The mammal within the animal kingdom

The animals in this group consist of many cells with a division of labour: one cell catches food, another anchors the body, a third supplies the sex cells. Between the cells sits a protein glue — collagen — and around them lies a shared outer skin, so that the whole behaves as one body. A body can therefore be far larger than a cell, and can have a quite different shape inside than out.

The body has a long axis, with a front end that goes first and a rear end that follows. Sense organs and nerve cells are piled up at that front end into something worth calling a head, and the gut runs all the way through from mouth to anus, so that eating and digesting happen at the same time. Between the outer layer and the gut lies a third germ layer, the mesoderm, and out of it muscle, skeleton and blood vessels are built.

Along the length of the body lies the notochord: a firm but flexible rod of fluid-filled cells inside a stiff sheath. Muscles pulling on either side of that rod do not simply fold the body double but make it wave from side to side like a whip, because the rod springs back every time. Swimming is therefore directed and fast instead of merely wriggling, and directly above the rod runs the nerve cord that sets the rhythm.

Around the notochord lie separate elements of cartilage or bone, linked up into a vertebral column: stronger than the rod alone, and with movable joints between the segments. At the front a skull closes around the brain, and the embryo has a tissue of its own, the neural crest, which supplies jaws, teeth and facial bones. The body thus has an internal frame that grows along with it, instead of an armour that has to be replaced again and again.

The foremost gill arches — cartilage hoops in the throat basket — do not stand upright here but tilt forward, and form hinged jaws. An upper and a lower jaw close against each other, carrying teeth made of the same tissue as the skin scales. An animal that can bite need not wait for whatever drifts past: it seizes prey, holds it and cuts it up, and hunting is therefore a way of life.

Inside the fleshy pectoral fin the bones lie in a fixed sequence: one bone against the shoulder, then two side by side behind it, and only then the fan of fin rays. That single bone is the humerus, the two behind it the radius and ulna — a bird's wing, a human arm. The head of the humerus is also convex and turns in a socket of the shoulder girdle, so the fin can prop and push instead of merely rowing.

At the end of the limb there is no fan of fin rays but a row of separate finger bones with joints between them, letting the animal push off against the ground. The shoulder girdle does not touch the skull, so there is a neck and the head can turn without the whole body turning with it. The ribs are heavy and hold the chest open, because out of the water a body presses down on its own lungs.

Three membranes lie around the embryo: the amnion with the fluid it floats in, the allantois that stores waste and handles breathing, and the chorion that encloses the whole. Around that is a leathery or chalky shell that keeps water in but lets oxygen through. Reproduction is therefore not tied to pond or pool, and the dry interior — dunes, steppe, high mountains — lies within reach as well.

In the skull roof there is a single opening behind the eye socket on each side, low down against the cheek arch. The jaw muscles attach to the rim of that hole and bulge out through it as they contract, so thicker muscles fit inside an equally light skull. Below the opening a narrow bridge of bone remains to take the pull. A reptile with two windows solves the same problem with two holes; a synapsid does it with one, and every mammal still carries that one window — the cheek arch you can feel above your own cheekbone is its lower edge.

The teeth in the upper jaw are no longer a row of identical points: incisors at the front, behind them a single pair of much larger canines, and smaller teeth again behind those. The bone of the upper jaw is thickened above that canine, because that is where the force of the bite concentrates, and the front teeth sit in deep sockets instead of shallowly on the jaw margin. An animal that can bite hard at one place in its jaw need not swallow its prey whole: it can kill it first and then cut it up. The distinction between incisor, canine and cheek tooth begins here.

The thigh bone no longer sticks out sideways but turns downward, and the socket in the hip therefore faces outward and down instead of straight out to the side. The leg carries the body like a pillar; the trunk no longer hangs slung between the legs. That saves so much muscular effort that walking and breathing stop getting in each other's way — a running lizard has to bend its trunk from side to side and squeezes its own lungs shut as it does so. The temporal window has also grown much larger, and the jaw muscle now also takes hold on the outside of the cheek arch, so the bite is harder and the head may be heavier.

