Having seen how the ancestors of turtles, ichthyosaurs and sirenians all conquered the sea, it will not be surprising to learn that the ancestors of whales and dolphins also did. Since before Darwin the assumption has always been that, if God created the world, evolution is excluded: species are fixed, and the world unchanging. Microevolution may be possible but not macroevolution, and least of all the evolution of whales from a terrestrial quadruped. Whatever is incredible for man must be impossible for God. Others consider rather that with Nature anything is possible. Thus we have the strange situation where believers in a deity with unlimited power believe that the transformation could not have happened, while those believing in no such person believe that it must have happened. One party says, “You only believe that whales evolved from land animals because your commitment to common descent forces you to.” The other says, “No, there really is evidence for the transition.”
Whales, dolphins and porpoises, collectively known as cetaceans, are mammals. They have an endothermic metabolism, a four-chambered heart separating oxygenated from deoxygenated blood, a single lower jawbone and three ossicles in the middle ear. They grow their young in a uterus and after birth suckle them with milk. Most still have some follicles, some even hair. All this amounts to a body plan that unites whales with other mammal groups, each of which have within the mammalian one a more specific body plan: for example, pangolins, primates, rodents. When one thinks of the cetacean body plan, one thinks of a streamlined marine mammal with flippers, no external hindlimbs, a short neck and a horizontal fluke. But this essence of what defines a cetacean came into existence over time. Their form changed, at the same time as every one of the systems distinctive of their mammalian nature changed. Certain types of teeth are also characteristic of mammals, but in cetaceans these changed beyond all recognition.
So we need a cast of mind that is not stuck in preconceived notions of what is, or is not, mutable; of what group is, or is not, related to another group. On the one hand, the phylogeny of the mammals that appeared in the early Cenozoic is, as biologists admit, ‘recalcitrant’: it resists the notion that they are all interrelated (Esselstyn et al. 2017). On the other, the fossil record shows that cetaceans did not begin as whales and dolphins, and molecular evidence suggests they are not without relatives among the living. Most probably they are related to the two extant members of the once more diverse hippopotamid family.
Hippopotamids arose in the mid-Miocene, much later than the cetaceans, so the relationship cannot be close. Moreover, there is no reason to think that hippos, although adapted to both land and water, are evolving into fully aquatic animals. They evolved so far and no further, from anthracotheres, a non-aquatic group within a much larger category known as even-toed ungulates or ‘artiodactyls’. Artiodactyls are hoofed animals whose weight is borne equally by two of their five toes, as opposed to odd-toed ungulates or ‘perissodactyls’, whose weight is borne by only the third toe. Their ankle bone is shaped like a double pulley. But the design of their foot does not by itself amount to a ‘body plan’, and artiodactyls (pigs, camels, ruminant animals such as deer) are not necessarily all related to each other: the feature could be convergent. Hippopotamids are not obviously related even to any of the other living artiodactyls (Boisserie et al. 2005).
The oldest group that can be identified as ancestral to modern cetaceans is the pakicetids. Although pakicetids are analysed as most closely related to ambulocetids and then to the pair consisting of remingtonocetids and protocetids, pakicetids and ambulocetids are largely contemporary in the fossil record. So too are the second pair, the remingtonocetids and protocetids. At the beginning of the Middle Eocene all four groups were contemporary. The points at which they diverged from each other must therefore lie further back in time, and pakicetids and ambulocetids, to take the two oldest groups, must have stemmed from an ancestor common to both of them.
A middle-Eocene group possibly related to cetaceans was a group called raoellids. They share similarities of teeth and a thickening of part of the bone enclosing the middle ear. In cetaceans this thickening is a modification for hearing under water. As cetaceans became more aquatic, air sinuses isolated the middle ear from the braincase by way of further adjustment for the change of medium. Raoellids are generally understood to be the sister group, with anthracotheres sister to cetaceans + raoellids. The raoellids split away no later than the Early Eocene and the anthracotheres some time before that, about the same time as most other modern mammals appeared in the fossil record. Raoellids predominantly inhabited the shallow waters overlying modern-day India and Pakistan. Anthracotheres and hippopotamids are decidedly an African group.
