Fossil Record (marine mammals)

 

The fossil record of marine mammals extends back more than 50 million (M) years (Fig. 1). Hundreds of species are known, with Cetacea numerically and taxonomically dominant and globally widespread, followed by Sirenia in mainly warm settings. Seals, sea lions, and relatives (Pinni-pedia) are locally abundant, but other marine carnivores (otters, amphicyonids) are rare and a marine sloth unique. Fossils occur in marine strata from nearshore to deep-ocean settings and. occasionally, freshwater habitats. Remains vary from near-complete skeletons through skulls and teeth to single and usually undiagnostic bones. The taxonomic framework, although not always firm, is adequate to review the diversity and spa-tiotemporal distribution of fossil marine mammals. Standard zoological techniques are used in taxonomy, classification, and analysis of function, whereas routine geological techniques are used to date fossils in terms of relative and absolute time scales and to interpret sedimentary environments. The fossil record shows patterns of evolution and extinction that link strongly with environmental change in the oceans.

Summary of geological age ranges for the main groups of fossil marine mammals.

Figure 1 Summary of geological age ranges for the main groups of fossil marine mammals.  

I. Occurrence, Environment, and Age

Fossil marine mammals occur in sedimentary rocks. Originally. remains accumulated in mud. silt, sand, or gravel, which, as flesh decayed, was buried and turned to rock through compaction and/or deposition of cementing minerals. Sedimentary rocks are recognized as discrete formations (genetically unified bodies of strata) and are named formally, e.g., the Calvert Formation, Maryland. Marine mammals come from strata including sandstone, mudstone, limestone, greensand, and phosphorite, most of which are marine rocks now exposed on land. Rare fossils have been recovered from the sea floor. Because broadly similar rock types may form at different times and places, sedimentary rocks must be dated to establish their time relationships.

Two correlated time scales, relative and absolute, are used in discussing the fossil record. The relative time scale has named intervals (epochs: Fig. 1) in an agreed international sequence: Eocene, Oligocene, Miocene, Pliocene, and Pleistocene. These epochs are usually subdivided into early, middle, and late. Stages (e.g., Aquitanian of Fig. 2) may provide finer subdivision. Typically, time intervals are recognized by a distinct suite of age-diagnostic fossils. The most reliable dates are based on oceanic microfossils with short time ranges, such as foraminifera, to allow correlation between ocean basins. Because of compounded errors of long-distance correlation, ages are rarely accurate to within 1 M years and manv fossils can be placed only roughly within a stage. Beyond the relative time scale, absolute dates in millions of years are needed to understand rates of processes. Absolute dates are obtained usually from radiometric analysis of grains of volcanic rock interbedded with strata that also contain age-diagnostic fossils.

II. Taxonomic Framework

Fossil marine mammals are classified on the evidence of skeletons. No other useful body parts preserve, and fossils have not yet produced biomolecules useful in molecular taxonomy. Skulls are by far the most versatile and thus important elements in classification, but teeth are taxonomically useful in most groups, and other bones (vertebrae, limb elements) have been used at times.

III. Cetacea

Fossils show that cetacean history extends back over 50 M years (Fig. 2). The earliest cetaceans—Archaeoceti—were small amphibious species that lived in fresh and brackish waters in the warm subtropical Tethys seaway between Eurasia, India, and Africa. By about 40 million vears before the present (Ma), archaeocetes included large and fully marine species that had spread to temperate latitudes. The earliest modern or crown group whales, species in the filter-feeding group Mysticeti, appeared at about the Eocene/Oligocene boundary, ~34-35 Ma. Odontocetes—echolocating toothed whales and dolphins— probably radiated about the same time. Odontocetes diversified dramatically to become the most speciose group of marine mammals. Most extant cetacean families were established before the end of the Miocene, 5-10 Ma, but no living species has a history clearly longer than 2 M years.

A. Archaeoceti

Knowledge of basal whales has expanded dramatically since the early 1980s, giving new insights into cetacean phylogeny, ecology, and distribution. As noted in Kellogg’s classic monograph, for many years, basal archaeocetes were known only from Protocetus atavus (Mokattam Formation, Middle Eocene, ~46.0 Ma; Egypt-Tethvs), which was represented by a skull and uncertainly related postcranial skeleton. Since the 1980s, new finds by Gingerich,  and others, especially in the eastern Tethys, have greatly increased diversity at the level of species, genus, and family. Basal archaeocetes are placed in the Pakicetidae, typified by the small Pakicetus inachus, from non-marine red beds of the Kuldana Formation (—49-49.5 Ma), Pakistan. Skull structure (Fig. 3B) indicates limited underwater hearing capabilities, and the teeth are more simple than those of many later forms. Pakicetus has been cited as evidence that the earliest cetaceans radiated slowly in productive shallow waters of the Tethys seaway between Asia and India. The pakicetid (and cetacean) record has been extended back to ~53.5 Ma, based on a fragmentary jaw of Himalayacetus subathuensis from India. Oxygen isotopes from Himalayacetus indicate marine rather than nonmarine habits. Other pakicetids include species of Ichthyolestes and Nalacetus, also from the eastern Tethys.

The family Protocetidae, now expanded beyond Protocetus, is an Early to Middle Eocene grade for species in which the skull has an enlarged supraorbital shield, the mandible has a large mandibular foramen, and hindlimbs are reduced; they lack the complex teeth and pterygoid sinuses of younger cetaceans. Some protocetids occur in the western Tethys to western central Atlantic, including the large protocetid Eocetus (Egypt, North Carolina), Pappocetus (Nigeria; possibly North Carolina), Natchitochia (Louisiana), and apparently Protocetus (Texas). Protocetids have a high diversity in the eastern Tethys, judging from the range of teeth from Pakistan and India, although few skulls are known. Babiacetus (~43.5 Ma) is known from teeth and jaws, whereas the slightly older (45.5-46 Ma) Gaviacetus, Takracetus, and Indocetus are represented by partial skulls. More complete is Rodhocetus kasrani (Domanda Formation, 46-46.5 Ma; Pakistan), in which the skeleton is strikingly intermediate between that of mesonychians (the terrestrial putative ancestors of cetaceans) and later whales. Cetacean features include the short neck vertebrae and more posterior vertebrae adapted for dorsoventral oscillation, but Rhodocetus retains a femur and sacrum. Rodhocetus kasrani is from deep rather than shallow water deposits, implying early colonization of offshore habitats. Another reputed protocetid, Georgiacetus vogt.lensis (McBean Formation, 40-41 Ma: Georgia), has somewhat elaborate cheek teeth, a pterygoid sinus in the skull base, and a reduced link between sacrum and pelvis; because there are otherwise basilosaurid features, Georgiacetus is perhaps better placed in the Basilosauridae.

Geological age ranges of major groups of Cetacea. Time scale shows absolute time, Epochs (e.g., Eocene, Oligocene) and their subdivisions (e.g., early, middle), and stages (e.g., Priabonian, Rupelian). Bars show age ranges for family level cetacean taxa, either clades or, where cladistic classification is lacking, grades. Accu racy of ranges varies between differen t groups and differen t time intervals. Inferred relationships follow Uhen and others for archaeocetes, and Muizon, Fordyce and Muizon, Barnes, Heyning, and Waddell et al. for odontocetes and mysticetes.

Figure 2 Geological age ranges of major groups of Cetacea. Time scale shows absolute time, Epochs (e.g., Eocene, Oligocene) and their subdivisions (e.g., early, middle), and stages (e.g., Priabonian, Rupelian). Bars show age ranges for family level cetacean taxa, either clades or, where cladistic classification is lacking, grades. Accu racy of ranges varies between differen t groups and differen t time intervals. Inferred relationships follow Uhen and others for archaeocetes, and Muizon, Fordyce and Muizon, Barnes, Heyning, and Waddell et al. for odontocetes and mysticetes.