Between the nasal cavity and the mouth a second palate closes over: a plate of bone that leads the inhaled air back from the nostrils, past the place where the food lies. An animal therefore need not break off chewing to take a breath — and so need not bolt its food whole either. Upper and lower teeth also meet each other instead of sliding past, so that food is genuinely broken down before it reaches the stomach. An animal that grinds its food gets more out of it, and getting more out of it is exactly what a body needs that stokes its own heat.

A reptile's lower jaw is made of a row of bones. Here the frontmost of them, the dentary, grows backward until it takes up almost the whole jaw, with a tall process reaching up against the skull. The bones behind it — articular, angular, quadrate — grow smaller instead, and sit as a small parcel in the angle of the jaw. The jaw muscle now attaches to the dentary itself and pulls straight upward rather than backward, so the bite is stronger while the old jaw joint has *less* force to bear. That is the precondition for everything that follows: only once that joint no longer has to carry anything can it begin to do something else.

At the back of the jaw in Probainognathus two hinges lie against each other: the old one, between the quadrate of the skull and the articular of the jaw, and a new one, in which the dentary props directly against the squamosal of the skull. As long as both are there nothing has to change all at once — and that is exactly why something *can* change: the new joint takes over the load, the old one is released. Those same two little bones move to the middle ear further along this chain; here they are still carrying a jaw.

These animals are small one and all, and being small costs heat: a body loses it across a surface that grows faster than the volume producing it. That is probably why the teeth grow finer here at the same time, why the nasal cavity carries ridges on which a mammal's thin bony scrolls sit that warm and moisten the inhaled air, and why the metabolism rises. What begins as a makeshift becomes a trade: a small animal that makes its own heat can hunt when it is cold, and fits wherever a large animal cannot reach.

A reptile replaces its teeth throughout its life, over and over, and because of that the upper and lower rows never meet each other exactly for long. Here it happens twice: a milk set and then a permanent set, and nothing further. Cheek teeth that come only once can be worn in against each other until their cusps fall into each other's basins, and such a set grinds rather than merely holding on. That it stops at twice hangs on something else in this group: a young animal getting milk need use no teeth at all in its first months, and can therefore wait for its permanent cheek teeth until the jaw has finished growing.

The old jaw joint has here come loose from the jaw altogether. The quadrate is now called the incus, the articular the malleus, and together with the stapes — the little bone a reptile already had — they carry the vibration of the eardrum through to the inner ear. Three bones in series work as a lever and turn a weak movement over a large membrane into a strong movement on a small window, and because of that a mammal hears high notes no reptile picks up. That migration can be followed step by step through the rock, from a jaw with two joints to an ear with three bones: there is little in all of palaeontology so well documented. Out of the skin comes hair, which holds in the warmth the animal makes itself, and among that hair lie glands that give off milk — probably beginning as skin glands that kept a parchment-shelled egg moist and free of mould. The platypus still uses them that way: no teat, but milk that stands on a patch of belly hair and that the young licks off.

The egg loses its shell and stays inside the mother's body; what comes out is a young animal. In the marsupials this happens very early and development continues at the teat; in the placentals the young stays inside far longer. Something is added in the ear: the cochlea coils into a spiral of more than a full turn, so a long sensory membrane fits into a small space and the animal can separate notes an egg-layer cannot tell apart. And the teat is there: milk no longer stands on the skin but comes through an opening the young latches onto.

The young is born after a few weeks, no larger than a bean, with hind limbs barely laid down but with forelimbs that already carry claws and a mouth that can only suck. Under its own power it crawls from the birth opening to the teat, fastens on, and does not let go for weeks or months. Development therefore happens largely outside the body rather than inside it, and that has an advantage often overlooked: in a bad year, breaking off the investment costs the mother almost nothing, whereas a placental has by then already put in months.

The placenta is half the young's and half the mother's: villi of the embryo push into the wall of the womb and come right up against the maternal blood vessels without the blood mixing. Oxygen, sugar and waste cross over. At the same time the mother's immune system has to leave a half-foreign creature alone for months, and that is the marvel, not the first part. The young can therefore stay inside until it is ready: a foal standing beside its mother within the hour, a young hare lying in a form with its eyes open and a coat on.

