Vertebrates, Phylogeny, and Philosophy

Edited by Kathryn M. Flanagan and Jason A. Lillegraven


Dr. George Gaylord Simpson, one of the most important evolutionary biologists of the Twentieth Century, was born on June 16, 1902 and died on October 6, 1984. His contributions to science include not only a modern synthesis of evolutionary thought, but original research on anthropology, mammalogy, paleontology, general biology, and statistics. His prolific writings were intended for scientific and non—scientific communities alike. He helped and encouraged many who now work in the fields of paleontology and evolutionary biology. Contributors to this book dedicate their efforts as tribute to his memory.

Included authors are colleagues, former students, and friends of Dr. Simpson’s. They represent but a few of the people he would have included in these categories. The book is intended to suggest only a sampling of the diversity of George Gaylord Simpson's impact on present vertebrate paleontology, from its most senior to its very junior participants.

Ms. Flanagan’s letter of invitation entreated the following from potential authors: "In the spirit of Dr. Simpson’s own writings, we encourage imaginative contributions that would be just a little different from items expected in a regular scientific journal." The title of the volume (Vertebrates, Phylogeny, and Philosophy) reflects that request. Though individual articles deal almost exclusively with fossil mammals, emphases cross the spectrum of evolutionary biology, including systematic paleontology, considerations of adaptation, ontogeny, analyses of evolutionary tem— po and mode, biogeographic procedure, and paleogeography. Philip Gingerich’s contribution stresses the crucial importance of solid empirical research to the foundations upon which theoretical/philosophical writings should be based. Mesozoic and Cenozoic taxa are considered, and two articles discuss the modern union of molecular biology, genetics, and paleontology. Most articles benefited directly from the pioneering writings of George Simpson, yet the breadth of concerns of this volume covers only a small fraction of the interests exhibited in his lifetime of evolutionary research.

Kathryn Flanagan served as principal correspondent with authors and reviewers. Jason Lillegraven had principal responsibility for manuscript editing and considerations of production.

We take this opportunity to thank the thirty-two authors for their contributions. Similarly, more than fifty individuals served as unpaid reviewers, and we give our most sincere thanks for their generosity of time and effort. Also, we thank Linda E. Lillegraven for creating the cover design.

  1. Page 1

    I find a proper introduction of George Gaylord Simpson in the citation for the honorary degree of Doctor of Science which the University of Arizona awarded him in 1982.

    “Simpson is not only a paleonologist rare amongst his colleagues, he has made himself master of all the disciplines involved in the synthetic theory of evolution and particularly of taxonomy which makes him a great biologist. He is not only a biologist but a man of science with the widest horizon and experience.”

    There is not adequate space to even outline Simpson’s magnificent scientific achievements. He did within 50 years travel to every continent and every state, usually accompanied with his partner-wife, Anne Roe Simpson, who not only made significant discoveries of vertebrate fossils but read and critiqued his writing. Surely a heroic task, for Simpson authored some 800 books and articles.

    An event destined to have great influence on the life of Simpson occurred in 183 1. In that year the 22 year old Charles Darwin enlisted on H.M.S. Beagle to participate in a voyage that took four years to encircle the globe. Early on that voyage a stop was made in lower South America which enabled Darwin to explore parts of Patagonia and bring back a significant collection of fossils. One century and one year later the youthful Simpson followed in Darwin’s Patagonian steps and beyond them, bringing back the most important collection of vertebrate fossils yet found there. “Attending Marvels” is the delightful and informative record of that trip and remains one of his most popular books.

  2. Page 3

    George Gaylord Simpson published some 21 books and monographs, 79 notes, and 271 research articles from 1925 through 1971. This primary literature totals 371 titles and 12,656 pages; 4,451 pages (35%) are devoted to mammals, and 2,363 pages (19%) are devoted to evolution. Simpson published primarily on Mesozoic and Paleocene mammals, but he also contributed significantly to the study of Eocene and Pleistocene mammals as well. Early work was concentrated on North American faunas, but interest later shifted to South America. Simpson published some 224 titles and 5,785 pages of empirical work, much of it during the first 20 years of his career. He published 109 titles and 6,675 pages of theoretical work. Research collections and museum support were important throughout Simpson’s working life. The concentration of empirical research early in Simpson’s career, with later emphasis on theoretical questions, affirms that observation and experience are important in generating ideas of lasting value.

