Witness To Creation

In his beautifully illustrated book Serengeti, wildlife photographer Mitsaki Iwago portrays immense herds of migrating wildebeests as they ford rivers or endure drought, often dying by the thousands, leaving behind carcasses that in turn make it possible for the region's scavengers, hyenas and vultures, to survive. The pictures serve as a vivid reminder that life and death are permanently entwined, a never-ending tango, sometimes painful, other times pleasurable, but in the absence of which no organism, neither you nor I, could exist.

Dinosaur paleontologists, it seems to me, are also scavengers of a sort. We labor in the graveyards of evolution, picking over bones, searching for clues to the lives of creatures that perished long ago. Until recently we were limited mostly to single specimens. On occasion, groups of up to twenty or thirty individuals have been found. And even less frequently, we've stumbled upon sites where the remains of hundreds of different kinds of dinosaurs have been washed together. But nothing like the Serengeti herds had been unearthed, preserved as fossils, until we discovered immense rookeries and monospecific bone beds in sediments of the coastal plain that lay between the Rocky Mountains and the Western Interior Seaway during the Cretaceous period.

Hadrosaurs, lambeosaurs, and ceratopsians, in particular, but others as well, including small theropods and hypsilophodontids, evidently lived together—nested in colonies, migrated in herds, or hunted in packs. If nothing else, our excavations over the past twenty years have shown that many kinds of dinosaurs were much more social than previously thought. I've already talked at length about nesting, and described what little we know about hunting, but migration warrants a closer look. Years ago, when we had only the Maiasaura data to consider, I conjectured that the duck-billed dinosaurs migrated along an east-west route, laying eggs and raising their young in the uplands, then moving to the lowlands to live. I was led to that viewpoint by a bias in the fossil record—the apparent absence of eggs in lowland sediments. But by the early 1990s, when we excavated several nesting grounds in the Judith River Formation outside Havre, it had become clear that the herding dinosaurs, both horned dinosaurs and duckbills, were following an overall north-south trend.

In all likelihood the mass movements were driven by seasonal variations, just as the migrations of plant-eating animals are today. Notwithstanding the comparatively warm temperatures that persisted throughout the Cretaceous period, permitting crocodiles to survive as far north as Alberta, the angiosperms, or flowering plants, that had begun to flourish by that time would have lost their leaves periodically. Add to that the impact that hundreds, even thousands, of foraging dinosaurs, each weighing between two and four tons, would have had on the local vegetation and it's hard to imagine how the herds could have survived without moving regularly, if not constantly. To remain in one place would have caused mass starvation.

Precisely how far the herds migrated is uncertain, though there seems to have been some measure of variation. To date, large groups of maiasaurs have been found only in Montana, and only in the Two Medicine Formation, suggesting that that particular dinosaur confined its migratory travels to the upland plains. By contrast, the remains of Edmontosaums, one of the largest duckbills, have been found in Wyoming, eastern Montana, Alberta, and the North Slope of Alaska, which is about as far north as one could have traveled then without crossing into Siberia. It's unlikely, of course, that the edmontosaurs of Alaska actually migrated to Wyoming, or vice versa. Distances of that magnitude are beyond the capacity of wingless creatures. But imagine, if you will, the time of the year when at extreme northern latitudes the sun scarcely rises above the horizon all day long. Would thousands of duck-billed dinosaurs, weighing five to six thousand pounds and ranging up to thirty-five feet in length, have remained there, milling about until the darkness lifted? I strongly doubt it. During that more temperate time there was no snow or ice to prevent them from moving. I think they would have followed the sun, migrating at least far enough southward to be able to see the vegetation they depended upon for survival.

Despite the uncertainty that remains regarding the direction and extent of migrations among duck-billed and horned dinosaurs, I have no doubt that they traveled in large social groups, very much like the bison of pre-Columbian North America and Africa's wildebeests, zebras, and kudus. That's the first thing the mass mortality sites tell us. They also tell us, as Ray Rogers determined in his convincing reconstructions of local environments in the Two Medicine Formation, that herding and migrating were survival strategies that also conferred disadvantages, all of which are variations on the following principle: Any threat to one was a threat to all. When drought or disease arrived, a flooded river had to be crossed, or violent volcanic eruptions occurred upwind in neighboring mountains, whole herds were placed at risk and consequently large numbers of animals perished at once.