The early cetacean groups were semi-aquatic, to differing degrees; that is, they could paddle in the sea (or walk on the sea bed) as well as walk on land. Pakicetids – so named because their bones were found in Pakistan, though they also occur in India – had heavier limb bones to offset their buoyancy, and their eyes faced upwards, suggesting a lurking life at the water surface close to river banks. Ambulocetids were larger, had large hands and feet relative to their upper limbs, and an elongate snout. Their eyes were less upward-facing. Remingtonocetids had longer necks, longer, more powerful tails, and ears more specialised for hearing under water. Their eyes had now moved to the side of the head. Their geographical range extended as far as Egypt, at a time when the proto-Mediterranean, Tethys, extended from the Atlantic all the way to China. India was converging on Asia from the south but still separated by ocean.
Protocetids – Maiacetus is a well-preserved example – had a nasal opening further back along the snout than their predecessors. Over time the opening continued to shift until it reached the top of the head, closest to the air/water interface, and became a blowhole. Although protocetids still had residual hooves, their long digits, probably webbed, would have made terrestrial motion ungainly. Like seals and sea lions, they most likely came ashore only to rest, mate and give birth. A shark bite in the carcase of one protocetid shows that the animal was swimming when attacked (Bianucci & Gingerich 2011). Originating in Indo-Pakistan, protocetids spread westwards along the north and west coasts of Africa. From there, like the sirenians, they somehow crossed the Atlantic and eventually reached the west coast of tropical South America (Lambert et al. 2019). A few thousand years seems ample time for a population to spread that far. The geological timescale imputes millions of years.
Basilosaurids, overlapping the protocetids chronologically, were markedly more whale-like. Best known is the 5-metre-long Dorudon, of Late Eocene age. The largest, Basilosaurus itself, grew to 15-20 m and hunted Dorudon calves. In contrast to protocetids, basilosaurids had flippers, a fluke at the end of the tail – now the main instrument of propulsion, as in modern cetaceans – and vertebrae that were both longer and more numerous. The neck, however, was shorter. In other respects basilosaurids were intermediate. Hindlimbs were present but greatly reduced, their teeth were still differentiated into incisors, canines, premolars and molars, and like most mammals they had only three phalanges per finger. Probably they had blubber, if only to help streamline their bodies. Later, blubber would have the additional function of providing insulation in cold water and a store of energy during long migrations. They also needed to have developed, by this stage, kidneys capable of processing salt water. Basilosaurids had a global distribution.
Let us briefly touch on the fascinating world of marine geochemistry. Researchers can establish how warm or cold oceans were by analysing the shells of single-celled organisms that lived on the seafloor, called foraminifera. An atom’s nucleus consists of protons and neutrons, and an element, chemically, is defined by how many protons it has; neutrons can vary. Elements which have more than one atomic weight are called isotopes. Oxygen has two main isotopes: oxygen-18, with 8 protons and 10 neutrons, and oxygen-16, with 8 protons and 8 neutrons. Researchers want to know the ratio of 18O to 16O in the calcium carbonate, CaCO3, making up the shells. If the carbonate precipitated from cold water, the ratio will be high; if it precipitated from warm water, the ratio will be low. Since the lighter isotope is much the more common, the fraction is multiplied by a thousand, as signified by the Greek letter delta. In principle, any δ18O value can then be translated into a temperature, with each unit representing a difference of 4.5° C. The foraminifera will tell us nothing about the temperature of the surface (the first two hundred metres), but they can tell us how cold the bottom waters were.
Consider now the graph encapsulating the oxygen-isotope record from the Late Cretaceous onwards, when the North Atlantic began to open. It is apparent that bottom-water temperatures were higher than in the present, and most of that time much higher. During the Late Cretaceous and Palaeocene values generally were in the range 0 to 1 ‰, equivalent to temperatures of 8-12°. Thereafter temperatures began to rise and, apart from a blip at the Palaeocene/Eocene boundary, reached a high around 53 Ma in radioisotope time. At that point bottom-water temperatures were 14-17°, coincident with vast eruptions of lava in the central North Atlantic (and probably other places). During the period 56.5–55.4 Ma, magmatism in the Faroe-Shetland Basin uplifted the crust by some 900 m (Hardman et al. 2018). There were also, around 53 Ma, huge surges of volcanism in the already opened South Atlantic. Vast amounts of CO2 bubbled up into the atmosphere, reaching 1400 ppm or more (compared with today’s 400 ppm). Temperatures were higher than at any other time in the Cenozoic, on land as well as in the sea. Thereafter bottom-water temperatures fell progressively to around 8° C. At the Eocene/Oligocene boundary they plunged further to 3-4° as within-plate volcanism (sea mounts and oceanic plateaus) suddenly slowed, resulting in cooling of the ocean crust, a marked drop in sea-level and a 600 m drop in the carbonate compensation depth (the depth at which carbonate dissolves, a function of the amount of CO2 dissolved in the ocean). Temperatures recovered somewhat in the mid Oligocene, then remained comparatively flat until the mid Miocene.