  3. Page 11

    The holotype specimen of Laopteryx prisca (YPM 1800), described by Marsh in 1881 as the cranium of a North American Jurassic bird, has been cited several times by subsequent authors as of unlikely avian identity or a possible pterosaur. Careful re-examination and illustration of the infrequently examined type specimen, and a review of the characters noted by Marsh, refute the avian identification and confirm later suspicions. A single tooth originally associated with the cranium, long thought to be lost, is also re-studied and illustrated, and is judged to be unrelated to Laopteryx.

  4. Page 21

    Endocranial casts of Mesozoic mammals and of some cynodonts are reviewed. New tentative reconstructions of brains of Probainognathus and Therioherpeton are given. It is claimed that the endocast of Amblotherium is an artefact. Brains of Mesozoic mammals were lissencephalic, with no flexure, had very large olfactory bulbs, relatively extensive cerebral hemispheres diverging posteriorly, and large paraflocculi. Within this pattern two types are designated: the cryptomesencephalic type (large vermis, no dorsal midbrain exposure, and no cerebellar hemispheres) which occurs in Triconodonta and Multituberculata; and the eumesencephalic type (wide cerebellum, cerebellar hemispheres, and large dorsal midbrain exposure) which occurs in Cretaceous Tribosphenida. Overlap of the midbrain took place in individual lines of the Tribosphenida at different times during the Tertiary. If advanced cynodonts (e.g., Probainognathus and Therioherpeton) had narrow cerebellum and exposed midbrain, then both types could develop from them: the cryptomesencephalic by overlap of the midbrain by an enlarged vermis, and the eumesencephalic by acquisition of enlarged cerebellar hemispheres. If, however, the midbrain was overlapped in advanced cynodonts, then they belong to the cryptomesencephalic type. If so, the eumesencephalic type would have developed from cryptomesencephalic by secondary exposure of the midbrain and acquisition of enlarged cerebellar hemispheres. This latter is less likely, as it would involve the reduction of an already expanded vermis. The expansion of cerebral hemispheres suggests that neocortex was possibly present in all Mesozoic mammals and in some cynodonts.

  5. Page 35

    The mammalian stapes is the subject of considerable investigation, but ambiguity remains with respect to the primitive condition in higher-level mammal clades and the pattern of subsequent modifications. We question the widely held belief that the strongly bicrurate, stirrup-like stapes represents the ancestral mammalian and therian state (Goodrich, 1930). The primitive stapes in the common ancestor of mammals and therapsids was probably columelliform and had a stapedial foramen for the passage of the stapedial artery. However, additional modifications are required to produce the bicrurate structure characteristic of many eutherians. Moreover, distributional evidence does not rule out the possibility that the columelliform-imperforate stapes seen in adult monotremes was ancestral for therians or a group comprising therians and monotremes.

  6. Page 55

    Upper molars of two new genera and species of mammals (Falepetrus barwini and Bistius bondi) are described from rocks of Late Cretaceous age of Montana, Wyoming, and New Mexico. Although these teeth are evolutionarily advanced in being fully tribosphenic, they have combinations of characters that preclude identification of the animals that bore them as either marsupials or eutherians. A review of available dental features useful in classification of known, fully-tribosphenic mammals from the Late Cretaceous suggests the presence of four principal groups: (1) Marsupialia; (2) Eutheria; (3) “deltatheridians;” and (4) others. Groups “ 1 ” and “2” are recognized as formal taxonomic units, defined by anatomically diverse suites of derived characters. Groups “3” and “4” are recognized informally and, as a grade, dubbed “tribotheres,” mammals with tribosphenic dentitions lacking documented specializations characteristic of either marsupials or eutherians. Although group “3” may represent an evolutionary clade equivalent in taxonomic rank to marsupials or eutherians, members of group“4” (including Falepetrus and Bistius) comprise a heterogeneous conglomeration whose members have uncertain relationships to members of groups “1-3.” In addition to the evolutionary radiations of contemporary marsupials and eutherians, the tribotheres provide evidence of at least a third, if not several, broad mammalian radiations during the Cretaceous. However, available dental criteria are inadequate to allow development of a useful, phylogenetically-based classification of the tribotheres.

