“So before everyone begins the big party for ‘Brontosaurus’ and celebrates this huge diversity of sauropod names, let’s hold our horses.”—Donald R. Prothero
The following post is by Donald Prothero, author of The Story of Life in 25 Fossils: Tales of Intrepid Fossil Hunters and the Wonders of Evolution:
Earlier this year, there has been a big buzz with the media reaction to a study by Emanuel Tschopp, Octavio Mateus, and Roger B.J. Benson analyzing the diplodocine sauropod dinosaurs and figuring out their classification and relationships. As I discuss in The Story of Life in 25 Fossils, since 1903 most paleontologists have regarded “Brontosaurus” as just another junior name for the dinosaur properly known as Apatosaurus.
The Tschopp et al. study itself is a landmark in careful anatomical work, analyzing the problem specimen by specimen (a total of 81 specimens used) rather than generalizing based on previous clusterings of specimens, and looking at far more anatomical evidence than any previous study. Naturally, the press missed the significance of the study completely, and focused on just one minor point: the idea that “Brontosaurus” is again a valid name. ALL the publicity, and all the reactions of the non-paleontological reporters and readers was focused on this rather trivial issue, which is not important to real paleontologists in any way (except that we always get asked about it by the general public). Most of the reaction by sauropod paleontologists who were interviewed were generally favorable, but others were more cautious. Almost all agreed that is the most thorough work on the subject written to date, and it will be the foundation on which all future analyses will be built. Similar reactions could be found on the SVPOW (“Sauropod Vertebra Picture of the Week”) website, which is the main forum for discussion by specialists and amateurs about sauropods.
If Tschopp et al. are correct, the diplodocids were incredibly diverse during the Late Jurassic. In fact, in the Morrison Formation alone (Late Jurassic, mainly Colorado-Utah-Wyoming), they record fourteen different species in nine genera (Suuwassea, Amphicoelias, Apatosaurus, Brontosaurus, Supersaurus, Diplodocus, Kaatedocus, Barosaurus, and Galeamopus) of diplodocines from a single formation that covers a limited geographic area and maybe 7-11 million years of time—plus several more specimens that have not been named yet. And that doesn’t count the non-diplodocid sauropods, including the monstrously huge Brachiosaurus, plus Camarasaurus, Haplocanthosaurus (the latter two are distantly related diplodocoids), and possibly several more. In a single quarry alone, Carnegie Quarry at Dinosaur National Monument (representing a single biological fauna and a short interval of time), they would recognize Apatosaurus louisae, Brontosaurus parvus, Diplodocus carnegii and D. hallorum, Barosaurus sp. among diplodocines, plus whatever taxon the specimen known as Carnegie Museum 3452 is (a sister-group to barosaurs), plus Camarasaurus and Haplocanthosaurus. That makes at least eight distinct species of huge sauropods from a single interval of time, all crowding together and sharing common resources.
I realize that this is the cutting edge of sauropod research, and the thoroughness and rigor of the analysis are impressive. But as a paleontologist who has published taxonomy on large terrestrial vertebrates over 35 years, I found some issues troubling. This is an incredibly high diversity of animals for any limited time and place. If we were talking about species of small-bodied creatures, such as insects or rodents, there would be no problem. But huge land animals need lots of room to roam and feed.
Ecological theory and empirical data show that larger species require bigger home ranges. The principle of competitive exclusion suggests that no two species can compete for the same resources, and that problem is magnified for larger animals, which rarely share territory with their own competing populations, let alone closely related genera and species. We know of no examples of large land vertebrates today that exist in high diversity, and compete for the same resources. Even before poaching reduced their numbers, only one really big mammal (the elephant) lived in the savannas of Africa (the most diverse large land-mammal fauna today), and even they have problems if their populations exceed the carrying capacity of the habitat.
The same goes for the prehistoric past. Among communities with large terrestrial mammals, there are no instances of more than two or three huge species (usually an elephant or mammoth) in a fauna at the same time as another huge species, competing for the same resources. Even the vaunted “mammoth steppe” with its incredibly high diversity for an Ice Age tundra-grassland, still supported no more than one species of mammoth at given time and place. Yet sauropods were many times larger than elephants—but a single time slice and locality (Carnegie Quarry) in the Morrison Formation (which had a similar semi-arid habitat to the African savanna, according recent analyses) allegedly supported eight different species of large sauropods.
What we are really talking about is the longstanding argument among taxonomists about “lumping” vs. “splitting”: “lumping” lots of different specimens in a single highly variable species vs. giving a different species name to every specimen that appears slightly different. The dispute goes back to the earliest days of taxonomy, when the pioneering collectors of the eighteenth and nineteenth centuries delighted on erected new species willy-nilly, as sort of a mark of accomplishment, and often driven by the pride of adding to their collection. Such practice has been likened to stamp collecting: great fun, but not very scientific. All the early dinosaurs were named this way, which is why O.C. Marsh named one of his first diplodocines from Como Bluff, Wyoming, Apatosaurus (a juvenile specimen that was incomplete), and then a few years later named a more complete adult specimen from the same locality Brontosaurus. The latter specimen became the famous mount at Yale Peabody Museum, soon copied by the American Museum in New York and Carnegie Museum in Pittsburgh with their fossils, and all these mounts got the name “Brontosaurus” and made it a household name. But as early as 1903, paleontologist and lumper Elmer Riggs realized that these specimens were extremely similar, and that Apatosaurus was a juvenile “Brontosaurus”. By the rules of priority, Apatosaurus is valid and “Brontosaurus” became invalid. So it remained until early this year.
