Whence Whales?

Hairless, legless and confined to the sea, whales make for unlikely mammals. But millions of years of evolution can yield surprising results. In the case of whales and their cetacean kin, it led to one of the most dramatic transformations known, producing fully aquatic mammals from terrestrial ones. For that reason, whale origins have long fascinated scholars.

The fossil record documents much of the whales' land-to-water transition. Determining which mammalian group gave rise to these leviathans, however, has proved difficult. Scientists agree that whales are actually highly specialized ungulates, or hoofed mammals. The question has been, to which ungulates are they most closely related? Traditionally, paleontologists have posited that whales descended from extinct hyenalike creatures called mesonychians, based on dental similarities between the two groups. But in recent years molecular biologists have put forth a different hypothesisbased on DNA from living animalsasserting that the ancestors of whales were instead artiodactylsa group whose extant members include hippopotamuses, pigs, camels and ruminants. Furthermore, several molecular studies have concluded that whales share a common artiodactyl ancestor with hippos and are thus more closely related to these animals than to any other living artiodactyl or to a mesonychian.

Now newly found fossils from Pakistan may resolve some of the confusion. According to two research teams, the ancient whale remains reveal telltale signs of artiodactyl ancestry. How the fossils bear on the proposed whale-hippo link, however, is less clear. The findings appear in the September 20 issue of the journal Nature and the September 21 Science.

Over the past decade paleontologists have unearthed a number of striking fossils from amphibious whalessuch as the bizarre-looking Ambulocetus, which moved easily between land and sea on well-developed limbsand wholly aquatic ones. The remains described in the Science report add substantially to that body of knowledge, revealing for the first time fully terrestrial whales.

J.G.M. Thewissen of the Northeastern Ohio Universities College of Medicine and his colleagues discovered the fossils in 50-million-year-old deposits in the Kala Chitta Hills of Punjab, Pakistan. The remains include several skulls, snout fragments and a number of postcranial bones representing two species of primitive whales known as pakicetids: wolf-size Pakicetus attocki and fox-size Ichthyolestes pinfoldi. Detailed analyses of the skull and locomotor skeleton, the team reports, reveal adaptations characteristic of land-dwelling mammals. In fact, many features, such as long, slender limb bones, indicate that the pakicetids were built to run.

The most important information provided by the Thewissen team's discovery lies in the well-preserved ankle bones, particularly the so-called astragali. Paleontologists have long suspected that such fossils would settle the debate over whale origins. Artiodactyls are characterized by certain features of their ankles that enable the flexibility required for running. Thus, if whales descended from artiodactyls and not mesonychians, the earliest whales should exhibit artiodactyl-like ankles. Thewissen uncovered some isolated fossil whale astragali a few years ago, but those bones were fragmentary and showed a puzzling combination of whale and artiodactyl features. The new fossils, in contrast, closely resemble artiodactyl astragali.

"For the first time, morphological evidence shows that artiodactyls are the closest relatives of the cetaceans," Christian de Muizon of the Museum of Natural History in Paris writes in a commentary accompanying the Science report. "The new fossils superbly document the link between modern whales and their land-based forebears," he concludes, "and should take their place among other famous 'intermediates,' such as the most primitive bird, Archaeopteryx, and the early hominid Australopithecus.

The specimens described in the Nature report also include ankle remains. University of Michigan paleontologist Philip D. Gingerich and his colleagues recovered partial skeletons of two new fossil whales, Artiocetus clavis and Rodhocetus balochistanensis, from 47-million-year-old rocks in Pakistan's Balochistan Province. Both of them, the team reports, have ankle bones bearing characteristics diagnostic of artiodactyls. For Gingerich, who previously espoused the view that whales evolved from mesonychids, the significance of the findings took a while to sink in. But after careful consideration, he finally concluded that whales evolved from artiodactyls. "Now I admit the possibility that hippos are a sideline of artiodactyls that might be closer to whales than any other living animals," he remarks.

"In the last few years, 15 or 20 DNA studies have come out supporting this artiodactyl connection," Gingerich notes. "Those weren't taken very seriously, but this finding shows that they need to be. If the studies are done well, the DNA that animals carry in their bodies today gives us a better picture of the past than we might have thought it did." He further asserts that although molecular techniques will never replace paleontology, "they will complement it and expand on what we can competently infer."

Gingerich's new fossils also provide insight into how these beasts got around. Unlike Pakicetus and Ichthyolestes, Rodhocetus and Artiocetus appear to have spent a fair amount of time in the water. "It's clear that these animals could hitch their way out of water and back in like sea lions do today," he observes. "But they were more aquatic than I realized." In the water, the whales' well-developed limbs probably would have functioned as paddles.

Although all the new fossils point to artiodactyls as the ancestors of whales, where on the family tree cetaceans belong in relation to hippos remains controversial. Whereas Gingerich's team considers the whale-hippo link a possibility, the analysis conducted by Thewissen and his colleagues indicates that cetaceans are not more closely related to one artiodactyl groupsuch as hipposthan another.

Resolving that matter will require further work. "Two other evolutionary transitions vital to our understanding of the relationship between whales and artiodactyls beg for elucidation: the precise ancestry of hippopotami and the origin of artiodactyls themselves," Kenneth D. Rose of Johns Hopkins University comments in a perspective article accompanying the Science report. "The answers seem likely to come only from an improved fossil recordperhaps from the same region that has yielded fossils showing that whales evolved from artiodactyls."