And judging from the fossils found in the bone beds, death was indiscriminate. Among maiasaurs, for instance, we found skeletons that vary from nine feet to the maximum known length in adults— about twenty-five feet. (The fate of the intermediate juveniles, those large enough to leave their nesting grounds, which we estimate were about three and a half feet long, yet were smaller than the smallest members of the herds, is unknown. We haven't found their remains anywhere.) Besides duck-billed and horned dinosaurs, another group that seems to have coevolved with the angiosperms are the tyrannosaurs. The evidence is persuasive, I believe, that these large carnivorous theropods scavenged the carcasses of the social plant eaters whenever presented with the opportunity, which is to say whenever groups of the latter succumbed to the weather and other environmental hazards. In short, tyrannosaurs followed the herds, or at least remained in their vicinity, just as hyenas and vultures do on the Serengeti Plain today. Consider the scene: a mile-long parade of maiasaurs, thousands of them, crossing a stream made swollen and raging fast by relentless seasonal rains. Around the herd, at an inconspicuous but perceptible distance, stand tyrannosaurs and other scavengers, and they are aroused, nervous with anticipation. Quite an image, isn't it? To an insatiable old bone scavenger like me, it's even more arresting than Iwago's impressive photographs.

I'm now going to ask you to return with me to the graveyards of the Two Medicine Formation, not the nesting grounds but the sites where large numbers of young and adult dinosaurs died at the same time. I want to show you something that profoundly and permanently altered the way I think about the late Cretaceous ornithopods and ceratopsians of North America. Pick any one of these bone beds—the gigantic group of maiasaurs buried by volcanic ash at the Willow Creek anticline; the horned dinosaur herds that fell victim to drought and opportunistic disease near Landslide Butte, along with the groups of hypacrosaurs and prosaurolophs that died en masse in that area as well; or the gryposaur and maiasaur kill sites near the Two Medicine River. What do you see when you walk through the graveyards? Well, for starters, everything that we have already discussed—evidence of social behavior, geological clues to the types of environments the dinosaurs inhabited, indications of their cause of death. But have we missed anything?

Let's step back for a moment and reconsider what lies before us, to see if there isn't something else the graveyards might reveal, just as I did one day at a place we call Thunder Dome, a barren, bowl-shaped hill in the Landslide Butte badlands. If it's true that all the bone beds in the upper Two Medicine Formation represent herds of one size or another, then what exactly does that say about the behavior of the herd beyond the fact that they lived together, sharing resources? What else, in other words, do herds offer to its members besides protection? While pursuing that line of thought a realization dawned on me: If a group of similar animals travels together chances are it also interbreeds. In biology a group of interbreeding organisms is called a population. Moreover, it's considered the level at which speciation takes place. We're going to try to avoid that term, of course. The taxonomic name we use to refer to herds of maiasaurs, hypacrosaurs, and gryposaurs isn't important. What's important is that within the herd, individual animals were mating and producing offspring.

We can't be certain of this, surely. But I think it's a reasonable assumption. The bone beds contain the remains of interbreeding populations, or at least populations that were capable of doing so. As I sat on Thunder Dome thinking about the graveyards of the Two Medicine Formation it further occurred to me that while this isn't exactly what Ernst Mayr had in mind when he proposed his operational definition, it's as close an approximation as one can hope for with respect to fossils. It's certainly closer than I'd ever imagined getting in the study of dinosaurs. About the reproductive behavior of two or three or five extinct plant-eating animals there's nothing anyone can say that isn't pure speculation. The same is true of small groups of carnivores, the Deinonychus packs that preyed on tenontosaurs, for instance. But very large aggregates of plant eaters that exhibit all of the other characteristics of migrating herds? That's another situation altogether. Because the monospecific bone beds are populations in the full, biological sense of the term, they offer information that individual specimens and other kinds of aggregates don't.