  7. Page 87

    The genus Catopsalis Cope includes eight species (C. matthewi, C. catopsaloides, C. joyneri, C. alexanderi, C. foliatus, C. utahensis, C. fissidens, C. calgariensis) spanning Late Cretaceous through late Paleocene/early Eocene time on two continents, Asia (first two taxa) and North America (last six taxa). A cladistic analysis of dental and palatal features within the Taeniolabididae (which includes Catopsalis, Kamptobaatar, Lambdopsalis, Prionessus, Sphenopsalis, and Taeniolabis) indicates that Catopsalis is a paraphyletic taxon, composed of no fewer than five independent monophyletic groups. Taeniolabis is a monophyletic taxon, and Lambdopsalis, Prionessus, and Sphenopsalis (individually monophyletic by monotypy) together form another monophyletic group. These two clades appear to have evolved from ancestors within the paraphyletic taxon Catopsalis; accordingly, the smallest monophyletic group including all Catopsalis species also includes Taeniolabis, Lambdopsalis, Prionessus, and Sphenopsalis. C. matthewi, the most primitive member of this clade, is returned to Djadochtatherium Simpson, previously considered a junior subjective synonym of Catopsalis. The relationships demonstrated among various members of the Taeniolabididae support the hypothesis of a Late Cretaceous taeniolabidid dispersal from Asia to North America. The data additionally suggest a second dispersal event, probably in the middle to late Paleocene, in which the ancestors of the Lambdopsalis/ Prionessus/Sphenopsalis lineage dispersed from North America back to Asia.

  8. Page 95

    Competitive displacement of one taxon by another in the fossil record may be indicated when: (1) an inverse correlation in diversity and, particularly, relative abundance can be demonstrated between the two groups through time; (2) aspects of their paleobiology suggest utilization of common resources; and (3) it can be shown that the two taxa evolved in allopatry prior to their sympatric association. Data from recent collections of Paleocene and Eocene mammals in the Western Interior of North America show marked inverse correlations both of generic diversity and relative abundance between multituberculates and rodents. The largest diminution in multituberculate diversity occurred in the latest Paleocene, near the Tiffanian-Clarkforkian boundary, not in the early Eocene as suggested previously. Reconstruction of diets, die1 activity patterns, locomotor habits, and body sizes of multituberculates and rodents suggests that both groups potentially utilized similar resources. The hypothesis that competitive exclusion may have played a role in the decline of multituberculates is strengthened by recent evidence that rodents evolved in Asia, immigrating to North America in latest Paleocene time. Evidence in support of alternative hypotheses employed to account for the decline and eventual extinction of multituberculates is wanting.

  9. Page 119

    Although species are the basic units of many paleontological and evolutionary studies, the term “species” applied to the fossil record does not convey the same concept to all workers. Simpson’s “evolutionary species” incorporated time into the species concept, but considered each non-branching lineage as a separate species; longer lineages with more continuous fossil records may require subdivision into successional species. One’s perception of paleontological species affects, and is affected by, evolutionary philosophy and models of how new species form and evolve. For example, if species actually arise abruptly and persist for much longer periods essentially unchanged (punctuated equilibria), discrimination of paleontological species should be a relatively simple matter. Alternatively, if there is continuous change within and between successive species (gradualism), species boundaries would be nebulous, and would have to be imposed arbitrarily. We summarize our study of omomyid primates and cite other supportive evidence which suggests that, where the record is sufficiently dense, gradual evolution (requiring arbitrary boundaries) is common between species and even genera.

