During the last 30 years paleontologists have uncovered a startling amount of fossil evidence which has illuminated the early evolution of whales. The earliest members of the cetacea looked nothing like the marine mammals we are familiar with today, and in December of 2007 a paper in Nature identified a small hoofed mammal called Indohyus as one of the closest relatives to the earliest whales. This hypothesis was supported by a subsequent study published a few months ago in the same journal.
One of the most interesting aspects of Indohyus was that parts of its skeleton were thickened through osteosclerosis. The earliest whales, the archaeocetes, exhibited this condition, as well, and it appears that such thickening is an adaptation to life in the water. Thicker bones would have acted as a kind of ballast which would have allowed the earliest whales, like Pakicetus and Ambulocetus, to expend more energy on swimming than on staying underwater. The trade off is that the bones become more brittle and it becomes more energetically expensive it is to move on land, so it seems probable Indohyus was already spending a good deal of time in the water.
Yet Indohyus did not have broad, paddle-like feet or a thick tail for swimming. How did it get around while submerged? It appears that Indohyus might have walked along the bottom of shallow bodies of fresh water in a manner similar to living hippos, which also have osteosclerotic skeletons. Given the size and general appearance of Indohyus, though, it has often been compared to the small swimming mouse deer of Africa and Asia. These hoofed mammals jump into the water when danger approaches and can walk along the bottom.
The connection seems obvious. Here we have a small, hoofed fossil mammal related to whales that was probably spending a considerable amount of time in the water and a small, living hoofed mammal that exhibits a behavior that could explain the start of the transition into the water. The early evolution of whales, then, might have begun as a result of Indohyus-like creatures diving into ponds or streams to escape predators.* This is the idea forwarded by a BBC story covering new research describing swimming mouse deer from Asia and an entry on Jerry Coyne's blog Why Evolution is True, but I'm not so sure.
*[Keep in mind that Indohyus was not the last common ancestor of all whales and that it may have differed significantly from the true ancestor of the earliest whales. Indeed, Indohyus and some of it's raoellid relatives lived at the same time as the earliest known fossil whales. Its use as a model for the ancestor of early whales is a hypothesis that has yet to be confirmed by further fossil evidence. Thanks to Neil for reminding me to make this point clear.]
The proposed transition is a little fuzzy. Coyne outlines it this way;
Clearly [mouse deer] can stay underwater for many minutes at a time. And avoidance of predators by jumping into the water and remaining submerged is obviously something that natural selection could favor. Ergo, hippos, and maybe whales eventually.
There could be some selective advantage to hiding in the water if it offers a survival advantage that allows certain individuals to survive long enough to reproduce, but as I thought about this the bones of Indohyus kept bothering me. As I stated before Indohyus was already exhibiting osteosclerosis as an adaptation to life in the water. How could jumping into a pool for a few minutes at a time select for this (especially if being submerged is not essential to escaping)?
If the bones of swimming mouse deer are thickened or osteosclerotic it could support the hypothesis mentioned above. Unfortunately I am not sure whether anyone has even looked into this question, and the most I could find was an uncited note in a recent review of whale evolution stating that "Hyemoschus [the African mouse deer] is not osteosclerotic and spends relatively little time in the water." This suggests that escaping predators by jumping into the water alone is not sufficient to explain the adaptations seen in Indohyus. Perhaps something else was to account for the thickening of bone seen in Indohyus and early archaeocetes.
Could feeding have played a more important role in the transition? From its dentition Indohyus appears to have been an omnivore and perhaps it occupied a niche near the water's edge. If its ancestors were increasingly going into the water for food, like plants or shellfish on the bottom, then heavier bones might be selected for as the individuals would expend less effort to reach a potentially rich food source. This nutritional advantage might then have been translated into a reproductive one.
Admittedly this scenario is just as speculative as the one I criticized. For support it would rely on 1) more fossils and Indohyus and closely related creatures, and 2) the ability to determine (perhaps through isotopes in the teeth) whether individuals were feeding/spending time in the water more often than on land. (Istope values for living mouse deer might be helpful for comparison if this were done.) Given that we would be trying to get at behavior of an animal that can no longer be directly witnessed, however, this would be a rather sticky task.