Two other archaeocete families are reported only from the Early and Middle Eocene of the eastern Tethys. First,

Archaeocete cetaceans. (A) Reconstruction of a pakicetid-grade archaeocete, by C. Gaskin, from the Geology Museum, University of Otago. (B) Skull and mandibles of Pakice-tus inachus (Eocene, Pakistati), lateral view, after Gingerich and Russell. (C) Skull and mandibles of Remingtonocetus harudiensis (Eocene, India), lateral view, after Kumar and Sahni; mandibular form is speculative. (D) Skull and mandibles of Basilosaurus cetoides (Eocene, Alabama), oblique lateral view of specimen in the U.S. National Museum of Natural History. (E) Skeleton of Zygorhiza kochi (Eocene, Alabama and Mississippi), reconstruction based on composite specimens, from Kellogg (1936); (F) Skull and mandible of Dorudon atrox (Eocene, Egypt), lateral view, slightly modified from Andrews. (G) Tooth of Dorudon or Zygorhiza species indeterminate (Eocene, New Zealand), medial view.

Figure 3 Archaeocete cetaceans. (A) Reconstruction of a pakicetid-grade archaeocete, by C. Gaskin, from the Geology Museum, University of Otago. (B) Skull and mandibles of Pakice-tus inachus (Eocene, Pakistati), lateral view, after Gingerich and Russell. (C) Skull and mandibles of Remingtonocetus harudiensis (Eocene, India), lateral view, after Kumar and Sahni; mandibular form is speculative. (D) Skull and mandibles of Basilosaurus cetoides (Eocene, Alabama), oblique lateral view of specimen in the U.S. National Museum of Natural History. (E) Skeleton of Zygorhiza kochi (Eocene, Alabama and Mississippi), reconstruction based on composite specimens, from Kellogg (1936); (F) Skull and mandible of Dorudon atrox (Eocene, Egypt), lateral view, slightly modified from Andrews. (G) Tooth of Dorudon or Zygorhiza species indeterminate (Eocene, New Zealand), medial view.

Ambulocetus natans (Kuldana Formation, 48.0-49.0 Ma; Am-bulocetidae) includes a substantially complete skeleton with a long-snouted skull and well-developed fore- and hindlimbs. Ambulocetus perhaps swam using pelvic paddling and dorsoven-tral undulations of the tail, comparable in style to some modern otters. A crocodile-like mode of predation in water is possible, but locomotion on land was probably clumsy. Second, the family Remingtonocetidae encompasses specialized long-snouted Middle Eocene species of Remingtonocetus (Fig. 3C), Andrewsiphius, and allies, known from at least partial skulls (~43.5-45 Ma). Despite previous suggestions, remingtono-cetids seem unrelated to the later radiation of odontocetes.

Basilosaurids, from the later Middle and Late Eocene, are the oldest cetaceans known beyond the Tethys. They are typified by the 15-m-long Basilosaurus cetoides (Fig. 3D), first described and named as a fossil reptile from Louisiana (Jackson Formation, Late Eocene, ~36-39 Ma; western North Atlantic). The large size of Basilosaurus and its elongate vertebrae are specialized features used to recognize a subfamily Basilosauri-nae. The latter include Basilosaurus isis (~39 Ma; Egypt, central Tethys), which has small but functional hindlimbs of ungulate-like character. Large later Eocene archaeocetes, presumably basilosaurines, have been reported from scattered localities worldwide (e.g., Northeastern Atlantic, proto-Southem Ocean, Southwest Pacific), indicating an expanding range for Cetacea.

The second subfamily of basilosaurids, the Dorudontinae, is a grade that includes smaller, more generalized, and somewhat dolphin-like species of Dorudon, Pontogeneus, Zygorhiza (Fig. 3E), Saghacetus, Ancalecetus, and perhaps Georgiacetus. These genera are rather similar to one another and are diagnosed on size, tooth form, and limb form. Dorudon is known from two species, the others one each. The typical species D. serratus is fragmentary, but others include some magnificent fossils (e.g., Dorudon atrox, Birket Qarun Formation, 39-40 Ma: Egypt; Fig. 3F). These formally named species are from the Tethys and western Central Atlantic, but tantalizing referred specimens (Fig. 3G) occur in other widely separated localities pointing, as for basilosaurines, to an early geographic spread.

Basilosaurids differ from more basal archaeocetes in having cheek teeth with complex denticles and expanded basicranial air sinuses. These features, which indicate more sophisticated feeding and hearing capacities, link basilosaurids closely with early odontocetes and mysticetes. Several dorudontines are equally plausible sister taxa to the Odontoceti + Mysticeti, but basilosaurines seem too specialized, in terms of large size and elongate vertebral bodies, to be directly ancestral to living cetaceans. No positively identified archaeocetes are known from Oligocene or younger rocks.

B. Mysticeti

Since the 1960s, the fossil record of mysticetes has expanded to reveal diverse toothed and toothless Oligocene species, which effectively “bridge the gap” between archaeocetes and baleen-bearing mysticetes such as cetotheres, balaenopterids, and right whales. Pivotal here is Aetiocetus cotylalveus (Aetiocetidae; Yaquina Formation, Late Oligocene; Fig. 4A) from Oregon.

Initially, this small cetacean was identified as an archaeocete because it has teeth, but other features, including the flattened triangular rostrum, indicate that it is an archaic mysticete. Ae-tiocetids are moderately diverse in their known range (North Pacific) and include species of Chonecetus and Morawanocetus from Japan. Because aetiocetids are Late Oligocene only, they are probably relict basal mysticetes that persisted after more crown-ward baleen-bearing mysticetes had appeared.

Older archaic toothed fossil mysticetes are more problematic. The enigmatic Llanocetus denticrenatus (Llanocetidae) was based on a fragmentary large toothed jaw and a brain cast from the La Meseta Formation, Eocene-Oligocene boundary (—35 Ma) of Seymour Island, Antarctica (Fig. 4B). Fragments of unnamed small toothed mysticetes have been described from the basal Oligocene of New Zealand (Southwest Pacific-marginal proto-Southem Ocean) (Fig. 4C). Also from the margins of the proto-Southern Ocean are two other notable toothed species: the specialized short-snouted Mammalodon colliveri (latest Oligocene or earliest Miocene, ~23-24 Ma, Australia; Maminalodontidae) and the enigmatic large Keken-odon onamata (Late Oligocene, —27-28 Ma, New Zealand; Kekenodontidae). Other toothed Cetacea formerly identified as archaeocetes and odontocetes from New Zealand, Australia, and France probably also belong in the Aetiocetidae, Llanocetidae, Mammalodontidae, and Kekenodontidae. Most are ar-chaeocete-like, with broad-based rostra and otherwise subtle mysticete characters, and they are more widespread and diverse than previously suspected. Their teeth were probably used in filter feeding, perhaps supplemented by baleen of which no trace has been reported. These fossils represent early branches in mysticete evolution, some of which persisted as relict taxa.

Remains of toothless and baleen-bearing mysticetes are relatively abundant in Miocene and younger strata worldwide, and some fossils (e.g., species of Mauicetus and Cetotheriopsis) occur in Late Oligocene sequences back to 29-30 Ma. Many of the fossils, particularly those older than Late Miocene (~12 Ma), lack the distinguishing skull features of right whales, balaenopterids, and grav whales and have been placed in the family Cetotheriidae. As commonly used, the Cetotheriidae is a grade or “waste basket” family (paraphyletic and probably poly-phvletic), including several different lineages of archaic mysticetes. More than 20 genera have variously been placed in the group. Some Miocene cetotheres are clearly close to balaenopterids, differing mainly in the more primitive structure of the frontal bone over the eye, and these fossils may indeed be on the lineage leading to rorquals. A few of the latter, e.g., idiot: etus and Plesiocetus, have been classified alternatively in the Cetotheriidae or Balaenopteridae.

