Origins of Marine Mammals

 

Whales, pinnipeds, and sirenians are examples of a major evolutionary phenomenon: the dramatic transformation of terrestrial amniotes (reptiles, birds, and mammals) to a more or less fully aquatic lifestyle. Indeed, marine mammals are far from unique, for this has happened several tens of times in evolutionary history. The whales are ecological replacements for the extinct ichthyosaurs, which were reptiles as fish-like as whales. The sea lions have a body form and swimming method broadly similar to those of the penguins, the extinct great auk (Pingtiimis impennis) and plesiosaurs. Other examples include the plotopterid birds, mosasaurs (aquatic lizards), at least two lineages of marine crocodilians, and the mollusc-eating placodont reptiles diat were analogous to the modern sea otter (Enhijdra lutris). Yet, however many times it has happened, it is still a huge transition from a terrestrial animal, adapted to live, feed, and breed on land, to an animal such as a whale, which is so notoriously fish-like that its relationships puzzle nonbiologists. Indeed, how whales could have arisen from land mammals was a question routinely posed by opponents of Darwin’s “Origin of Species,” and he himself regarded it as a crucial issue.

The problem can be approached today by asking what phylogenetic route the transition took: what are the actual relationships? Then what can we say about the animals that were the actual intermediates? What did they look like and in what habitats did they live? What was, in fact, the ecological route of the transition? And, in parallel, does the evidence make functional sense that corroborates the hypothesis? This article focuses on general issues common to all lineages.

I. How the Transition Can Be Studied

It is a central tenet of Darwinian evolutionary theory that relatively gradual changes can lead to major transitions. Under natural selection, a species can gradually change if this is advantageous (more correctly, advantageous through increasing the frequency of the individual gene alleles under selection). The degree of adaptation need not be perfect. An animal only has to be sufficiently well adapted to gain some advantage; even quite a poor degree of adaptation can be enough. For a land animal, the occasional ability to swim, however poorly, can be critical in escaping predators or locating new food resources. Furthermore, adaptation is “”fuzzy”: animals have some flexibility in how they use their adaptations, allowing natural selection to start driving change in either direction. There is almost no such thing as a purely land mammal, for most mammals can at least swim in an emergency. This provides the potential basis for the evolution of a more aquatic lifestyle.

The transition from a land animal to, say, a whale will have taken place by a relatively gradual process widi many intermediate forms. For this transition from land to sea to be possible, natural selection should be able to act along the whole range of the spectrum, and each intermediate form across the spectrum should be reasonably well adapted. In other words, there should be no insurmountable adaptive barrier in the way. (The available evidence on the origin of marine mammals is not sufficiently fine-grained for worthwhile discussion of the fine detail of this transition—for instance, whether morphological change was truly gradual or was concentrated in bursts during the formation of new species, as the hypothesis of punctuated equilibria predicts.)

The first line of evidence for demonstrating the viability of a transition is direct, from actual fossils of the transitional forms (or more probably close relatives). They may not (yet) be available, but are enormously valuable when they are. Early whales are an excellent example.

The second line of evidence for the viability of a transition is the existence of a spectrum of biological analogues, unrelated species that demonstrate a complete range of adaptations from fully terrestrial to fully marine. Their survival is evidence that their various lifestyles, and therefore the stages in the transition. are viable.

The third type of evidence is to some extent an independent test. It comprises theoretical biomechanical analyses that confirm that particular hypotheses of evolutionary transitions are functionally possible. Conversely, a hypothesis that required a functionally implausible intermediate form would thereby be considerably weakened.

II. Environment of the Transition

What resources drove natural selection during the land-sea transition? The most likely one is food. Less obvious but also very important are shelter from predators and shelter from extremes of temperature. Different resources occur in different environments; e.g., seaweed occurs on the shore and in seawater, but basking in the sun may only be practical on land. It is likely that the transition was driven by the move between environments in order to exploit resources.

The fact that the transition is driven by nondirectional natural selection has the major implication that selection and therefore evolution can in principle simply halt, or reverse direction, for greater or lesser periods, within any one lineage. This has undoubtedly happened. Indeed, certain forms are truly amphibious, relying on a highly specific combination of land and water resources to enable a lifestyle that is otherwise impossible. For instance, the hippopotamus feeds on land at night, but rests in the water during the day, whereas the river otter (Lutra Intra) forages in water but comes out to warm up, rest, and breed. To speak of these as being stuck at an intermediate stage of the evolution of aquatic forms risks missing the truth, which is that such amphibious forms are highly adapted in their own right. Crocodilians likewise evolved a highly efficient amphibious freshwater and estuarine lifestyle, which remains the main crocodilian way of life. However, some crocodilian lineages became fully marine and others reverted to a fully terrestrial lifestyle. Suggestions that some modem terrestrial groups are secondarily terrestrial, derived from amphibious or aquatic forms, are therefore entirely reasonable in principle and can only be assessed on actual evidence. Such suggestions have been made for snakes, proboscideans, and human beings, for instance.