The most important new information by far is the degree of individuation within the population. By that I mean the degree of morphological difference that thousands of animals might display and still be able to reproduce. To be specific, within the graveyards lay cross sections of the range of variation of crucial skeletal characters, for example, the length of ceratopsian horns and the shape of their neck shields. What is the value of such information? To illustrate, let's say that the horns on fully grown adults in one population range between thirty and thirty-six inches. If you find another ceratopsian outside the bone bed whose horns fall within the established limits for that character and that in every other important respect resembles the skeletons in the bone bed, then very probably it's a member of the same population. This is even more likely if the lone ceratopsian is recovered from the same sedimentary layer, which means that it lived at the same time. In short, knowing how much key characters vary within a population enables one to form hypotheses about the hereditary relatedness of different groups of dinosaurs.

Dinosaurs have never been subjected to this kind of analysis before. No one has made the attempt. In paleontology, population studies have been restricted almost exclusively to marine invertebrates, and for good reason: The fossils are small—indeed, tiny— by comparison with just about every kind of vertebrate, and thus relatively easy to excavate. What's more, there are tremendous numbers of them in the geologic record. Finally, the environments in which they were deposited—sea-bottom silt and mud—tend to be very stable, yielding sedimentary rock in which the evolutionary time frame is more clearly and completely preserved. Stephen Jay Gould, who has devoted a great deal of attention to ancient snails, among other seagoing creatures, is probably the best known student of extinct populations and surely one of the most provocative. His research will always be more quantitatively rig orous than mine. As dinosaurs go, the bone beds of the Two Medicine Formation are unusally populous, yet under the best circumstances, any conclusion I might reach about them will be measured in terms of tens of individuals. Invertebrate paleontologists, on the other hand, have access to thousands, even tens of thousands of the same kinds of specimens. Snail shells, I'm afraid, will always greatly outnumber ceratopsian skulls.

Even so, I think there's a decisive advantage to studying populations of dinosaurs. Like all other large vertebrates, ceratop-sians retain characters whose functions are readily recognizable. Consider the horns and neck shields I spoke of earlier. Such prominent anatomical features are known collectively as sexual ornamentation, and their purpose, like that of elaborate elk antlers and brightly colored peacock plumage, is to attract prospective mates. The presence of horns on ceratopsians, then, provides an opportunity to observe sexual selection, the evolutionary process by which certain kinds of ornamentation change through time under the influence of environmental pressure. But what constitutes sexual ornamentation in snails? How do we determine which male features the female finds attractive? I'm not saying it's impossible to study sexual selection in snails but it's surely much harder than in dinosaurs. Indeed, morphological variation is so much easier to detect and measure in dinosaurs than in snails that any ideas we infer from them will be appreciably less equivocal and hence a good deal more relevant to our understanding of evolutionary processes. A sample that runs to ten thousand may give one statistical confidence in what one says about the sample, but if the specimens are so difficult to study that what can be said is inconsequential, what difference do the numbers make?

In the years since that modest epiphany on Thunder Dome, another dimension has been added, the last that we required to complete the reconstruction of dinosaur lives on the coastal plain. The work that Ray Rogers and I have done, characterizing environmental changes that took place during the late Cretaceous, especially events tied to the periodic advance and retreat of the Western

Interior Seaway, has provided a more detailed historical framework for understanding the dinosaurs that inhabited the coastal plain. Since the 1960s, certain geologists and paleontologists have speculated about correlations between the dynamics of the seaway and the evolution of terrestrial organisms living immediately west of the seaway (marine organisms were likely affected as well, and thus have been included in some of the speculations, but that's not our concern here). None of the proposed models for what transpired on the coastal plain during the late Cretaceous has been convincing, however, because no one had found terrestrial populations of sufficient size and distribution in the fossil record to support the models.

This changed when we discovered large dinosaur bone beds in the Two Medicine Formation. Not only did we have populations in which the evolutionary impact of environmental pressure could be observed, we had, thanks to Ray's work, populations whose positions in the overall historical context were clear. And position in this instance is of the utmost importance, because the bone beds represent crucial points during the application of environmental pressure along the coastal plain—namely, very near to a maximum point of regression, when the seaway had shrunk to its smallest size, and very near to a maximum point of transgression, when the seaway had increased to its largest size, flooding most of the land east of the Rockies.