Strictly, cetotheres are typified by the Middle Miocene (12-13 Ma) Cetotherium rathkii from Ukraine (Paratethys), which is known only from an incomplete skull rather different in structure from those of living mysticetes (Fig. 4D). For example, the upper jaw (rostrum) thrusts back into bones of the braincase with a sharp narrow triangular apex, almost obscuring the nasal bones. The Pliocene Herpetocetus from the Yushima Formation of Japan shows a similar structure. Eventually, the Cetotheriidae will be defined as a elade for Cetotherium rathkii and its close relatives, although this will leave many other archaic mysticetes (e.g., Fig. 4E) uncertainly placed. This matter, the exact family level identity of many fossil mysticetes, is a key problem area in cetacean phylogeny.

Mysticete cetaceans. (A) Skull of Aetiocetus cotylalveus (Oligocene, Oregon), oblique dorsolateral view of holotype. (B) Field site showing excavation of ribs of the archaic mysticete Llanocetus denticrenatus (Eocene/Oligocene boundary. Antarctica). (C) Tooth of Llanocetus-?ifce archaic mysticete (?Llanocetidae) (Oligocene, New Zealand), from Fordyce. (D) Skull and mandibles of Cetotherium rathkii (Miocene, Ukraine), dorsal view, holotype skull and uncertainly associated mandibles, from Van Beneden ami Gervais. (E) Broken skull, mandibles, and associated elements of an undescribed Mauicetus-/ifce "cetothere" (Oligocene, New Zealand), dorsal view; specimen in Geology Museum, University of Otago.

Figure 4 Mysticete cetaceans. (A) Skull of Aetiocetus cotylalveus (Oligocene, Oregon), oblique dorsolateral view of holotype. (B) Field site showing excavation of ribs of the archaic mysticete Llanocetus denticrenatus (Eocene/Oligocene boundary. Antarctica). (C) Tooth of Llanocetus-?ifce archaic mysticete (?Llanocetidae) (Oligocene, New Zealand), from Fordyce. (D) Skull and mandibles of Cetotherium rathkii (Miocene, Ukraine), dorsal view, holotype skull and uncertainly associated mandibles, from Van Beneden ami Gervais. (E) Broken skull, mandibles, and associated elements of an undescribed Mauicetus-/ifce “cetothere” (Oligocene, New Zealand), dorsal view; specimen in Geology Museum, University of Otago.

Fossil balaenopterids, like their living relatives, have a distinctive skull structure in which the frontal bone above the eye is depressed to house the large muscles that close the lower jaw. Fossils such as Megaptera miocaena (Sisquoc Formation, California, northeast Pacific) indicate that Megaptera (subfamily Megapteri-nae, humpback whales) had diverged from the Balaenoptera lineage (subfamily Balaenopterinae, rorquals) by the Late Miocene (~12 Ma). Other Late Miocene and younger records of Megaptera, such as M. hubachi (Figs. 5A and 5B), are known. Although

Mysticete cetaceans. Skull and mandibles of Megaptera hubachi (Pliocene. Chile), after Dathc (1983, Zeitschrift fur geologische Wissenschaften 11): (A) skull, dorsal view; and (B) skull and mandibles, lateral view. (C) skull of an undescribed species of Balaenoptera (Pliocene, New Zealand), oblique dorsal view; specimen in Museum of New Zealand. Skeleton and ear bones o/Balaena mystieetus (extant, Arctic), from Van Bencden and Gervais (1868-1880). (D) lateral view of skeleton and (E) lateral (left), anterior (middle), and internal (right) views of ear bones, with pcriotic above and tympanic bulla below. (F) isolated tympanic bulla of PBalaena primigenia (Pleistocene?, Britain), internal view, from Van Bencden and Gervais.

Figure 5 Mysticete cetaceans. Skull and mandibles of Megaptera hubachi (Pliocene. Chile), after Dathc (1983, Zeitschrift fur geologische Wissenschaften 11): (A) skull, dorsal view; and (B) skull and mandibles, lateral view. (C) skull of an undescribed species of Balaenoptera (Pliocene, New Zealand), oblique dorsal view; specimen in Museum of New Zealand. Skeleton and ear bones o/Balaena mystieetus (extant, Arctic), from Van Bencden and Gervais (1868-1880). (D) lateral view of skeleton and (E) lateral (left), anterior (middle), and internal (right) views of ear bones, with pcriotic above and tympanic bulla below. (F) isolated tympanic bulla of PBalaena primigenia (Pleistocene?, Britain), internal view, from Van Bencden and Gervais.

Megaptera is rather divergent from Balaenoptera, this does not necessitate a much older split between these two groups. For Balaenoptera and close relatives, die oldest described fossils are also Late Miocene (~12 Ma), with less certain Middle Miocene records. There are many records of later Miocene, Pliocene, and Pleistocene Balaenoptera fossils (Fig. 5C).

The oldest fossil gray whale, from the Pleistocene (~0.5 Ma), gives no obvious clue to the origins of Eschrichtius robustus. Fossils do not support the notion of links between the gray whale and cetotheres: equally, they do not discount relationships with Balaenoptera. Geologically voung gray whale fossils indicate that these animals occurred in the North Atlantic.

Mysticetes diversified greatly in the Oligocene, before about 25 Ma, yet the earliest described right whale (Balaenidae) is the Earlv Miocene Morenocetus parvus (~20 Ma) from Patagonia. Morenocetus anchors the Balaenidae (represented by living Balaena and Eubalaena; Figs. 5D and 5E) as the oldest family of living mysticetes. Jaws are not known for Morcnocctus, but other aspects of the skull fit the balaenid pattern well. The later Miocene record of right whales is patchy, with only fragmentary fossils reported, but better Pliocene (2-5 Ma) and Pleistocene (<2 Ma) records include many nominal species (e.g., B. primigenia, Fig. 5F) and other published records of partial skeletons. A nearly complete skeleton of a large Early Pliocene (4-5 Ma) balaenid close to Balaena mysticetus is known from the Yorktown Formation, Atlantic Coastal Plain. There is no published fossil record of Neobalaenidae to indicate the origins of the pygmy right whale, Caperea marginata, from its presumed balaenid ancestors.

C. Odontoceti

Odontocetes are much more diverse in terms of taxa and structure than mysticetes. The oldest named species that are well dated are Late Oligocene (<30 Ma). Reportedly older species (Early Oligocene, ~30-34 Ma) are undescribed or of less certain age. The most archaic described odontocete is Archaeodelphis patrius, based on a fragmentary skull of uncertain origin and possible Oligocene age. Uniquely, this enigmatic species barely shows evidence of nasofacial muscles, which, in other odontocetes, are implicated in echolocation. Despite its reputedly ancestral position, because its skull base is somewhat specialized, Archaeodelphis is perhaps not directly on a lineage leading to living species. Almost as archaic is Xenorophus, a bizarre Late Oligocene genus containing one or two species with specialized facial structures superimposed on a rather primitive skull (Fig. 6A).

There is not yet consensus about the widely discussed Agorophiidae, a group previously linked with the squalodon-tids. Strictly, the family includes only the type species of Agorophius (Agorophius pygmaeus; Cooper Marl, Late Oligocene, >24 Ma, South Carolina), for which most of the holotype skull (Fig. 6B) is lost. Some paleocetologists prefer a grade family Agorophiidae, including Archaeodelphis, Xenorophus, and varied fragmentary archaic odontocetes. There is no evidence at present to regard the Agorophiidae as a basal clade of odontocetes.

