 1972 that a second specimen was found but, after that, excavation of fossil beds in the area picked up and it became clear by at least 1992 that it was merely one member of a group of similar animals. All of which, incidentally, have been found in the same small area of the southeastern US coast and (so far as we can tell) never spread to the other side of the Atlantic, let alone into other oceans.

In 2008, Mark Uhen officially named that group as the Xenorophidae. At the time, it included Xenorophus itself, Archaeodelphis, which had actually been discovered slightly earlier, in 1921, and Uhen's newly described Albertocetus. Since then, five further genera have been added to the xenorophids, suggesting a group that was reasonably common and diverse, and that may be significant in cetacean evolution.

A new analysis of Xenorophus fossils, including a recently uncovered one that was unusually complete was published earlier this week, expanding on what we already knew about these animals. For example, while it is difficult to estimate the overall size of a cetacean from the skull alone, since bodily proportions varies more than it does amongst say, cats, here there's enough that the authors estimate that the animal would have between 2.6 and 3 metres (8½ and 10 feet) in length. 

The shape and wear of the teeth suggest that, as in modern odontocetes, the teeth did not grind against one another as they do in most land mammals, and that the diet of Xenorophus was unusually variable - mostly regular fish, but with some individuals feeding primarily on either soft-bodied squid or tougher food items such as small sharks. The snout is relatively long, similar to that of a modern spinner dolphin and thus proportionately longer than some other odontocetes, and one juvenile fossil indicates that this shape appeared earlier in life than we might think. This suggests rapid strikes to snap at fish, using the teeth to piece and hold them. This is perhaps unsurprising, but it's worth noting that another member of the family, Inermorostrum, had no teeth at all, presumably feeding on squid by simply sucking them out of the water. So feeding styles had diversified surprisingly early on in whale evolution, even among such a small and localised group.

Despite these similarities with modern dolphins, Xenorophus also shows several primitive features. The teeth are not all identical in shape, the bones of the snout are differently arranged, the anchor for the jaw muscles suggests a more powerful, but less rapid bite, and so on. But, bringing us back to the topic at the start of this post, there is good reason to suppose that it could use ultrasound to echolocate. This evidence comes from the new specimen, identified as belonging to a different species of the same genus as earlier ones and includes the fact that it did have the air sinuses that odontocetes need to home in on the sound of their sonar pings returning. 

Similarly, the skull shows the asymmetric pattern seen in living whales that is thought to be associated with their directional hearing, and there is a small hole in the skull that matches one seen only in odontocetes, and through which the nerve to the melon and its surrounding structures passes. This fits with what we already knew of another xenorophid, Echovenator, which clearly has the structures required to hear ultrasound, including fine structures of the inner ear that (since they're on the inside of the bone) have yet to be examined in Xenorophus itself.

This, and other features, such as the number and size of the teeth, imply rapid evolution at the very base of the odontocete family tree, making them look quite different to their older, non-echolocating, ancestors after a relatively short period of time. Yet there are also several differences, suggesting that the xenorophids are a very early branch among the odontocetes, splitting off before the last common ancestor of all living species - which, in turn, means that they left no descendants. 

That may not be terribly surprising, given how early they lived. But the apparently simple picture is confused by the 2017 discovery of the fossil whale Olympicetus from Washington state. While this has no living descendants either, it has several features that suggest it is closer to the living odontocetes than it is to the xenorophids - its ancestor split off before theirs did. Which wouldn't be odd... except that examination of its inner ear revealed that it couldn't hear ultrasound. 

Of course, it's possible that it had somehow lost the ability. But why would it do that? A more likely explanation is that it, and its common ancestor with the xenorophids, already had the directional hearing adaptations that are a prerequisite for sonar, but hadn't yet developed the full suite. In which case, the xenorophids must have developed ultrasound independently - the feature evolved twice. 

The xenorophids did not last long, dying out at the end of the Oligocene 23 million years ago. By which time, the ancestors of all living odontocetes also had the same ability. But there seems to have been a time, very early on, when two different groups of whale both adapted their pre-existing abilities at directional hearing to do something entirely new for marine mammals and shared the same oceans.

[Illustration from Boessenecker and Geisler 2017, available under CC-BY-4.0 license.]