  10. Page 131

    The fossil record of lorisiforms in Asia is currently restricted to specimens recovered from a half-dozen localities of Miocene age (13 Ma to 7 Ma) in the Siwalik Group of northern Pakistan and in related deposits of India. More than one lorisid taxon is represented in the Pakistan material, but Nycticeboides simpsoni Jacobs, 1981 is currently the only named species. A partial skeleton of Nycticeboides, although poorly preserved, possesses diagnostic lorisid synapomorphies of the auditory region and the vertebral column. The fact that Nycticeboides was a small animal is important for understanding its ecology. A primate frugivore with the M1 dimensions of Nycticeboides should have a body weight of only about 500 g according to commonly-used regression statistics. However, if Nycticeboides was mostly insectivorous, and its molar teeth scaled to body size in the manner characteristic of highly insectivorous primates and non-zalambdodont insectivores, then it may have weighed much less than this estimate.

  11. Page 163

    Work of the past fifty years, or so, has shed much additional light on the phylogeny and history of the Order Rodentia. Classical views have been challenged, and new concepts invoked, or older views rediscovered. Of these new concepts, hystricomorphy, rather than protrogomorphy, as a primitive state for rodents seems difficult to accept, and is not really new. Hystricognathy versus sciurognathy as the fundamental division of the Rodentia seems perilous if pardlelism is as important a phenomenon as is frequently suggested. The argument for multiserial incisor enamel, rather than pauciserial, as the primitive incisor kind is very persuasive, but perhaps more work is needed on Eocene rodent enamel. Punctuated equilibrium, if really a new idea, seems promising in explaining the obscure origin of most rodent groups, but gradualism is evident in many specific lines of descent in rodents. Virtually excluding temporal consideration from phylogenetic studies seems extreme, as does cladistic analysis when it excludes parallelisms and paraphyletic groups. In spite of recent work, the gap between Eocene groups such as the Paramyidae and Ctenodactyloidea, and the Oligocene and later families remains considerable, and largely unexplained. Extraterrestrial collision events in this case can hardly be regarded as pertinent for rodents.

  12. Page 177

    A new species of machaeroidine creodont, Machaeroides simpsoni, is described from the Lostcabinian and Gardnerbuttean of the Wind River Formation, Natrona County, Wyoming. The sabertooth adaptation is fully developed in this species, which has a combination both of more primitive characters and a more derived sabertooth condition than in the Bridgerian species, M. eothen. Some morphological features of Machaeroides indicate that the machaeroidines are more closely allied with the oxyaenids than with the limnocyonids.

  13. Page 183

    Three species of artiodactyls (Diacodexis sp. cf. D. secans, Simpsonodus chacensis [new genus and species], and Wasatchia grangeri) are reported from the San Jose Formation, New Mexico; these three and an additional four taxa (Simpsonodus sp., W. pattersoni, Bunophorus sinclairi, and Hexacodus pelodes) occur in the Debeque Formation, Colorado. The new diacodexeid genus, Simpsonodus, includes D. chacensis and W. lysitensis. Wasatchia pattersoni, new species, the most derived known species of Wasatchia, is morphologically suitable as an ancestor to Bunophorus, with which it overlaps stratigraphically. Diacodexis sp. cf. D. secans is morphologically and apparently phylogenetically intermediate between D. secans, on the one hand, and Hexacodus and Antiacodon on the other. The early Eocene record of diacodexeids is robust enough to indicate: (1) the gradual and continuous divergence of D. secans and Diacodexis sp. cf. D. secans in the mosaic distribution of derived character states; and (2) from these taxa, respectively, the gradational origin of other diacodexeids/leptochoerids and antiacodontids/homacodontids. In these instances, microevolutionary processes appear to account for the macroevolutionary pattern.

  14. Page 197

    A rodent fauna was collected from three localities in the Almagre facies of the San Jose Formation, San Juan Basin, New Mexico. The fauna contains approximately 350 isolated rodent teeth, predominantly from the Sciuravidae. The taxa include Knightomys depressus, Pauromys sp., Lophiparamys debequensis, Paramys copei, Paramys excavatus, Mattimys kalicola, and two new species (Apatosciuravus jacobsi and Knightomys cf. K. minor). Microparamys reginensis has been placed in the genus Knightomys based on discovery of the upper dentition and additional specimens of lower dentition. This is the first record of Mattimys kalicola, Lophiparamys debequensis, and Pauromys in the San Jose Formation.

    This rodent fauna is early Eocene (Wasatchian Land Mammal Age), and correlates closely to the Lysitean “subage” based on presence of Lophiparamys debequensis and Knightomys cf. K. minor.