Among living odontocetes, sperm whales (Physeter macrocephalus, Figure 7A-B; and Kogia spp.) represent a basal radiation, yet sperm whales have a poor early fossil record. The fragmentary Ferecetotherium kelloggi is reportedly Late Oligocene (23+ Ma; Maikop? Formation), providing the earliest record for a living family of odontocetes. Early Miocene sperm whales, such as Diaphorocetus and Idiorophus (Figs. 7D and 7E) from the South Atlantic, show the characteristic basined facial bones of later sperm whales, but retain a more primitive braincase, which links thein firmly with other odontocetes. Many named species of fossil sperm whales are based on isolated teeth (e.g., species of Scaldicetus; Fig. 7C), which, however, reveal little about the actual animal. The oldest Kogiidae are later Miocene species from the eastern Pacific, including Praekogia from Isla Cedros (Almejas Formation, ~6 Ma), and the large narrow-skulled Scaphokogia from the Pisco Formation of Peru.

Beaked whales, Ziphiidae, have a disappointing early record based on isolated and worn rostra (“beaks” or upper jaws) but few diagnostic skull remains. The oldest reported ziphiid is Squaloziphius emlongi (Clallam Formation, latest Oligocene or earliest Miocene, ~23 Ma, Washington), which, equally, may be a eurhinodelphinid. Other supposed ziphiids actually represent the platanistoid dolphin group Squalodelphinidae. Most of the described ziphiid fossils are resistant isolated rostra of Middle Miocene age or younger (<16 Ma). Many of these are named, mostly as species of Belemnoziphius, Choneziphius (Fig. 7E), and Mesoplodon, despite the problem of determining exactly what sort of ziphiid they are. Ziphiid beaks and ear bones (e.g., Figs. 7H-7J) are common cetacean fossils at unconformities and among deep sea dredgings. Perhaps the most unusual occurrence of a fossil cetacean is that of a large ziphiid in freshwater Miocene strata of Kenya.

Odontocete cetaceans. (A) Skull of Xenorophus sloani (Oligocene, South Carolina), dorsal view, after Kellogg (1923) and after Whitmore and Sanders (1977, Systematic zoology 25). (B) Skull and teeth of Agorophius pygmaeus (Oligocene, South Carolina), dorsal and lateral views of skull and detail of one tooth

Figure 6 Odontocete cetaceans. (A) Skull of Xenorophus sloani (Oligocene, South Carolina), dorsal view, after Kellogg (1923) and after Whitmore and Sanders (1977, Systematic zoology 25). (B) Skull and teeth of Agorophius pygmaeus (Oligocene, South Carolina), dorsal and lateral views of skull and detail of one tooth.

Odontocete cetaceans. Skeleton and ear bones o/Physeter macrocephalus (extant, cosmopolitan), from Van Beneden and Gervais: (A) lateral view of skeleton and (B) internal view of ear bones, periotic above and tympanic bulla below. (C) Tooth of Scaldicetus macgeei (Pliocene, Australia). Skull of Idiorophus patagonicus (Miocene, Patagonia), from Lydekker (1894), (D) anterior view and (E) lateral view. Incomplete skull of Choneziphius planirostris (Miocene-Pliocene, Belgium), from Van Beneden and Gervais. (F) dorsal view and (G) lateral view. Earbones of Mesoplodon bidens (extant, Atlantic), from Van Beneden and Gervais: (H) periotic, internal; (I) tympanic bulla, internal; and (J) tympanic bulla, dorsal.

Figure 7 Odontocete cetaceans. Skeleton and ear bones o/Physeter macrocephalus (extant, cosmopolitan), from Van Beneden and Gervais: (A) lateral view of skeleton and (B) internal view of ear bones, periotic above and tympanic bulla below. (C) Tooth of Scaldicetus macgeei (Pliocene, Australia). Skull of Idiorophus patagonicus (Miocene, Patagonia), from Lydekker (1894), (D) anterior view and (E) lateral view. Incomplete skull of Choneziphius planirostris (Miocene-Pliocene, Belgium), from Van Beneden and Gervais. (F) dorsal view and (G) lateral view. Earbones of Mesoplodon bidens (extant, Atlantic), from Van Beneden and Gervais: (H) periotic, internal; (I) tympanic bulla, internal; and (J) tympanic bulla, dorsal.

Living platanistids are known only from one species of Platanista from rivers in India and Pakistan, but fossil relatives are diverse. Some extinct long-beaked marine dolphins from the western North Atlantic, including Zarhachis (Early-Middle Miocene, —14-16+ Ma) and Pomatodelphis (Late Miocene, 6-9 Ma), have facial crests, which place them close to Platanista in the Platanistidae. The marine Allodelphis (Jewett Sand, Early Miocene, ~20 Ma; California) is a more archaic possible pla-tanistid that lacks facial crests. Clearly, platanistids invaded freshwater late in their history.

Related to the platanistids is the extinct and more archaic family Squalodelphinidae, which is based on the long-beaked Squalodelphis (Early Miocene, Mediterranean). Southern squalodelphinids include several species of Notocetus (Fig. 8A) from the latest Oligocene and Early Miocene ( — 18-24 Ma; New Zealand and Patagonia) bordering the Southern Ocean; formerly, some were identified wrongly as beaked whales.

A small heterodont dolphin with a slightly asymmetrical skull, Waipatia maerewhenua (Late Oligocene, ~25-26 Ma; New Zealand), typifies a family Waipatiidae. Waipatia (Figs. 8B and 8C) shows skull features that indicate an ability to echolo-cate. Other previously enigmatic odontocetes, such as Sulako-cetus and Microcetus from the Tethys-North Atlantic, may also be waipatiids. These animals probably he close to the base of the platanistoids.

Probably the best known of the platanistoids are the shark-toothed dolphins, Squalodontidae, which are geographically widespread medium to large odontocetes with long rostra and conspicuous, robust, triangular heterodont teeth (Figs. 8D and 8E). Squalodontids were long implicated in the phylogeny of living odontocetes, but recently have been included among the platanistoids. Squalodontids are based on Squalodon gratelupi (Early Miocene, ~20 Ma. eastern North Atlantic; Fig. 8E) and other clearly related species. Other supposed species of Squalodon, which were named from fragments, apparently represent other families of odontocetes or even mysticetes. Probably, the squalodontids include only Squalodon, Kelloggia, Eosqualodon, and Phoberodon. The broad-beaked Prosqualodon (latest Oligocene-Early Miocene, ~18-~24 Ma. marginal Southern Ocean) may represent a related separate family, Prosqualodontidae.

Odontocete cetaceans. (A) Skull and mandibles of Notocetus vanbenedeni (Miocene, Patagonia), from Lijdekker. Waipatia niaerewhenua (Oligocene, New Zealand): (B) skull, lateral view; and (C) teeth and ear bones. (D) Skull and mandibles of Squalodon-like squalodontid (Oligocene, New Zealand), lateral view. (E) Incomplete skull of Squalodon gratelupi (Miocene, France), lateral view, from Van Beneden and Gervais. (F) Skeleton of Eurhinodelphis cocheteuxi (Miocene, Belgium), lateral view, from Abel (1909).

Figure 8 Odontocete cetaceans. (A) Skull and mandibles of Notocetus vanbenedeni (Miocene, Patagonia), from Lijdekker. Waipatia niaerewhenua (Oligocene, New Zealand): (B) skull, lateral view; and (C) teeth and ear bones. (D) Skull and mandibles of Squalodon-like squalodontid (Oligocene, New Zealand), lateral view. (E) Incomplete skull of Squalodon gratelupi (Miocene, France), lateral view, from Van Beneden and Gervais. (F) Skeleton of Eurhinodelphis cocheteuxi (Miocene, Belgium), lateral view, from Abel (1909).

