The East African Hippopotamus.
If the idea of whales being mammals has always seemed a bit wild, then you'll probably be surprised to learn that the giant aquatic beasts are pretty closely related to the hippopotamus.
Scientists have been wrangling over these relations for centuries.
One theory had been that hippos were related to pigs. Yet mounting evidence suggested they are closer to whales. A new study concludes that a four-footed semi-aquatic mammal that thrived for some 40 million years was a common ancestor to both whales and hippos.
"The problem with hippos is, if you look at the general shape of the animal it could be related to horses, as the ancient Greeks thought, or pigs, as modern scientists thought, while molecular phylogeny shows a close relationship with whales," said Jean-Renaud Boisserie, a post-doctoral fellow at the University of California, Berkeley. "But cetaceans - whales, porpoises and dolphins - don't look anything like hippos."
To complicate matters, there is a 40-million-year gap between fossils of early cetaceans and early hippos.
Boisserie and colleagues in France say they've filled in the gap with fossils of a "water-loving animal" that evolved into two groups, early cetaceans and a group of four-legged animals called anthracotheres. The pig-like anthracotheres, which developed at least 37 distinct genera, died out less than 2.5 million years ago, leaving only one line: the hippopotamus.
The analysis puts whales within a large group of cloven-hoofed mammals called Artiodactyla. That makes them relatives of cows, pigs, sheep, antelopes, camels and giraffes, too.
The idea of whales and hippos being related has gained steam in recent years. Boisserie's team analyzed new and previous hippo, whale and anthracothere fossils to pin down anthracotheres as the missing link between hippos and cetaceans, they say.
"Our study is the most complete to date, including lots of different taxa and a lot of new characteristics," Boisserie said. But leaving the case not quite shut, he added: "Our results are very robust and a good alternative to our findings is still to be formulated."