The deep ocean was much warmer than it is today because the rate of seafloor spreading was much faster and within-plate volcanism more voluminous. Volcanic seafloor spreading was faster because the mantle beneath the crust was hotter, and it was hotter because radioactive elements were producing more heat, at faster rates of decay. Like northern Canada, Antarctica in the Late Cretaceous and early Cenozoic had temperate forests, despite the lack of sunlight half the year. However, as the Cenozoic progressed, Antarctica became separated from Australia and South America, a circumpolar current isolated it from the warm global ocean, and the surface Southern Ocean disproportionately cooled.
In ice-free conditions seawater is one δ18O unit lighter than when ice caps sequester a significant amount of frozen, previously evaporated seawater. Part of the rise in carbonate values from the mid Miocene onwards, especially during the Pliocene, is therefore probably due to the effect of ice-cap growth on the δ18O of the water from which the carbonate precipitated. With atmospheric CO2 above 400 ppm and seafloor spreading still vigorous, it is difficult to see how there could be substantial polar ice caps until after the mid Miocene, when bottom-water temperatures again began to fall.
Throughout the Cenozoic bottom-water temperatures correlate with sea-level. Sea-floor spreading being more rapid, the ocean crust was warmer, less dense and more buoyant, and ocean basins consequently shallower, forcing the oceans to spill onto the continents. In the Eocene, sea-level was 180 m higher than today. Consider, for example, the Wadi Al-Hitan or ‘Valley of the Whales’, so-named because of the many basilosaurids found there. Now 250 km south of Egypt’s coastline, in the mid Eocene it was shallow-marine. The present situation is exceptional. Exposed continental area is at an all-time maximum, firstly because 70 metres’ worth of ocean is locked up in ice sheets and glaciers, and secondly because the rate of sea-floor spreading is at an all-time low.
How warm the oceans are affects many things: climate at high latitudes, the mobility of current systems, the amount of evaporation (and therefore rainfall), the amount of oxygen dissolved in the ocean, and the availability of nutrients. Most nutrients – whatever elements go into building a cell, but especially trace elements – enter the sea through mid-ocean spreading centres. When an animal composed of these elements dies, it sinks and gradually decomposes, so that nutrients are most abundant in the colder and denser depths. There they are trapped, leaving food chains at the surface starved. Productivity is greatest along coasts where winds interact with the major current systems to generate upwelling of the nutrient-rich waters, at high latitudes, where winds are strong and the density gradient is weakest. Around the Antarctic conditions are ideal. A zone of upwelling boosts the production of phytoplankton, which feeds the proliferation of tiny copepods and krill, which are a favourite food of whales, amongst others.
Modern cetaceans divide into two groups, distinguished by their feeding apparatus: baleen whales, or Mysticeti, and toothed whales, Odontoceti. Mysticetes first appear shortly before the sudden cooling at the Eocene/Oligocene boundary. Odontocetes appear later, but presumably they split from basilosaurids about the same time as the mysticetes. How it came about that certain basilosaurids, some time before the end of their own lineage, split into two new lineages is not known.
Baleen is the name for the array of filaments fringing the upper jaw that enables mysticetes to sieve huge quantities of krill and copepod in preference to fish. Early mysticetes, such as the 8-metre-long Llanocetus, had no baleen; they had widely spaced teeth and well-developed gums. Later species lost their teeth and resorted to suction feeding. Still later species developed keratinous plates in their gums, unrelated to teeth, and baleen formed as the tips frayed. Teeth buds still form in mysticete embryos.