Could the origin of whales have begun with little Indohyus-like creatures diving into the water to avoid becoming lunch? Perhaps, but we would still need an explanation as to why they stayed in the water and what selected for the osteosclerosis that became an important adaptation to aquatic life. That the bones of modern swimming mouse deer are not osteosclerotic suggests that there was something else going on, but just what that was may be nearly impossible to determine. There are important details yet to be ironed out, and we should be careful when appealing to present behaviors to explain past transitions.
As you stated in this post, Hippos are the closest living relative of whales and are also osteosclerotic. However, it might help our understanding if we knew why Hippos evolved to spend so much time in the water. If they started going into the water to avoid predators (a la the swimming mouse deer), that might support the BBC report. On the other hand, if hippos entered the water to feed, that might support your line of thought. Given the little I know about hippos, they don't seem too worried about predators (I recall reading about one that killed a crocodile once) and do seem concerned with food. However, their lack of predators could be a result of their larger size which evolved after their entry into water. Are there any studies on hippo evolution or lifestyles that might be relevant to whale evolution?
Curiously, while the two extant genera of hippos (Hippopotamus and Choreopsis) have rather different diets, both primarily feed on terrestrial vegetation. That doesn't necessarily indicate that their ancestors didn't feed on aquatic vegetation. However, it does suggest that the adaptive value of aquatic habits and pachyostosis (or osteosclerosis or pachyosteosclerosis or whatever) is not necessarily linked to aquatic foraging.
Tapirs, which do feed at least partly on aquatic plants if I recall, also exhibit thickening of the bones.
Although adult Hippopotamus are supposed to have no natural predators (aside from humans), juveniles are certainly at risk to predation by crocodiles, large cats etc., as probably are pygmy hippos. Predator avoidance *might* play a role here.
Off the top of my head, the avoidance of flying insect pests and thermoregulation (hippos forage at night) are two other reasons hippos might want to spend so much time in the water. Thermoregulation may sound wonky, but it is interesting that superficially "hippo-like" or "tapir-like" animals were fairly abundant and widespread in the Eocene, when global temperatures were much warmer than they are today (I'm thinking anthracotheres, pantodonts, basal proboscideans etc. as well as the ancestors of modern hippos and tapirs).
You might also argue that wallowing relieves them from having to expend so much energy to support their massive body on land -- but that argument gets a little circular though I suppose it might have a "positive reinforcement" aspect to it if the ancestors of hippos originally took to the water for some other reason.
Regardless, and contra Coyne, we know from the fossil record that whales DID NOT go through a hippo-like phase in their evolution. It's also important to keep in mind that raoellids were apparently closely related to--but contemporaneous with and not ancestral to--pakicetids the earliest group thought to be on the "main-line" to whale evolution. Both whales and hippos may have ultimately sprung from a raoellid-like/chevrotain-like ancestor but we don't really have the fossil evidence to prove that yet.
Thanks Neil, as ever, for an excellent comment. I probably need to update my post to make sure people understand the last point you made, especially (i.e. that we're assuming Indohyus is a good model for the earliest whales).
Thank you for the details about osteosclerosis in aquatic/semi-aquatic mammals. I don't think the ideas I presented here entirely solve the problem, either, but I was trying to think of a scenario where thickening of bones could be selected for and could be supported by further evidence. The "mouse deer jumped into the water and explains how whales took to the water" scenario just seemed too flimsy to me.
At least we have some idea where to look and what to look for now, though. Hopefully discoveries made in the coming years will help flesh out the earliest part of this transition a bit.
In the case of Hippopotamus amphibius, which remains in the water during the day and emerges to forage terrestrially at night, I believe avoiding the heat and desiccation is also a major factor. That wouldn't explain things for Hexaprotodon liberiensis (a.k.a. Choeropsis liberiensis), which I believe is less extensively aquatic than Hippopotamus but probably more similar to the ancestral morphotype for hippopotamids.