The splitters were dominant in the early twentieth century as well, but by the mid-twentieth century, the trend shifted. Paleontologists such as George Gaylord Simpson, who was also a pioneer in statistics in paleontology, argued that we should think of fossil samples like we think of modern populations of animals, and determine how many valid species there really are based on living biological analogues. Since then, most of the trend has been to lump the many invalid species created willy-nilly by pre-1940s paleontologists, and reduce the chaos of noise and bad taxonomy to get at a true biological signal of diversity.
The famous and powerful paleontologist Henry Fairfield Osborn was a hyper-splitter who named nearly every specimen with the slightest differences from any other specimen a new species. His huge monographs on mammoths and elephants (1936) and the two-horned brontotheres (1929) are monuments of oversplit and just plain incompetent taxonomy, and both have been completely replaced by modern taxonomy that abolished most of his invalid names. The hundreds of amazing skulls of the late Eocene brontotheres of the Big Badlands of South Dakota and adjacent states were once divided into many genera and dozens of species.
Today, Matt Mihlbachler and colleagues recognize only one valid species, Megacerops coloradensis, and all the other famous names (still in the kiddie books) like Titanotherium, Brontotherium, Brontops, Allops, Menops, and others, are no longer taken seriously. When I started working on fossil rhinoceroses over 35 years ago, nearly every specimen with slightly different teeth was another species. But a large quarry sample of a single population from Trigonias Quarry in the upper Eocene beds of Colorado showed that all those variations in the crests and cusps of the upper premolars are due to intrapopulational variability, and since my 2005 rhino monograph all those long invalid names have vanished. Likewise, the gigantic indricothere rhinoceroses of Asia are now all in the genus Paraceratherium (replacing junior names like Baluchitherium and Indricotherium) for exactly the same reason: these huge animals would have been much like elephants with small populations roaming large distances to find enough treetops to browse on, and it’s extremely unlikely that more than one species could co-exist in the same place and time.
This is not to say that every recent study automatically leads to lumping. The hominid fossil record was once grossly oversplit, when only a handful of fossils were known and each one got its own name. Then during the 1960s, paleoanthropologists went in for the trend toward lumping and perhaps overlumping on the grounds that only one human species lives on the planet today. But by the 1970s and 1980s, so many different fossils had appeared that there are more and more taxa that appear to be valid. Another example is a new paper I just published, where I recognize a whole new subfamily of extinct peccaries, the Hesperhyinae, with seven distinct genera (four new) and seven species (two new). But I based this analysis on a bunch of new unstudied specimens. In addition, I could show that there were male and females in each species (their tusks are distinctive), and I had large population samples of several species. I also accounted for the high degree of variability in the teeth, which are not as diagnostic as once thought (just as in the rhinos and so many other mammals). In addition, none of the species overlap in time or space, which is a strong grounds for distinguishing them—just as the overlap in the same quarry is grounds for questioning whether more than one species is present.
Could this apparent high diversity of sauropods in the Morrison Formation also be due to oversplitting and failure to account for high intrapopulational variability? The biological and ecological arguments suggest this. To cite a similar example, Mihlbachler et al. (2004) did a careful study of African giraffes (long-necked, large-bodied animals a bit like sauropods), and found that they showed huge variability not only in cranial appendages (their horn-like ossicones), but in the proportions and shapes of the skull, features of the neck, etc.—precisely the same features that are used to justify so many different sauropod taxa. Yet giraffes are all one species (Giraffa camelopardis), with local geographic subspecies, but no one argues that they require different species or genera for these differences.
Reading Tschopp et al. (2015) carefully, it appears that they have apparently accounted for juveniles and changes in shape due to growth of immature individuals, which is reassuring. But another possible source of variability might be sexual dimorphism, which I do not see mentioned at all in their paper. That is a tough issue, since it’s hard to tell whether fossils are truly male and female in most cases (except for those like deer where we have good modern analogues). A recent paper attempting to demonstrate sexual dimorphism in stegosaurs was harshly criticized on a number of grounds (not the least of which it used undescribed, unpublished specimens in private collections and not available for other scientists to study—a big no-no).
The dinosaur paleontologists are aware of the problem, of course, and there are a number of people who have speculated as to how explain so many closely related gigantic herbivores in the same place and time, when the resource base of the dry savannas of the Morrison Formation were a very meager food base to support even one gigantic herbivore. The most recent effort, by Button, Rayfield and Barrett (2014), argues that the skulls of these fossils are just different enough in their feeding mechanics that they could have specialized on eating different types of trees and foliage. Such studies assume that the taxonomy is valid and there really were a large number of huge species living together, and therefore this has to be explained somehow. Maybe that’s plausible, but it still doesn’t fully address of how so many different species of gigantic animals, with monstrous appetites, made a living at the same time and place on just the foliage of tree tops and maybe some ferns. Remember: this is the Late Jurassic, when there were no flowering plants or fruits, no grasses, and most of the trees were conifers and cycads with tough fibrous slow-growing needles and fronds, not rapidly-growing nutritious leaves and fruits of flowering plants.
So before everyone begins the big party for “Brontosaurus” and celebrates this huge diversity of sauropod names, let’s hold our horses. Think outside the box for a moment. Is there any way to look at the sample of specimens from Carnegie Quarry as a single population of diplodocines (possibly some are males and some are females), or are the anatomical features completely incompatible with this idea? Keep the example of the extreme variability of giraffes in mind. Until someone has convincingly addressed the issue, I’m sticking with what I wrote in The Story of Life in 25 Fossils. I will continue to put “Brontosaurus” in quotes and not follow the latest media fad, nor will I overrule Riggs (1903) and put the name in my writings as a valid genus.