Species in the extinct family Eurhinodelphinidae (sometimes called Rhabdosteidae) have dramatically long rostra and quite specialized skulls. The early record is patchy, with Oligocene forms restricted and poorly known. This, and problems of interpreting skull bones, means that relationships are obscure; they could lie with delphinoids or, equally, platanistoids. Early and Middle Miocene eurhinodelphinids such as Eurhinodelphis (Fig. SF) and Agyrocetus are widespread. One Late Oligocene species occurs in freshwater strata of central Australia. Eurhinodelphids are related to another extinct group, the Early Miocene Eoplatanistidae.

Early true dolphins represent the grade family Kentriodon-tidae (Delphinoidea). Kentriodontids are geographically widespread small to medium-sized animals with largely symmetrical skulls, including Kentriodon (Fig. 9A) and Hadrodelphis (both Calvert Formation, Early to Middle Miocene, 14-16+ Ma). Rare Late Oligocene (>23 Ma) kentriodontids indicate ancient origins for the lineage leading to delphinids. The geologically youngest kentriodontids were contemporaneous with early delphinids from the Late Miocene (~10—12 Ma), animals characterized particularly by markedly asymmetrical skulls.

Odontocete cetaceans. (A) Skeleton of Kentriodon pemix (Miocene, Maryland), lateral vieu: of speculative reconstruction, from Kellogg (1927). (B) Skull p/Lomacetus gins-burgi (Miocene, Peru), dorsal view. (C) Periotie of Delphinapterus sp. (Pliocene, North Carolina), ventral view. (D) Skull of Albireo whistleri (Miocene, Isla Cedros), lateral view.

Figure 9 Odontocete cetaceans. (A) Skeleton of Kentriodon pemix (Miocene, Maryland), lateral vieu: of speculative reconstruction, from Kellogg (1927). (B) Skull p/Lomacetus gins-burgi (Miocene, Peru), dorsal view. (C) Periotie of Delphinapterus sp. (Pliocene, North Carolina), ventral view. (D) Skull of Albireo whistleri (Miocene, Isla Cedros), lateral view.

Like delphinids, porpoises (Phocoenidae) and white whales (Monodontidae) range back to the Late Miocene, although no clear ancestors have been identified. The record of porpoises is better than that of monodontids, originating with Salumipho-caena stocktoni from the eastern North Pacific (Monterey Formation, 10-11 Ma; California). Fossil ear bones and skulls indicate a high diversity for phocoenids in the latest Miocene and earlier Pliocene of the eastern Pacific (e.g., Lomacetus and Pis-colithax spp.; Fig. 9B). A more sparse record of monodontids reveals that this group was also important in warm equatorial waters of the eastern Pacific until well into the Pliocene, with fossils such as Denehola present in Peru and Cedros Island. In the Atlantic, Early Pliocene (~4.5 Ma) Delphinapterus occurs in North Carolina (Yorktown Formation; Fig. 9C), and there are reports of geologically young (Late Pleistocene, <0.5 Ma) Delphinapterus and Monodon from midlatitude North Atlantic shores.

While most of the Late Miocene and younger fossil delphi-noids are of modern aspect, two extinct families are known from the latest Miocene and earlier Pliocene of the eastern Pacific. The Albireonidae is based on the latest Miocene porpoise-like Albireo whistleri (Fig. 9D) from Cedros Island, whereas two species of tusked “walrus whales” (Odobenoce-tops; Odobenocetopsidae) are from the Pisco Formation, Peru, The latter are uncannily like the living walrus, Odobenus ros-marus, in skull form.

Several groups of river dolphin lie close to the Delphi-noidea. Among members of these groups, pontoporiid fossils occur in freshwater strata in South America and Mio-Pliocene marine rocks bordering the eastern North Pacific and western North Atlantic. The small short-beaked Brachydelphis mazeasi (Pisco Formation, ~ 12-15 Ma) from Peru represents a distinct subfamily (Pontoporiidae: Brachydelphininae) related to the long-beaked Peruvian fossil Pliopontos littoralis (~4-5 Ma) and the living franciscana (Pontoporia blainvillei). Marine long-beaked species of Parapontoporia from Isla Cedros (Almejas Formation, ~5 Ma) and southern California (San Diego Formation, ~2-3 Ma) may be related to Pontoporia, although ear bone anatomy points to relationships with living baiji (Lipotes vexillifer). The problematic Prolipotes, based on a fragment of possibly Miocene lower jaw from freshwater deposits in China, is too incomplete to confirm a relationship with Lipotes. Many fossils have been placed in the family Iniidae, with the living Inia geoffrensis, but most of these belong elsewhere. Fossil iniids include the South American Late Miocene (“Mesopotamiense” horizon, >5 Ma) Ischyrorhynchus and probably Saurodelphis and may include the marine Goniodel-phis hudsoni (Early Pliocene, 4-5 Ma) from Florida.

IV. Marine Carnivora

The record of marine Carnivora starts about the Oligocene/Miocene boundary (~25 Ma). Pinnipeds are most speciose and geographically widespread aquatic carnivores, well represented in northern temperate regions. Related closely is the fossil marine amphicyonid, Kolponomos. The only marine ursid, the polar bear, Ursus maritimus, has no significant fossils and probably originated only a few tens of thousands of years ago. Otters have a meager but notable fossil record.

A. Pinnipeds: Seals, Walruses, Fur Seals, and Sea Lions

As with Cetacea, the fossil record of seals, sea lions, walruses, and relatives has expanded greatly in recent decades, particularly through finds around the North Pacific. Fossil pinnipeds are less diverse than the Cetacea in terms of species, genera, and families, and the geological record is shorter, extending back to the late Oligocene (—25 Ma) (Fig. 10). Furthermore, the lack of consensus about family level taxonomy and indeed about pinniped monophyly versus diphyly complicates a review of these animals; a genus, or different species on one genus, may be placed in one of several different taxa, some of which are acknowledged grades, and others of which are clades of varying family level rank (e.g., family or subfamily).

Pinniped diphyly was favored strongly from the 1960s to 1980s (Fig. 10A). Of the two living groups and their fossil relatives, the eared seals—fur seals and sea lions—supposedly originated from a bear-like stock in the North Pacific. The walrus was placed tentatively with the eared seals to form the group Otarioidea (or Otariidae). In contrast, the earless or true seals (Phocidae, or Phocoidea) were thought to be a distinct and different group related to mustelids and originating in the North Atlantic. By the 1990s, some anatomically based cladis-tic analyses of pinnipeds concluded that seals and relatives are monophyletic, with the walrus and Phocidae being the most crown-ward taxa, and Otariidae more basal in the group Pinni-pedimorpha (Fig. 10B). Classification is still volatile, and communication is hampered because taxonomic names have quite different meanings to different workers. The cladistic methods seem clear and repeatable, supporting the notion of monophyly. Conversely, a good case has been made on morphological grounds that some key characters used in cladistics are ambiguous, arguably including significant reversals and convergences, undermining claims of monophyly. Further, the monophyletic pattern plotted against time (Fig. 10C, lower) produces long “ghost lineages,” whereas the diphyletic pattern reveals long-ranging paraphyletic groups (Fig. 10C, upper). Some recent molecular taxonomies have corroborated anatomical cladistics and recognized the living pinnipeds as monophyletic, but have supported “traditional” views in placing the walrus closer to otariids than to phocoids. There is not yet a consensus.

The small marine carnivore Enaliarctos mealsi is an archaic species based on skulls and a skeleton from the Jewett Sand of California (about Oligocene/Miocene boundary; ~23 Ma) (Fig. 11A). The species typifies the Enaliarctinae (or, alternatively, Enaliarctidae), is widely regarded as a basal otariid, and may lie at the stem of all pinnipeds. Other enaliarctines of comparable age have been named from fossils from California and Oregon (e.g., Pteronarctos goedertae, Astoria Formation, Middle Miocene, —16 Ma) (Fig. 11B), and an undetermined Ena-liarctos-like species occurs in Japan (late Early Miocene, —17 Ma). Also from California, and roughly contemporaneous with Enaliarctos, is Pinnarctidion bishopi, an archaic pinniped also initially placed in the Enaliarctinae. A second species, P. rayi, from coastal Oregon (Early Miocene, >19 Ma) includes skulls and postcranial bones, which, alternatively, have been allied with Phocidae.

