Sinonyx, a wolf-sized hyena-like, land-dwelling mammal with hoofs from the order Condylarthra, which gave rise to artiodactyls ( even-toed hoofed mammals), perissodactyls ( odd-toed hoofed mammals), proboscideans (Trunk-nosed mammals), from the late Paleocene, about 60 million years ago. The characters that link Sinonyx to the whales, thus indicating that they are relatives, include an elongated muzzle, an enlarged jugular foramen (a natural opening or perforation through a bone or a membranous structure), and a short basicranium (underside of the skull). (Zhou and others 1995). The tooth count was the primitive mammalian number (44); there were different types of teeth like mammals today. The molars were very narrow shearing teeth, especially in the lower jaw, but possessed multiple cusps. The elongation of the muzzle is often associated with hunting fish - all fish-hunting whales, as well as dolphins, have elongated muzzles. These features were atypical of mesonychids, indicating that Sinonyx was already developing the adaptations that later became the basis of the whales' specialized way of life. Length 1.5 meters/ 5 feet.
The next fossil in the sequence, Pakicetus, is the oldest cetacean (marine mammal), and the first known archaeocete (ancient whale). It is from the early Eocene of Pakistan, about 52 million years ago (Gingerich and others 1983). Although it is known only from fragmentary skull remains, those remains are very diagnostic, and they are definitely intermediate between Sinonyxand later whales. This is especially the case for the teeth. The upper and lower molars, which have multiple cusps, are still similar to those of Sinonyx, but the premolars have become simple triangular teeth composed of a single cusp serrated on its front and back edges. The teeth of later whales show even more simplification into simple serrated triangles, like those of carnivorous sharks, indicating that Pakicetus's teeth were adapted to hunting fish.
A well-preserved cranium shows that Pakicetus was definitely a cetacean with a narrow braincase, a high, narrow sagittal crest, and prominent lambdoidal crests. Gingerich and others (1983) reconstructed a composite skull that was about 35 centimeters long. Pakicetus did not hear well underwater. Its skull had neither dense tympanic bullae nor sinuses isolating the left auditory area from the right one - an adaptation of later whales that allows directional hearing under water and prevents transmission of sounds through the skull (Gingerich and others 1983). All living whales have foam-filled sinuses along with dense tympanic bullae that create an impedance contrast so they can separate sounds arriving from different directions. There is also no evidence in Pakicetus of vascularization of the middle ear, which is necessary to regulate the pressure within the middle ear during diving (Gingerich and others 1983). Therefore, Pakicetus was probably incapable of achieving dives of any significant depth. This paleontological assessment of the ecological niche of Pakicetus is entirely consistent with the geochemical and paleoenvironmental evidence. When it came to hearing, Pakicetus was more terrestrial than aquatic, but the shape of its skull was definitely cetacean, and its teeth were between the ancestral and modern states. Length 1.5 meters/ 5 feet.
In the same area that Pakicetus was found, but in sediments about 120 meters higher, Thewissen and colleagues (1994) discovered Ambulocetus natans, "the walking whale that swims", in 1992. Dating from the early to middle Eocene, about 50 million years ago, Ambulocetus is a truly amazing fossil. It was clearly a cetacean, but it also had functional legs and a skeleton that still allowed some degree of walking on land. The conclusion that Ambulocetus could walk by using the hind limbs is supported by its having a large, stout femur. However, because the femur did not have the requisite large attachment points for walking muscles, it could not have been a very efficient walker. Probably it could walk only in the way that modern sea lions can walk - by rotating the hind feet forward and waddling along the ground with the assistance of their forefeet and spinal flexion. When walking, its huge front feet must have pointed laterally to a fair degree since, if they had pointed forward, they would have interfered with each other.
The forelimbs were also intermediate in both structure and function. The ulna and the radius were strong and capable of carrying the weight of the animal on land. The strong elbow was strong but it was inclined rearward, making possible rearward thrusts of the forearm for swimming. However, the wrists, unlike those of modern whales, were flexible.
It is obvious from the anatomy of the spinal column that Ambulocetus must have swum with its spine swaying up and down, propelled by its back feet, oriented to the rear. As with other aquatic mammals using this method of swimming, the back feet were quite large. Unusually, the toes of the back feet terminated in hooves, thus advertising the ungulate ancestry of the animal. The only tail vertebra found is long, making it likely that the tail was also long. The cervical vertebrae were relatively long, compared to those of modern whales; Ambulocetus must have had a flexible neck.
Ambulocetus's skull was quite cetacean (Novacek 1994). It had a long muzzle, teeth that were very similar to later archaeocetes, a reduced zygomatic arch, and a tympanic bulla (which supports the eardrum) that was poorly attached to the skull. Although Ambulocetus apparently lacked a blowhole, the other skull features qualify Ambulocetus as a cetacean. The post-cranial features are clearly in transitional adaptation to the aquatic environment. Thus Ambulocetus is best described as an amphibious, sea-lion-sized fish-eater that was not yet totally disconnected from the land life of its ancestors. Length 3 meters/ 10 feet.