What was the environment of the transition from land to sea? There are various possible environments where land, fresh water, and sea variously meet:

1. Freshwater river

2. Lake or other body of standing fresh water

3. Estuary with at least some saltwater influence

4. Sandy open seashore

5. Rocky open seashore

6. Unusual habitats, especially highly productive ecosystems rich in food; modem examples include mangrove swamps, the Florida Everglades, and the regularly flooded lowland riverine forests of the Amazon Basin. It must not be forgotten that, quite apart from the effects of human activity today, ecosystems in the past may have been very different, even without allowing for the “wild card” of unpredictably unusual habitats without any modern analogue.

These different environments pose varying advantages and problems to animals that exploit them. Freshwater rivers and lake shores, however, offer good opportunities for amphibious forms such as hippopotamuses, which rest in the water but feed on land at night. However, because of the risk of stranding during occasional droughts, only the largest rivers are suitable for fully aquatic animals such as odontocetes. Conversely, rocky shores offer plenty of fish and cephalopods to animals such as river otters, and molluscs and sea urchins to animals such as sea otters. However, life in salt water poses problems of salt excretion and thermoregulation. River otters lack blubber and use air-filled fur for insulation. They risk death if their coats are not washed in fresh water, as the salt water mats their coats when it dries. They are therefore dependent on the availability of standing pools of fresh water on shore.

An example of a hypothesis of simple transition from land to sea through a single habitat, rocky seashores, can be formulated using modern analogues:

1. Primarily terrestrial carnivore/carrion feeder, not normally a swimmer, but foraging on the shore above water, e.g., stoat (Mustela erminea).

2. Feeding facultatively on rocky shores above water and in rock pools, and possibly diving for food. Some swimming ability, but not much more than most terrestrial mammals, e.g., American mink (Mustela vison).

3. Shallow-water swimmer, foraging routinely in inshore water, for example, benthic invertebrates, remaining on land otherwise (e.g., for resting, thermoregulation, and breeding). Swimming by forceful kicks of webbed hind feet linked to flexion and extension of back, supplemented by dorsoventral undulation of a flattened tail, e.g., river otter (pursuit predator) or the archaic whale Ambulocetus (probably an ambush predator).

4. Shallow-water swimmer and diver, remaining in the water for much of the time and specializing in aquatic food. Swimming somewhat more specialized for aquatic movement, notably the development of bone ballast in buoyancy control, which slows the animal, e.g., the sea otter.

5. Fully specialized swimmer, feeding in open water but perhaps resting and breeding on beaches inaccessible to land predators, with increased development of a caudal fin to provide most or all of the thrust, e.g., archaic protocetid whales such as Rodhocetus and the living gray seal (Hali-choerus grypus) (in which the hind feet are modified to the functional equivalents of caudal fins).

6. Fully aquatic, not coming to land at all, e.g., short-beaked common dolphin (Delphinus delphis).

This testable hypothesis is given as an illustrative example based solely on carnivorous mammals on rocky temperate coastlines, and alternatives can be developed for other groups and other environments. Nevertheless, note that this hypothesis contains two critical elements:

• Ecological plausibility—these are stages in a gradual spectrum, exploiting resources as they become newly accessible

• Functional plausibility—again, these are stages in a gradual spectrum, with a clear explanation for how the development of caudal fin propulsion might occur, which can be reinforced by biomechanical analyses

Second, there may be no one single environment for the transition. Because the transition may take place in stages, natural selection can operate in successively different ways depending on the environment of the time. This has major implications, as discussed later.