The themes raised by these developments are the most difficult in all of paleontology, more than that, in the life sciences as a whole, and for that reason they often are avoided or taken up only by those of a theoretical turn of mind. These themes include the emergence, diversification, and disappearance of new characters and, ultimately, new kinds of organisms, as well as the ecological factors that affect such evolutionary changes. Even more uncommon is any attempt to specify the rate, direction, and mechanisms of evolutionary change, especially in the case of dinosaurs. Indeed, the bone beds of the Two Medicine Formation provide the first opportunity to do so. Viewed from the proper vantage point, those graveyards can be seen as a series of stages upon which the very processes of life itself become visible. Sitting atop Thunder Dome, we are granted the surpreme privilege of observing creation at work.

Thunder Dome, then, is where we finally gained enough resolving power to bring the evolutionary landscape into focus. This was made possible, it bears repeating, by the discovery of several large dinosaur bone beds that could be treated as populations, which in turn permitted us to observe transformational sequences, or change through time, at the level at which such change actually takes place—among individuals. But it wouldn't have happened without our also having developed a much more detailed map of the Two Medicine Formation, which enabled us to identify the precise locations of those populations in the geologic column.

Besides placing too much faith in the Linnaean taxonomic system, paleontologists have been too quick to embrace the system geologists use for classifying sedimentary rock. They sometimes forget that the development of that system was driven not by the search for fossils but by the search for fossil fuels. When petroleum geologists set out to describe the sedimentary rocks of Montana, they divided them into units, or formations, that made sense geologically. And as an aid to understanding the geological history and composition of the region, their maps and cross sections are very valuable—to the paleontologist as much as to anyone else. But as a method for categorizing organisms, geological formations can be misleading, in the same sense that the concept of species is misleading. They fail to resolve the geologic record to a fine enough scale to permit transformation sequences among fossils to become apparent. Unfortunately, some paleontologists continue to publish papers with titles like, "The Vertebrate Fauna of the Judith River Formation," implying that all of the animals catalogued in the paper lived at the same time, thereby further implying associations that do not exist while obscuring others that do. The Judith River Formation spans about five million years. A whole lot of evolution can go on in five million years.

The reason there's been so much speculation about the relation ship between the Western Interior Seaway and terrestrial life along the Rocky Mountain Front, however, is that even when dinosaurs are lumped together by formation, the fauna that lived before the transgressions, especially the first and third, differ so markedly and with such consistency from those that lived after the associated regressions that it seems the rising and falling sea was somehow responsible for the shifts, a veritable engine of evolution. The pattern is simple, with the greatest degree of biological change following the most extensive environmental disruption, and that was the transgression represented today by the shales of the Colorado Group. (You might want to flip back to the diagram on page 47 as we take a second walk through this period of geological history.)

Picture once again the large-scale environmental events that took place on the coastal plain during the Cretaceous period. About 97 million years ago the seaway started to rise, gradually flooding the lowlands and, eventually, the uplands as well. The Colorado expansion continued for ten million years, at its greatest extent even invading the valleys of the newly forming Rockies. When the seaway stopped rising and began to recede, all of the land east of the Rocky Mountain Front was under water. There was no coastal plain 87 million years ago. All of the terrestrial habitat along the Rocky Mountain Front had been converted into aquatic habitat. The regression that followed occurred relatively quickly, over a period of about four million years. The second transgression, which started 83 million years ago and reached its zenith 79.6 million years ago, was comparatively small. But the third, which started 75.4 million years ago and reached its zenith about 74 million years ago, also significantly flooded the plain (unlike the first transgression, though, the third did not cover it entirely).

