In ancient Greek mythology Proteus was the herdsman of Poseidon’s seals, a god who could change himself into any creature on earth – leopard, tree, snake, water, anything. Although some may not like the comparison, in the modern world-view Nature is thought to possess a similar capacity: leopards, rhododendrons, mushrooms, sharks, oysters, mosses, sauropods, sparrows, bacteria, all have sprung from Nature’s limitless capacity to adapt to new environments and conjure up new forms.

The truth of this view depends on establishing that life began in the sea, which four-footed vertebrates subsequently abandoned because they found life easier on land. There is little support for such a transition, either now or in the past. The closest Nature comes today to making the transition is when fish called mudskippers use their forefins to hop or walk across tidal flats. Drawing on air either trapped in their gills or absorbed through their skin, they can survive out of the water for several days, but they are still fish, and it is doubtful whether they provide any analogy to what is thought to have happened 380 million years ago. The fossil record is also unforthcoming. The clearest examples all go in the other direction – animals that left the land and plunged into the sea, such as ichthyosaurs, turtles, marine snakes, penguins, whales. These were all originally terrestrial animals. Manatees and dugongs – collectively known as seacows – are another such example.

Christopher Columbus saw manatees on his first voyage to the Americas and, being reminded of another mythical creature, called them sirens. ‘Not half as beautiful as they are painted’ was his verdict on their grey, snub-nosed bodies. Seacows are thus technically known as sirenians, the name for an order comprising, today, just three manatee and one dugong species. In the past they were both more diverse and more widespread. In recent times man has contributed to their decline. To differing degrees they are all endangered, as a result of hunting, destruction of habitat, boating and pollution. Steller’s Sea Cow, a species of dugong that could attain a weight of over 6 tons and a length of 9 metres, was hunted to extinction in the 18th century, a few decades after it was discovered.

Manatees live in the tropical, shallow, vegetated waters of the Caribbean, the Gulf of Mexico, the Amazon river basin and the coast of West Africa. Rarely wandering far from fresh water, they have a pronounced, bristly, upper lip that they use for grabbing overhanging mangrove leaves and submerged seagrasses. While their eyesight is weak, this is compensated for by excellent hearing and tactile sensitivity, enabling them to navigate in waters that are often murky. They swim by undulating their bodies and paddle-shaped tails.

Dugongs live in the more open coastal waters off eastern Africa, western India, Malaysia, Indonesia and northern Australia. They look similar to manatees, but their bodies are smaller, their tails are fluked, like those of whales and dolphins, and their snouts are longer and more down-turned (a specialisation for bottom-feeding). They feed mainly on seagrasses.

Sirenians, like whales and dolphins, are mammals. Being mammals, they immediately suggest a terrestrial origin, since distinctively mammalian features such as sweat glands and hair – largely lost in sirenians – only make sense in relation to a terrestrial existence. The hair of manatees (and probably also dugongs) has a sensory rather than insulating function: an array of bristles and whiskers on the snout provide tactile information while others over the rest of the body sense variations in water flow and topography. More than 200,000 nerve fibres transmit the data to an equally specialised part of the brain. Over time there have been many other adaptations. The pelvis, for instance, has been reduced and is no longer attached to the backbone. The hindlimbs have gone completely. The flippers are much-modified five-fingered forelimbs – manatees indeed still have fingernails. Their air-breathing also points back to a previous terrestrial existence. Sirenians breathe through nostrils kept shut when under water by valves, but their lungs have become thinner and longer so that they lie along the backbone instead of the rib cage, thereby distributing buoyancy over the whole body.

Sirenians are all adapted specifically to shallow water – they do not have the ability to dive deep. They even sleep under water, drowsily surfacing for air every few minutes.

The rise of seagrasses

The rise of the sirenians is closely associated with the rise of their principal diet, the sea-grasses. Sea-grass meadows are among the most productive plant communities, serving as nurseries for fish and molluscs and as feeding grounds for predatory fish, sea birds, sirenians and turtles. Sea-grasses are flowering plants, or angiosperms, but despite their name they are not closely related to the terrestrial grasses. Fossils of them are sparse, the oldest dating back to the Late Cretaceous; a few more are known from the Eocene. Originating from four distinct terrestrial lineages, they are an example of convergent evolution, and provide ‘dramatic evidence of the adaptive capacity of flowering plants to evolve and survive in extreme environ- ments’ (Waycott et al. 2006). Sea-grasses flourished perhaps especially during the Late Oligocene and Early Miocene, when plate-tectonics produced vast areas of shallow-water habitat.

The oldest known sirenian is Prorastomus, from Early Eocene Jamaica. Another species, Pezosiren, is a slightly later but better preserved example of the same family, also from Jamaica. Pezosiren was quadrupedal, but specialisations such as retracted nostrils and heavy ribs for ballast suggest that it spent most of its time in the water. By the Late Eocene sirenians had little use for hindlimbs and were fully aquatic. In tandem with the increasing opportunities to live by angiosperm vegetation on the coasts, they continued to multiply as the Cenozoic progressed, reaching peak diversity in the Miocene – about the same time as sea-grasses were at their most extensive and diverse. One of the reasons why sirenians are now in danger of extinction is that sea-grass meadows are in steep decline.

