NEWS AND VIEWS motions. Nonetheless, there remains a storm, the "hard wind" will eventually die adept at swimming. The long tail suggests small possibility that the result is simply away, and the Universe on large scales will that this creature lacked a fluke, and from a statistical fluke. be restored to the tranquil state it seemed the shape of the lumbar vertebrae it In situations like this, there is no substi­ to possess just a few short months ago. 0 seems that the creature swam by undula­ tute for more data. And more data will tions of its vertebral column. The large certainly come, although they will take August E. Evrard is in the Department of paddle-like hind feet were thrust through several years to gather. In the meantime, Physics, University of Michigan, Ann Arbor, the water when the back was flexed and cosmologists will scramble to find ways to Michigan 48109, USA. Nick Kaiser is in extended. Unlike seals, however, the accommodate this new result. like pic­ the Canadian Institute for Theoretical Astro• plane of undulation was not mediolateral nickers sent scurrying by a sudden squall. physics, University of Toronto, 60 St but dorsoventral. as in living cetaceans. Most would hope that. as in a thunder- George Street, Ontario, Canada M5S 1A1. The forelimbs were probably involved in steering but not propulsion. It seems PALAEONTOLOGY ------. _. - ~-~~ logical that may have gone through a seal-like amphibious stage early in their evolution. but the direct evidence for such Whales leave the beach a phase, as represented by , is extraordinary. Michael J. Novacek A seaward shift beyond Ambulocetus is dramatically exemplified by the THE mammalian clade, to which we be­ -artiodactyl link is supported bl whale, Rodhocetus, now described by long, primarily represents the baroque anatomical and molecular evidence4- . Gingerich et al. 2. Rodhocetlls is geologi­ extravagance of terrestrial adaptations ~ On the morphological side, further re­ cally younger than Ambulocetlls and also burrowing, galloping, hopping, bipedal finement comes from the study of a shows a more specialized apparatus for walking and tree climbing, not to mention group of lumbering four-legged mam­ swimming. Its short cervical vertebrae, gliding and powered flight, the probable mals, , that thrived in the large unfused sacral vertebrae and re­ extensions of a lifestyle in the trees. But a Early Cenozoic and were especially di­ duced femur are derived features associ­ 4 7 few returned to the aquatic verse in central Asia • . There is a general ated with adeptly swimming archaeocetes milieu of their ancestors. In thc view that mesonychids represent the in­ and modern whales. At the same time, case of the Order , the whales and termediate lineage between whales and Rodhocetus retains features of the ver­ dolphins, this return to the sea was so other ungulates7 (although this allocation tebral column and pelvic girdle that hark 2 decisive that adaptations to swimming, could stand some re-examination). At any back to terrestrial relatives . So it was not diving and feeding match or surpass those rate, mesonychids have served as a bau­ amphibious like Ambulocetus, but did in fishes and sharks. It was a big evolution­ plan for the evolutionary modifications show a mosaic of aquatic and terrestrial ary plunge, one that has eluded docu­ leading to cetaceans. traits. The new specimen also casts light mentation for many decades. Until recent years, the nature of this on early whale habitats, in that the Things, though, are changing. New dis­ transition was a matter of inference based was found in rocks that indicate deep­ coveries of early fossil whales, such as on comparison of whales with their puta­ water (neritic) environments. To date, those of Thewissen et al. (published in tive terrestrial relatives. The gap was filled early archaeocetes have been limited to l Science ) and Gingerich et al. (page 844 by a series of discoveries of fossil near-shore deposits, and so Rodhocetus 2 of this issue ), provide striking evidence archaeocete whales~. These forms show demonstrates early variations in both for both the phylogenetic connections of small pelvic girdles and hindlimbs, where­ locomotion and distribution. Indeed, the cetaceans as well as the fascinating mosaic as modern adult whales retain only ves­ diversified lifestyles revealed by these of early adaptive experiments in the tran­ tiges of pelvic and hindlimb elements early fossils suggest that whale evolution sition from land to water. This expanding embedded in the flank musculature. Un­ got off to a very quick start. fossil casebook on the origins of whales is fortunately, the adaptive role of these Amblilocetus, Rodhocetus and other one of the triumphs of modern vertebrate archaeocete hind appendages was uncer­ more aquatically specialized archaeocetes palaeontology. tain. They seemed of doubtful use in cannot be strung in procession from ances­ Intuition may weigh against the notion locomotion, but might have functioned as tor to descendant in a scala naturae. H that something so 'fishy' as a whale may copulatory guides • Nonetheless, these fossils are real data on have closest kinship with land mammals. A leap in insight comes, however, with the early evolutionary experiments of Aristotle, however, was not fooled by Thewissen and colleagues' discoveryl of whales. They powerfully demonstrate intuition; he pointed out that whales and Ambulocetus, a truly remarkable fossil transitions beyond the reach of data, dolphins, unlike bony fish, are live­ from 52-million-year-old sequences in whether molecular or morphological, bearing (viviparous). John Ray, in his Pakistan. Ambulocetus can be clearly derived from living organisms alone. classic treatise3 of 1693, was more em­ allied with archaeocetes and other ceta­ phatic: "For as except as to the place in ceans, based on features of the middle ear, Michael J. Novacek is at the American which they live, the external form of the muzzle, skull roof and teeth 1.4. Behind the Museum of Natural History, Central Park body, the hairless skin [sic], and the skull, however, the specimen shows an West at 79th Street, New York, New York progressive or swimming motion, they extraordinary combination of features 10024, USA. [cetaceans] have almost nothing in com­ that anticipate, but do not fully embody, mon with fishes, but in remaining [charac­ the aquatic adaptations of cetaceans. Both 1. Thewissen. J. G. M , Hussain, S. T. &Arif. M. Science 263. ters] agree with viviparous quadrupeds". front and hindlimbs are well developed, 210--212 (1994) Modern systematists have gone some with flexible elbows, wrists, knee joints 2. Gingerich. P. D .. Raza, S. M .. Arif, M .. Anwar, M. &Zhou. X. Nature 368. 844~847 (1994). way in refining this concept. The Order and digits. The hand is large and elongate 3. Ray, J. Synopsis Methodica Animalium Quadrupedum et Cetacea seems to be related to the great and the hind foot is huge. Toes end in serpentini Generis. 8 (London. 1693). hooves as in mesonychids and other ungu­ 4. Berta. A. SCience 263. 180~181 (1994). radiation of herbivorous ungulate mam­ 5. Milinkovitch. M. C., Ortl. G. & Meyer. A. Nature 361, mals, and specifically the cloven-hoofed lates, and the tail is notably long. 346-348 (1993). (even-toed) artiodactyls, an order whose Thewissen et al. propose that Ambu­ 6. Novacek. M.J. Nature356, 121~125 (1992). 7. VanValen. L. Bull. Am. Mus. nat. Hist.132.1 (1966). modern members include deer, antelope, locetus was amphibious; that it was poss­ 8. Gingerich. P. D., Smith. B. H. &Simons. E. L. Science 249. camels, pigs, giraffes and hippos. The ibly awkward but mobile on land, and 154-157 (1990). NATURE . VOL 368 . 28 APRIL 1994 807

© 1994 Nature Publishing Group