Let's also remind ourselves of the geological structures these events left behind. The terrestrial sediments deposited during the first transgression are known as the Cloverly Formation in the eastern part of the coastal plain and as the Kootenai Formation in the western part; the marine sediments are called the Colorado Shale. The terrestrial sediments deposited during the second transgression are known as the Eagle Sandstone in the eastern part of the coastal plain and the lower Two Medicine Formation in the western part; the marine sediments are called the Claggett Shale. The terrestrial sediments deposited during the third transgression are known as the Judith River Formation in the eastern part of the coastal plain and the upper Two Medicine Formation in the western part; the marine sediments are called the Bearpaw Shale. Above the Bearpaw Shale, consisting of terrestrial sediments deposited during the last regression of the seaway and afterward, are, to the east, the Hell Creek Formation and, to the west, the St. Mary River Formation.

Here's the pattern that, for some time now, has captured the imagination of paleontologists: In the Cloverly Formation we find two groups of ornithopods—tenontosaurs and hypsilophodon-tids; nodosaurs, which were a primitive type of ankylosaur, or armored dinosaur; the small theropod Deinonychus and the large theropod megalosaur; and various sauropods. These are the animals that lived on the coastal plain prior to the first and largest transgression. By the time the seaway receded, millions of years later, the cast changed substantially. Dinosaurs found in the Judith River Formation, which contains the remains of creatures living in the lowland plains prior to the third and final transgression, include hypsilophodontids as well as theropods—tyrannosaurs, for instance, and raptors.

But alongside these are entirely new kinds of dinosaur, pachy-cephalosaurs and their relatives the ceratopsians, as well as hadrosaurs and lambeosaurs. Where did the horned dinosaurs, in particular, come from? And how about the duckbills and crested duckbills? Obviously they are related to the tenontosaurs of the Cloverly Formation, but how exactly? The faunal changes that followed in the wake of the Bearpaw transgression weren't as striking but significant nonetheless. There are no traces of nodosaurs or lambeosaurs in the Hell Creek Formation, and new kinds of tyran-nosaur, hadrosaur, and ceratopsia have appeared, once again raising questions about a possible relationship between the activity of the seaway and the fate of dinosaurs on the coastal plain.

It appears, then, that we have three distinct fossil assemblages that don't vary within themselves but differ greatly from each other. What's more, they are separated by varying periods of time, measured in the millions of years. Taken at face value, this is precisely the kind of pattern that inspired the punctuated equilibrium theory. The entire fossil record in fact looks the same—groups of very different organisms separated stratigraphically from each other, a record in which the intervals of silence are longer and more numerous than the intervals that offer information. In this instance, however, there's unmistakable evidence of environmental events of a scale and an intensity sufficient to account for the apparent jumps in evolution. Those who subscribe to Eldridge and Gould's perspective might venture to propose that the "speciation events" that presumably took place along the Rocky Mountain Front during the Cretaceous period, though undetectable, were governed by the dynamics of the Western Interior Seaway. But that's all they would venture. And even at that, at the level of resolution represented by the three fossil assemblages, the proposal is untestable. The fossil record doesn't provide enough of the right kinds of information.

I wasn't satisfied with our ideas about the evolution of dinosaurs on the coastal plain. They were little more than guesswork, guesswork of a sophisticated order, to be sure, and with which I was largely in agreement, but guesswork all the same. In particular, I was convinced that we didn't need to conjure up a deus ex machina to account for the three bursts of novelty in the fossil record. If we studied the terrestrial sediments in greater detail—increased the resolution, as it were—we would most likely uncover transformation sequences, that is, evidence of variation among populations within the faunal assemblages of each formation that would in turn reveal the manner in which the different populations are related. Even more important, we would at long last be able to hear the music that played during the evolutionary waltz on the coastal plain—in other words, actually identify the mechanisms that linked the lives of dinosaurs to the rhythms of the sea.