The significance of sirenians

Since sirenians still retain vestiges of their terrestrial ancestry, it would be futile to try to argue that they were created with the forms they have now. They are a wonderful example of evolution. The important question is: what kind of evolution? Were their present forms the pre-ordained result of genetic programming, or the fortuitous outcome of innumerable genetic mutations? Transformations equally wonderful can occur in a single lifetime, as when a caterpillar metamorphoses into a butterfly, so we know that programmed evolution is possible. Undirected evolution is an altogether more speculative proposal, if only because it must have occurred over lengths of time stretching far beyond what we could observe.

Here are four reasons for concluding that the evolution was genetically controlled:

• First, so far as we can tell, today’s manatees and dugongs lie at the end of the group’s evolutionary journey. They had completed the transition from land to sea already by the end of the Eocene, subsequent to which the pace of anatomical change was slow and the degree of change relatively minor (continuing diversification in dentition and digestion, for example). The rapidity with which species effected the crossing is the phylogenetic equivalent of the rapidity with which individual caterpillars, or tadpoles, can metamorphose: it suggests a directed rather than random process.
• Second, the journey was unidirectional. The sea was an extreme environment, yet the earliest sirenians did not merely dip their toes in it, finding it unbreathable and too salty to warrant further experiment. Something in their biology urged them to plunge all the way, like herds of migrating wildebeest, as if the end was already known and genetically provided for. Today, having, like the sea-grasses, largely exhausted their capacity for innovation, they now have nowhere to go. Like so many creatures facing perilous futures, they must either hold on to what remains of their habitats or go extinct.
• Third, natural selection seems to have played no significant role. There is no evidence that the sirenians became adapted to their forbidding new environments through being involved in a bloody competition for resources. Nor is it easy to conceive how the complex genetic changes underlying the adaptations – from the development of nasal valves to the transformation of their tails into paddles – could have been the result of haphazard and unorchestrated mutations, with every step giving the animals greater ‘fitness’ and incrementally leading towards their final forms. Moreover, in the Darwinian story everything depends on the animal’s leaving more offspring than its competitors. Sirenians are extremely slow breeders: intervals between births, when usually only one calf is born, are typically 2-5 years.
• Fourth, their migration from land to sea was not unique, even amongst mammals. Pinnipeds (seals, sea lions and walruses) and cetaceans (dolphins, porpoises and whales) are groups that also descended from land animals, and they completed their journeys over similar timescales: cetaceans about the same time as the sirenians, pinnipeds slightly later. Although they evolved from independent lineages, some of their adaptations – for example, the caudal fluke of cetaceans and dugongs, the flippers and reticulated kidneys of all three groups – were exactly the same. These parallel phenomena tend to reinforce the view that the evolution of the sirenians was not an undirected process.

Recolonisation

The rise of the seacows is one of the many fascinating chapters that make up the story of Earth’s recolonisation. Terrestrial life at the end of the Hadean had been blotted out. The oceans, newly formed in the cataclysm, had to be stocked either from life that had previously existed underneath the land or from animals that survived on the Ark. In the former case, the animals had been aquatic from the beginning. Although their yet-to-be-performed feats of transformation would be equally extraordinary, their theatre would remain the sea. In the latter case, the animals were air-breathing walkers and crawlers of the land; if they were to be enlisted for the task, they had to adapt themselves, somehow, to another world.

So in the Palaeozoic and Mesozoic a variety of reptiles, and in the Cenozoic three groups of mammals, took the sea. At their appointed times the ancestral animals launched forth to exploit the new opportunities. Sea-levels and global temperatures peaked at the beginning of the Eocene. It was about then that cetaceans first ventured into the water, starting from the margins of a Mediterranean Sea that stretched from the Americas to South East Asia, plunging deeper and deeper. At the same time sirenians began to colonise the sea-grass meadows and mangroves that were spreading through the Caribbean. Then as temperatures continued to fall, in the Oligocene, pinnipeds began to wade into the coastal waters around North America, from where they later took on the even more chilling challenges of the Arctic and Antarctic. The oceans were divided amongst them in such a way that they did not have to compete with each other.

These changes were rapid, but they did not happen overnight. They must have taken place over thousands of years. But the increasing length of time required by the geological and palaeontological evidence as one approaches the present is part of the story. Radioisotope time was slowing down. The true length of time represented by one million radioisotope years in the Eocene was much longer than, say, the same number in the Carboniferous. In reality the Eocene may have come less than half way through the total length of time since the cataclysm. That would be one reason why ‘molecular clock’ estimates of the time when major taxonomic groups diverged are often much older than first appearances in the fossil record.

Nonetheless, a considerable amount of evolutionary history for the ancestors of the sirenians is still implied, even allowing for the fact that large-scale evolution is not continuous. Prior to the Eocene, the fossil trail grows cold. This is not because sirenians were created in the Palaeocene, or because of an incomprehensibly incomplete fossil record, but because, in the early stages of recolonisation, the animals themselves were sparse on the ground. Even when they do appear, fossils are initially rare. Ideas about ancestry consequently become more speculative as one goes back in time. That proviso made, genetic and morphological relationships suggest that seacows are part of a group called the Paenungulates, sharing a common ancestor with elephants (Proboscidea) and hyraxes (Hyracoidea). The transforma- tions documented by the fossil record are but the tip of the iceberg.