Fur seals and sea lions, [Otariidae (or Otariinae)] appeared by the Late Miocene (>9 Ma) in the North Pacific, as shown by Thalassoleon and Pithanotaria (Fig. 11C). The two living otariid groups, the fur seals and sea lions, have a short fossil record around the Pacific and Southern Ocean. Fossil northern fur seals (arctocephalines) include the small Late Pliocene (2-3 Ma) Callorhinus gilmorei (Japan, California, Baja, California), which is close to the living North Pacific C. ursinus. A larger, southern fur seal is the Late Pliocene (—2 Ma) Hydrarctos lo-niasiensis from Peru, which is possibly allied to living fur seals in the genus Arctocephalus. Among sea lions (otariines), North Pacific fossil Eumetopias is known from teeth and some post-cranial bones of latest Pliocene age (—2.5 Ma, Japan), and there are reports of Pleistocene Zalophus. In the south, the extinct Neophoca palatina is based on a single skull of middle Pleistocene age (—0.5+ Ma, New Zealand).

The diverse cluster of medium to large extinct pinnipeds known as allodesmines is typified by the rare Allodesmus ker-nensis and A. kelloggi (Fig. 11D) and more common A. gracilis (Fig. 11E) from Sharktooth Hill, California (Middle Miocene, 13-14 Ma). Allodesmines range temporally from the early Middle to Late Miocene, and widely across the North Pacific (Japan, and Washington to Baja California, Mexico). At least eight species have been described in four genera (Allodesmus, Brachyallodesmus, Atopotarus, and Megagomphos). They were large-eyed, deep-diving animals that swam by forelimb propulsion. Allodesmus and its relatives have been placed variously (as a family or a subfamily) in the Otariidae or alternatively close to the true seals, Phocidae.

The single species of living walrus is the sole survivor of a once diverse group. Walrus origins may lie with the Early Miocene (16-22 Ma) genus Desmatophoca, which contains two large rare species from the Northeast Pacific (Fig. 11F). Alternatively, Desmatophoca has been allied with the phocid radiation. Another group implicated in walrus origins is the North Pacific Miocene Imagotariinae. Of these, the basal Neotherium was small, but most imagotariines were medium to large archaic sea lion-like forms. Middle Miocene ( — 13-16 Ma) fossils include species of Prototaria (Japan) and Proneotherium (Oregon); Imagotaria (California) is a representative Late Miocene form. Though on the walrus lineage, these were not particularly mollusk eaters but probably had a generalized piscivorous diet.

The North Pacific group Dusignathinae includes extinct forms regarded variously as walrus-like sea lions (“pseudo-walruses”) or, alternatively, true walrus relatives (with sister group status in Fig. 11C, upper and lower). These animals, including Dusignathus and the very large Gomphotaria (known from a nearly complete skeleton), range from Late Miocene (5-8 Ma) to Late Pliocene (2-3 Ma). Dusignathines differ from odobenines in that both the upper and the lower canines are enlarged as tusks.

Different hypotheses of relationships among pinnipeds, and geological age ranges of major groups of pinnipeds. (A) One concept of pinniped diphijly, based on Barnes and others. (B) One concept of pinniped monophyltj, based on Berta and others. (C) Geological age ranges of major groups of pin nipeds, plotted against a standa rd time scale (for time scale details, see Fig. 2). Hypotheses ofdiphyly (tipper) or monophyly (lower) are used to show inferred relationships. Bars show age ranges for taxa, which are mostly families and subfamilies but sometimes genera. Clades are in-filled bars, grades or paraphyletic or stem groups are open bars, and genera or key species within genera are dots.

Figure 10 Different hypotheses of relationships among pinnipeds, and geological age ranges of major groups of pinnipeds. (A) One concept of pinniped diphijly, based on Barnes and others. (B) One concept of pinniped monophyltj, based on Berta and others. (C) Geological age ranges of major groups of pin nipeds, plotted against a standa rd time scale (for time scale details, see Fig. 2). Hypotheses ofdiphyly (tipper) or monophyly (lower) are used to show inferred relationships. Bars show age ranges for taxa, which are mostly families and subfamilies but sometimes genera. Clades are in-filled bars, grades or paraphyletic or stem groups are open bars, and genera or key species within genera are dots.

Pinniped carnivores. (A) Skeleton of Enaliarctos mealsi (Oligocene/Miocene boundary, California), lateral view, after Berta et al. (B) Skull of Pteronarctos goedertae (Early Miocene, California), lateral view, after Barnes (1989, Contributions in science, Nat. Hist. Mus. of LA County 403). (C) Skeleton of Pithanotaria starri (Late Miocene, California), lateral view, after Kellogg (1925). (D) Skeleton of Allodesmus kelloggi (Middle Miocene, California), lateral view, after Mitchell. (E) Skull and mandibles of Allodesmus gracilis (Middle Miocene, California), lateral view, after Barnes and Hirota (1995, The Island Arc 3[4]). (F) Skull of Desmatophoca brachycephala (Early Miocene, California), lateral vietv, after Barnes. (G) Skull and mandibles of Protodobenus japonicus (Early Pliocene, Japan), lateral view, after Horikawa (1995, The Island Arc 3[4]).

Figure 11 Pinniped carnivores. (A) Skeleton of Enaliarctos mealsi (Oligocene/Miocene boundary, California), lateral view, after Berta et al. (B) Skull of Pteronarctos goedertae (Early Miocene, California), lateral view, after Barnes (1989, Contributions in science, Nat. Hist. Mus. of LA County 403). (C) Skeleton of Pithanotaria starri (Late Miocene, California), lateral view, after Kellogg (1925). (D) Skeleton of Allodesmus kelloggi (Middle Miocene, California), lateral view, after Mitchell. (E) Skull and mandibles of Allodesmus gracilis (Middle Miocene, California), lateral view, after Barnes and Hirota (1995, The Island Arc 3[4]). (F) Skull of Desmatophoca brachycephala (Early Miocene, California), lateral vietv, after Barnes. (G) Skull and mandibles of Protodobenus japonicus (Early Pliocene, Japan), lateral view, after Horikawa (1995, The Island Arc 3[4]).

Many extinct genera have been placed with the large, tusked living walrus, Odobenus rosmarus, in the Odobenidae. Late Quaternary fossils (<1 Ma) of walrus are known around the margins of the North Atlantic and North Pacific, with Odobenus sp. also reported from the Late Pliocene (2-2.7 Ma) of the North Pacific. Other extinct species of odobenine (large-tusked) walruses lived in northern waters outside polar regions, in situations known to be warmer than today. For example, the Late Pliocene Valenictus chulavistensis, which is related closely to Odobenus, is known from skulls and associated skeletons from southern California (San Diego Formation, 2-3 Ma), whereas teeth referred to the extinct Trichecodon huxleyi have been reported from the earlier Pliocene (Bone Valley Formation, 4-5 Ma) of Florida and Pleistocene (Red Crag, <2 Ma) of Britain, south of the inferred range of the living walrus. Protodobemis japonicus (Fig. 11G), from Japan, may be as old as 5 Ma and thus is the oldest reported odobenine walrus.

In contrast to the various classifications of otariids and odobenids, true seals (Phocidae) are well unified by specialized swimming features. This unity might mislead one into thinking that phocid paleontology and phylogeny are straightforward. The fossil record of phocids is patchy, and few basal or archaic taxa are known. The oldest reported specimens, from South Carolina, are, surprisingly, from the Late Oligocene (>23 Ma). If the concept of pinniped monophyly is correct, with the Pacific Oligo/Miocene Enaliarctos at the stem of pinnipeds and phocids much further crown-ward, this early Atlantic record for phocids has major implications for the geography and timing of pinniped evolution. Phocids are a negligible component of North Pacific fossil faunas (with sporadic Phoca-like fossils from the Pleistocene and, rarely, Pliocene), but are significant in the Atlantic and. lesserlv, the Southern Ocean.