  15. Page 221

    The fauna of the Skyline channels, West Texas, of early Duchesnean age, contains the following taxa: Simidectes magnus, Harpagolestes sp., Hyaenodon cf. H. vetus, Mahgarita stevensi, ?Leptotomus coelumensis (new species), Duchesneodus cf. uintensis, ?Hyracodon sp. indet., Amynodontopsis bodei, Protoreodon pumilus, Agriochoerus sp. indet., Leptoreodon sp., and Hendryomeryx sp. Simidectes is the most common form; this apparently is its last appearance, along with Harpagolestes and Leptoreodon. It is the first appearance of Hyaenodon, ?Duchesneodus, and Hyracodon.

  16. Page 237

    The seven species of Pseudhipparion are, from oldest to youngest, as follows: unnamed species; P. retrusum; P. curtivallum, new combination; P. hessei, new species; P. gratum; P. skinneri, new species; and P. simpsoni, new species. They range from late Barstovian through latest Hemphillian in the Great Plains and the Gulf Coastal Plain. Both the phylogenetic relationships among these species and the phylogenetic position of this genus among other hipparionine horses are discussed.

    A notable feature of Pseudhipparion evolution is its prevailing tendency toward dwarfing. All crown dimensions except unworn height decrease through time, although analysis is complicated by the fact that sample means are larger in the Great Plains than in contemporaneous Gulf Coast samples, in keeping with Bergmann’s Rule. Rates of change in several dental measurements between various Pseudhipparion species pairs, calculated over intervals of about one million years, have a mean value of 0.11 darwins, which is equal to or greater than in other hipparionine species pairs. In P. simpsoni, the final late Hemphillian species, root formation was delayed ontogenetically, producing extremely high-crowned (incipiently hypsodont) cheek teeth and incisors. Potential crown heights are 85 mm in upper premolars and 110 mm in upper molars; and the enamel patterns are greatly simplified. Such extreme hypsodonty was attained during no more than 1.5 million years within the late Hemphillian at a rate of at least 0.58 darwins, roughly six times the normal rate of crown height increase in hipparionine horses.

  17. Page 273

    New data from the middle and late Miocene Siwalik deposits of Pakistan provide accurate estimates of real temporal durations of extinct species of rhizomyid rodents. Most early rhizomyid species survive on the order of millions of years, with at least two spanning about five million years, and display apparent stasis in most characters. Average species duration for all Rhizomyidae of the Potwar Plateau is about 1.2 million years, a figure in line with other estimates for all Mammalia. Three closely related species show sharp differences in hypsodonty, while other traits remain static in each species or change slowly within the clade, on a scale above the species level. Evolution of this lineage shows at least one step in a staircase pattern, with descendants replacing ancestors, and entails an abrupt morphological change that provides a nonarbitrary definition for species boundaries. One ancestral morphotype appears to survive for a short time with its daughter species. Whereas early nonburrowing Rhizomyidae display longterm stasis, later species, some with burrowing adaptations, are shorter lived and at least one rhizomyine shows rapid, perhaps continuous phyletic change.

    “The time at which an evolving population became different enough from its ancestry to be called a different species cannot, even in theory, be a precise, naturally defined date unless the new, descendant species arose in a single, abrupt step.”

    G. G. Simpson, 1953, p. 35

  18. Page 287

    Abelmoschomys simpsoni, n. gen. et sp., from the latest Clarendonian of Florida, is the earliest known species of Neotropical sigmodontine rodent. Neotropical sigmodontines initially evolved and diversified in North America in the late Miocene. This group is derived from the North American genus Copemys, and forms the sister group of the peromyscines. The Neotropical sigmodontines entered South America in the Pliocene at about the time of formation of the Panamanian land bridge. This hypothesis is supported by evidence from physiology, karyology, molecular systematics, comparative anatomy, and paleontology, and is not contradicted by parasite data. The present day diversity of sigmodontines in South America is that expected for its continental area. Taxonomic frequency rates necessary to produce the more than 200 species of South American sigmodontines as determined by the geometric growth equation range between 0.68 and 0.82 per million years. Although these rates are higher than those previously reported for Cenozoic mammals, they are comparable to rates for other muroid radiations. These high rates may account for discrepancies between divergence times calculated by the molecular clock versus the fossil record.