It is also possible, and even likely, that the transition from land to water moved through a series of habitats, not just one. This may be important in easing the direct transition from land to sea. For instance, the rocky shore model given earlier does not offer any obvious answers to the problem of the origin of sirenians, which are marine herbivores. Individual red deer (Cervtts elaphus) facultatively eat intertidal marine algae or “seaweed,” and interestingly the North Ronaldsay breed of domestic sheep, which has long been confined to the shore of North Ronaldsay in the Orkney Islands, is now so adapted to a diet containing at least some algae that it cannot survive on grass alone. However, there are no living specialized mammalian grazers on algae comparable to (say) the Galapagos marine iguana (Amblyrhynchus cristatus). This might be explained on two grounds: (1) that the transition did go through a rocky shore phase, and it just happens that no analogues are alive today, in which case the extinct Pliocene aquatic sloth Thalas-socntis may be one such analogue; or (2) that the transition took place somewhere else, presumably in fresh water or estuaries because sandy beaches are devoid of edible plants. This leads to another possible spectrum, again with some modern analogues, but this time moving through geographical as well as niche space:

1. Riverine animal, walking (but not normally swimming) in water and feeding on land, e.g., hippopotamus, or land animal wading in fresh water to feed on aquatic plants.

2. More aquatic, swimming, feeding on freshwater plants in lakes and rivers.

3. Aquatic, swimming form, feeding in estuaries.

4. Fully aquatic, feeding in estuaries and nearshore, e.g., manatees (Tricheclws spp.).

5. Fully aquatic, feeding largely or wholly on marine plants, e.g., dugong (Dugong dugon) and Steller’s sea cow (Hydro-damalis gigas).

Note that this offers one explanation of how land animals can survive in the salt-rich “desert” of the sea, by way of intermediate forms able to cope with increasingly brackish water (although alternatively a marine mammal might be derived from desert-living land mammals, which are already adapted to retaining body water, e.g., by excreting urine much saltier than the blood).

The environment, or environments, of transition also bias the fossil record. Even if an animal dies, escapes scavenging, and is fossilized, it is not necessarily buried where it lived in life. For instance, water currents might have moved the carcass. To summarize a complex situation, corpses in estuaries and rivers are apt to be swept out to sea. Carcasses of animals that lived on rocky coasts, which are being eroded, tend not to be preserved in situ as there is no sediment to bury them in. Carcasses on sandy coasts also tend not to be preserved because these coasts are relatively high-energy environments that disrupt the corpse. If the carcasses are buried, this may be because they have drifted offshore and sunk to the sea floor, or possibly because open sea animals have been stranded on the shore and preserved there. Only animals in rivers and lagoons have much chance of being preserved where they lived, and even then riverine animals are quite likely to be washed away downriver. These biases are borne in mind by competent paleontologists and can be partly mitigated by functional analysis (which would reveal anomalies such as an amphibious animal in an open-sea deposit).

III. Functional Issues

The functional problems involved in the transition from land to water, such as swimming, feeding, and sensory physiology, are mostly discussed elsewhere in other entries. Here is a case study. focusing on buoyancy control, and showing how this one function interacts with several others, such as swimming and feeding.

Land mammals breathe air and may also rely on air-filled fur for heat insulation, especially in temperate and polar regions. This air makes swimming and diving difficult if the buoyancy provided by air in the lungs (and any fur) is not precisely compensated by the denser parts of the body such as bone. This wastes at least part of the swimming effort, which goes to staying up or down in the water rather than thrusting the animal forward. Animals well adapted to aquatic life have proper adaptations to buoyancy control separate from swimming, which can thus be optimized for forward movement. However, locomotion interacts with respiration here. Air in large lungs is useful as an oxygen store, but only for animals that do not swim fast (because the added bulk slows them down) or dive deeply (because larger lungs compress more with depth, making buoyancy changes even greater, and because of the risk of the “bends” if high-pressure air in the lungs is allowed to remain in respiratory exchange with the bloodstream). To compensate for the air in their lungs, marine mammals may have stones (gastroliths) in the stomach or massively enlarged bony skeletons for ballast. Alternatively, they may reduce the size of their lungs and replace air-filled fur with naked skin underlain by an insulating layer of subcutaneous fat or blubber. In practice, of course, all functions and organs are interdependent. For instance, sirenians can only tolerate their large lungs and correspondingly massive skeletons because they feed on plants that do not swim away, whereas odontocetes have to be much faster swimmers to catch their prey, and their smaller lungs are coordinated with the evolution of much lighter skeletons. These changes appear to occur in predictable patterns in the evolution of marine mammals and their reptilian and avian analogues:

1. Land animal, relatively buoyant in water, e.g., stoat.

2. Amphibious animal with little or no bone ballast, because it slows it too much on land, but remaining too buoyant in water for efficient locomotion, e.g., American mink, river otter, polar bear (Ursus maritimus), desmostylians, Tlialassocnus (sloth), Ambidocetus (primitive whale), and Galapagos marine iguana. Tied to living on land (or ice) for most of the time.

