by Brian Switek
The evolution of the amniotic egg—a self-contained pond, surrounded by an outer shell for developing vertebrates to grow in—around 330 million years ago allowed vertebrates to move away from the water and proliferate into a variety of forms. Among this radiation of amniotes were the early members of a group called the Synapsida, such as the 306-million-year-old Archaeothyris. If we could see Archaeothyris alive today we might be tempted to call it a lizard, but its skeleton exhibited some telltale characteristics that distinguish it from true reptiles. One of the most important was the presence of a single opening behind the eye socket in the skull called the temporal fenestra. This single opening, used as a site for jaw muscle attachments, is a defining feature of synapsids and is seen in members of this group to this day. (Even you and I have modified versions of this structure.) The earliest reptiles did not have this opening, and many of the later reptiles either had two temporal fenestrae or descended from ancestors with two temporal fenestrae.43 The presence, position, and abundance of this major skull opening provides a quick and dirty way to separate early relatives of mammals from reptiles.
From early synapsids like Archaeothyris evolved other forms like the approximately 280-million-year-old Eothyris, another mock lizard with two pairs of large, canine-type teeth sticking out of its upper jaw. This differentiation in teeth would be further selected for among synapsids, as seen in the pelycosaurs.44 These creatures, such as Dimetrodon, flourished during the early Permian, and included an array of forms from sail-backed predators to massive herbivores with comically small heads, such as Cotylorhynchus. Most, however, died out long before the massive extinction that rocked the planet at the era’s close. While the relationships among the pelycosaurs are currently undergoing revision, members of a particular subgroup called sphenacodontians closely resemble a lineage of later synapsids with even more specialized features, the therapsids.
The therapsids are distinguishable from their pelycosaur ancestors by having larger temporal fenestrae, limbs that were held more verti-cally beneath their bodies, and distinct incisor, canine, and molar teeth. Although the exact details of their origins from the sphenacodontians are still being worked out, the earliest known therapsids evolved during the Middle Permian around 267 million years ago and quickly diversified into several different groups that replaced the more archaic pelycosaurs.45
FIGURE 48 - A restoration of Dimetrodon , an apex predator during the early part of the Permian. Despite its appearance, it was a synapsid and more closely related to us than to reptiles.
Some of the earliest therapsids were the carnivorous biarmosuchians, predators that were clearly slightly modified versions of the sphenacodontian pelycosaurs. Their altered body plan made them more efficient hunters, however, and it was from creatures like this that other types of strange therapsids evolved. There was a mixed group of herbivores and carnivores known as the dinocephalians (“terrible heads”), the herbivorous anomodonts to which Owen’s Dicynodon belonged, and the mostly carnivorous group known as the theriodonts, which included the terrible, saber-fanged gorgonopsians. As a whole, these therapsid groups evolved a wider diversity of forms than that seen in their pelycosaur forebears: the anomodont Suminia was one of the first vertebrates to live in the trees, herbivorous dinocephalians like Moschops had stout heads reinforced with bone for literal head-to-head competition; and large gorgonopsians like Inostrancevia stripped the flesh off their prey with an advanced set of dental cutlery. The middle of the Permian was the heyday of synapsids, but many of these groups would soon be wiped out.
About 251 million years ago the earth suffered the worst mass extinction in its history. Most synapsid groups present at the time were mowed down by the catastrophe. Even so, some synapsids persisted through the event, and among the survivors were the cynodonts.
Cynodonts were the small therapsids that Broom had initially proposed as mammal ancestors, and even though they evolved shortly before the Permian extinction they somehow survived it. They underwent their own diversification in the wake of the extinction and, while most of the cynodonts faded into oblivion, at least one group among them was ancestral to the earliest mammals. By the time the earliest “true” mammals evolved during the early Jurassic, however, dinosaurs had already taken over as the dominant large vertebrates on land. Mammals started off as small, shrewlike creatures, and most would never grow larger than a house cat during the Mesozoic. The first mammals had evolved in a world overrun by dinosaurs, and the so-called “Age of Mammals” only began after non-avian dinosaurs became extinct. Indeed, descendants of the earliest synapsids survive to this day; you and I belong to the group. We are quite different from many of our early synapsid ancestors, but we retain ancient traits.
If we were to try to construct a pathway from sphenacodontian pelycosaurs through therapsids to the cynodonts and then modern mammals, the development of the mammalian inner ear might initially appear to be an insurmountable obstacle. How could the delicate association of miniscule bones (the incus, malleus, and stapes) that transmit sound from the outside world to our inner ear have evolved?
The bones of the mammalian inner ear did not appear out of nowhere. Mammals have only a single lower jaw bone, the dentary, but many early synapsids had multiple jaw bones that articulated with the back of the upper jaw in a way very different from that seen in modern mammals. These “extra” bones in the lower jaws of early synapsids would eventually become the components of the mammalian inner ear.
This connection between the mammalian jaw and ear had been known since the time Bain was digging for fossils in the Karoo desert. In 1837, the German anatomist C. B. Reichert noticed that during the embryonic development of a pig fetus, its lower jaw forms from a more flexible precursor called Meckel’s cartilage. The formation of the lower jaw bone occurs through the transformation of part of the cartilage into bone, yet not all of the cartilage goes into the lower jaw. Some of the posterior part of the cartilage turns to bone and migrates into the inner ear, becoming the auditory bone called the “malleus.” Through development, ear and jaw were connected.
Given that Reichert published this finding more than twenty years before the idea of evolution hit the scientific mainstream, however, it was not immediately recognized as relevant to the origins of mammals. It was the later accumulation of fossil evidence that would cause paleontologists to look back at and test what Reichert’s observations had hinted at.
The story of the mammalian ear actually began long before the first synapsids or even the first amniotes. Tens of millions of years earlier, when there were no vertebrates on land, part of the gill arch of fishes was used to support the braincase of the skull. This bone was called the hyomandibular, and by the time the first tetrapods had evolved this bracing structure had been modified to assist in multiple functions. In the early tetrapods Acanthostega, for example, the hyomandibular was a small but stout bone that offered support to the skull, may have played a role in respiration (as it does in living lungfish), and was placed right next to the external ear opening on the head. This bone was in just the right position to conduct sound waves from the outside world to the skull. It had changed so much that it was given a new name, the stapes, and it is one of the chief components of the mammalian inner ear.
Early synapsids and pelycosaurs inherited stapes from their amniote ancestors, but they lacked the other ears bones seen in their later relatives. Skull shapes had changed quite a bit at this point, though, and the sound-conducting stapes was in direct contact with a bone called the quadrate in the upper jaw of creatures like Dimetrodon. This is important, as the quadrate bone was the part of the skull that the lower jaw articulated with, meaning that synapsids with this arrangement could hear through their jaws.
In mammals the quadrate bone is known by another name, the incus, but this auditory ossicle did not immediately detach from the lower jaw and migrate into the inner ear. In many therapsids, what we call the incus and the stapes were still connected to the lower jaw, and the lowe
r jaw itself was undergoing a major reorganization. By setting up a sequence of skulls running from the pelycosaurs through the therapsids to the cynodonts, it could be seen that the dentary bone makes up more and more of the lower jaw as we look from the geologically older forms to the younger ones. In some cynodonts, in fact, the rear portion of the dentary even extends upward and backward to create a second contact with the upper jaw between a bone behind the dentary called the surangular and a bone of the upper jaw called the squamosal. The evolution of this second joint not only strengthened the jaw (with the force of chewing borne on one lower jaw bone), but allowed some lineages of cynodonts to switch between jaw joints.