In the middle Eocene (46-7 million years ago) Rodhocetus took all of these changes even further, yet still retained a number of primitive land features (Gingerich and others 1994). It is the earliest archaeocete (ancient whale) of which all of the thoracic, lumbar, and sacral vertebrae have been preserved. The lumbar vertebrae had higher neural spines than in earlier whales. The size of these extensions on the top of the vertebrae where muscles are attached indicate that Rodhocetus had developed a powerful tail for swimming.
Elsewhere along the spine, the four large sacral vertebrae were unfused. This
gave the spine more flexibility and allowed a more powerful thrust while
swimming. It is also likely that Rodhocetus had a tail fluke, although
such a feature is not preserved in the known fossils: it possessed features -
shortened cervical vertebrae, heavy and robust proximal tail vertebrae, and
large dorsal spines on the lumbar vertebrae for large tail and other axial
muscle attachments - that are associated in modern whales with the development
and use of tail flukes. All in all, Rodhocetus must have been a very good
tail-swimmer, and it is the earliest fossil whale committed to this manner of
The pelvis of Rodhocetus was smaller than that of its predecessors, but it was still connected to the sacral vertebrae, meaning that Rodhocetus could still walk on land to some degree. However, the ilium of the pelvis was short compared to that of the mesonychids, making for a less powerful muscular thrust from the hip during walking, and the femur was about 1/3 shorter than Ambulocetusís, so Rodhocetus probably could not get around as well on land as its predecessors (Gingerich and others 1994).
Rodhocetus's skull was rather large compared to the rest of the skeleton. The premaxillae and dentaries had extended forward even more than its predecessorsí, elongating the skull and making it even more cetacean. The molars have higher crowns than in earlier whales and are greatly simplified. The lower molars are higher than they are wide. There is a reduced differentiation among the teeth. For the first time, the nostrils have moved back along the snout and are located above the canine teeth, showing blowhole evolution. The auditory bullae are large and made of dense bone (characteristics unique to cetaceans), but they apparently did not contain the sinuses typical of later whales, making it questionable whether Rodhocetus possessed directional hearing underwater.
Overall, Rodhocetus showed improvements over earlier whales by virtue of its deep, slim thorax, longer head, greater vertebral flexibility, and expanded tail-related musculature. The increase in flexibility and strength in the back and tail with the accompanying decrease in the strength and size of the limbs indicated that it was a good tail-swimmer with a reduced ability to walk on land. Length 3 meters/ 10 feet.
The particularly well-known fossil whale Basilosaurus represents the next evolutionary grade in whale evolution (Gingerich 1994). It lived during the late Eocene and latest part of the middle Eocene (35-45 million years ago). Basilosaurus was a long, thin, serpentine animal that was originally thought to have been the remains of a sea serpent (hence it is name, which actually means "king lizard"). Its extreme body length (about 15 meters) appears to be due to a feature unique among whales; its 67 vertebrae are so long compared to other whales of the time and to modern whales that it probably represents a specialization that sets it apart from the lineage that gave rise to modern whales.
What makes Basilosaurus a particularly interesting whale, however, is the distinctive anatomy of its hind limbs (Gingerich and others 1990). It had a nearly complete pelvic girdle and set of hindlimb bones. The limbs were too small for effective propulsion, less than 60 cm long on this 15-meter-long animal, and the pelvic girdle was completely isolated from the spine so that weight-bearing was impossible. Reconstructions of the animal have placed its legs external to the body - a configuration that would represent an important intermediate form in whale evolution.
Although no tail fluke has ever been found (since tail flukes contain no bones and are unlikely to fossilize), Gingerich and others (1990) noted that Basilosaurus's vertebral column shares characteristics of whales that do have tail flukes. The tail and cervical vertebrae are shorter than those of the thoracic and lumbar regions, and Gingerich and others (1990) take these vertebral proportions as evidence that Basilosaurus probably also had a tail fluke.
Further evidence that Basilosaurus spent most of its time in the water comes from another important change in the skull. This animal had a large single nostril that had migrated a short distance back to a point corresponding to the back third of the dental array. The movement from the forward extreme of the snout to the a position nearer the top of the head is characteristic of only those mammals that live in marine or aquatic environments. Length 15 meters/ 50 feet.
Dorudon was a contemporary of Basilosaurus in the late Eocene (about 40 million years ago) and probably represents the group most likely to be ancestral to modern whales (Gingerich 1994). Dorudon lacked the elongated vertebrae of Basilosaurus and was much smaller (about 4-5 meters in length). Dorudonís dentition was similar to Basilosaurusís; its cranium, compared to the skulls of Basilosaurus and the previous whales, was somewhat vaulted (Kellogg 1936). Dorudon also did not yet have the skull anatomy that indicates the presence of the apparatus necessary for echolocation (Barnes 1984). Length 5 meters/ 16 feet.
Basilosaurus and Dorudon were fully aquatic whales (like Basilosaurus, Dorudon had very small hind limbs that may have projected slightly beyond the body wall). They were no longer tied to the land; in fact, they would not have been able to move around on land at all. Their size and their lack of limbs that could support their weight made them obligate aquatic mammals, a trend that is elaborated and reinforced by subsequent whale taxa.