The typical phocid genus Phoca has been used widely and often wrongly, for diverse fossil seals and even some fossil cetaceans, but Phoca in the strict sense is not certainly older than Pleistocene. Phoca and its living relatives comprise a group Phocini (or Phocinae), which also includes fossils as old as Middle Miocene (14-16+ Ma). For example, the Middle Miocene presumed phocine Leptophoca lenis, which was based on an isolated humerus, is known from many skeletal elements from the Calvert Formation (~15 Ma; USA). Leptophoca, Prophoca, Cn/ptophoca, and related genera are all from the Atlantic shores of North America and Europe or from the Paratethys, which was an ancient shallow and even brackish northeastern extension of the Mediterranean Sea.

Fossils from a second group of phocid seals, the Monachi-nae, include a scatter of Monachus-like genera based on fragmentary North Atlantic fossils, such as “Monotherium” from the Calvert Formation (~15 Ma) related to living Monk seals (Monachini of some authors). Other excellent Late Miocene to Early Pliocene specimens are allied to living Lobodon car-cinophaga (Fig. 12A) in the Lobodontini. Lobodontines include Piscophoca pacifica and the elongate, slender-bodied Acrophoca longirostris (both Pisco Formation, Peru, eastern Pacific) and, in the Atlantic, the Late Miocene Properiptychus argentinus (Argentina). From the margins of the Southern Ocean come the well-known Homiphoca capensis (earliest Pliocene. ~5 Ma) (Figs. 12A and 12B) from South Africa and more fragmentary Homiphoca-like fossils from southern Australia. The only notable fossil of a living Antarctic lobodontine seal is a jaw identified as Ommatophoca rossi from New Zealand (latest Pliocene, >2 Ma; southwest Pacific) (Fig. 12D).

B. Amphicyonidae

The extinct Early Miocene carnivore Kolponomos (Amphicyonidae) is known from two North Pacific species from the Clallam and Nye Formations of Oregon and Washington. The skull is massive, with forward-directed eyes, a narrow snout, and broad low-crowned teeth, which indicate a crushing feeding habit, probably on shelled invertebrates (Fig. 13A). Other than the skull, the skeleton is known poorly, but was presumably bear like. Kolponomos is interpreted as amphibious, not a strong swimmer, and living nearshore. It has been placed close to the base of the pinnipeds.

Pinniped carnivores. (A) Skull and mandibles of Lobodon earcinophaga (extant, Antarctica), lateral view, from Gray (1846). Homiphoca capensis (earliest Pliocene, South Africa): (B) skull and mandibles of lateral view, after de Muizon and Hendey (1980, Annals of the South African Museum 82[3]) and (C) right temporal bone (ear region ), ventral vieio. (D) Mandible of Ommatophoca rossii (latest Pliocene, Neiv Zealand), lateral view.

Figure 12 Pinniped carnivores. (A) Skull and mandibles of Lobodon earcinophaga (extant, Antarctica), lateral view, from Gray (1846). Homiphoca capensis (earliest Pliocene, South Africa): (B) skull and mandibles of lateral view, after de Muizon and Hendey (1980, Annals of the South African Museum 82[3]) and (C) right temporal bone (ear region ), ventral vieio. (D) Mandible of Ommatophoca rossii (latest Pliocene, Neiv Zealand), lateral view.  

C. Sea Otters (Mustelidae: Lutrinae)

Sea otters are known as fossils from around the North Pacific and North Atlantic. Despite their presence in marine strata, none of the species was clearly an obligate marine mammal. A key diagnostic feature is blunt molar teeth, which indicate a durophagous (shell-crushing) diet. The giant Late Miocene (~7 Ma) otter, Enhydritherium terraenovae, is known from fragmentary specimens from marine strata in California and Florida and from an informative articulated skeleton from freshwater strata in Florida (Fig. 13B). The latter represents an animal with powerful neck muscles, forelimbs that probably helped in swimming, and hindlegs suitable for terrestrial locomotion. In younger fossils, such as Enhydra macrodonta (Late Pleistocene, California), limb structure indicates hindliinb swimming, with forelimbs used to manipulate food.

V. Tethytheres: Sirenians and Desmostylians

Tethytheres are herbivorous, mostly large mammals that include elephants, sirenians, and the extinct Desmostylia. Sirenians and desmostylians have a significant marine fossil record.

A. Sirenia

Living sirenians are obligate aquatic mammals that dwell in shallow subtropical and tropical waters, both marine and freshwater. The fossil record, which is mostly from the Northern Hemisphere (Fig. 14), shows that the two living families (Trichechidae and Dugongidae) have a long and diverse history. Two extinct families of sirenians are based on fossils. The fossil record of sirenians, as documented extensively by Domning, is judged reasonably complete and probably gives a reliable guide to the evolution of the group.

The oldest and most archaic sirenian is Frorastomus sirenoides (sole member of the Prorastomidae), named by the famous anatomist Richard Owen for a skull, mandible, and vertebra from the Early and Middle Eocene (~50 Ma) of Jamaica. A few other specimens are known. This occurrence in the western tropical Atlantic is surprising, as tethytheres are thought to have evolved in the eastern Tethys. Frorastomus is generally intermediate in structure between other tethytheres and later Sirenia, although perhaps it is not directly ancestral to any known later sirenian. Its notable sirenian features include the inflated rostrum, pachyostotic skull (with dense bone), retracted enlarged nares, and five premolars.

Before the end of the Eocene, sirenians occupied warm waters from the western tropical Atlantic through the Tethys to the western Pacific. Many of the fossils represent species of Proto-siren (extinct family Protosirenidae; Middle-Late Eocene), distinguished from Frorastomus by details of the ear and a more down-turned rostrum. Protosiren occurs in Egypt and India and presumably ranged through the middle Tethys. Some specimens retained short hindlimbs and were probably amphibious rather than completely aquatic. Protosirenids are plausible structural ancestors for the two extant families: Dugongidae and Trichechidae. Despite the abundance of protosirenid fossils and the presence of effectively a circumglobal equatorial seaway (Tethys-Caribbean-Pacific), there are no reports of Eocene or Early Oligocene sirenians from outside Indonesian west Pacific.

The earliest dugongids—the halitheriines (paraphyletic subfamily Halitheriinae; Eocene-Pliocene)—were contemporaneous with protosirenids, and overlap in range, being known from Middle Eocene to Oligocene rocks in Egypt (e.g., Eosiren, Fig. 15A). Dugongids are characterized by a loss of hindlimbs and by changes in the ear region of the skull, but relationships among archaic forms are not fully resolved. Fossils indicate that, apart from the hydrodamalines (see later), the family is subtropical-tropical.

In the Oligocene, halitheriine dugongids occurred beyond the shrinking Tethys sea, with significant fossils of Halitherium in Europe and the western North Atlantic. Halitherium (Germany: Fig. 15B) includes some of the very few Early Oligocene marine mammals known. Geologically, younger halitheriines include species in the widespread Early-Middle Miocene Metaxijtherium (Fig. 15C), characterized by strongly down-turned snouts and small upper incisor tusks, and interpreted as generalized bottom feeders. Metaxijtherium is reported from Europe (eastern Atlantic/Mediterranean), the western Atlantic, and the North to equatorial Pacific. Presumably, Pacific halitheriines, with their significant Miocene record, originated from animals that spread westward through the Central American seaway. One halitheri-ine, Metaxijtherium crataegense, is reported both from the western Atlantic and the eastern tropical Pacific. Older Pacific records of dugongids are fragmentary and enigmatic; thev include an indeterminate Late Oligocene dugongid from Japan and an unnamed Early Miocene (18-25 Ma) halitheriine from Oregon.