3. Primarily aquatic form with bone ballast in the form of larger, or denser, bony skeletons. Swimming can be decoupled from buoyancy control and made more efficient, e.g., sea otter and some primitive whales. Typical of animals living close inshore. Nile crocodile (Crocodylus niloticus) uses gastroliths but is not quite so fully aquatic.

4. Further evolution into more specialized fully aquatic forms, three different groupings being identifiable depending on locomotor and feeding adaptations, which drive selection in different ways, even reversing it:

a. Slow-swimming predators (or herbivores) on sessile food such as plants or benthic invertebrates: full development of bone ballast and large lungs that act as air stores and give rapid loss of buoyancy during a dive as the increasing water pressure compresses the lungs, enabling the animals to become negatively buoyant and walk on the bottom, e.g., placodonts (extinct reptiles feeding on hard-shelled benthic invertebrates) and sirenians. This is an essentially shallow water adaptation, linked to inshore and coastal waters.

b. Predators on mobile food—an apparently intermediate body plan of underwater fliers with hydrofoil-shaped swimming limbs, using gastroliths and/or bone ballast for buoyancy control: plesiosaurs, penguins, and sea lions. Also permitting open sea lifestyles, but (in penguins and sea lions) allowing them to retain hindlimbs relatively unmodified for use when breeding on land (it is not clear what plesiosaurs did, as their two pairs of hydrofoil “wings” are essentially identical). Also permits open sea lifestyles, although these animals are for some reason more coastal in lifestyle and distribution than the cetaceans.

c. Predators on mobile food—fast swimming with caudal fins: selection for improved speed and acceleration leads to reversal of selection to give extensively lightened skeletons and smaller lungs (which mean less change of net buoyancy with depth, making deep diving more efficient); e.g., ichthyosaurs and cetaceans. Permits open sea lifestyles.

IV. The Problem of the Intermediate

The problems of functional change are not a simple matter of moving from terrestrial to aquatic adaptations. “Amphibious” forms have to tolerate an intermediate level of adaptation, develop a doubly functional anatomy (which may itself involve a level of compromise), or remain poorly adapted in one of the two environments. This is not in itself a problem if the benefits compensate for the disadvantages, as it is better to be an inefficient swimmer (in terms of energy cost of locomotion) with access to abundant aquatic food than to starve on land.

However, a net increase in energy intake is not sufficient. Animal species do not normally exist in isolation, but have to compete with one another for resources. It is not clear how intermediate forms can survive at all under these circumstances. At any one time and place an intermediate form will always risk being outcompeted by a land specialist or by a sea specialist, if not both. For instance, an intermediate form will be a slower and less energy-efficient swimmer than a specialist marine form, and a slower and less efficient runner than a specialist land form. River otters are not quite as nimble as comparable land mustelids, and sea otters are more aquatic but still clumsier on land. The dilemma can only be resolved if, as seems likely, intermediate forms rely on “refuges” free of competition. These refuges might be real physical ones, in space and time, or virtual ones, in ecological “niche space.” Possible refuges and examples of animals exploiting them include:

1. Ecological isolation in an area inaccessible to land or water predators and competitors: possibly sea otters in rocky shallow water.

2. Specialization in a dietary resource not exploited by similar animals: sea otters feeding on hard-shelled shallow water invertebrates.

3. Time averaging of predator avoidance, e.g., getting into the water when a land carnivore is around, but getting out when a shark arrives. The risk that both will be around at the same time so that escape becomes impossible is a multiple of the two separate probabilities that each will be around, and is therefore smaller than either.

4. Time averaging of food or other resource, e.g., hippopotamuses feeding on land at night, but retreating to water during the day.

5. Combining complementary resources not available to pure land or pure water specialists, e.g., using land to bask in the sun to gather heat, but wading and diving in cold water to feed on algae, as adult Galapagos marine iguanas do (perhaps only relevant to those mammals that are small enough and live in cold enough sea water to have problems with thermoregulation, notably river otters on northern temperate seacoasts).

6. Geographical isolation to escape land predators, perhaps on oceanic islands, e.g., Galapagos marine iguana.

7. Isolation in evolutionary time. Modern marine mammals evolved as replacements for the Mesozoic marine reptiles, the last of which went extinct in the mass extinction at the end of the Cretaceous. The mammals were thus sheltered from competition by these direct analogues in a way that would not be the case today (at least before human interference). However, this is only a partial explanation, as early Tertiary marine mammals had to contend with predatory fishes and crocodiles in fresh water and in the sea. together with land predators.