There is no bone and no tooth by which you can recognise an afrotherian: a golden mole looks like a mole, an elephant shrew like a shrew, a hyrax like a marmot — and none of those likenesses means kinship. What they do share sits in the genetic material: the same inserted pieces of DNA at exactly the same places, which arise only once and are inherited thereafter. One thing is nevertheless visible in the skeleton, and it is an odd one: in very nearly every other placental the thoracic and lumbar vertebrae together number nineteen or twenty, and in several afrotherians that count runs well above it.

The front incisors have no root and grow throughout life, in a rodent a few millimetres a week. The enamel sits only on the front face; the back is softer dentine and wears faster, so the tooth sharpens itself into a bevelled chisel as long as the animal keeps gnawing. Behind those incisors lies a bare gap without teeth, and there a rodent can fold its cheeks inward: it gnaws through wood or soil without taking in the chips. If the opposing tooth is lost, nothing checks the growth any more — the tooth curls round and finally grows into the skull, and that is the end of the animal.

In every carnivoran two opposing cheek teeth are reshaped into blades: the last premolar above and the first true molar below. They do not meet face to face but slide past each other edge on, exactly like the two blades of a pair of shears, cutting through tendon and hide where an ordinary cheek tooth would slip off. That pair is the carnassials, and from that pair alone you can read what a carnivoran actually eats: in the cat they are tall and razor-sharp, in the badger and the bear they are flattened into a grinding surface, and in the seal the teeth behind them have become pointed hooks that hold a fish but grind nothing.

An ungulate stands neither on the sole of its foot nor on its toes, but on their nails: the nail has become a hoof, a horn sleeve around the last toe bone. The wrist and ankle bones are fused into a single long rod, and the muscles that move the limb sit high against the body with long tendons running down — the end of the limb therefore weighs next to nothing and swings along like a whip, which saves energy at every step. The two orders each do it their own way: in the horse one toe carries everything, in the deer two side by side, and it is that difference — odd or even — that gives them their names.

Almost every mammal has an ankle bone (the astragalus) with a single rounded joint surface on top, working like a hinge. In the even-toed ungulates that bone has such a rounded surface on BOTH ends — a double pulley — letting the foot not only swing up and down but also push off smoothly in a leap. The feature is so distinctive that for decades it counted as the clinching proof of this order. When the palaeontologist Philip Gingerich dug up the first complete hind legs of an early whale (Rodhocetus, from the Eocene of Pakistan) in 2001, this very bone turned out to be there too: the double pulley, down to the detail. That settled a long-running dispute at a stroke — geneticists had already placed the whales among the even-toed ungulates for years, but many palaeontologists doubted it until this fossil spelled it out in black and white.

Well into the twentieth century, hippopotamuses were placed beside pigs and whales were treated as a separate order, at most related to extinct wolf-like ungulates called mesonychids. In the early 1990s the first DNA comparisons (Irwin & Arnason, Gatesy and others) gave a different answer: of everything alive today, the hippopotamus is the whale's closest relative. Many palaeontologists were sceptical — no fossil brought the two any closer together than to the other even-toed ungulates — until later finds of early whales with a double-pulley ankle (see the previous node) and of a small, deer-like animal called Indohyus showed the genetics had been right. It now stands as one of the best-documented cases of DNA predicting a kinship that the animals' outward appearance hid completely.

The oldest whales (Pakicetus, some 50 million years old) still had four functional legs and walked about on land, though they were found at the water's edge. In Ambulocetus, a few million years later, the hind legs are already reshaped into paddles; in the first protocetids a streamlined tail appears; and in the basal whales of 40 to 34 million years ago (Basilosaurus) the hind legs have shrunk to two small, functionless bones that never touch the ground again — some specimens even still carry them as a clearly recognisable rudiment, tucked away inside the body. From four-legged land animal to a mammal fully bound to water in some twenty million years: few transitions in the whole of mammalian evolution are recorded this closely, step by step, in the fossil record.