    “Is it of any interest to anyone but a mouse fancier? (The answer is yes.)”

    George Gaylord Simpson (1980, p. 196)
  19. Page 305

    Two species of very hypsodont bovids are recorded from the middle Miocene (Astaracian/Vallesian) Beglia Formation of west-central Tunisia. They are presumed to be members of the Rupicaprini, a tribe that may have had Mediterranean origin. The Tunisian species show very early development of hypsodonty and loss of the basal pillars. The Rupicaprini are shown to have an important center of evolution in the Mediterranean region throughout the later Tertiary.

  20. Page 317

    Biogeographers have used biotic resemblances and differences to distinguish discrete biogeographic areas. One of the most useful and commonly applied measures of biotic resemblance is a simple binary similarity coefficient developed by G. G. Simpson (1936). In the present study I apply the Simpson Coefficient of faunal similarity to three distributional data sets (at several taxonomic levels): (1) Recent North American mammals; (2) Recent global mammals; and (3) Early Eocene mammals of North America, Europe, and Asia. These analyses yield the following results.

    Faunal realms of large geographic scale can be distinguished both by within-realm faunal similarities/between realm faunal dissimilarity, and by relatively sharp gradients of change of similarity across faunal realm boundaries.

    Geographic distribution patterns of Recent North American mammals show several important trends: (1) an inverse relationship between faunal similarity and both longitudinal and latitudinal separation between sites; (2) a latitudinal asymmetry in similarity comparisons; (3) endemism of western North American faunas; and (4) differentiation of lower level faunal provinces is the result of complex interplay between major latitudinal climatic gradients and less influential (frequently longitudinal) regional climatic, tectonic, and geographic factors.

    Significant differences in taxonomic diversity between localities makes the Simpson Coefficient a more useful measure of faunal resemblance than other binary similarity coefficients, both for Recent and fossil assemblages.

    High similarities for Early Eocene faunal comparisons indicate presence of a single North American-European faunal realm; low resemblances of Asian Early Eocene faunas both to European and North American faunas indicate a distinct Asian faunal realm.

    Available, but incomplete, evidence from faunal resemblances indicates that Ellesmere Island was part of a single, continuous European/North American faunal realm.

  21. Page 339

    Investigation of organismal evolution has been augmented in the last few decades by the development of techniques permitting the genetic material to be studied directly across extant taxa. The relationship among evolutionary processes governing genes, chromosomes, and organisms is explored here. While their rates and modes of evolution are usually greatly dissimilar, genotypes, karyotypes, and phenotypes trace the same phylogenetic path, and can therefore be used to illuminate different aspects of that path. A useful metaphor to express the relationship among these diverse evolutionary systems is the Ptolemaic epicycle, as each epicycle (i. e., evolutionary system) represents a unique detour off the main orbit (i. e., phylogeny). The hominoid primates are taken as an illustration of how these different systems can be used to track aspects of the common phylogeny.

  22. Page 351

    From the introduction of the synthetic theory of evolution in the late 1930s and 1940s until the 1960s, the study of evolution was cast largely within its tenets. Starting in the 1960s changes, primarily introduced by paleontologists, began to have an impact. Although gradualism and phenotypic selection are still prominent, ideas of stasis, punctuation, and heirarchical selection have suggested that revisions of the synthetic theory may be necessary. Molecular biology and molecular evolution have developed in parallel with studies of phenotypic evolution during the last forty years. For the most part these two lines of investigation have remained discrete, although an increasing number of investigators is applying data and concepts from both fields. This convergence is particularly apparent in studies of rates of molecular clocks and the relationship between changes in the genome and changes in the phenotype. As a better understanding of the relationship between the genome and the phenotype emerges, an enlargement or possible abandonment of some aspects of the synthetic theory of evolution may prove necessary. Thus, the synthetic theory is presently being reshaped by data and concepts originating in areas as diverse as paleontology and molecular biology.

Purchase Chapters

Recommended Reading