Other iruirine carnivores. (A) Skull and mandibles of Kolponomos new-portensis (Early Miocene, Oregon), lateral view, after Tedford et al. (1994, Proc. San Diego Mus. Nat. Hist. 29). (B) Skull and mandibles of Enhydritherium terraenovae (Late Miocene, Florida), lateral view, after Lambert (1997).

Figure 13 Other iruirine carnivores. (A) Skull and mandibles of Kolponomos new-portensis (Early Miocene, Oregon), lateral view, after Tedford et al. (1994, Proc. San Diego Mus. Nat. Hist. 29). (B) Skull and mandibles of Enhydritherium terraenovae (Late Miocene, Florida), lateral view, after Lambert (1997).

Geological age ranges of families and subfamilies within the Sirenia plotted against a standard time scale (for time scale details, see Fig. 2). Clades are in-filled bars, and grades or paraphijletic or stem groups are open bars. Inferred relationships follow the work of Do inning (1996).

Figure 14 Geological age ranges of families and subfamilies within the Sirenia plotted against a standard time scale (for time scale details, see Fig. 2). Clades are in-filled bars, and grades or paraphijletic or stem groups are open bars. Inferred relationships follow the work of Do inning (1996).

In the Early to Middle Miocene, a significant new lineage of dugongids, the hydrodamalines, radiated in the North Pacific. This group, characterized primarily by a large body size, probably arose within Metaxijtherium. Early fossils include two species of Du-sisiren, wliile the cold-tolerant Hydrodamalis is known from the circum-North Pacific Late Miocene to Pliocene H. cuestae and its descendant, Hydrodamalis gigas. The latter, Steller’s sea cow, was exterminated by humans; it was a huge animal that, unlike other sirenians, lived in exposed cold high-latitude waters where it fed on kelp. Steller’s sea cow has significant Quaternary records.

The dugong lineage arose by the Late Oligocene, when the Dugonginae branched away from the halitheriines. The oldest and most archaic dugongines (Crenatosiren) and greatest known diversity are in the western North Atlantic-Caribbean region, and perhaps the group originated there. Dugongines are, however, known as fossils from the eastern North

Atlantic-Mediterranean (Rytiodus; Early Miocene) and the eastern North Pacific (Dioplotheiium; Early-Middle Miocene). Beyond a few fragments of Pleistocene age, there are no notable fossils for the living western tropical Pacific Dugong dugon; however, phylogenetic analyses indicate a significant long but unrecorded history for the lineage leading to this species.

Manatees (Trichechidae), characterized by a secondarily reduced rostrum, probably evolved from Prototherium-Uke dugongs in the later Eocene or Early Oligocene. The oldest trichechids are the European Miosireninae-the Late Oligocene Anomotherium and Early Miocene Miosiren. By the Middle Miocene (~14 Ma), archaic manatees, species of Potamosiren (Trichechinae), occupied fresh waters in South America and since then have been restricted mainly to inshore waters. The Late Miocene to Pliocene (~4-6 Ma) Ribodon, from the lower La Plata basin, shows the distinctive manatee-like character of supernumerary molar teeth, interpreted as an adaptation for feeding on abrasive freshwater grasses. In turn, aquatic plants perhaps evolved in response to changing South American drainages caused by the uplift of the Andes. An Early Pliocene (~5 Ma) Ribodon is known from North Carolina at about the time that dugongs disappeared from the Caribbean-western North Atlantic, and fossils close to Trichechus appear in the Late Pliocene (2-3 Ma). Such fossils indicate that manatees ecologically replaced dugongs in the Caribbean-western North Atlantic, although whether by competition is uncertain. There is no fossil record to reveal the dispersal of Trichechus eastward to Africa.

B. Desmostylia

Desmostylians are large extinct North Pacific marine mammals that were first described, in the 1880s, on the basis of isolated molar teeth with distinctive bundles of high columnar cusps (Desmostylus; Fig. 15F). In 1953, they were placed in their own group, which is the only extinct order of marine mammals. Currently they are regarded as tethytheres allied with, but distinct from, Sirenia. Desmostylians have been interpreted as amphibious, hippopotamus-like, shallow-water mammals that fed on algae and seagrasses. They are known from subtropical to temperate coasts of North America and Japan, but did not range into the South Pacific.

Sirenians and desmostylians. (A) Skull of Eosiren libyca (Eocene, Egypt), lateral view, from Andrews (1906). (B) Skeleton of Halitherium schinzii (Oligocene, Germany), lateral view, from Romer (1945). (C) Skull of Metaxytherium arctodites (Miocene, Baja California Sur), lateral view, after Aranda-Manteca et al. (1994, Proc. San Diego Mus. Nat. Hist. 29). (D) Field site showing excavation of holoti/pe of Metaxytherium arctodites (Miocene, Baja California Stir). Desmostylus hesperus (Middle Miocene, Japan): (E) skull, oblique lateral view; (F) cheek tooth, oblique occlusal view; and (G) skeleton, lateral view, after Inuzuka et al. (1995, The Island Arc 3[4J).

Figure 15 Sirenians and desmostylians. (A) Skull of Eosiren libyca (Eocene, Egypt), lateral view, from Andrews (1906). (B) Skeleton of Halitherium schinzii (Oligocene, Germany), lateral view, from Romer (1945). (C) Skull of Metaxytherium arctodites (Miocene, Baja California Sur), lateral view, after Aranda-Manteca et al. (1994, Proc. San Diego Mus. Nat. Hist. 29). (D) Field site showing excavation of holoti/pe of Metaxytherium arctodites (Miocene, Baja California Stir). Desmostylus hesperus (Middle Miocene, Japan): (E) skull, oblique lateral view; (F) cheek tooth, oblique occlusal view; and (G) skeleton, lateral view, after Inuzuka et al. (1995, The Island Arc 3[4J).

The most archaic desmostylian is Behemotops, with a single species (B. proteus) based on mandibles and teeth of Late Oligocene age (~25-29 Ma) from Washington. Behemotops also occurs in Hokkaido, Japan. Tooth form in Behemotops is reminiscent of a group of terrestrial Asian tethytheres, the archaic proboscidians called Anthracobunidae.

Among the four other recognized genera, Cormvallius (Late Oligocene) and Vanderhoofius (Middle Miocene) are known only from western North America. Both species of Desmostylus (Early to Middle Miocene) (Figs. 15E-15G), which are distinguished on the basis of teeth and skulls, occur on both the eastern and the western coasts of the Pacific. Similarly. Pale-oparadoxia (Early to Middle Miocene), with three to four species, also occurs on both Pacific coasts. Desmostylians have not, however, been reported from the Southern Hemisphere.

Unifying features for desmostylians include a skull with a long broad muzzle and prominent tusk, dorsally protruding eyes, a shortened neck, a broad sternum, and many details of the postcranial skeleton. The matter of body stance is contentious; it has been reconstructed, using skeletal form and fossilized posture, as reptilelike, with the limbs extended laterally. Alternatively, stance has been interpreted as the familiar quadrupedal form of mammals, with the body well off the ground and limbs more or less under the body. Irrespective of stance, desmostylians were probably large and slow moving on land and swam much like polar bears.

C. Edentata

One of the most unusual fossil occurrences is that of abundant ground sloths from marine Pliocene strata (~4 Ma) of Peru. The extinct marine sloth, Thalassocnus nutans, is from a rich assemblage of marine vertebrates. Although the sloth shows 110 clearly obligate aquatic features, its structure would allow swimming comparable to that of otters. Because die adjacent Peruvian coast was a desert, die sloth probably ate seaweeds or seagrasses.

Next post:

Previous post: