Written in Stone: Evolution, the Fossil Record, and Our Place in Nature

by Brian Switek

  Falconer described these extinct behemoths in more detail in his uncompleted monograph series Fauna Antiqua Sivalensis. Following in Cuvier’s footsteps, Falconer’s work was more descriptive than theoretical, but he identified some interesting patterns among the fossils he had found. The most immediately apparent was that the elephants of the past were far more diverse than living forms. Modern elephants were only the “ragged remnant representation of the rich garment of life with which the continent was formerly clothed.”

  Extinction had swallowed all of these fossil species, but the force that brought them into existence was a mystery. Falconer rejected the popular idea that complete faunas sprung forth instantaneously in the wake of catastrophes. It was an insult to the Creator to think that the world required such constant revisions. The continuity of the fossil record suggested something different.

  By the 1850s the remains of prehistoric elephants had been dug up from locations all over the northern hemisphere for almost the entire span of the Cenozoic (at least as it was then understood). There was the American mastodon, an extremely large variety from Europe with recurved tusks sticking out of its lower jaw named Deinotherium, and the numerous species Falconer had dug up in the Siwalik Hills, but the most widely distributed type was the mammoth. In a study published in 1845, the French geologist Étienne d’Archiac proposed that mammoths ranged from western Europe through Asia, over the Bering Strait into western Canada, and down into the southwestern United States. While they lived during the latter parts of the Cenozoic, divisions of time known as the Pliocene and Pleistocene, as a single species they would have persisted over a huge expanse of time.

  FIGURE 68 - The skull of Deinotherium, one of the first fossil proboscideans ever discovered.

  Falconer suspected that there was something amiss with this pattern, and in his reexamination of fossil elephants he found that other naturalists had perhaps been a bit overzealous in lumping the various fossils into just one species. Using delicate dental characteristics as his guide, Falconer was able to distinguish several species in Europe alone. There was not just one species of mammoth but several closely related types from different places and times.

  What was even more startling, however, was that some of the species Falconer identified seemed to represent intermediate forms. The grinding teeth of the different species could be organized into a linear progression from the rough, bumpy teeth of the mastodon to the flat, crenulated molars of living elephants. Not all the fossil teeth fit comfortably within this arrangement, but the transitional features Falconer identified still created a continuous series of forms.

  Falconer had created what would appear to be an evolutionary series, yet he doubted that any of Europe’s fossil elephants could have been ancestral to later types. His reasoning was that none of the elephant species he had studied showed any variations that would reveal them “approaching,” or smoothly grading into, another species in form. Every species was an island that remained unchanged for long periods of time. This, in Falconer’s mind, ruled out evolution by natural selection as proposed by Charles Darwin. But Falconer did not rule out evolution altogether; he rejected special creation and believed that mammoths certainly were “the modified descendants of earlier progenitors.” The only question was what, if not natural selection, had precipitated this change.

  Many other paleontologists were in accord with Falconer. Even though a variety of fossil elephants were known by the end of the nineteenth century, the early evolution of the Proboscidea, the group to which elephants and their extinct relatives belonged, was entirely unknown.

  Life was not the same everywhere, however, and due to the contingent facts of history the leading paleontologists had primarily worked and studied near areas that preserved relatively recent fossil elephants. The sought-after early ancestral types would be found elsewhere. During the late nineteenth century, England and France gained increasing control of the Egyptian government, though they were resented by the Egyptian people. In 1882 this tension escalated and erupted into the Battle of Tel el-Kebir, in which British soldiers were dispatched to protect the crown’s interests. The British overran the Egyptian soldiers, allowing England to take control of the country and opening the land to visiting geologists.

  At the turn of the twentieth century, the chronically ill British paleontologist Charles Andrews, seeking a warm climate on doctors’ orders, set out to search for fossils in the de facto colony. Just a few years before, Andrews’s colleague Hugh Beadnell had found numerous fossils in the Fayum desert, about eighty-one miles south of Cairo, and with Beadnell’s help Andrews was to find many more.

  Most of the fossils of the Fayum were from much older parts of the Cenozoic than were preserved back in Europe or in Hugh Falconer’s old stomping grounds in India. The deposits were of Eocene and Oligocene ages with many fossils spanning the boundary between the two at about thirty-seven to twenty-eight million years old. In contrast to the modern, arid conditions, though, during this time the Fayum had been home to tropical forests and swamps, and in the vicinity of those swamps lived some of the earliest proboscideans.

  Andrews and Beadnell found several early elephant relatives. There was Barytherium, an animal the size of an Asian elephant that had two sets of four very small tusks oriented upright in its jaws. This was an arrangement similar to that of Moeritherium, a smaller Fayum proboscidean that had an enlarged set of incisor teeth but a build like a hippo. Their neighbors, Palaeomastodon and Phiomia, were decidedly more elephantlike. These forms had longer upper and lower tusks, and situated further back on the skull a nasal opening that undoubtedly supported a rudimentary trunk.

  FIGURE 69 - The skull of Palaeomastodon, with a restoration of its head by Charles R. Knight.

  FIGURE 70 - The skull of Moeritherium, one of the earliest proboscideans known from the Fayum desert.

  While there was some disagreement about the exact placement of these creatures in relation to other fossil elephants, they seemed to solve the mystery of the group’s early evolution. Proboscideans had originated with stout, hippolike forms such as Moeritherium before evolving into the much more elephantlike Palaeomastodon. From there elephant evolution exploded, with a profusion of four-tusked creatures such as Gomphotherium giving rise to mastodons, mammoths, and eventually modern elephants. Thus the great march of elephant evolution seemed all but complete. In a 1914 summary of elephant evolution the American paleontologist Erwin Hinckly Barbour stated,The genealogy of [elephants] is now so well known to naturalists, that it is interesting to note in the writings of Cope and others of twenty-five years ago, that the intermediate proboscideans are entirely lost, and the phylogeny of the order absolutely unknown. As a reward of zeal, the genetic gaps are being filled so rapidly, that ultimately knowledge of the history of the Proboscidea must be as well known as that of the Equidae [horses].

  But not all the known fossil proboscideans fell into this straight line of elephant development. Barytherium, known from only a handful of bones at the time, did not fit the pattern that was expected and was pushed off to the side. Likewise Deinotherium, the large “anchortusked” elephant relative discovered in the early part of the nineteenth century, was also seen as an evolutionary dead end only distantly related to living elephants.

  Then there were the “shovel-tuskers.” During the 1920s Barbour described Amebelodon, a proboscidean with a long, scooplike lower jaw that lived about nine to six million years ago during the Miocene. Even more widely dispersed was a similar form that lived around the same time, Platybelodon, that also had lower jaws and teeth shaped like a shovel. At the time naturalists took this to mean that both Amebelodon and Platybelodon lived in wetlands, where they used their jaws to scoop up soft water plants, but they were so strange that clearly they could not have been ancestral to any of the mammoths or ancestors of living elephants.65,66

  This led paleontologists to use two competing sets of imagery when mapping the ancestry of elephants. The public was presented with an
image of straight-line elephant evolution from smaller, generalized ancestors to larger, specialized descendents. The truth of evolution could not be denied in the face of such illustrations. Yet these images were often accompanied by wildly branching bushes of elephant diversity that housed all the disparate lineages of the trunk-bearing herbivores.

  FIGURE 71 - The evolution of elephants as envisioned during the early twentieth century.

  These seemingly competing forms of imagery visually complemented how many paleontologists viewed evolution during the early twentieth century. Rather than being affected by natural selection, evolution was believed to be driven by internal forces toward particular endpoints. There was a “March of Progress” from the primitive to the advanced, words that were just substitutes for “lower” and “higher,” which was thought to be the main stem of evolution. Any creature that could not be slotted into this trajectory was considered to be a dead-end offshoot, and these creatures typically received comparatively little attention.

  A perceived side effect of the internal perfecting principle, however, was that many lineages made attempts at reaching the same goals. If they failed, like Deinotherium and Platybelodon, it was because they had somehow been knocked off the progressive path evolution was proceeding along. Any traits they shared with similar creatures could be explained away under the aegis of “parallel development.”

  This view was most explicitly outlined by the American paleontologist H. F. Osborn, especially in his massive, posthumously published 1936 monograph on fossil elephants simply titled Proboscidea. As envisioned by Osborn, fossil elephants could be arranged in a palm-frond pattern. From one central point there was an explosion of elephant types, each proceeding in a straight line from primitive to advanced forms during the course of time. There were a few branching points here and there, but the overall picture was one of a diverse collection of elephants all struggling upward toward a particular endpoint.

  Unfortunately for Osborn, even though he had been a powerful force in paleontology in the early twentieth century, by the time of his death in 1935 his evolutionary views were outmoded, even among paleontologists who had worked under him at the American Museum of Natural History, such as W. K. Gregory and W. D. Matthew. The new evolutionary synthesis grounded in evolution by natural selection was emerging.

  The task of reinterpreting Osborn’s evolutionary scheme would be undertaken by G. G. Simpson, one of the chief architects of the new evolutionary synthesis, in his comprehensive 1945 monograph On the Principles of Classification and the Classification of Mammals. In the case of proboscideans, Simpson identified that there was an early radiation of bizarre forms, and while most of them became extinct twenty-three million years ago, at least one branch formed the base for the array of mastodons, shovel-tuskers, mammoths, and ancestors of living elephants. Elephant lineages were not competing with each other to reach an end goal, but a series of evolutionary divergences within a nested hierarchy that could be traced back into the Eocene. Proboscideans were a widespread, prolific group that flourished until very recently, and echoing the statements Falconer made a century before, Simpson wrote: Although this order has living representatives, familiar to all, the two surviving genera are relicts of a dying group.... The order had a much greater role in faunal history than one would dream from this impoverished representation, and it formerly occurred in bewildering numbers and variety over the whole of all the continents, except Australia, and on a number of islands.

