Illustration by Luci Betti-Nash, courtesy Stony Brook University Moeritherium, a 37 million-year-old amphibious relative of elephants

This is pretty mindblowing stuff. The ideas include the theory, based on DNA evidence, that the elephants could be related to actual aquatic, fully marine, animals like manatees. The BBC:

DNA evidence suggests that elephants are related to seagoing manatees and dugongs, and another land-based mammal, the rabbit-like hyrax.

This led to the theory that elephants and their extinct relatives may have evolved from a water-dwelling ancestor.

Scientists in the UK and the US looked at fossil teeth of two species that belong to an extinct family of mammals related to the elephant and, more distantly, the sea cow. They lived in northern Egypt during the Eocene Epoch, about 37 million years ago.

Now- Mammals started out as purely land dwelling animals. Visualize a series of situations where a land dwelling animal becomes a marine animal, and its relatives become elephants. Some family. This is even more extraordinary because the elephant family is well documented, in many forms, including mammoths, mastodons, and a range of other Ice Age relatives, some of which are quite bizarre, but recognizable elephant types. Researchers have found distinct resemblances between Moeritherium and living animals with amphibious lifestyles:

Dr Erik Seiffert told BBC News: "The isotopic pattern preserved in their teeth is very similar to that of living aquatic mammals.

"It supports the hypothesis that, at some point early in the evolution of elephants, these animals were very dedicated to either a fully aquatic or amphibious lifestyle - they probably spent most of their life in water."

A hippo-like lifestyle is considered the probable scenario, for Moeritherium with a guess that drying up of habitat started the move to elephant-dom, but according there’s no actual fossils of aquatic or semi aquatic elephants to back it up, according to researchers. Interesting logic, because they might also try looking for the post-aquatic/amphibious intervening forms between Moeritherium and modern elephants, which you’d think could be relevant, too. The relationship between elephants and aquatic animals is fascinating, and very significant. This research is important, but it seems to be missing a basic evolutionary principle. The divergence of any species is based on a process of separation. It doesn’t follow that species with a common ancestor which have become quite different produce a single continuum. They produce several continuums, leading back to the ancestral form . Manatees and dugongs, for example, are ancient in their own right, and coexisted with ancient elephants. Moeritherium is a sort of mid-point. For an effective evolutionary continuum the trail has to lead forward and back. An “aquatic elephant”, as such, may not have existed. Elephants, as elephants, don’t need flippers, and don’t eat sea grass. An ancestor of elephants which was aquatic, however, would be a staggering discovery. It would mean a whole cycle of land/aquatic or amphibious/land dwelling. Even the idea of that is new ground. A fully environment-driven evolutionary path like that would also confirm a lot of Darwin's basic principles about natural selection. Amphibious or aquatic, it's quite a CV for a species. I guess we’ll have to wait and see what this produces.