The rise of the mysticetes and subsequent evolution of baleen can be interpreted as a pre-ordained response to the changes in ocean ecology. Since basilosaurids died out at the start of the Oligocene, it may be that mysticetes prospered because of greater tolerance of cold. The group today is not diverse, just 15 species, most of which belong to the rorqual family, such as the blue whale, the fin whale (the second largest animal on earth) and the humpback whale. Diversity declined with the widespread glaciation that set in near the end of the Pliocene, but those species that survived benefitted from the superabundance of zooplankton and krill to reach lengths of up to 30 metres (Slater at al. 2017). Large body size facilitated body-heat retention and long-distance migration. Mysticete navigational skills are prodigious (Travis et al. 2011). Humpback whales regularly travel distances of 5000 km between their low-latitude breeding grounds and high-latitude summer feeding grounds. Having set a straight-line course, they deviate from it by less than 1 degree.
Toothlessness and suction feeding also evolved in the toothed whales, at least seven times (Peredo et al. 2018). Odontocetes are much more diverse than mysticetes. Among them are sperm whales, beaked whales (which are almost toothless), belugas, narwhals (closely related to belugas), porpoises, river dolphins and marine dolphins, the biggest dolphins being the orcas. Their most distinctive attribute is echolocation, rather like the sonar in modern submarines. However, while sonar may be of comparable capability, the technology in submarines did not evolve by any natural process, it cannot assemble itself, it cannot replicate itself, and it is useful only when linked to a separate control centre, the human brain. All odontocetes echolocate, and a primitive echolocation apparatus seems to have been present even in the earliest known odontocetes.
Most cetaceans have a dorsal fin. Grey whales, right whales, bowhead whales, belugas, narwhals and the finless porpoise do not, and of course their terrestrial ancestor did not. The fin is uncannily similar to the dorsal fin in sharks and ichthyosaurs, so its subsequent appearance in cetaceans should make one think. Very probably it is an example of modular expression, whereby certain genes during development switch on or off an entire complex of instructions. Another example would be hindlimbs. Forelimb and hindlimb buds both initially develop in embryo, but, as with mysticete teeth, the hindlimb buds then get reabsorbed – but we have already discussed this phenomenon in relation to snakes.
- First and foremost, food resources: adaptations to increase access to the fish in the sea.
- Ease of movement: consider that sea turtles make migrations of hundreds of kilometres or more, whereas a similar-sized tortoise will not migrate even a few kilometres.
- Relatively stable temperatures: a concern for ectotherms more than for endotherms, but water temperature varies much less through the year or by latitude than air temperature does.
But non-human animals do not spend their time reflecting on their best options and instructing their genomes to evolve accordingly. An ‘adaptation’ is potentially an advantage only after it has come into being. The question of how it came into being still has to be addressed.
Natural selection, a surmise at best, explains nothing. Heavier bones do not make the species contemplating life in the sea more likely to survive. If it does not lose buoyancy, it sticks to the coast. The same with the nasal opening: the shift from the front of the snout to the rear was a matter of convenience rather than survival. To become more adapted to an unfamiliar environment, such as the sea, was to become less adapted to that which the species was previously well adapted to. Losing the ability to run on land was ostensibly a high price to pay for the ability to paddle. Anthracotheres, like hippopotami and most artiodactyls, were herbivores: paddling out to sea was not an attractive option. Their relatives were forced to live off the sea, and they succeeded in doing so only because concomitant changes in teeth, taste-buds and digestive system made it possible. Some whales lost their teeth – again, not something any right-minded animal would have chosen.
Artiodactyls are of course not turtles, and it is doubtful whether any would have thought it a change for the better to have to migrate thousands of kilometres to find food. Stable temperatures also do not seem to have weighed heavily in the decision. Cetaceans adapted themselves to temperatures that other animals raised in the tropics would have deemed intolerable. Procuring the food one finds palatable takes some effort. Beaked whales plunge from waters warmed by the sun to depths of 1400 m or more, where they cannot see (except by echolocation), where the temperatures are not much above freezing, and where the pressure is crushing. The record is a dive to 2992 metres – all, of course, on the oxygen taken up when at the surface (Schorr et al. 2014). It was not natural selection that effected the most spectacular evolutionary transformation in Earth history, over the course of just ‘10 million years’. Cetaceans were compelled to make a new life for themselves by their genomes, by whoever all foreknowingly made their genomes.