  The Proboscidea had not sprung from nothing, however. They were mammals that would have to be anchored within the mammalian family tree. At first glance elephants seem to be different from almost every other kind of mammal. Their long-ago grouping with other thick-skinned mammals, such as rhinos and hippos, as “pachyderms” crumbled when viewed from an evolutionary perspective. With the development of comparative anatomy, however, naturalists began to notice similarities between living elephants and another group of odd mammals, the sirenians, or the aquatic mammals popularly known as manatees and dugongs. While Georges Cuvier insisted that the sirenians were herbivorous relatives of whales, his peer Blainville had identified persisted. Males of both manatees and elephants have internal testes (“the sinews of his stones are wrapped together,” per the description in Job), for example, and females of both groups develop breasts that are closer to their chests than their abdomen. And while elephants took the feature to great extremes, some sirenians also had enlarged, tusklike second incisors. Along with a few other features, such as eyes oriented more forward on the head, these specialized characteristics solidified the relationship between elephants and sirenians. Comparisons of elephant and sirenian DNA show that they are more closely related to each other than any other type of mammal.

  But, as recognized by Falconer and Simpson, the organisms inhabiting the world today are remainders of past radiations, many branches of which have become extinct. Manatees and elephants are each other’s closest living relatives, but both are also closely allied to two groups that have long been extinct: the embrithopods and the desmostylians. The most famous representative of the former, among the numerous fossils Andrews and Beadnell found in the Fayum desert, was an early Oligocene mammal with two immense facial horns called Arsinoitherium . The latter were marine mammals that lived from about thirty million years ago until about seven million years ago. Found primarily at sites bordering the modern Pacific Ocean, these mammals had heads like hippos but bodies more similar to those of sea lions. As with Arsinoitherium and its close relatives, their exact relationships to elephants and manatees are unknown, but the desmostylians do share specialized features that associate them more closely with those groups than with other mammals. Proboscideans, sirenians, desmostylians, and embrithopods can all be placed into a group called the Tethytheria, and each of these groups evolved and dispersed around the edge of the ancient Tethys sea.

  Frustratingly, however, the earliest representatives of each group have been elusive. Moeritherium and Palaeomastodon, creatures thought to represent the beginning of proboscidean evolution, were too specialized to be the earliest elephant ancestors. The first animals that could have rightly been called proboscideans would have been small, lacked large tusks, and had no trunk to speak of. In 2009 paleontologist E. Gheerbrant described what is probably the oldest proboscidean yet found. The sixty-million-year-old remains from Morocco came from a rabbit-sized animal Gheerbrant named Eritherium. It possessed key proboscidean characteristics, such as slightly enlarged second incisors, but there is no sign that it had a trunk. At present it is impossible to know whether it was ancestral to any later species, but given its age it probably represents the form of some of the early proboscideans that evolved in the wake of the extinction of the dinosaurs.

  It was likely a creature like Eritherium that gave rise to the diversity of elephants found in the Fayum desert. Reflected in Simpson’s classification, the relatively early radiation of proboscideans covered a wide range in body size and possessed different arrangements of tusks and trunks. What is especially curious, though, is that some of these early forms appear to have spent a good deal of time in the water.

  Ever since the time of its discovery, the vaguely hippo-shaped Moeritherium has been thought of as a swampdwelling creature. Dead animals can be washed into bodies of water, however, and many fossil animals that were once thought to be aquatic have been shown in later studies to prefer life on land or in the trees. In the case of Moeritherium and its relative Barytherium, though, the best evidence that they spent much of their time in the water comes not from their overall body shape but their teeth.

  FIGURE 72 - The partial skull of Eritherium as seen from the front (top) and underneath (front of the skull facing up).

  In 2007 A. G. Liu, Erik Seiffert, and Elwyn Simons looked at the details of carbon and oxygen isotopes retained within the teeth of Moeritherium and Barytherium found in thirty-seven-million-year-old deposits in the Fayum desert. The amount and type of these isotopes were influ
enced by what kind of food the animals ate (carbon) and even whether they spent a good deal of time in the water (oxygen). (Similar techniques have been used to determine the pattern in which early whales moved further from shore and whether prehistoric horses were grazers or browsers.) This is possible because enamel is often less altered during fossilization than bone, thus preserving a record of the chemical components inside the teeth.

  What the team found was that the carbon and oxygen isotope values of Moetherium and Barytherium were more similar to each other than any other mammals found from the same deposits. The oxygen isotope values in each creature, especially, showed little variation, a pattern seen in aquatic or semi-aquatic animals where much of the isotopic oxygen comes in through the animal’s skin from the surrounding water. Terrestrial animals, which are not typically submerged, have more variable oxygen isotope values as the isotopic oxygen in their bodies comes from a wider variety of sources. While the authors were cautious in their assessment, stating that some factors (such as the amount of water an individual animal drank) could influence the isotope values, it appears that both genera were semi-aquatic and frequently ate water plants. Barytherium stayed closer to shore, chewing on the tough plants that grew at the water’s edge, while the isotope values from Moeritherium suggest that it spent more time in the water and fed on more succulent plants that could only be reached by wading farther out.