This shift in jaw joint construction was essential to the later development of the inner ear. As the synapsids evolved, the muscles used in chewing became increasingly attached to the dentary, thus selecting for a larger dentary and smaller post-dentary bones. Even so, the bones behind the dentary in cynodonts still played a minor role in jaw support and movement in species where the articulation between the dentary and the squamosal bone of the skull had not been fully established. In species where the new jaw articulation was more firmly in place, however, the post-dentary bones were relieved from their roles in jaw movement, thus allowing them to be further coopted for hearing. One particular genus of cynodont, Diarthrognathus (roughly, “two-jointed jaw”), is virtually caught in the act of this transition, and in such cynodonts the dentary extends all the way back to contact the squamosal, successfully having switched to the kind of jaw joint seen in all mammals.
But even by the time the cynodonts had evolved a mammal-type lower jaw, they did not have inner ears like modern mammals. The forerunners to the tiny bones in our own ears had been shrunk down from parts of the more reptilelike lower jaw possessed by early synapsids, but they had not yet migrated inside the skull into the inner ear. Instead they became associated with the modified version of the angular bone that formed a ring that housed the tympanic membrane, a bit of soft tissue that acted like a drumskin to take in sound and transmit it to the ear bones. This ring fit into a notch that was positioned just behind the dentary in the lower jaw.
A recently described early mammal provides near-perfect evidence of this arrangement. In 2007 Luo Zhe-Xi and colleagues described the nearly complete specimen of an approximately 125-million-year-old mammal they called Yanoconodon allini. It was so well-preserved that the scientists were able to identify the malleus and incus in the fossil. These tiny bones were still connected to the lower jaw by a bit of ossified Meckel’s cartilage but had migrated farther from the jaw than in similar mammals that have been studied. In other words, what once were parts of the lower jaw were by now used for hearing and were only barely connected to the lower jaw. Yanoconodon allini confirmed what Reichert had glimpsed in the development of pigs, and confirmed that mammals, as Stephen Jay Gould once put it, have “an earful of jaw.”46
FIGURE 49 - The skeleton of Yanoconodon allini as preserved in the rock and compared to a restoration.
FIGURE 50 - The right lower jaw of Yanoconodon allini as seen from the inside and top. The ring on inner ear bones are connected to the lower jaw by way of a bar of ossified cartilage.
We owe another peculiar trait to our synapsid forebears: a secondary palate. It is what you are touching when you stick your tongue onto the top of your mouth, and it acts as a bony divider between your nose and mouth. This partition is important to an active lifestyle. Without it you would have to hold your breath while you ate your dinner.
Almost everyone, at one time or another, has been admonished by their parents to chew their food carefully. Careless masticatory behavior could cause an improperly chewed chunk of food to get lodged in the back of the throat, clogging the windpipe and putting us in mortal peril. We can blame this vulnerability on our early tetrapod ancestors. In early tetrapods the tubes used for breathing and swallowing food shared a common vestibule, the mouth. This arrangement meant that food could become clogged in the air-intake line with relative ease, and this is a danger all tetrapods have had to live with since. As if that were not enough, early tetrapods faced another challenge. They breathed by gulping air into their mouths, and in order to swallow their prey they would have to temporarily stop breathing.
This remained a problem even as the first amniotes evolved. Even though they had enlarged nasal openings, air still had to travel through the mouth to the windpipe, and air intake would be blocked while their mouths were full. This may not have presented much of a problem for small prey, but larger morsels posed more of a challenge. These creatures literally had to remember to breathe while they ate.
Many of the early synapsids encountered the same problem. The pelycosaurs and many of their therapsid descendants lacked the nasal/oral divider, yet it appears this feature evolved three separate times during synapsid history. This involved the extension of bones nestled near the nasal opening between the tooth rows of the upper jaw, and among other groups it evolved among the cynodonts.
Precisely why the secondary palate evolved and what function it might have initially had has been difficult to determine. Perhaps, by allowing synapsids to chew and swallow their food quickly, it allowed these animals to become endothermic (and hence more active). Then again, the secondary palate might have strengthened the upper jaw to withstand greater bite forces, allowing the synapsids to tackle a wider array of food. Whatever the reason for its evolution in several lines of synapsids, however, its presence in modern mammals is attributed to a quirk of history.
As Broom had hypothesized, the ancestors of the first true mammals were to be found among the cynodonts, the small synapsids who were among the survivors of the worst mass extinction ever suffered by life on earth. It was a catastrophe of unimaginable scale that no species could have prepared for. During such an event natural selection did not just work on the level of the individual animal, but on species and even entire groups of organisms. Some species already possessed features that allowed them to survive, while others suffered under the weight of such heavy selection pressures that they were extinguished entirely.47 While the Permian world had been ruled by synapsids, they would not prevail again for more than 150 million years.
Extinction came quickly, in an event so catastrophic that paleontologist Michael Benton dubbed it “when life nearly died.” Nearly seventy percent of all terrestrial vertebrates vanished, while over ninety-five percent of the known marine fossil species disappeared forever. It was as close as evolution has ever come to having its “redo from start” button pressed.
This is a relatively new understanding. Until recently it was believed that the Permian mass extinction occurred in an orderly manner in which each species dropped out of “life’s race” one by one. This was the classic uniformitarian view that had been supposed since Darwin’s time, but paleontologists have recently learned that large-scale catastrophes can happen. What we have come to find in the latest Permian and earliest Triassic rocks is that life on land and in the sea was almost wiped out virtually overnight.
It is often easier to understand the evolution of a species than its extinction. Every species carries a record, albeit a partial one, of its evolution in its biology, but how can we account for the disappearance of a species? Contrary to the beliefs of some early twentieth-century naturalists, extinction is not the result of a species reaching “old age” and disappearing at the end of the time natural forces have allotted them. Instead, mass extinctions are the effect of ecological interactions that cannot be observed by humans directly. We can see the damage, the wounds, but identifying the implements used during the massacre is extraordinarily difficult. To figure out what caused the Permian extinction, though, we need to look closely at the ancient crime scenes that record the tragedy.
The end-Permian mass extinction was most deeply felt in the seas. At that time the global ocean supported a wide array of coral reefs. Fish and the spiral-shelled ammonites congregated in the water above, while snails, single-hinged shelled mo
llusks called brachiopods, and other invertebrates practically swarmed on the bottom. These reef communities were not whittled away piecemeal but were eradicated in a geological instant. They were replaced by a wasteland inhabited by only a handful of species, which proliferated in the wake of the event. It was like looking at a forest recently leveled by fire; a small collection of fast-breeding “crisis” species lived where a more complex collection of organisms once thrived. Some groups, such as the last of the trilobites, were wiped out entirely, while others, such as brachiopods and the frond-like crinoids, persist to this day but never regained their previous levels of diversity.
Things were nearly as bad on land, which is best seen in the transition across the Permian-Triassic boundary in South Africa’s Karoo desert. Around 251 million years ago this area was host to a collection of thirty-four genera of vertebrates, who lived along the horsetail-filled river margins and among the forests created by the treelike plant Glossopteris . Small cynodonts chased lizards and insects through the underbrush, while the far larger gorgonopsid predator Rubigea feasted upon the herbivorous Dicynodon and armored reptiles called pareiasaurs. The anomodont herbivores were much more abundant than their predators, however, and this slice of earth history is known as the Dicynodon Zone.
Much like the reefs offshore, the Dicynodon Zone represented a complete, complex ecosystem with various levels of interaction between its members. Near the top of the Dicynodon Zone, however, the abundance of vertebrates begins to dwindle, and on the other side of the boundary the “recovery” vertebrate community only boasts seventeen genera. Most of these are entirely different from the ones that had lived in the same place before. Approximately eighty-eight percent of the vertebrate genera disappeared during the transition, and the early Triassic fauna was made up of different members. The most numerous vertebrate was Lystrosaurus, a survivor from the end of the Permian and a relative of Dicynodon. It was also so numerous in the earliest Triassic strata that this complementary part of the series is called the Lystrosaurus Zone. It did not have as much to fear as its earlier relative, though. The large gorgonopsid predators did not survive. Other fossil localities in places like the Ural Mountains of Russia tell a similar story. Complex ecological webs were erased and replaced extremely abruptly. How could this have happened?