We couldn't search for fossils everywhere at once, of course, so the first task was to select the most promising sediments to explore. From a practical standpoint, the evolutionary changes associated with the Colorado transgression appear too difficult to study, at least for the time being. For one thing, the stratigraphic distance between the Cloverly and Judith River formations is so great, representing twelve to fifteen million years, that other factors beside the advance and retreat of the seaway may have contributed to the changes that occurred between the older and younger faunal assemblages. Determining whether that's the case and sorting through all of the possibilities would require an enormous amount of research. For another, apart from a few notable exceptions, the Cloverly Formation has yielded a small number of taxa fossils; the Kootenai Formation, very few; and the valley sediments, representing the area to which the animals that survived the flooding of the coastal plain would have retreated, no Cretaceous dinosaur remains at all (not in Montana or Alberta, at any rate). Likewise, the intervening lower Two Medicine Formation and Eagle Sandstone—which are, respectively, the upland and lowland terrestrial sediments deposited the first time the seaway retreated and the plain expanded—have produced very little relevant fossil material.

In rock deposited more recently, by contrast, dinosaur bones are plentiful. Most of the skeletal remains that Gilmore and Brown collected, for example, are from the upper Two Medicine Formation, and during the past twenty years I and my crews have recovered thousands upon thousands of specimens from the same sediments. Indeed, thanks to that work, which was motivated at first by a desire to find eggs and babies, and the fact that the formation as a whole is well preserved—some two thousand feet representing six to seven million years—in north-central Montana, I found myself in a position to examine evolutionary processes that corresponded with the Claggett regression, starting 79.6 million years ago and ending 75.4 million years ago, about four million years later, and the Bearpaw transgression, when the seaway west

Rocky Mountains _

Cross section showing the top of the Two Medicine Formation, and the relationship of the four centrosaurine skulls. D is Pachyrhinosaurus. The arrows show the direction the shoreline was moving.

reversed course and began to rise again, reaching its maximum point of incursion onto the coastal plain 74 million years ago, about one and a half million years later. To be specific, our excavations at Landslide Butte and the Two Medicine River enabled me to resolve the upper Two Medicine Formation into several groups of fossils, each representing an upland population, some of which are separated from the others by various lengths of time.

As you may recall from the fieldwork chapters, the Landslide Butte sites include three new groups of horned dinosaur (from three stratigraphically discrete layers in and around Canyon Bone Bed, one of which is identical to the sediments of nearby Dino Ridge). To avoid the impression of taxonomic hierarchy, I'm going to refer to these new horned dinosaurs as centrosaurines, because that's the family to which they and their apparent relatives belong. Included among the other Landslide Butte sites are one each of Hypacrosaurus stebingeri (Lambeosite) and

Prosaurolophus blackfeetensis (Westside Quarry). The dinosaurs at the Two Medicine River sites are Gryposaurus latidens (Hillside Quarry) and Maiasaura peeblesorum (West Hadrosaur Bone Bed).

The three centrosaurine bone beds at Landslide Butte, dating to a several-thousand-year-long period about 74 million years ago, coincided with the final stage of the Bearpaw Transgression. Those bone beds therefore represent a chronological series of dinosaur populations being subjected to environmental pressure—significantly reduced land surface in the face of an advancing sea—at its most extreme. Although the exact point when the flooding stopped and the water began to fall again has not been located, it's been estimated that at that time the plain was only thirty to fifty miles wide, as little as one-eighth of its original size. Let's begin there, then, at Canyon Bone Bed and Dino Ridge, among the centrosaurines that, along with all of the other animals inhabiting the coastal plain about 74 million years ago, were being crowded into a progressively smaller habitat and thus, as time went by, competing for fewer and fewer resources.

Although back in chapters 4 and 5 I described the skeletal variation exhibited by the three different groups of horned dinosaur, I'm going to quickly recap that information now, since it is the foundation for my argument about the evolutionary consequences of the Bearpaw Transgression. Don't forget that each group represents a population that perished together, which allowed us to determine the maximum variation of any particular character within the population. Also keep in mind that we scoured the area over a period of five years, collecting thousands of specimens, without finding any evidence for overlap among the three populations. Each was recovered from a separate stratigraphic horizon, which is to say that they lived at different times. This gave us a basis for comparing characters that varied from one population to another through time. And in the case of the horned dinosaurs of Landslide Butte, the characters of interest are features of the skull, the horns and shields especially.

The most primitive of the three specimens has a long, forward-

0 0

Post a comment