It is entirely possible that more than one type of refuge was exploited during the evolution of any one marine mammal lineage, perhaps even at the same time.

V. The Limits of Adaptation

So far the discussion has been based on what is evolutionarily, ecologically, and functionally plausible—roughly, what is permitted by the ecological context, the laws of physics, and the need for gradual transformation through a series of equally viable ancestors. However, the adaptation of marine mammals is also limited by phylogenetic constraint: the concept that an evolutionary lineages change is limited in direction or in scope by inherited features. Indeed the need to evolve along a continuous spectrum of viable forms is in itself such a constraint, as suggested by the repeated patterns in the evolution of buoyancy control outlined earlier. But does the inherited body plan itself pose constraints? A good example seems to be marsupial “whales,” or rather their nonexistence, which at first sight seems surprising. Convergent evolution with placental mammals has generated, for instance, marsupial “mice,” “cats,” and “wolves,” so why not “seals” and “whales”? This is probably because marsupial young are born at a very immature stage compared to placental mammal young and develop within the mother’s pouch or hanging from the teat so that they would drown the moment their mother went swimming. Marsupials cannot therefore easily evolve aquatic forms. The only well-attested semiaquatic marsupial is the yapok or water opossum (Chi-ronectes minimus), which forages for prey such as fish and crustaceans in fresh water. In the female, the waterproof pouch can be closed off with a sphincter to keep water out, and the young can tolerate the ensuing low oxygen levels for many minutes. However, this is an impractical solution for a lifestyle involving swimming for more than brief periods. The marsupial mode of reproduction therefore appears to be a phylogenetic constraint impeding the development of fully aquatic lifestyles beyond the semiaquatic stage. Because placental mammals carry their young for much longer periods, they have been able to bypass this constraint.

Indeed, it can be argued that many marine mammals (and marine amniotes more generally) are, or were, not fully aquatic. The pinnipeds, the penguins, marine turtles, and most seabirds (for instance), remain critically dependent on breeding on land or on floating ice, which has major impacts on their biology. They have to retain at least some ability to move on land, compromising full adaptation to life in water. Even so, they are extremely vulnerable to land predators when they go ashore to breed. They therefore have to swim or fly, sometimes over global distances, to suitable rookery areas in sometimes inhospitable environments. The spectacular concentrations sometimes resulting at these rookeries provide dense populations of potential mates and competitors, which may drive intense competition between males, selecting for massive sexual dimorphism, as in elephant seals (Mirounga spp.). Plainly something about their various modes of reproduction prevents young being born live, or if it is born live, from surviving in water. These animals therefore still occupy an evolutionary refuge of the kind outlined earlier in the form of predator-free, land breeding areas.

Whales and sirenians do give birth to live young in water and have therefore—or concurrently—lost the ability to move on land, freeing them to become entirely adapted to life in water, just as the ichthyosaurs did more than a hundred million years before. Whales and sirenians, then, seem to be the only truly marine mammals with no remaining link to the land that their ancestors left millions of years before. Nevertheless, the contrast between ichthyosaurs and dolphins reveals evidence of phylogenetic constraint in what are superficially functionally optimal animals. They are superficially very similar, under the same selective pressures of a fully aquatic lifestyle, with fishlike bodies with dorsal and caudal fins and long narrow snouts for snapping at fish and cephalopods underwater. They even give, or gave, birth to live young underwater (fossils of gravid female ichthyosaurs have been found). Still, they differ in ways that reflect their differing ancestries. Ichthyosaurs were (presumably) derived from lizard-like land reptiles that ran by using lateral undulations of their bodies to increase the forces and movements of their legs, and hunted by sight rather than smell. Ichthyosaurs accordingly had tails that beat from side to side with vertically oriented caudal fins. They also had huge eyes for seeing in deep or silt-laden water. In contrast, dolphins were ultimately derived from small nocturnal insectivorous mammals rather shrew-like in habits, with high-frequency hearing and rudimentary sonar (as in rats and blind people) and whose running was enhanced by dorsoventral flexion and extension of the body. Accordingly, dolphins swim with up and down beats of their tails, which bear horizontally oriented caudal fins. Instead of the huge eyeballs of ichthyosaurs, the domed foreheads of dolphins contain the sonic lenses and huge brains needed to send sonar pulses and process the resulting data.

Thus even the most adapted of the marine mammals show the signs of their origins.

Next post:

Previous post: