CO CO Ui-ivLrtSUY OF rH ILLINOIS LIBRARY a. AJ URBANA-CHAMPAIGN co GEOLOGY The person charging this material is re- sponsible for its return to the library from which it was withdrawn on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. To renew call Telephone Center, 333-8400 UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN i in i ii m iniu DEC $ L161— O-1096 FIELDIANA Geology Published by Field Museum of Natural History New Series, No. 12 SYSTEMATICS OF THE EXTINCT SOUTH AMERICAN MARSUPIAL FAMILY POLYDOLOPIDAE LARRY G. MARSHALL December 30, 1982 Publication 1339 SYSTEMATICS OF THE EXTINCT SOUTH AMERICAN MARSUPIAL FAMILY POLYDOLOPIDAE FIELDIANA Geology Published by Field Museum of Natural History New Series, No. 12 SYSTEMATICS OF THE EXTINCT SOUTH AMERICAN MARSUPIAL FAMILY POLYDOLOPIDAE LARRY G. MARSHALL Department of Geosciences University of Arizona Tucson, Arizona 85721 Accepted for publication April 16, 1982 December 30, 1982 Publication 1339 Library of Congress Catalog Card No.: 82-82735 ISSN 0096-2651 PRINTED IN THE UNITED STATES OF AMERICA CONTENTS List of Illustrations vi List of Appendices viii Abstract 1 Introduction 2 Abbreviations 2 Acknowledgments 2 Historical Review 5 Systematics 15 Superfamily Polydolopoidea 15 Family Polydolopidae 15 Subfamily Polydolopinae 16 Polydolops 16 P. clavulus 17 P. kamektsen 20 P. rothi 22 P. winecage 23 P. serra 25 P. thomasi 34 P. mayoi 44 P? sp. indet 48 Summary of Evolution of Polydolops 48 Pseudolops 51 P. princeps 51 Amphidolops 54 A. yapa 55 A. serrula 56 Eudolops 60 £. tetragonus 61 E. hernandezi sp. nov 70 Subfamily Epidolopinae 73 Epidolops sf. . . . . 74 E. ameghinpi 74 Evolution of Polydolopidae 82 Literature Cited 89 LIST OF ILLUSTRATIONS 1. Map of South America showing some vertebrate fossil localities discussed in text 3 2. Map of Chubut Province, southern Argentina, showing some fossil localities discussed in text 4 3. Chronology of South American Paleogene land mammal ages 6 4. Phylogenetic affinities of Polydolopidae as conceived by Ameghi- no 7 5. Cladograms showing possible phylogenetic relationships of poly- dolopoids, phalangeroids, and caenolestoids 10 6. Postcanine cheek tooth nomenclature used for polydolopids by various workers 10 7. Labial and occlusal views of right mandibular rami of living Austral- asian "diprotodonts" 11 8. Labial and occlusal views of right mandibular rami of living and fossil South American "pseudodiprotodonts" 13 9. Polydolops clavulus Ameghino, 1902b (Casamayoran) 18 10. Polydolops clavulus Ameghino, 1902b (Casamayoran) 18 11. Polydolops clavulus Ameghino, 1902b (Casamayoran) 19 12. Polydolops clavulus Ameghino, 1902b (Casamayoran) 19 13. Polydolops kamektsen Simpson, 1935a (Riochican) 21 14. Polydolops kamektsen Simpson, 1935a (Riochican) 21 15. Polydolops rothi Simpson, 1936 (Riochican) 22 16. Polydolops rothi Simpson, 1936 (Riochican) 23 17. Polydolops winecage Simpson, 1935a (Riochican) 24 18. Polydolops winecage Simpson, 1935a (Riochican) 24 19. Polydolops serra Ameghino, 1902b (Casamayoran) 26 20. Polydolops serra Ameghino, 1902b. (Casamayoran) 27 21. Polydolops serra Ameghino, 1902b (Casamayoran) 28 22. Polydolops serra Ameghino, 1902b (Casamayoran) 29 23. Polydolops serra Ameghino, 1902b (Casamayoran) 29 24. Polydolops serra Ameghino, 1902b (Casamayoran) 29 25. Polydolops serra Ameghino, 1902b (Casamayoran) 30 26. Polydolops serra Ameghino, 1902b (Casamayoran) 31 27. Polydolops serra Ameghino, 1902b (Casamayoran) 32 28. Polydolops serra Ameghino, 1902b (Casamayoran) 33 29. Polydolops serra Ameghino, 1902b (Casamayoran) 34 30. Polydolops thomasi Ameghino, 1897 (Casamayoran) 35 31. Polydolops thomasi Ameghino, 1897 (Casamayoran) 36 32. Polydolops thomasi Ameghino, 1897 (Casamayoran) 37 33. Polydolops thomasi Ameghino, 1897 (Casamayoran) 38 34. Polydolops thomasi Ameghino, 1897 (Casamayoran) 40 35. Polydolops thomasi Ameghino, 1897 (Casamayoran) 43 36. Polydolops mayoi Odreman Rivas, 1978 (Mustersan) 45 37. Polydolops mayoi Odreman Rivas, 1978 (Mustersan) 46 38. Polydolops mayoi Odreman Rivas, 1978 (Mustersan) 47 39. Polydolops? sp. indet. (Deseadan) 49 40. Cladogram showing probable phylogenetic relationships of the known species of Polydolops 50 41. Pseudolops princeps Ameghino, 1902b (Casamayoran) 52 42. Amphidolops yapa (Simpson, 1935a) (Riochican) 55 43. Amphidolops yapa (Simpson, 1935a) (Riochican) 56 44. Amphidolops serrula Ameghino, 1902b (Casamayoran) 56 45. Amphidolops serrula Ameghino, 1902b (Casamayoran) 57 46. Amphidolops serrula Ameghino, 1902b (Casamayoran) 57 47. Amphidolops serrula Ameghino, 1902b (Casamayoran) 58 48. Amphidolops serrula Ameghino, 1902b (Casamayoran) 59 49. Amphidolops serrula Ameghino, 1902b (Casamayoran) 60 50. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 62 51. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 63 52. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 64 53. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 65 54. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 66 55. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 66 56. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 67 57. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 67 58. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 68 59. Eudolops tetragonus Ameghino, 1897 (Casamayoran) 69 60. Eudolops hernandezi sp. nov. (Casamayoran) 71 61. Eudolops hernandezi sp. nov. (Casamayoran) 72 62. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 75 63. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 76 64. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 77 65. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 78 66. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 79 67. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 80 68. Epidolops ameghinoi Paula Couto, 1952 (Riochican) 81 69. Comparison of lower dentitions of various species of Polydolopidae showing relative size and proportions of teeth 83 70. Comparison of upper dentitions of various species of Polydolopidae showing relative size and proportions of teeth 84 71. Size distribution of various species of Polydolopidae as indicated by relationship of length vs. width of P3 85 72. Size distribution of various species of Polydolopidae as indicated by relationship of length of Pt vs. length of M, 86 73. Size distribution of various species of Polydolopidae as indicated by relationship of length vs. width of M2 87 74. Cladogram showing probable phylogenetic relationships of better- known genera of Polydolopidae 88 LIST OF APPENDICES 1. Measurements of lower cheek teeth of some species of Polydolops 92 2. Measurements of cheek teeth of Polydolops serra 92 3. Measurements of upper cheek teeth of Polydolops thomasi 93 4. Measurements of lower cheek teeth of Polydolops thomasi 93 5. Statistics for some cheek tooth dimensions of Polydolops thomasi 95 6. Measurements of lower cheek teeth of Polydolops mayoi 95 7. Summary and distribution of some diagnostic characters of known spe- cies of Polydolops 96 8. Measurements of cheek teeth of Amphidolops serrula 99 9. Measurements of some cheek teeth of species of Eudolops 100 10. Measurements of upper cheek teeth of Epidolops ameghinoi 101 11. Measurements of lower cheek teeth of Epidolops ameghinoi 102 12. Statistics for some cheek tooth dimensions of Epidolops ameghinoi 104 13. Summary and distribution of some diagnostic characters of better-known genera of Polydolopidae 105 ABSTRACT Members of the extinct marsupial family Polydolopidae (superfamily Polydolopoidea) are known in South America from beds of Riochican (middle to late Paleocene), Casamayoran (early Eocene), and Mustersan (middle Eocene) age in Patagonia, southern Argentina; from beds of Riochican age in Brazil; and apparently from beds of Deseadan (early Oligocene) age in Bolivia. Five genera and 13 species, grouped in two subfamilies, are recognized. The Epidolopinae include Epidolops ameghinoi Paula Couto, 1952, from Riochican beds in Brazil, and an unnamed and undescribed epidolopine (apparently referrable to Epidolops) from Riochican beds in southern Argentina. The Polydolopinae in- clude Polydolops clavulus Ameghino, 1902b (Casamayoran); P. kamektsen Simpson, 1935a (Riochican); P. rothi Simpson, 1936 (Riochican); P. winecage Simpson, 1935a (Riochican); P. serra Ameghino, 1902b (Casamayoran); P. thomasi Ameghino, 1897 (Casamayoran); P. mayoi Odreman Rivas, 1978 (Mus- tersan); Pseudolops princeps Ameghino, 1902b (Casamayoran); Amphidolops yapa (Simpson, 1935a) (Riochican); A. serrula Ameghino, 1902b (Casamayoran); Eudolops tetragonus Ameghino, 1897 (Casamayoran); and £. hernandezi sp. nov. (Casamayoran) — all from Patagonia, southern Argentina. A specimen from beds of Deseadan age in Bolivia is tentatively referred to Polydolops? sp. indet. The species and genera are readily distinguished on the basis of absolute and relative size, structure of cheek teeth, presence or absence of various teeth, and formation by certain teeth of the "plagiaulacoid" or bladelike elements in the dentition (e.g., P- and P8 / P;i and trigonid of M, — Polydolops, Eudolops; P' / P, and trigonid of M, — Amphidolops; P' / P, — Epidolops). An attempt is made to demonstrate that the large procumbent lower tooth in polydolopids is the canine and not an incisor. A historical review is given of polydolopid systematics. Nomenclature and relationships of included taxa are stabilized and clarified. Taxonomic limits of the Polydolopidae are defined; the group is shown to be monophyletic in ori- gin. It is concluded that the Polydolopoidea and Caenolestoidea evolved in- dependently from didelphoidlike ancestors in South America: any spe- cializations shared by these groups, either with one another or with Austral- asian Phalangeroidea, are the result of convergence in evolution. 2 FIELDIANA: GEOLOGY INTRODUCTION This study is a detailed systematic review of the extinct South American marsupial family Polydolopidae Ameghino, 1897 (superfamily Polydolopoidea). Revised diagnoses are given for the family, subfamilies, genera, and species, along with a list of included taxa and synonymies. Consideration is given to previous views of the relationships of polydolopoids with other marsupial groups to demonstrate that the included taxa are monophyletic relative to other superfamilies. The families and superfamilies of Marsupialia are in need of a rigorous cladistic analysis to clarify their phylogenetic relationships, and this study is but a step to that end. It is limited in scope to stabilization of nomenclature, clarification of the phylogenetic relationships of included taxa, and definition of the known taxonomic limits of the Polydolopidae. During the course of this study I was able to examine firsthand all pertinent known materials, including type and referred specimens. This work includes discussion and description of some new materials, but is essentially based on a reappraisal of previously known specimens and literature; it represents an at- tempt to bring together in one place a modern, expanded, and synthetic treat- ment of these animals. The fossil localities mentioned below (figs. 1, 2) are shown on maps and discussed in greater detail in various papers summarized in Marshall, Hoffstetter & Pascual (in press). The chronology and usage of South American Paleogene land mammal ages (fig. 3) follows Marshall, Butler, Drake, & Curtis (1981) and Marshall, Hoffstetter, & Pascual (in press) and references therein. ABBREVIATIONS Abbreviations used in the text, figure captions, and appendices of mea- surements are: C, canine; ca, approximate measurement; CV, coefficient of variation; I, incisor; L, length; M, molar; N, number; OR, observed range of sample; P, premolar; s, standard deviation of sample; W, width; and x, mean. All measurements are in millimeters (mm). Abbreviations used for specimens from institutional collections are: AC, Amherst College, Amherst, Massachusetts; AMNH, American Museum of Nat- ural History, New York; DGM, Divisao de Geologia e Mineralogia do De- partamento Nacional da Producao Mineral, Rio de Janeiro, Brazil; FMNH, Field Museum of Natural History, Chicago; MACN, Museo Argentino de Ciencias Naturales "Bernardino Rivadavia", Buenos Aires, Argentina; MLP, Museo de La Plata, La Plata, Argentina; MMP, Museo Municipal de Ciencias Naturales de Mar del Plata "Lorenzo Scaglia", Mar del Plata, Argentina; MNHN, Museum National d'Histoire Naturelle, Paris, France; MNRJ, Museu Nacional e Uni- versidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil; and PU, Princeton University, Princeton, New Jersey. ACKNOWLEDGMENTS I am indebted to the following individuals for allowing me to study speci- mens at their respective institutions: D. Baird (PU); J. Bonaparte (MACN); M. Coombs (AC); R. Pascual (MLP); L. Price and D. Campos (DGM); G. Scaglia (MMP); F. L. de Souza Cunha (MNRJ); and M. C. McKenna and R. H. Tedford 80° 60° 40° -0°^ >£* o°- -20° \1> \La Sal la J 20°- 1 Sao Jose de Itaborai' -40° fli V 40°- 0 600 l_ i .,„_j v\\\\\\ ^ Fig. 2 J km Ai«^ 80° ^S6* 60° 40° 20° Fig. 1. Map of South America showing some vertebrate fossil localities (filled circles) discussed in text. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 5 (AMNH). The figures of specimens were drawn by Elizabeth Liebman; the stereophotos were made by Ron Testa from original specimens or epoxy casts. Initial stages of this study were made possible by three grants (1329, 1698, 1943) from the National Geographic Society, Washington, D.C.; completion was made possible by National Science Foundation grant DEB-7901976. HISTORICAL REVIEW Polydolopids were first discovered by the eminent Argentine collector Carlos Ameghino during his ninth trip to Patagonia, southern Argentina, during the austral summer of 1895-96. The specimens collected by Carlos were described by his elder brother Florentino, who in 1897 named two monotypic genera, Polydolops (P. ihomasi) and Eudolops (E. tetragonus), and placed them in a new marsupial family: the Polydolopidae. These taxa were based on specimens col- lected from the Barranca (cliff) south of Lago Colhue-Huapi (fig. 2) from a level which Florentino recognized as yielding the Notostylops fauna, the basis of the Casamayoran land mammal age, conventionally assigned to the early Eocene (fig. 3). Subsequent trips by Carlos to Patagonia resulted in collection of addi- tional specimens of Casamayoran polydolopids, and these too were described and named by Florentino (see Ameghino, 1902a, b, 1903, 1904a, b, 1906). By 1906 Florentino had named a dozen genera now recognized as polydolopids, and these (along with Odontomysops — of uncertain affinity, see Simpson, 1948, p. 68) he placed in various families in the Allotheria ( = Mul- tituberculata) as follows: Suborder Allotheria Family Odontomysopidae Odontomysops Family Promysopidae Promysops Propolymastodon Family Polydolopidae Polydolops Eudolops Pseudolops Pliodolops Amphidolops Orthodolops Archaeodolops Anadolops Family Neoplagiaulacidae Anissodolops Ameghino (e.g., 1903) (see fig. 4) regarded the Polydolopidae as descendants of one group of caenolestoids, the "Garzonidae" ( = Caenolestinae, see Mar- shall, 1980a), and ancestors of another caenolestoid group, the "Abderitidae" ( = Abderitinae, see Marshall, 1980a). He also derived some multituberculates and rodents from polydolopids. Ameghino thus considered polydolopids re- lated to multituberculates, rodents, and caenolestoids, groups now regarded as belonging to the Prototheria, Eutheria, and Metatheria, respectively. Such a relationship was based on the common possession by at least some caenoles- toids, polydolopoids, and multituberculates of a "plagiaulacoid" shearing tooth (see Simpson, 1933), a procumbent gliriform tooth, and/or various similarities in cheek teeth morphology. Many of Ameghino's contemporary workers and Millions of years ago 25- 30- 35- 40- 45- 50- 55- 60- 65 Epoch LU z LU O o o LU z LU O o LU LU z LU O o LU _l < Q. South American Land Mammal Ages Colhuehuapian Deseadan Divisaderan ■'/////////////& Mustersan V/////////////7, Casamayoran Riochican Z Fig. 3. Chronology of South American Paleogene land mammal ages. Hatching denotes hiatuses in knowledge of fossil land mammal faunas. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE CD CO Polydolopidae CD n CD -Q \imysopidi 1/ Polydolopidae / Microlestidae Coenolestidae Garzonidae Microbiotheridae Fig. 4. Phylogenetic affinities of Polydolopidae as conceived by Ameghino (1903, sim- plified from chart on p. 188). some subsequent ones followed his views and classified the various polydolopid taxa among the Multituberculata. Gregory (1910, pp. 211-14), however, regarded the resemblances between polydolopids, multituberculates, and rodents as a result of convergence in evo- lution. He tentatively suggested that polydolopids were a specialized branch of caenolestoids. This view was championed by Simpson (1928, 1930, 1933), who concluded (1928, p. 13): The family Polydolopidae represents an early and specialized side branch of the group Caenolestoidea, more or less intimately allied to the true Caenolestidae and like the latter ultimately derived from primitive Cretaceous polyprotodont marsupials similar to or belonging in the family Didelphidae in a broad sense. 8 FIELDIANA: GEOLOGY Accordingly, Simpson (1930) classified them (along with several taxa of un- certain or other affinity) as follows: Order Marsupialia Superfamily Caenolestoidea Family Caenolestidae Family Polydolopidae Amphidolops Anadolops Anissodolops Archaeodolops Eomannodon Eudolops Odontomysops Orthodolops Pliodolops Polydolops Promysops Propolymastodon Pseudolops During 1930-31 and 1933-34, Simpson made substantial collections of early Tertiary mammals, including polydolopids, from various localities in Patagonia, southern Argentina. Most of these polydolopids were of Casamayoran age, but a few were from older beds which Simpson named the Rio Chico Formation, namesake of the Riochican land mammal age now known to encompass middle and late Paleocene time (Marshall et al., 1981) (fig. 3). Simpson described and named these specimens and others in a series of subsequent papers (1935a, b, 1936, 1948, 1964, 1967b). In 1948 he reviewed the Polydolopidae in volume one of The Beginning of the Age of Mammals in South America, and at that time included consideration and measurements of all known pertinent specimens. Based on progressive knowledge of the affinities of included taxa, Simpson (1948) classified them as follows: Order Marsupialia Superfamily Caenolestoidea Family Caenolestidae Family Polydolopidae Amphidolops (incl. Anadolops) Eudolops (incl. Promysops, Propolymastodon) Polydolops (incl. Pseudolops, Pliodolops, Orthodolops, Anissodolops, Archaeo- dolops) Seumadia Simpson (1948) noted that classification of the Patagonian specimens at the generic level proved difficult. The available material was scanty in number of specimens and fragmentary in preservation. Associated upper and lower den- titions were unknown, and association of any sort was uncommon. Most im- portantly, the types are not all comparable; numerous names, often applied to fragmentary materials, cannot always be rejected and often cannot be validated. Of the four genera recognized by Simpson, Amphidolops, Eudolops, and Polydolops were known from referred partial upper and lower dentitions, while Seumadia was known only from an isolated M:i (his M4). Paula Couto (1952, 1961, 1970) described a new genus, Epidolops, from fissure filling of Riochican age in the Itaborai Formation in a limestone quarry at Sao Jose de Itaborai, northwest of Niteroi, capital of the state of Rio de Janeiro, Brazil (fig. 1). The described material includes a nearly complete skull and MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 9 associated mandible, making Epidolops the most completely known of all polydolopids. All species and specimens of polydolopids described so far are of Riochican or Casamayoran age; the group was long believed to have gone extinct at the end of Casamayoran time. Recently, however, Odreman Rivas (1978) described a new species of Polydolops from beds of Mustersan (middle Eocene) age in Patagonia, and Patterson & Marshall (1978) recorded a fragmentary polydolopid of indeterminate affinity from the Deseadan (early Oligocene) fauna from Salla, Bolivia (fig. 1). Pascual (1980a) assigned two genera, Prepidolops and Bonapartherium (the latter tentatively), to the Polydolopoidea, although he later (1980b, 1981) re- garded them as specialized Didelphoidea, a view followed here. These taxa are from the Lumbrera Formation in Salta Province, northwestern Argentina, and are regarded as Casamayoran in age. Pascual & Bond (1981) most recently rec- ognized two subfamilies of Polydolopidae, the Polydolopinae and Epidolopinae (the latter to include only Epidolops). This grouping is followed and further documented below. After this study was completed, polydolopoids were discovered in beds of Eocene age on Seymour Island, Antarctica (see Simon, 1982). These specimens are being described by Dr. M. O. Woodburne, University of California- Riverside. Based largely on the work of Simpson (1928, 1948) and Paula Couto (op. cit.), the marsupial affinities of polydolopids have been firmly established. The fea- tures which indicate their marsupial affinity are the joint occurrence of an inflected angle on the mandibular ramus and a cheek tooth formula of Pl-3 and Ml-4 (Marshall, 1979). Yet the affinities of polydolopoids with other marsupial groups, namely, caenolestoids and Australasian phalangeroids, have been the subject of long dispute. Four views exist regarding the possible phylogenetic relationships of these groups (fig. 5): 1) phalangeroids, polydolopoids, and caenolestoids are a monophyletic group (fig. 5a); 2) phalangeroids and polydolopoids are a monophyletic group relative to caenolestoids (fig. 5b); 3) polydolopoids and caenolestoids are a monophyletic group relative to phalangeroids (fig. 5c); and 4) polydolopoids and caenolestoids represent independent lines of evolution from didelphoidlike ancestors in South America, any specializations they share among themselves or with Australasian phalangeroids being the result of con- vergence in evolution (fig. 5d). The controversy stems from uncertainties re- garding homologies of the "plagiaulacoid" shearing tooth, the lower pro- cumbent gliriform tooth, reduced antemolar teeth, and various similarities in cheek tooth morphology. The issues have been clarified in recent years by additional fossil materials and by study of various aspects of soft part anatomy of living caenolestoid and phalangeroid taxa, including structure of brain, sperm morphology, and rates of evolution of blood proteins. I will first review data bearing on the controversy and then consider which of the four possible scenarios of affinity appears best supported by the data. 1. Cheek tooth formula and homology of " plagiaulacoid" tooth. — I recognize the permanent primitive marsupial dental formula so far known as If, C{, Pj, Mf, and accordingly label the three premolars as PI, P2, P3 and the four molars as Ml, M2, M3, M4 (fig. 6). This is the classic system for serial designation of the 10 FIELDIANA: GEOLOGY phalangeroids polydolopoids caenolestoids t phalangeroids polydolopoids caenolestoids caenolestoids polydolopoids phalangeroids phalangeroids polydolopoids caenolestoids Fig. 5. Cladograms showing possible phylogenetic relationships of polydolopoids, phalangeroids, and caenolestoids. Solid circles mark relative times of acquisition of de- rived characters (if any) which would indicate the unity of two (b,c) or all three (a) groups, or the independent acquisition of derived characters by each group (d). Ameghino, 1903 Ml M2 M3 M4 M5 M6 M7 Simpson, 1948 PI P2 P3 Ml M2 M3 M4 Paula Couto, 1952; this author PI P2 P3 Ml M2 M3 M4 Fig. 6. Postcanine cheek tooth nomenclature used for polydolopids by Ameghino (e.g., 1903), Simpson (e.g., 1948), Paula Couto (1952), and this author. permanent cheek teeth in marsupials, and its use implies homology within the Marsupialia but not necessarily with other therian groups. (Ameghino [e.g., 1903] numbered the cheek teeth as one series, calling them all molars [fig. 6], while Simpson [e.g., 1948] designated the "plagiaulacoid" tooth in polydolopoids the Ml on the assumption that polydolopoids were specialized caenolestoids [fig. 6]). In most phalangeroids there are four permanent molars (fig. 7), although in some groups (i.e., Burramyidae) the M4 has been lost. In those phalangeroid groups having a "plagiaulacoid" tooth it is always the P3, although the P2 may Fig. 7. Labial and occlusal views of right mandibular rami of living Australasian "di- protodonts". Top, Petaurus breviceps (FMNH 60954); middle, Trichosurus vulpecula (FMNH 60933); bottom, Hypsiprymnodon moschatus (FMNH 60953). Scale = 10 mm. After fig. 21.3a in Marshall, 1980b. 11 12 FIELDIANA: GEOLOGY also show bladelike specializations (Archer, 1978). All caenolestoids have four permanent molars (fig. 8, top). The "plagiaulacoid" tooth, in those forms hav- ing it, is the Ml (e.g., Abderites; fig. 8, bottom); in some forms (e.g., Parabder- ites) the P3 is bladelike as well (Marshall, 1976, 1980a). It was long believed that the "plagiaulacoid" tooth in polydolopoids was the Ml. This view was based on the fact that this tooth was followed by three molariform teeth, the same arrangement as occurs in abderitine caenolestids (fig. 8, bottom). Based on specimens of Epidolops from Riochican age in Brazil, Paula Couto (1952) demonstrated that the "plagiaulacoid" tooth in polydolopoids is the P3 (fig. 8, middle) and not the Ml as previously believed. This knowledge is based on the fact that species of Epidolops retain a small M4 lost in all other known polydolopoids. Thus, the main "plagiaulacoid" tooth in phalangeroids and polydolopoids is the P3, while in abderitine caenolestids it is the Ml. Loss of the M4 has occurred in some phalangeroids and in all polydolopoids except Epidolops. 2. Homologies of large lower procumbent gliriform tooth. — Ride (1962) re- viewed data bearing on the identity and homology of the large lower pro- cumbent tooth in caenolestoids and phalangeroids. He concluded that in phalangeroids this tooth was probably the I:j (or possibly even the L,) (fig. 7), while in caenolestoids (fig. 8, top) it was either the I, or L. The procumbent tooth in polydolopoids is regarded as an incisor by all earlier workers. I believe there is sufficient evidence, however, to suggest that this tooth is the canine. In the type of Epidolops ameghinoi (see p. 74), there are three procumbent single-rooted teeth in the lower jaw. Paula Couto regarded the smaller two anteromedial ones (1952, fig. 3) as incisors, and the posterolabial larger one as the canine. I concur with these identifications but refrain from identifying the homology of the two incisors. A similar situation occurs in the type (MACN 10340a) of "Promysops acuminatus" ( = Eudolops tetragonus, see p. 61) in which there are alveoli of three antepremolar teeth. The most distal of these is very large; the two mesial ones are small and set one above the other, the more distal of the two being the smaller. This arrangement in relative size and placement of the antepremolar teeth is identical to that seen in Epidolops ameghinoi. Assuming these teeth are homologous in the two species, then the larger distal alveolus in the type of "P. acuminatus" can be regarded as a canine and the two smaller, mesial ones as incisors. It is my contention that in the evolution of polydolopoids the incisors continued to decrease in size in some lineages relative to the canine and that, as seen in species of Polydolops (see p. 16), all the incisors are eventually lost, leaving the canine as the large pro- cumbent tooth. An intermediate stage in this trend exists in Epidolops ameghinoi and "Promysops acuminatus." An alternative view regarding the homology of the large procumbent tooth in polydolopoids is that it represents the I3 or L,. This interpretation is possible if it is assumed that the tooth in Epidolops identified by Paula Couto as the canine is really a posterior incisor (I:{ or L;). If this were the case, then the canine was lost very early in polydolopoid evolution and the I;! or L. became large at a very early time in the history of this group. I do not prefer this view, although some workers may favor it as a viable alternative to the interpretation followed here. In summary, the large procumbent lower tooth in phalangeroids appears to be the I3 (or I4); in caenolestoids it is either the I, or L>, while in polydolopoids it is apparently the canine. M.P3 Fig. 8. Labial and occlusal views of right mandibular rami of living and fossil South American "pseudodiprotodonts". Top, Caenolestes fuliginosus (FMNH 53303) (living); middle, Epidolops ameghinoi (DGM 321-M) (Riochican); bottom, Abderites meridionalis (MACN 2037) (early Miocene-Santacrucian). Scale =10 mm. After fig. 21.3b in Marshall, 1980b. 13 14 FIELDIANA: GEOLOGY 3. Brain structure. — The controversy of phalangeroid, polydolopoid, and caenolestoid affinity was in part clarified by Abbie's (1937) study of the brains of marsupials. He demonstrated that in all marsupial groups studied, except for Australasian phalangeroids, two large fiber bundles interconnect pallial struc- tures of the two cerebral hemispheres. This same pattern of commissural con- nections occurs in monotremes, and by application of the principles of com- monality and outgroup comparison it probably also occurred in the therian ancestors of marsupials and placentals (Marshall, 1979). In phalangeroids a third bundle of neocortical commissural fibers, the fasciculus aberrans, is added, a condition regarded as apomorphic (derived) for that group of marsupials (Mar- shall, 1979). Caenolestoids lack this structure (Obenchain, 1925) and as such show no special affinity to phalangeroids, as least as far as the state of this character is concerned. 4. Sperm pairing and morphology . — Biggers & DeLamater (1965) demon- strated that a high percentage of pairing of spermatozoa in the epididymes occurs in each of the species of didelphids and caenolestids tested, but no evidence of this phenomenon was observed in the Australasian marsupials studied. As inferred elsewhere (Marshall, 1980a), sperm pairing is a derived condition among the Mammalia, indicating that all living American marsupials form a monophyletic group relative to Australasian taxa. Biggers & DeLamater also demonstrated that the structure of sperm in Caenolestes was distinct from that of other American and Australasian marsupials tested. 5. Serology. — Kirsch (in Hayman et al., 1971) compared blood sera of two species of caenolestids with members of all families of living Marsupialia, except Thylacynidae. He concluded (1977) that the two caenolestids are more similar to each other than they are to any other marsupial group. Furthermore, his data suggested a tripartite split among living marsupials: caenolestoids, didelphoids, and Australasian taxa. These data suggest that the inferred relationships of caenolestoids, pha- langeroids, and polydolopoids (fig. 5, cladogram d) best express their relative affinity. While phalangeroids and polydolopoids are synapomorphous, having P3 as the "plagiaulacoid" tooth, this feature is here regarded as a convergent feature in evolution. This view is supported by the fact that a phalangeroid- polydolopoid special relationship is not supported by any additional synapomorphous features. Indeed, the gliriform tooth in phalangeroids and polydolopoids is apparently not homologous (although it could be so inter- preted by some worker), and loss of the M4 in some members of each group is regarded as a convergent feature. Structure of the brain, sperm, sperm pairing, and serology, as well as knowl- edge of homology of the "plagiaulacoid" tooth and the large lower procumbent tooth, do not support a special relationship between phalangeroids and caenolestoids. Caenolestoids and polydolopoids are in turn distinguished by apparent differences in homology of the "plagiaulacoid" and gliriform tooth. All in all, these features suggest an independent evolution of each of these groups from a more generalized, possibly didelphoidlike ancestor. Apart from the supposed homology of the "plagiaulacoid" and gliriform tooth, and of the reduced antemolar teeth, other features shared by caenoles- toids and polydolopoids were recognized by early workers as supporting their close affinity. Simpson (1948, pp. 51-52) noted that the "proliferation of cusps" in the upper molars of polydolopoids seems to overlie a caenolestoid ancestry in MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 15 which these teeth are broad and subquadrate, with a median basin, outer and inner cuspate crests, and a bilobed lingual surface. In addition, the lower molar pattern as seen in the generalized structure of M3 in polydolopoids resembles abderitine caenolestoids: "trigonid and talonid poorly differentiated, trigonid with one internal and two external cusps, talonid much enlarged, basined with cuspidate rim." Based largely on these features — but at the same time re- cognizing that the "plagiaulacoid" shearing specializations evolved in- dependently in the two groups — subsequent workers continued to accept the view that polydolopids were specialized caenolestoids. It was, however, recognized that polydolopids were too specialized to be ancestral to any known caenolestoid despite the fact that polydolopids occur earlier in the fossil record: in polydolopids the shearing tooth is more spe- cialized and the antemolar teeth are subject to greater reduction and loss than in abderitine caenolestoids. These features clearly indicate an ancient divergence between caenolestoids and polydolopoids. Based largely on knowledge of these facts, Pascual & Herrera (1973, 1975) argued that there is no convincing evidence to indicate that polydolopoids are any closer phylogenetically to caenolestoids than either group is to didelphoids. These workers concluded that there is, however, sufficient evidence indicating an independent origin of caenolestids and polydolopids from didelphoidlike ancestors. These observations suggest that 1) caenolestoids (as demonstrated by features of known fossil and living forms) cannot be regarded as structural ancestors of polydolopoids; and 2) the successive grades of specialization within caenoles- toids (i.e., Caenolestinae-Palaeothentinae-Abderitinae) represent an adaptive process independent of that of polydolopoids (Simpson, 1971, p. 113). There is also sufficient evidence to indicate that abderitine caenolestoids did not evolve until after the disappearance of polydolopoids, suggesting at least partial ecological replacement of polydolopoids by abderitine caenolestoids. In view of these facts and based largely on the work of Pascual & Herrera (1973, 1975), Clemens & Marshall (1976, p. 9) allocated the polydolopids to rank of superfamily, the Polydolopoidea. This action left the Caenolestoidea (sensu Marshall, 1980a) an uncluttered, cohesive, and monophyletic taxon. These views are followed here. SYSTEMATICS Superfamily POLYDOLOPOIDEA (Ameghino, 1897, p. 496) Clemens & Marshall, 1976, p. 79 Family POLYDOLOPIDAE Ameghino, 1897, p. 496 (Includes Promysopidae Ameghino, 1902a, p. 36) Revised Diagnosis. — Dental formula J$, C\, P^i M^; sectorials P2 and P3 (or just P) above, P:i and trigonid of M, (or just P:)) below; lower C (where known) well developed, single rooted, procumbent, and gliriform in some taxa; where known zygomatic arch broad, rostrum brachycephalic; symphysis ligamentous, unfused in adult; large inflected angular process and large palatal vacuities present; postcranial material and mode of locomotion not known. Known Range. — Riochican, Casamayoran, and Mustersan age beds in Patagonia, southern Argentina; Riochican beds in Brazil; and apparently Deseadan beds in Bolivia. 16 FIELDIANA: GEOLOGY Subfamily POLYDOLOPINAE (Ameghino, 1897, p. 496) Pascual & Bond, 1981, p. 483 Diagnosis. — Polydolopids of very small to very large size; dental formula 1^, C{, Pjij, M|; molars multitubercular with cuspidate rims; sectorials P2 and P'! (or just P!) above, P;i and trigonid of M, below; P3 aligned in same anteroposterior axis as molar series; M4 absent; relative size of Ml to M3 subequal, or slight decrease from Ml to M3; molar basins shallow or deep and with lightly to strongly wrinkled enamel; anterior edge of M, trigonid elevated well above rest of molars and bladelike. Known Range. — Riochican, Casamayoran, and Mustersan age beds in Patagonia, southern Argentina, and apparently Deseadan beds in Bolivia. Polydolops Ameghino, 1897 Polydolops Ameghino, 1897, p. 497. Pliodolops Ameghino, 1902b, p. 41. Orthodolops Ameghino, 1903, p. 130; 1904b, p. 257 (said to be new in 1904, but publica- tion in 1903 was prior and valid). Anissodolops Ameghino, 1903, p. 148; 1904b, p. 257 (said to be new in 1904, but publi- cation in 1903 was prior and valid). Archaeodolops Ameghino, 1903, p. 150; 1904b, p. 257 (said to be new in 1904, but publication in 1903 was prior and valid). Type of Polydolops. — P. thomasi Ameghino, 1897, p. 497. Type of Pliodolops. — P. primulus Ameghino, 1902b, p. 41. Type of Orthodolops. — O. sciurinus Ameghino, 1903, p. 131. Type of Anissodolops. — A. serrifer (Ameghino, 1902b, p. 39). Type of Archaeodolops. — A. clavulus (Ameghino, 1902b, p. 40). Revised Diagnosis. — Polydolopids of small to large size, known dental for- mula Is, Q, Pj, M§; upper incisors, canine, and upper and lower Pis unknown; sectorial shear blades include P2 and P8 / P3 and trigonid of M, ; P2 well- developed, two-rooted blade, cutting edge serrated especially along dorsopos- terior edge, prominent but narrow arced rib extending from anterior edge of posterior root to tip of crown on both labial and lingual surfaces, crown larger and wider than P:i; P:i well-developed, two-rooted blade, cutting edge strongly serrated especially along anterodorsal edge, prominent but narrow arced rib extending from anterior edge of posterior root to tip of crown on both labial and lingual surfaces; basins of upper and lower molars shallow to well developed, with weakly wrinkled enamel; M1 with six primary cusps labial to basin and from one to five stylar cuspules labial to them; M2 with five to six primary cusps labial to basin and one medially placed stylar cuspule labial to them; M:i trian- gular shaped with apex pointing posteriorly, edges somewhat elevated but without distinct cusps; lower canine large, single rooted, procumbent; P-, very small, crown without cusps, flattened, irregularly oval in outline, two rooted or a grooved root, sometimes absent; P:i large, "plagiaulacoid" blade, cutting edge ranges from unserrated to distinctly serrated, small but distinct anterobasal cuspule usually present, weak medial rib present on labial and lingual surfaces; M, with anterior edge of trigonid elevated well above rest of molars and bladelike, labial side of talonid with three to four distinct cusps, lingual side with two to four distinct cusps; M2 with three to four distinct cusps on labial side, three to six distinct cusps on lingual side; M;j with four distinct cusps along both labial and lingual sides. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 17 Comments. — In this study I recognize Ameghino's genera Pliodolops, Or- thodolops, Anissodolops, and Archaeodolops as junior subjective synonyms of Polydolops. These generic synonymies were first proposed and justified by Simpson (1948, pp. 53-54) and are corroborated by the present study. The only deviation I make from Simpson's arrangement is in tentative retention of Pseudolops as valid; he somewhat tentatively included this genus in synonymy with Polydolops. Reasons for this move are discussed under the "Comments" section (p. 51) of Pseudolops princeps. Seven named species are here included within Polydolops: clavulus, kamektsen, rothi, winecage, serra, thomasi, and mayoi. Of the known genera of Polydolopidae, Polydolops is taxonomically the most diverse and has the longest known geologic range (Riochican through Mustersan). Of the known genera from Patagonia, Polydolops is represented by the largest number of known specimens; whether for this reason or another, the better-known species show a marked degree of variation. The known variation within and between species generally involves size, proportions, and /or minor details of cusp development (e.g., number and placement of cusps or cuspules). The included species are thus structurally quite similar, although in size they range from the smallest to some of the largest of known species of Polydolopidae. Five of the seven named species of Polydolops are known only from partial isolated lower dentitions. As a result, the possibility exists that, with better knowledge of the upper and ante- premolar dentitions of some of these more incompletely known species, further generic subdivision may prove warranted. For the present, however, it is con- venient to include these seven species within the genus Polydolops. Polydolops clavulus Ameghino, 1902b. Figures 9-12; Appendix 1. Polydolops clavulus Ameghino, 1902b, p. 40; Simpson, 1948, p. 61. Archaeodolops clavulus Ameghino, 1903, pp. 150, 174, figs. 75, 103; 1904b, p. 257. Type. — MACN 10356, greater part of a left mandibular ramus with alveolus of a large procumbent tooth (here regarded as the canine) and P2-M, complete (originally with M2 present, but lost since specimen was figured by Ameghino, 1903, figs. 73, 103). This specimen was selected lectotype by Simpson (1948, p. 61). Hypodigm. — The type (or lectotype) and MACN 10360, a fragment of a right mandibular ramus with P:J-M2 complete and alveoli of M:! (possibly same indi- vidual as MACN 10356). Horizon and Locality. — Casamayoran beds, Patagonia; no other data. Age. — Casamayoran. Revised Diagnosis. — Smallest known species of Polydolops; average ratio of P.t L / W = 1.17; P-, with prominent narrow arced rib extending from anterior edge of posterior root to tip of crown on labial surface, but with only a hint of rib on lingual surface and only slightest hint of serrations on cutting edge; M, with three distinct labial and two distinct lingual cusps; M2 with three distinct labial and three distinct lingual cusps; ratio of M> L / W = 1.06. Comments. — Polydolops clavulus is known from a relatively complete left and right mandibular ramus, possibly of the same individual. This species is easily distinguished from other Polydolopidae by its small size, structure of the P.(, and the presence of only two distinct lingual cusps on the M, . J I I 1 t Fig. 9. Polydolops clavulus Ameghino, 1902b, p. 40 (Casamayoran). MACN 10356 (lec- totype), greater part of a left mandibular ramus with alveolus of C, and P2-M, complete: a, labial; b, occlusal; c, lingual views. Scale = 6 mm. Fig. 10. Polydolops clavulus Ameghino, 1902b, p. 40 (Casamayoran). Stereopairs of MACN 10356 (lectotype), greater part of a left mandibular ramus with alveolus of C, and P2-M, complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 18 Fig. 11. Polydolops clavulus Ameghino, 1902b, p. 40 (Casamayoran). MACN 10360, a fragment of a right mandibular ramus with P:1-M2 complete and alveoli of M:1: a, labial; b, occlusal; c, lingual views. Scale = 7 mm. Fig. 12. Polydolops clavulus Ameghino, 1902b, p. 40 (Casamayoran). Stereopairs of MACN 10360, a fragment of a right mandibular ramus with P;,-M2 complete and alveoli of M3: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 19 20 FIELDIANA: GEOLOGY The alveolus of a large procumbent tooth, here regarded as the canine, is preserved in the lectotype (MACN 10356). The diastema is sufficiently well preserved to show that there were no other teeth between this tooth and the P2. Ameghino (1903, fig. 75) figured MACN 10356 and erroneously restored a small vestigial tooth in the P, (his P2) position. I could find no evidence, however, that a P, was ever present. Ameghino originally referred clavulus to the genus Polydolops, but later (1903, p. 150; 1904b, p. 257) made it the genotype of Archaeodolops. He distinguished Archaeodolops as having a P:! somewhat smaller, crown less oval and more pointed, with a trenchant border but lacking denticles and ribs, and possessing a P, similar in structure to that of P2. As noted above, there is no evidence that a P, was present in either of the two known specimens. Apart from its smaller size, there is no character which can be used to separate clavulus at the generic level from the other species here included within Polydolops. I thus follow Simpson (1948, p. 54) and return clavulus to Polydolops, where Ameghino him- self first placed it. Polydolops kamektsen Simpson, 1935a. Figures 13, 14; Appendix 1. ?Polydolops kamektsen Simpson, 1935a, p. 5, fig. 2; 1948, p. 63, pi. 5, figs. 3, 4. Type. — AMNH 28525, a fragment of a left mandibular ramus with alveoli of M, , and M2 complete (figured by Simpson, 1935a, fig. 2; 1948, pi. 5, figs. 3, 4). Hypodigm. — Type only. Horizon and Locality. — Rio Chico Formation at Canadon Hondo, Chubut Province, southern Argentina; collected by members of the Scarritt Expedition to Patagonia in 1931. Age. — Riochican. Revised Diagnosis. — Polydolops species of small size; M2 with three distinct labial and four distinct lingual cusps; ratio of M2 L / W = 1.15; M2 larger than in P. clavulus and smaller than in P. rothi, P. winecage, and P. serra. Comments. — This species is known only from the type, AMNH 28525, a fragment of a left mandibular ramus with a complete but worn M2. It is not a right mandibular ramus as recorded by Simpson (1935a, p. 5; 1948, p. 63). He tentatively placed kamektsen in the genus Polydolops, noting (1935a, p. 5) that: It is morphologically so distinctive that it probably belongs to a new genus, but the specimen is inadequate for definite decision on this point. All the morphological features in this specimen, however, suggest that it be- longs in Polydolops. In fact, the M2 of kamektsen is virtually identical to that of P. rothi (i.e., they both have three distinct labial and four distinct lingual cusps [of the latter there are three smaller posteriorly and one larger anteriorly] and simi- lar ratios of M2 L / W [1.15 and 1.12, respectively] [see Appendix 7, points 7, 8, 9]). P. rothi differs from P. kamektsen in being about 18-20% larger in known linear tooth dimensions (Appendix 1). The CV of combined linear tooth di- mensions of these species (Appendix 1) is 13.73 (L) and 15.56 (W). These values are about twice those recorded for the same dimensions in the large sample of P. thomasi (Appendix 5), so there is little doubt that the species kamektsen and rothi are distinct. P. kamektsen cannot be compared with P. winecage as the M2 is not known in that species; however, judging from the M2 alveoli in kamektsen, this Fig. 13. Polydolops kamektsen Simpson, 1935a, p. 5 (Riochican). AMNH 28525 (type), a fragment of a left mandibular ramus with alveoli of M,, and M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. Fig. 14. Polydolops kamektsen Simpson, 1935a, p. 5 (Riochican). Stereopairs of AMNH 28525 (type), a fragment of a left mandibular ramus with alveoli of M, , and M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. 21 22 FIELDIANA: GEOLOGY tooth is considerably smaller than occurs in winecage. P. kamektsen is larger in known linear tooth dimensions than P. clavulus (Appendix 1) and is smaller than P. serra (Appendix 2). Polydolops rothi Simpson, 1936. Figures 15, 16; Appendix 1. Polydolops rothi Simpson, 1935b, p. 14 (nomen nudum); 1936, p. 71, fig. 1; 1948, p. 62, fig. 7. Type. — MLP 11-122, a partial left mandibular ramus with alveoli of P2 and P:!-M2 complete (figured by Simpson, 1936, fig. 1; 1948, fig. 7). Hypodigm. — Type only. Horizon and Locality. — Rio Chico beds at Pan de Azucar west of Gaiman, Chubut Province, Argentina. Age. — Riochican. Fig. 15. Polydolops rothi Simpson, 1936, p. 71 (Riochican). MLP 11-122 (type), a partial left mandibular ramus with alveoli of P2, and P3-M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 23 EM Fig. 16. Polydolops rothi Simpson, 1936, p. 71 (Riochican). Stereopairs of MLP 11-122 (type), a partial left mandibular ramus with alveoli of P2 and P3-M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. Revised Diagnosis. — Polydolops species of medium size, ratio of P3 L/W = 1.38; ratio of M2 L/W = 1.12; readily differentiated from P. clavulus and P. kamektsen in being larger in comparable linear tooth dimensions; differentiated from similarly sized P. winecage and P. serra in P3 having a broad instead of narrow rib on labial and lingual surfaces, which is in a more medial position, and in cutting edge being unserrated; differentiated from P. serra in M2 having only three distinct labial cusps; readily differentiated from P. thomasi and P. mayoi in being markedly smaller in size. Comments. — Like the other known Riochican species of Polydolops, P. rothi is known only from its type, MLP 11-122, a partial left mandibular ramus with P3-M>. P. rothi is easily distinguished from the similarly sized P. winecage and P. serra in structural differences in the P3: the cutting edge is simple and not serrated, and the labial and lingual ribs are broad and more medially placed. P. rothi is larger in comparable linear tooth dimensions than P. kamektsen, but in structural aspects these species are virtually identical. Thus, of known species of Polydolops, P. rothi is structurally and apparently phylogenetically most closely related to P. kamektsen. Polydolops winecage Simpson, 1935a. Figures 17, 18; Appendix 1. Polydolops winecage Simpson, 1935a, p. 4, fig. 1; 1935b, p. 12; 1948, p. 63, fig. 8. Type. — AMNH 27893 (cast), a fragment of a left mandibular ramus with P3 and M,; original, with no catalogue number, in Feruglio Collection, University of Padua, Italy. Hypodigm. — Type only. Horizon and Locality. — Rio Chico beds, about six meters below lowest true ash bed of the Casamayoran horizon, Bajo de la Palangana, Chubut Province, Argentina. Age. — Riochican. Revised Diagnosis. — Polydolops species of medium size; ratio of P3 L / W = 1.10; P3 with small but distinct anterobasal cuspule, a prominent but nar- row arced rib extending from anterior edge of posterior root to tip of crown on 24 FIELDIANA: GEOLOGY Fig. 17. Polydolops winecage Simpson, 1935a, p. 4 (Riochican). AMNH 27893 (cast of type), a fragment of a left mandibular ramus with P:! and M, complete: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. I I I Fig. 18. Polydolops winecage Simpson, 1935a, p. 4 (Riochican). Stereopairs of AMNH 27893 (cast of type), a fragment of a left mandibular ramus with P3 and M, complete: a, labial; b, lingual views. Scale = 10 mm. labial surface and a similar but less prominent rib occurring on lingual surface, and cutting edge distinctly serrated; M, with three distinct labial and three distinct lingual cusps; readily differentiated from P. clavulus and P. kamektsen in being markedly larger in size, and from P. thomasi and P. mayoi in being mark- edly smaller in size; differentiated from similarly sized P. rothi in having a relatively broader P:! with a serrated edge, and from P. serra in having a slightly longer and wider P;! and a slightly shorter M, . MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 25 Comments. — Polydolops winecage is known only by its type, AMNH 27893 (cast), a fragment of a left mandibular ramus with P3-M, . Of the other known species of the genus, P. winecage is most similar to the Casamayoran P. serra. The known linear tooth dimensions of these species (Appendices 1, 2) are of similar size: the P.t has a small but distinct anterobasal cuspule, a prominent but narrow arced rib extending from the anterior edge of posterior root to tip of crown on the labial side, a similar but less prominent rib on the lingual side, and a distinctly serrated cutting edge; the M, has three distinct labial and three distinct lingual cusps. These species differ in P. winecage having a slightly longer and wider P;, and a slightly shorter M, . These observed size and pro- portional differences, however, are minor and based on comparison of but a few specimens. Some of these observed differences will probably "disappear" with the discovery and comparison of large sample sizes. These species are indeed very similar, and I here recognize P. winecage as the probable direct Riochican ancestor of P. serra. Polydolops serra Ameghino, 1902b. Figures 19-29; Appendix 2. Polydolops serra Ameghino, 1902b, p. 39; Simpson, 1948, p. 61. Pliodolops primulus Ameghino, 1902b, p. 41; 1903, p. 109, figs. 27, 65, 116; 1906, p. 359, fig. 201; Schlosser, 1923, p. 433, fig. 544D. Polydolops primulus Simpson, 1948, p. 62. Amphidolops serrifer Ameghino, 1902b, p. 42. Polydolops serrifer (misprint for Amphidolops serrifer) Ameghino, 1904b, p. 257. Anissodolops serrifer Ameghino, 1903, p. 148, fig. 72; 1904b, p. 257; 1906, p. 360, fig. 203. Polydolops bocurhor Simpson, 1948, p. 62, pi. 6, figs. 1, 2. Lectotype of P. serra. — MACN 10341, a fragment of a left mandibular ramus with Pa (double rooted) and P:! complete, and anterior edge of M, (selected lectotype by Simpson, 1948, p. 61). Type of P. primulus. — MACN 10353, a right maxillary fragment with M'~2 (figured by Ameghino, 1903, figs. 27, 65, 116; 1906, fig. 201; Schlosser, 1923, fig. 544D). Type of A. serrifer. — MACN 10359, an isolated right M2 (figured by Ameghino, 1903, fig. 72; 1906, fig. 203). Type of P. bocurhor. — AMNH 28427, a fragment of a right maxillary with M{'2 (figured by Simpson, 1948, pi. 6, figs. 1, 2). Hypodigm. — The four types and AMNH 28408, an isolated right M2 missing anteriormost edge; AMNH 28409, an isolated right M'; AMNH 28412, an iso- lated right M'; AMNH 28425, an isolated right M1; AMNH 28426, an isolated left M'; AMNH 28429, an isolated right M,; probably AMNH 28885, an isolated right M'; MACN 10361, a fragment of a right mandibular ramus with M,.2; MACN 10363, a fragment of a left mandibular ramus with posterior edge of M2 and with M;1 complete (latter two specimens are probable syntypes of P. serra). Horizon and Locality. — All specimens are from the Casamayoran beds of Chubut Province, Patagonia, southern Argentina. MACN 10341 is without specific locality data; the other MACN specimens are from the Barranca south of Lago Colhue-Huapi and were collected by Carlos Ameghino. AMNH 28425, 28426, 28427, and 28429 are from Cahadon Vaca and were collected by members of the Scarritt Expedition in 1930; AMNH 28408, 28409, and 28412 are from a location 1.5 leagues northeast of Cabeza Blanca and were collected by members of the Scarritt Expedition in 1931; AMNH 28885 is from Lomas Blancas and was collected by members of the Scarritt Expedition in 1931. 1 % i JJ in rH II CO O re u z ui u «i < £ S 01 > c re id 3 >-. bC o C E W re re ia u 3 o> ro u u o ci xT xT ■ y > T3 «s s OS i i i i i Fig. 20. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). Stereopairs of MACN 10341 (lectotype), a fragment of a left mandibular ramus with P2 and P:) complete, and anterior edge of M, : a, labial; b, lingual views. Scale = 5 mm. Age. — Casamayoran. Revised Diagnosis. — Medium-sized species of Polydolops; ratio of P3 L/ W =1.18; P:, with small but distinct anterobasal cuspule, prominent but narrow arced rib extending from anterior edge of posterior root to tip of crown on labial surface and similar but less prominent rib present on lingual surface, and with distinctly serrated cutting edge; M, with three distinct labial cusps and three to four distinct lingual cusps; M2 with four distinct labial cusps and four to six distinct lingual cusps; average ratio of M2 L/W = 1.21; M;) with four distinct labial cusps and four distinct lingual cusps; M1 with six primary cusps labial to basin and two to five stylar cuspules; M2 with six primary cusps labial to basin and one stylar cuspule; differs from similarly sized P. winecage in having a slightly shorter and narrower P:t and a slightly longer M, , from similarly sized P. rothi in P:! having a distinctly serrated cutting edge, and from P. thomasi in its 14-30% smaller size in comparable linear tooth dimensions. Comments. — Polydolops serra is the only medium-sized species of Polydolops known from beds of Casamayoran age. Next to the larger P. thomasi it is the most abundant species of Polydolops known, represented by 13 specimens col- lected from four known localities. P. serra is also the only species of Polydolops other than P. thomasi which is known from referred upper and lower dentitions. Ameghino's (1902b) original description of Polydolops serra included discus- sion of size and morphology of Pj-M:t (his M.,.7). Simpson (1948, p. 61) noted that no specimen now in the Ameghino Collection has all these teeth, although there are three specimens (MACN 10341, 10361, 10363) which are clearly con- specific and bear Ameghino's label as being of P. serra. Taken together, these Fig. 21. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). MACN 10361, a frag- ment of a right mandibular ramus with M,.2: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. 28 Fig. 22. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). Stereopairs of MACN 10361, a fragment of a right mandibular ramus with M)4: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. Fig. 23. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). MACN 10363, a frag- ment of a left mandibular ramus with posterior edge of M2 and with M3 complete: a, labial; b, occlusal; c, lingual views. Scale = 4 mm. Fig. 24. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). MACN 10359 (type of Amphidolops serrifer), an isolated right Mr. a, labial; b, occlusal; c, lingual views. Scale = 3 mm. 29 30 FIELDIANA: GEOLOGY three specimens include P>-MA as recorded in Ameghino's original description. Simpson concluded, and I agree, that these specimens are probably syntypes; he selected MACN 10341, a fragment of a left mandibular ramus with P2 and P:i complete and anterior edge of M, present, as lectotype. Of the other two speci- mens, one (MACN 10361) includes M,., and the other (MACN 10363) includes part of M2 and all of M:!. ft ig itf ] Fig. 25. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). Stereopairs of MACN 10359 (type of Amphidolops serrifer), an isolated right M2: a, labial; b, occlusal; c, lingual views. Scale — 3 mm. Amphidolops serrifer was erected by Ameghino (1902b, p. 42) on an isolated right M2 (MACN 10359). This species was later made type of Anissodolops (Ameghino, 1903, p. 148); it has no special relationship with Amphidolops serrula (see below). The genus Anissodolops was, as Simpson noted (1948, p. 54), based on an error. Ameghino reversed the inner and outer surfaces of the type tooth in his original diagnosis. Thus, his description of the labial surface is based on the lingual, and visa versa. In reality, the molar structure of Anissodolops serrifer is readily comparable to that of Polydolops serra (see below). Ameghino made no mention in the literature that he ever compared the types of these species, and this synonymy was first recognized by Simpson (1948, p. 61). The type of serrifer (MACN 10359) has four distinct labial cusps (three large and subequal ones posteriorly and one small one anteriorly) and four distinct lingual cusps (three small ones of subequal size posteriorly and one large one anteriorly). The tooth is unworn and the basin is distinctly wrinkled. In MACN 10361, a probable syntype of P. serra, the M2 has four distinct labial cusps as in MACN 10359, but in contrast to the latter there are six distinct lingual cusps (five small ones of subequal size posteriorly and a large one anteriorly). The M2 is also more worn in MACN 10361 than in MACN 10359, but it is still determinable that the basin was distinctly wrinkled. Apart from differences in relative size and number of distinct labial cusps, these specimens are comparable and may be referred without reservation to the same species. I therefore follow Simpson (1948, p. 61) and recognize A. serrifer as a junior synonym of P. serra. Unfortunately, there are no upper and lower dentitions of P. serra known to have been found in direct association. Primarily for this reason, no upper den- titions have been assigned to P. serra; the species has previously been charac- terized on lower dentitions only. During the course of this study I deemed it appropriate, for reasons discussed below, to recognize Polydolops primulus and Polydolops bocurhor as synony- mous, P. primulus having priority. P. primulus is known only from upper den- titions of Casamayoran age, and is a medium-sized member of the genus. Furthermore, seven known specimens were assignable to this "upper denti- tion" species P. primulus, while six were assignable to the "lower dentition" MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 31 I 1 i I I 1 Fig. 26. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). MACN 10353 (type of Pliodolops primulas), a right maxillary fragment with M1- complete: a, labial; b, occlusal; c, lingual views. Scale = 6 mm. species P. serra. The upper cheek teeth of P. primulus are structurally similar to those of the slightly larger Casamayoran species P. thomasi. In comparable linear tooth dimensions P. primulus differs from P. thomasi by the following percentages: M' 18-25%, M2 14-20%, M3 25-30%. In turn, comparable linear tooth dimensions of P. serra differ from those of P. thomasi by the following percentages: P;i 25%, M, 25-30%, M2 15-20%. Thus, the upper dentitions re- ferred to P. primulus and the lowers referred to P. serra are 1) from similarly aged faunas; 2) represented by approximately the same number of specimens; 3) both 32 FIELDIANA: GEOLOGY from medium-sized members of the genus and the only medium-sized repre- sentatives of Casamayoran age; 4) both structurally similar to those of the larger P. thomasi; 5) both different in comparable linear dimensions from those of P. thomasi by percentages ranging from 14 to 30 percent (P. primulus, 14-30%; P. serra, 15-30%). When I placed Mx.2 (MACN 10361) of P. serra and M12 (AMNH 28427) of P. bocurhor together, they were of comparable size and occluded per- fectly. I therefore formally recognize P. primulus (and its junior synonym P. bocurhor) as the upper dentition of P. serra. The senior synonym of the three, based on page priority, is P. serra. Pliodolops primulus was erected by Ameghino (1902b, p. 41) on a right maxil- lary fragment with M1"2 (MACN 10353). As noted by Simpson (1948, p. 54) the original diagnosis of Pliodolops primulus is descriptive and not comparative, and the distinction from Polydolops of this genus and species rests mainly on the greater development of the stylar cuspules on its M1. Comparative study of available upper dentitions of Polydolops shows that, apart from size, some minor structural differences do exist; these are here regarded as warranting specific value only. For example, the M's of P. primulus have six primary cusps labial to the basin and from two to five stylar cuspules labial to these; the M2 has six primary cusps labial to the basin and one stylar cuspule labial to these. In the comparative sample of P. thomasi the M1 has (as in P. primulus) six primary cusps labial to the basin and from one to four stylar cuspules; the M2 has five primary cusps labial to the basin and (as in P. primulus) one stylar cuspule. Pliodolops is thus regarded as a junior subjective synonym of Polydolops. Simpson (1948, p. 62) diagnosed primulus as follows: "External cuspules of [M1] large. [M1] markedly and [M2] very slightly wider than long." He assigned it to the genus Polydolops. In the same paper Simpson erected a new species, Polydolops bocurhor, and diagnosed it as follows: "External cusps of [M1] as in P. primulus. [M2] much wider than long, [M1] nearly equidimensional and rela- tively longer than in P. primulus." These species were each based on a right maxillary fragment with M1"2 and are thus readily comparable point for point. Structurally the types are virtually identical. In MACN 10353 (type of P. primulus) the M1 has six primary cusps labial to the basin and four additional stylar cuspules labial to these. In AMNH 28427 (type of P. bocurhor) there are also six primary cusps labial to the basin, but there are five stylar cuspules (four in the same position as in MACN 10353 and a small additional one at the posterolabial corner of the tooth). Fig. 27. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). Stereopairs of MACN 10353 (type of Pliodolops primulus), a right maxillary fragment with M1'2 complete: a, lingual; b, occlusal; c, labial views. Scale = 5 mm. Fig. 28. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). AMNH 28427 (type of P. bocurhor), a fragment of a right maxillary with M12: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. 33 34 FIELDIANA: GEOLOGY Fig. 29. Polydolops serra Ameghino, 1902b, p. 39 (Casamayoran). AMNH 28885, an isolated left M3: a, labial; b, occlusal; c, lingual views. Scale = 2.5 mm. Four other M's similar to AMNH 28427 and MACN 10353 are known (AMNH 28409, 28412, 28425, 28426); these all have the same six primary cusps as in those specimens, but they have stylar cuspules, ranging from two in AMNH 28409 to five in AMNH 28412. These six M's are all comparable in size (Appendix 2) and have no features which I regard as warranting specific separation of any one of them from the others. The stylar cuspules occur in the positions shown in Appendix 7; the highest numbers are the result of addition to the pos- terolabialmost corner of the tooth. Two WPs are known, one in each of the types. Both have six primary cusps labial to the basin and one mediolabial stylar cuspule (see Appendix 7, charac- ters 16, 17). Structurally these teeth are virtually identical and are readily differ- entiated only by the greater width (3.6 mm) of AMNH 28427 compared with the narrower width (2.9 mm) of MACN 10353. Considering, however, the similarity in structure and size of the M's of the types and the similarity in structure and length of the M2s, I choose to attribute the differences in widths of the M2s to variation among individuals in a single population. I do not regard the dif- ferences in widths of the M2s as warranting specific recognition. For these reasons I recognize P. bocurhor as a junior synonym of P. serra. Polydolops thomasi Ameghino, 1897. Figures 30-35; Appendices 3-5. Polydolops thomasi Ameghino, 1897, p. 497, fig. 73; 1898, p. 185; 1902b, p. 38; 1903, p. 141, figs. 63, 68, 101, 102, 104; 1904a, p. 43, fig. 26; 1906, p. 358, fig. 198; Schlosser, 1923, p. 433, figs. 544A, B, C; 1925, p. 26, figs. 37A-C; Simpson, 1948, p. 58, pi. 6, figs. 3, 4; 1964, p. 9; 1967b, p. 8. Polydolops fur Ameghino, 1902b, p. 39. Polydolops crassus Ameghino, 1902b, p. 39. Orthodolops sciurinus Ameghino, 1903, p. 131, figs. 54, 106; 1904b, p. 257 (said to be new in 1904, but publication in 1903 was prior and valid); 1906, p. 367, fig. 220. Polydolops simplex Ameghino, 1903, p. 185, fig. 119; 1904b, p. 256 (said to be new in 1904, but publication in 1903 was prior and valid). MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 35 Polydolops thomasi thomasi Simpson, 1948, p. 58, pi. 2, fig. 8, pi. 4, figs. 1, 2; Paula Couto, 1952, p. 18, fig. 5B; Pascual & Bond, 1981, pi. I, figs. A, B; pi. II, figs. A, B. Polydolops thomasi crassus Simpson, 1948, p. 59, pi. 2, fig. 9. Polydolops thomasi paahi Simpson, 1948, p. 59, pi. 5, figs. 1, 2; 1964, p. 9; 1967b, p. 9. Polydolops thomasi mara Simpson, 1948, p. 61, pi. 4, figs. 3, 4. Pseudolops princeps Ameghino, 1902b, p. 40 (partim); 1903, p. 149, figs. 73, 108 (partim). Polydolops princeps Simpson, 1948, p. 62 (partim). Lectotype of P. thomasi. — MACN 10338, a right maxillary with F-M2 (figured by Ameghino, 1897, fig. 73; 1903, figs. 63, 102, 104; 1904a, fig. 26 [top]; 1906, fig. 198 [top]; Schlosser, 1923, figs. 544A, B; 1925, figs. 37A, B; Simpson, 1948, pi. 2, fig. 8; Pascual & Bond, 1981, pi. I, figs. A, B). Type of P. fur. — MACN 10342, a fragmentary right mandibular ramus with posterior half of M,, M2 complete, M:) missing lingual side. Fig. 30. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). MACN 10337, a right mandibular ramus with P2-M2 complete and M:) missing posterolingual corner: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. Fig. 31. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). Stereopairs of MACN 10337, a right mandibular ramus with P2-M2 complete and M:, missing posterolin- gual corner: a, labial; b, occlusal; c, lingual views. Scale = 20 mm. 36 Fig. 32. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). MACN 10338 (type), a right maxillary with P*-M*: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 37 38 r \^a FIELDIANA: GEOLOGY Fig. 33. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). Stereopairs of MACN 10338 (type), a right maxillary with F--M2: a, labial; b, occlusal; c, lingual views. Scale = 15 mm. Type of P. crassus. — MACN 10349, a fragment of a left mandibular ramus with P3-M;i (figured by Simpson, 1948, pi. 2, fig. 9). Type of O. sciurinus. — MACN 10336, a right mandibular ramus with P2-M, broken, M->.3 complete (M,.3 are figured by Ameghino, 1903, figs. 54, 106; 1906, fig. 220). Type of P. simplex. — MACN 10335, a relatively complete left mandibular ramus with roots of P2, crowns of P3 and M> (figured by Ameghino, 1903, fig. 119). Type of P. t. paahi. — AMNH 28434, a right mandibular ramus with P3-M2 (figured by Simpson, 1948, pi. 5, figs. 1, 2). Type of P. t. mara. — AMNH 28921, a left mandibular ramus with P3-M2 (figured by Simpson, 1948, pi. 4, figs. 3, 4). Hypodigm. — The seven types and AC 3171, a fragment of a left mandibular ramus with M,.3; AMNH 28420, an isolated right M2 missing posterolabial corner; AMNH 28424, an isolated right M1; AMNH 28428, an isolated right M2; AMNH 28440, a left maxillary with F-M3 (figured by Simpson, 1948, pi. 6, figs. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 39 3, 4); AMNH 28443, an isolated left P3; AMNH 28444, a right mandibular ramus with P2-M, (figured by Simpson, 1948, pi. 4, figs. 1, 2; Paula Couto, 1952, fig. 5B); AMNH 28449, an associated isolated left P- and M2;AMNH 28920, an iso- lated right M2; AMNH 28924, an isolated right M1; AMNH 28925, an isolated right M'; AMNH 28926, an isolated right M2; AMNH 28927, an isolated right M,; AMNH 28930, a fragment of a mandibular ramus with P.,; AMNH 28931, a fragment of a left mandibular ramus with M2; AMNH 28934, an isolated left P3; FMNH P14717, a left mandibular ramus with M,.2 complete and labial side of M;1 present; MACN 10332c, an isolated lower right M2 (figured by Ameghino, 1903, figs. 73c-d); MACN 10332e, a right maxillary fragment with M1"2 (figured by Ameghino, 1903, figs. 73a, b); MACN 10332f, anterior edge of a P3?; MACN 10332g, anterior edge of a P3? (MACN 10332c, e-g were original cotypes of Pseudolops princeps); MACN 10337, a right mandibular ramus with P2-M2 com- plete and M3 missing posterolingual corner (figured by Ameghino, 1903, figs. 68, 101; 1904a, fig. 26 [bottom]; 1906, fig. 198 [bottom]; Schlosser, 1923, fig. 544C; 1925, fig. 37C; Pascual & Bond, 1981, pi. II, figs. A, B); MACN 10343, a fragment of a right mandibular ramus with M,.;!; MACN 10350, a fragment of a right mandibular ramus with M,.2 complete, but very worn; MACN 10351, a fragment of a left mandibular ramus with P:!-M2 complete and with roots of M3; MACN 16382, a fragment of a left mandibular ramus with P2 complete, P3 missing tip of crown, M,.2 complete and with alveoli of M3; MACN 18469, a fragment of a left mandibular ramus with P2.3 complete; MACN 18472, a frag- ment of a left mandibular ramus with M,_2 complete (same individual as MACN 18469); MLP 59-11-24-38, an isolated right M2; MLP 59-11-24-659, an isolated right M,; MLP 59-11-24-660, a fragment of an isolated M'; MLP 59-11-28-81, a fragment of a left mandibular ramus with posterior half of P3 and M2 complete; MLP 59-11-28-82, a fragment of a left mandibular ramus with P3 (base missing), no indication of a P2 on diastema, tooth clearly missing; MLP 59-11-28-83, a fragment of a left mandibular ramus with alveoli of P2 (two rooted), P3 complete; MLP 59-11-28-84, a fragment of a left mandibular ramus with roots of P3-M, and M3, and with M2 complete; MLP 59-11-28-85, a fragment of a right mandibular ramus with roots of P3 and with M,_2 present, but very worn and broken along lingual side (probably same individual as 59-11-28-^84); MLP 66-V-4-21, a fragment of a left ramus with P3-M2; MLP 66-V-4-22, a fragment of a right mandibular ramus with M,.3 very worn and broken; MLP 66-V-4-23, a frag- ment of a left mandibular ramus with alveoli of P2, and P3-M, complete; MLP 66-V-4-24, a fragment of a left maxillary with M1"2; MLP 66-V-4-26, a fragment of a left maxillary with P'-M1 complete; MLP 66-V-4-27, a fragment of a right mandibular ramus with P3-M,; MLP 66-111-28-28, a fragment of a left man- dibular ramus with M2.3; MLP 77-VI-13-9, an isolated left M,; MLP 77-VI-14-7, an isolated right M2; MLP 77-VI-14-8, an isolated left P3; MNHN No. 3, a fragment of a left mandibular ramus with M3. Horizon and Locality. — All specimens are from Casamayoran beds in Chubut Province, central Patagonia, southern Argentina; their localities of collection are as follows: Afloramiento entre Aguada La Escondida y Aguada de Batistin, Paso de Indios MLP 59-11-24-38 (collected by MLP personnel in February 1959); Aguada de Batistin, Laguna de La Bombilla, Paso de Indios MLP 59-11-28-81, 59-11-28-82, 59-11-28-83, 59-11-28-84, 59-11-28-85, 66-V-4-21, 66-V-4-22, 66-V-4-23, 66-V-4-24, 66-V-4-26, 66-V-4-27, 69-111-28-28 (collected by MLP personnel from beds of upper Casamayoran age — from "tobas rosadas", same level as Fig. 34. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). AMNH 28440, a left maxillary with F'-NP: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 40 MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 41 Oxybunotherium praecursor — see Pascual, 1965, p. 59); Barranca south of Lago Colhue-Huapi AC 3171 (collected by W. Stein in January 1912), AMNH 28434, 28440, 28449 (collected by members of Scarritt Expedition to Patagonia in 1930; AMNH 28434 was collected from below main fossil level in Casamayoran beds), FMNH P14717 (collected by G. F. Sternberg during First Marshall Field Expedi- tion to Patagonia in 1922-24), MACN 10332 (c, e-g), 10335, 10338, 10342, 10349 (specimens collected by Carlos Ameghino), MACN 18469, 18472 (specimens collected by L. G. Marshall and O. Gutierrez, February 1975), MLP 77-VI-13-9, 77-VI-14-7, 77-VI-14-8 (collected by E. Herrera from "Altura km 163 de la ruta 26 a Colonia Sarmiento," from the Great Barranca), MNHN No. 3 (labeled "Couche inf" [i.e., lower bed] and is thus from a relatively low level at Tour- nouer's "Cerro Negro" locality [see Tournouer, 1903; Simpson, 1967a], or from the Barranca south of Lago Colhue-Huapi [Simpson, 1964, p. 9; 1967b, p. 8]); Canadon Vaca AMNH 28420, 28424, 28428, 28443, 28444 (collected in 1930 by members of the Scarritt Expedition to Patagonia); Cerro Guacho, Lago La Bom- billa, Dept. Paso de Indios MLP 59-11-24-659, 59-11-24-660 (collected by MLP personnel in January 1959); Rinconada de Los Lopez AMNH 28920, 28921, 28924, 28925, 28926, 28927, 28930, 28931, 28934 (28921 and 28930 are from near the puesto of Don Mariano) (collected by members of the Scarritt Expedition to Patagonia in 1930); no data MACN 10336, 10337, 10343, 10350, 10351; uncertain MACN 16382. Age. — Casamayoran. Revised Diagnosis. — Polydolops species of large size and largest Casamayoran species of genus known; average ratio of P3 L / W = 1.18; P3 with small but distinct buttressed anterobasal cuspule, prominent but narrow arced rib ex- tending from anterior edge of posterior root to tip of crown on labial surface, equally prominent rib occurring on lingual surface, and cutting edge distinctly serrated; labial side of M, with three to four distinct cusps, lingual side with three; labial side of M2 with four distinct cusps, lingual side with four to five, ratio of M2 L / W = 1.23; labial and lingual sides of M:, each with four distinct cusps; P- with prominent but narrow arced rib extending from anterior edge of posterior root to tip of crown on both labial and lingual sides, cutting edge distinctly serrated, especially along posterior edge; P' smaller than P2, structur- ally similar except that cutting edge distinctly serrated, especially along anterior edge; M1 with six primary cusps labial to basin and from one to four stylar cuspules; M2 smaller than M1, with five primary cusps labial to basin and with one stylar cuspule; readily differentiated from smaller P. serra by comparable linear tooth dimensions 14-30% larger in size; differentiated from larger Mus- tersan P. mayoi in smaller P3 L, M-> L and W, and by average of most comparable tooth dimensions 12-16% smaller. Comments. — Polydolops thomasi is the largest and most abundant poly- dolopid in the Casamayoran fauna, and is represented by some 50 known specimens. P. thomasi was founded on a right maxillary with P2-M- and a lower "incisor". In the original description it is implied but not explicitly stated that these specimens were found in association. The maxillary (MACN 10338) is recognized by Simpson (1948, p. 58) as the type, and in this study I formally regard it as the lectotype. The lower "incisor" was not mentioned by Simpson (1948, p. 58), I could not locate it in the MACN collection, and its present whereabouts and identity are unknown. Assuming, however, that the lower 42 FIELDIANA: GEOLOGY "incisor" referred to by Ameghino is in reality the large lower procumbent tooth of P. thomasi, then it is almost certainly a canine (see p. 12). Ameghino (1902b, p. 38) redefined Polydolops thomasi on the basis of a re- ferred lower dentition(s) in which he described M,.3 (his M.w). It is not certain which specimen(s) Ameghino used for this description, but it was possibly based on MACN 10343, a fragment of a right mandibular ramus with M,.;J, as this is the only specimen in the Ameghino Collection which preserves these teeth alone. Ameghino, however (1903, figs. 68, 101), figured and referred to P. thomasi a right mandibular ramus (MACN 10337) with P2-M2 complete and M;! missing the posterolingual corner. It could be assumed that this was the, or one of the, specimen(s) upon which his 1902 description was based, yet if this were the case it seems unlikely that he would have purposefully omitted description of P2 and P3. Within P. thomasi I include as junior synonyms Polydolops fur Ameghino, 1902b, p. 39; Polydolops crassus Ameghino, 1902b, p. 39; Orthodolops sciurinus Ameghino, 1903, p. 131; and Polydolops simplex Ameghino, 1903, p. 185. These synonymies were first proposed by Simpson (1948, pp. 58-59) and are ratified in the present study. In addition, I include within P. thomasi some of the syntypes (but not holotype) of Pseudolops princeps Ameghino, 1902b, p. 40. The taxonomic disposition of the syntypes of P. princeps is discussed under the "Comments" section for that species (see p. 51); justification for the other synonymies is given by Simpson (1948). Ameghino compared P. fur, P. simplex, and O. sciurinus only with P. thomasi in the original description of each of these species; he compared P. crassus only with P. fur. He was thus aware of the close affinity of these "species". In all cases the primary features used by Ameghino to distinguish them included slight differences in overall size or relative shape of the teeth and/or mandibular ramus; in some cases mention was made of differences in number of labial and/or lingual cusps on the molars. There is little doubt, however, that the types of these "species" and the other specimens referred here to P. thomasi represent but a single species. The observed size (Appendices 4, 5) and structural dif- ferences can be attributed to sex, age, individual, geographic, and demonstra- ble time differences of the horizons from which these specimens were collected. The sample is thus of heterogeneous origin. The specimens are all from beds of Casamayoran age but not from a single level; all are from the same general area but from at least six different localities. MLP 59-11-28-82 and 59-11-28-83 are from beds of upper Casamayoran age (see p. 39). These specimens are larger in M, width, the latter in M, length, than are any of the other specimens referred here to P. thomasi. Most of the other specimens referred to P. thomasi are from beds which, on the basis of associated fauna, are regarded as somewhat older than those containing these two MLP specimens. Elsewhere (p. 48) I opt for the view that P. thomasi is the direct Casamayoran ancestor of the larger Mustersan species P. mayoi. Assuming the existence of such a lineage, then the implied size increase of P. thomasi leading to P. mayoi was already begun in some upper Casamayoran age specimens of P. thomasi. In addition, the sample size of P. thomasi is the largest for any known species of Polydolops; some teeth are represented by as many as 24 measurable specimens (Appendix 5). This and other factors do well to account for the considerable but not excessive size variation seen in the P. thomasi sample; CV values range from 1.46 to 13.33 for linear tooth dimensions, with sample sizes MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 43 ranging from 2 to 24. Of the 16 tooth dimensions listed in Appendix 5, two (P W and M, W) have CV values greater than 10. For M, W the sample size is 21 and the CV value of 10.85 is not considered extreme. For P W, however, the sample size is only three and the CV value is 13.33. This indicates that P W is extremely variable. In all cases, width of a particular tooth is more variable than its length (Appendix 5). Aspects of tooth structure and variability are sum- marized in Appendix 7. Simpson (1948, p. 57) noted some discontinuity in the distribution of variable characters within the lower dentitions of the P. thomasi sample which he studied. These differences involved absolute and relative size of teeth and of the mandibular ramus. In recognition of these demonstrable differences he was able to separate the more complete specimens into four groups, which he elected to recognize as subspecies and defined as follows: Polydolops thomasi thomasi. — Length P-M3 in type 17 mm. In referred lower jaws, length M, 5.1-5.2, length M-, 3.9-4.1, ratio of LM,: LM, 1.25-1.33. Width M, 3.8-3.9. M, pointed. Mandible slender. (1948, p. 59) Polydolops thomasi crassus. — Length M, 4.3 mm., slightly (probably not significantly) longer than P. t. thomasi. Ratio LM,:LM.> 1.14, lower than in available specimens of P. t. thomasi. Width M, 5.0, markedly greater than in P. t. thomasi. M, pointed. Mandible very stout. (1948, p. 59) Polydolops thomasi paahi. — Length M, 5.4-5.8, somewhat greater than in either P. t. thomasi or P. t. crassus. Length M-, 4.4, greater than in P. t. thomasi. Ratio LM, :LM2 1.32, about as in P. t . thomasi. Width M, 4.4-4.6, intermediate between P. t . thomasi and P. t . crassus. Shape of M, about as in those subspecies. Mandible as in P. I . thomasi. (1948, p. 59) Polydolops thomasi mara. — Length M, 5.7-5.8, comparable to P. t . paahi. Length M2 3.9, comparable to P. t. thomasi. Ratio LM, :LM2 1.49, larger than in other subspecies. Width 4.5-4.8, comparable to P. t . paahi. Apex of M, more rounded than in other subspecies, its highest part not a point of the upper end of the carinae but the rounded edge posterior to this. Mandible about as in P. t. thomasi. (1948, p. 61) One of the great drawbacks of this arrangement involves its incompatability with the subspecies concept of Recent populations. Biologists tend to recognize subspecies as mutually exclusive entities, such that only one subspecies can i 1 tj 9 fe KJ M ' M Fig. 35. Polydolops thomasi Ameghino, 1897, p. 497 (Casamayoran). Stereopairs of AMNH 28440, a left maxillary with P-M1: a, labial; b, occlusal; c, lingual views. Scale = 15 mm. 44 FIELDIANA: GEOLOGY exist for a prolonged period of time in any one area. Thus, subspecies are generally allopatric. Of the four recognized subspecies of P. thomasi, three are known from two localities (Barranca south of Colhue-Huapi; Rinconada de Los Lopez) and two from another (Canadon Vaca) (Simpson, 1948, p. 58, table 6). There are no published data documenting that these subspecies came from dif- ferent horizons or were temporally isolated; it can only be assumed that at least three of the four subspecies occurred sympatrically, in both space and time — a highly unlikely possibility given knowledge of living populations. For this reason alone I opt to reject this subspecies arrangement. There is a definite strong point, however, in Simpson's arrangement. He noted (1948, p. 57) that the most complete specimens which he had for study belonged to a different subspecies from each locality (e.g., P. t. thomasi, Cana- don Vaca; P. t. paahi, Colhue-Huapi; P. t. mara, Rinconada de Los Lopez). He concluded, and I agree, that this may indicate the existence of a geographic and temporal separation. This conclusion is highlighted by the fact that relative (much less absolute) ages of Casamayoran local faunas have yet to be established. Nevertheless, available faunal studies do suggest that time dif- ferences exist between some local faunas (Pascual, 1965). Polydolops mayoi Odreman Rivas, 1978. Figures 36-38; Appendix 6. Polydolops mayoi Pascual & Odreman Rivas, 1971, p. 381 (nomen nudum); Odreman Rivas, 1978, p. 31, pi. 1, figs. A-D. Type. — MLP 69— III— 24— 1, a right mandibular ramus with base of a large procumbent tooth (here regarded the canine), alveoli of P2, P-s missing tip of crown, M, missing anteriormost edge, and M2 missing anterolingual corner. Hypodigm. — The type and MLP 52-XI-4-176, a fragment of a right man- dibular ramus with P:i-M2 complete; MLP 59-11-28-95, a fragment of a left mandibular ramus with roots of P;i and with M,.2 nearly complete (no evidence of a P2 on diastema); MLP 59-11-28-96, a fragment of a right mandibular ramus with P3-M2 very broken; MLP 59-11-28-135, a fragment of a left mandibular ramus with M,_2 very broken and with roots of M:i; MLP 61-VIII-3-329, a frag- ment of an edentulous right mandibular ramus with posterior alveolus of P;! and alveoli of M,.3; MLP 69-111-24-12, an isolated right P3. Horizon and Locality. — All specimens are from Mustersan age beds in Chubut Province, central Patagonia, southern Argentina. MLP 69— III— 24— 1 and 69— III— 24— 12 are from La Gran Hondonada ( =el Pozon), Departamento de Tehuelches; MLP 59-11-28-95, 59-11-28-96, and 59-11-28-135 are from Cerro Conhue, Tobas entre Laguna de La Bombilla y El Toro Hosco, Camino Aguada de Batistin, Paso de Indios; MLP 61-VIII-3-329 is from Cerro Bagual, Cerro Mentira; MLP 52-XI-4-176 was collected by S. Roth from "T.i.C.B." ( = Ter- ciario inferior de Canadon [or Cerro] Blanco). Age. — Mustersan. Revised Diagnosis. — Polydolops species of very large size; only known polydolopid of Mustersan age; average ratio of P3 L/ W = 0.99; P3 with small but distinct anterobasal buttressed cuspule, prominent but narrow arced rib ex- tending from anterior edge of posterior root to tip of crown on labial surface, an equally prominent rib occurring on lingual surface, cutting edge distinctly ser- rated; M, with three to four distinct labial cusps and three distinct lingual cusps; M2 with four distinct labial cusps and four distinct lingual cusps; average .2 £~ c c E J-2&- 5 «£ s> g o c £ ° ■5 £ So ■- ! JZ Si 5 3 00 « w £. u £ ^-° L- 00 _ . tv -7 ■£ I en J. £ K£ ha V4- 60 t > ~ «> 2 — 60g 60S C £ C 60 <« o -5 c .2 r 2 - 45 46 FIELDIANA: GEOLOGY Fig. 37. Polydolops mayoi Odreman Rivas, 1978, p. 31 (Mustersan). Stereopairs of MLP 69— III— 24— 1 (type), a right mandibular ramus with base of C, alveoli of P2, P3 missing tip of crown, M, missing anteriormost edge, and M2 missing anterolingual corner: a, labial; b, occlusal; c, lingual views. Scale = 20 mm. ratio of M2 L/W = 1.09; differs in some comparable linear tooth dimensions from smaller P. thomasi by 12-16%. Comments. — Polydolops mayoi was erected by Odreman Rivas (1978) on two specimens from Mustersan age beds in Patagonia. The type, MLP 69 —III— 24— 1 , is a greater part of a right mandibular ramus with base of a large procumbent tooth (here regarded the canine), a long toothless diastema, alveolus of a P2, and with P:i-M2 relatively complete but very worn. This specimen is readily distin- guishable from undoubted Casamayoran specimens of Polydolops thomasi in the greater P;! W and greater M2 L and W; these three measurements are well outside MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 47 Fig. 38. Polydolops mayoi Odreman Rivas, 1978, p. 31 (Mustersan). Stereopairs of MLP 52-XI-4-176, a fragment of a right mandibular ramus with P3-M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. the range of the P. thomasi sample (Appendices 5, 6). These three dimensions also differ from the average of the P. thomasi sample by 21-26%. In linear tooth dimensions of P;i L and M, L and W, the type of P. mayoi falls within the range of the P. thomasi sample. The second specimen referred by Odreman Rivas to P. mayoi is MLP 69— III— 24-12, an isolated right P:). This specimen is smaller in length and width than the type (Appendix 6) and falls within the range of the sample of P. thomasi (Appendix 5). In fact, this specimen was (and is here) assigned to P. mayoi because it is found in Mustersan age beds and was collected from the same locality and horizon as that species. Had it been found by itself, it would have probably been assigned to P. thomasi, from which it is indistinguishable. Two additional specimens here referred to P. mayoi (MLP 52-XI-4-176, MLP 59-11-28-95) have, like the type, P3 L and M2 L and W either outside of or at the extreme of the range of the P. thomasi sample. In P:i L and M, L and W these specimens are, however, within the range of the P. thomasi sample (compare measurements of P. mayoi specimens in Appendix 6 with statistics of P. thomasi samples in Appendix 5). The average linear tooth dimensions of the four P. mayoi specimens given in Appendix 6 were compared with the averages of the same tooth dimensions of P. thomasi (Appendix 5). In P3 L the averages differ by less than 1%, but for the other comparable dimensions P. mayoi is larger than P. thomasi by the following percentages: P:i W 16%, M, L 16%, M, W 12%, M, L 13%, M2 W 22%. Thus, for average measurements of comparable linear tooth dimensions other than P^ L, P. mayoi is approximately 12-22% larger than P. thomasi. In addition, P. mayoi usually has a much wider P:t and a longer and wider M2 than does P. thomasi. The Mustersan species of P. mayoi can thus be distinguished from the smaller- sized Casamayoran species P. thomasi by its larger overall average size and, in some cases, larger size of particular linear tooth dimensions. In features other than size, P. mayoi and P. thomasi are very similar. In both, P3 has a small, distinct buttressed anterobasal cuspule, a prominent but narrow arced rib extending from anterior edge of posterior root to tip of crown on labial side, an equally prominent rib on lingual side, and a distinctly serrated cutting 48 FIELD1ANA: GEOLOGY edge; M, has three to four distinct labial cusps and three distinct lingual cusps; M2 has four distinct labial cusps (three large posteriorly [middle one is smallest] and one smaller anteriorly) and at least four distinct lingual cusps (three small posteriorly and one large anteriorly). Because of the striking similarity in structure of these species and their occurrence in beds of different and succes- sive ages, I recognize P. thotnasi as the probable direct Casamayoran ancestor of the slightly larger Mustersan species P. mayoi. Poly do lops? sp. indeterminate. Figure 39. A single specimen of an apparent polydolopid, PU 21998, is known from Deseadan age beds in the Salla-Luribay Basin, about 90 km southeast of La Paz in the Serrania of Sicasica, Bolivia (Patterson & Marshall, 1978, p. 90, fig. 23). PU 21998 is a fragment of a left mandibular ramus with base of a large procumbent tooth (here regarded the canine) and roots of P2 and P3. It is the latest known polydolopid and the only one of Deseadan age. Description. — Base of large procumbent tooth (canine) is narrow transversely (3.0 mm), deep dorsoventrally (5.5 mm), and projects laterally from jaw at an angle of 35° relative to a line drawn through centers of P2 and P:i; a large, edentulous diastema separates procumbent tooth from P2; P2 is very small, situated directly anterior to P;i, and double rooted (anterior root smaller than posterior); base of P2 is 2.7 mm long and 1.8 mm at posterior end; as shown by its roots, P3 is a large elongate tooth, its base about 9.6 mm long and anterior root longer (5.8 mm) than posterior (3.8 mm), but both are of about equal breadth (4.0 mm); mandibular symphysis is smooth, suggesting that rami were loosely joined and probably capable of movements independent of each other; a small mental foramen occurs 4.0 mm below base of P2; depth of ramus below middle of P:5 on labial side 14.0 mm, breadth 6.8 mm. Comparisons. — PU 21998 compares best with the large Mustersan species Polydolops mayoi. The PU specimen differs from the type of P. mayoi (MLP 69— III— 24— 1) in its significantly larger size (P;! L in P. mayoi ranges from 5.1-5.3 mm, see Appendix 6; base of P:i in PU 21998 is 9.6 mm in length) and in location of the mental foramina (in P. mayoi, one occurs 3 mm below diastema midway between the large procumbent tooth and P2 and another of comparable size occurs 6 mm below anterior root of M,, while in PU 21998 one occurs 4 mm below base of P2). The preserved similarities are such that PU 21998 appears simply to represent a large Deseadan species of Polydolops; I have tentatively referred it to this genus. An alternative possibility is that PU 21998 represents a large palaeothentine caenolestid of some sort, or is even a member of some other group. The fragmentary nature of this specimen permits such speculation, although polydolopoid affinities seem the safest interpretation based on the evidence at hand. Summary of Evolution of Polydolops The distribution of some diagnostic characters of the seven species of Polydolops is summarized in Appendix 7; the probable phylogenetic re- lationships of these species are shown in the cladogram in Figure 40. The cladogram is based primarily on consideration of the distribution of some character states in Appendix 7, as summarized in the caption to Figure 40. u rr 0> — ri D 5. ,_i g 1 | IB E -r. o c II 01 u "d rj g £ a (A to 2 rv . O 0) > o r3 3, o a, — u 9 tn ■a o 3 49 Fig. 40. Cladogram showing probable phylogenetic relationships of the known species of Polydolops. Only relative positions of pre-Riochican common ancestors are indicated. Numbers mark distribution of character states as summarized in Appendix 7. 1, Very small size; P:j lacks anterobasal cuspule, prominent but narrow arced rib ex- tending from anterior edge of posterior root to tip of crown on labial side, only hint of rib on lingual side; cutting edge of P:i unserrated; M, with three distinct labial cusps, two distinct lingual cusps; M2 with three distinct labial cusps, three distinct lingual cusps. 2, Cutting edge of P:i with only slight serrations in unworn teeth. 3, Increase in size; P:i with small but distinct anterobasal cuspule; M, with three distinct lingual cusps; M2 with four distinct lingual cusps. 4, Apparent retention of character states present in common ancestor represented by point 3. 5, Apparent retention of character states present in common ancestor represented by points 3 and 4. 6, Increase in size; Pt with broad but distinct straight rib extending from notch between roots to tip of crown on labial and lingual sides. 7, Increase in size; P3 with rib on lingual side more pronounced; cutting edge of P3 distinctly serrated; M2 with four distinct labial cusps; distinct lingual cusps on M2 sometimes exceed four. 8, Apparent retention of character states present in common ancestor represented by point 7. 9, M, with three to four distinct lingual cusps (anterior two are fused basally in some forms). 10, Increase in size: Pi with rib on lingual side prominent; M, with three to four distinct labial cusps. 11, Increase in size. 50 MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 51 The criteria employed in selection of the morphoclines for the characters used are as follows (see Appendix 7): Character 1 — The general trend in polydolopid evolution is assumed to have been from very small to very large size. Character 3 — Structure of P:i in primitive polydolopids is assumed to have been characterized by lack of a distinct anterobasal cuspule and presence of a narrow arced rib extending from anterior edge of posterior root to tip of crown on labial side with only hint (or lack) of rib on lingual side. Character 4 — Cutting edge of P8 primitively with no or only slight hint of serrations. Character 5 — M, with three (or less) distinct labial cusps in ancestral polydolopid. Character 6 — M, with two (or less) distinct lingual cusps in ancestral polydolopid. Character 7 — M2 with three (or less) distinct labial cusps in ancestral polydolopid. Character 8 — M-> with three (or less) distinct lingual cusps in ancestral polydolopid. Utilization of these morphocline series is complementary, one with the other; the final result as shown in the cladogram in Figure 40 is judged to be a reason- able approximation of the phylogenetic relationships of the species of Polydolops. Pseudolops Ameghino, 1902b Pseudolops Ameghino, 1902b, p. 40. Type. — P. princeps Ameghino, 1902b, p. 40. Diagnosis. — As for type and only known species. Known Range. — Casamayoran of Patagonia, southern Argentina. Pseudolops princeps Ameghino, 1902b. Figure 41. Pseudolops princeps Ameghino, 1902b, p. 40; 1903, p. 149, figs. 73, 108. Polydolops princeps Simpson, 1948, p. 62. Lectotype. — MACN 10332a, an isolated left P3 (figured by Ameghino, 1903, fig. 108). Hypodigm. — Type only. Horizon and Locality. — Casamayoran beds from Barranca south of Lago Colhue-Huapi, Chubut Province, Patagonia, southern Argentina; collected by C. Ameghino. Age. — Casamayoran. Revised Diagnosis. — Polydolopid of large size; P'J L = 5.9 mm, W = 4.3 mm; weak and narrow arced rib extending from anterior edge of posterior root to tip of crown; blade distinctly serrated along entire edge. Comments. — Pseudolops princeps was based on a number of isolated teeth and tooth fragments, and one jaw fragment with two teeth. These cotypes include a large isolated left P8 (MACN 10332a, figured by Ameghino, 1903, fig. 108, as M4); a slightly smaller isolated P or M1 (MACN 10332b, figured by Ameghino, 1903, fig. 108, as M'); an isolated right M2 (MACN 10332c, figured by Ameghino, 1903, figs. 73c-e, as M:>); the tip of a lower left procumbent tooth (MACN 10332d); a right maxillary fragment with M1"2 (MACN 10332e, figured by Ameghino, 1903, fig. 73a, b, as Mv,;); and the anterior edges of two left P3s (MACN 10332f-g). As the association of some or all of these teeth (except for the M12) is doubtful (see below), Simpson (1948, p. 62) elected as lectotype the larger P! (MACN 10332a). 52 FIELDIANA: GEOLOGY Fig. 41. Pseudolops princeps Ameghino, 1902b, p. 40 (Casamayoran). MACN 10332a (lectotype), an isolated left P!: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. Simpson, (1948, p. 62) noted that some of these teeth might belong to P. thomasi. I agree and here regard all of them except MACN 10332a, b, and d as referrable to that species. MACN 10332c, the isolated lower M2, measures L = 3.9 mm, W = 3.2 mm. These dimensions are well within the range of measurements of the P. thomasi sample (Appendix 5). Also, as in most known specimens of P. thomasi, there are four distinct labial cusps (three larger posteriorly and one smaller anteriorly), and four distinct lingual cusps (three smaller posteriorly, and one larger an- teriorly). MACN 10332e, the right maxillary fragment with M12, has the following measurements: M1 L = 4.3 mm, W = 4.5 mm; M2 L = 3.4 mm, W = 3.7 mm. The M1 has six distinct cusps labial to the basin and four accessory cingular cuspules labial to these. The M2 has five distinct cusps labial to the basin and one accessory cingular cuspule. In both size and morphology these teeth fall within the range of variation of the P. thomasi sample (Appendix 5). MACN 10332f-g are the specimens which Ameghino (1902b, p. 40) described as "molar 4 inferior", and which Simpson (1948, p. 62) called P8. In reality, this "tooth" consists of the anterior edges of two lower left P:!s, the broken edges of which were erroneously joined together with wax in the belief that they repre- sented the anterior and posterior halves of the same tooth. This resulted in a hybrid tooth of basic size and shape of a relatively complete lower P;i. I have separated these tooth fragments and have removed the wax. These two speci- mens demonstrate a significant point: more than one individual was repre- sented among the cotypes of P. princeps. This fact confirms the doubt expressed by Simpson that all of the cotypes may not have been found in direct association nor represent one and the same individual. MACN 10332d, the tip of a lower left procumbent tooth, has an anterolingual beveled wear surface. This tooth is smaller but structurally similar to the tip of the right lower "incisor" which Ameghino figured (e.g., 1903, fig. 22) and referred to Propolymastodon caroli-ameghinoi ( = Eudolops tetragonus, see p. 61). MACN 10332d is thus probably a polydolopid, but there is no solid evidence to warrant its assignment confidently to P. thomasi, P. princeps, or to any other known species. This stems from the fact that there are no known jaws of Polydolops which retain a complete procumbent tooth, making assignment of such isolated teeth tentative at best. MACN 10332b is the specimen which Ameghino (1903, fig. 108) called M* and which Simpson (1948, p. 62) called M, . The identity of this tooth is very un- MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 53 certain. It compares best with upper right M's of species of Abde rites. A right M1 of the Colhuehuapian species A. crispus is figured by Marshall (1976, p. 61, fig. 2) and a right M1 of the Santacrucian species A. meridionalis is figured by Mar- shall (1976, p. 69, fig. 5). MACN 10332b is larger (L = 4.8 mm, W = 3.5 mm) than known specimens of A. crispus, in which a specimen has a length of 4.6 mm and a width of 2.6 mm, and in A. meridionalis, in which three specimens have a range of 4.2-4.6 in length and 2.1-2.3 mm in width (Marshall, 1976). Structurally, however, MACN 10332b is virtually identical to specimens of Ab- derites. The anterior edge of MACN 10332b is very narrow and has two or three narrow but distinct parallel ribs along its anterolabial and anterolingual sides. The dorsalmost edge is relatively flat and unserrated, with but a slight indica- tion of a shallow concave wear surface. The posterior half of the tooth is broad and the crown height at the posterolingual corner is much less than occurs at the posterolabial end; a distinct wear facet occurs on the medioposterolingual sur- face. The labial and lingual surfaces, except for the narrow ribs along the an- terior edge, are smooth and without ornamentation. Beds of Colhuehuapian age occur at the Barranca south of Lago Colhue- Huapi, the locality where the cotypes of P. princeps were collected. In view of this fact and the demonstrated fact that at least some of the cotypes of P. princeps are from different individuals and species, I opt for the possibility that MACN 10332b represents a specimen of the caenolestid genus Abderites. It may repre- sent a large individual of A. crispus or A. pristinus, species collected from the Colhuehuapian beds at the Barranca. The possibility thus exists that MACN 10332b is derived from the Colhuehuapian, was washed out of these beds, and came to rest on the Casamayoran beds at the bottom of the Barranca. This specimen may have been collected in association with the other cotypes, but it in fact represents an element derived from the Colhuehuapian fauna. Alternatively, this tooth may represent an example of an abderitine in the Casamayoran — which, based on present knowledge, seems unlikely, since the subfamily is first known from the Deseadan (Marshall, 1976) — or it may repre- sent a polydolopid which converged upon abderitines. If an abderitine, then the tooth is an M1; if a polydolopoid, it is most likely a F. In any event, this tooth is not referrable to the same species as the type of P. princeps, nor can it even be tentatively referred to any other known polydolopoid species. The lectotype of Pseudolops princeps, the upper left F (MACN 10332a), is quite different from any other known polydolopid. It differs from Eudolops tetragonus in being larger and having a serrated cutting edge, and from Polydolops thomasi in being larger and having structural differences of the mediolabial and mediolingual vertical ribs. The posterior edge of MACN 10332a where it abuts the M1 is broad and flat, while anteriorly there is a narrow triangular-shaped flat area along the lower half of the crown which presumably marks contact with the P*. This inferred area of contact with the V1 is relatively smaller than that in P. thomasi, and is proportionately similar in size to that in Eudolops tetragonus. Overall, the lectotype of P. princeps is equally as distinct from P. thomasi as it is from £. tetragonus, a fact which supports separation of these species at the generic level. It is thus appropriate to retain Ameghino's original generic name, Pseudolops, for this species. It must be noted that premolars are unknown for Amphidolops serrula; the possibility exists that MACN 10332a represents the F of that species. Judging from the roots of the F in AMNH 28929 (A. serrula), they are quite large, 54 FIELDIANA: GEOLOGY measuring 5.0 mm in length and 4.0 mm in width and approaching the size of P. princeps (L = 5.0 mm, W = 4.0 mm). In this specimen of A. serrula, however, there is no indication that a P2 was ever present, while in P. princeps there is at least a suggestion of its presence. Clarification of the relationships of these species will come only with the discovery of more complete upper dentitions in which premolars and molars are found in direct, undoubted association. Amphidolops Ameghino, 1902b Amphidolops Ameghino, 1902b, p. 42. Anadolops Ameghino, 1903, p. 186; 1904b, p. 258 (said to be new in 1904, but publica- tion in 1903 is prior and valid). Seumadia Simpson, 1935a, p. 5. Type of Amphidolops. — A. serrula Ameghino, 1902b, p. 42. Type of Anadolops. — A. thylacoleoides Ameghino, 1903, p. 186. Type of Seumadia. — S. yapa Simpson, 1935a, p. 6. Revised Diagnosis. — Medium- to large-sized polydolopids; known dental formula It, Ct, P^> M|, P2 apparently absent; sectorial shear blades apparently P3 above, P3 and trigonid of M, below; P3 large, two rooted, posterior root twice as wide as anterior root (structure of crown unknown); upper molars low crowned and with strongly wrinkled, shallow basins; M1 significantly larger than M2; M1 with two parallel rows of labial cusps and one row of lingual cusps (each row has about five or six cusps); anterior end of M1 slightly narrower than posterior end; M2 semiquadrate in shape and with anterior end slightly broader than posterior, labial side slightly higher than lingual side, former with about four indistinct cusps fused basally and latter with about three indistinct cusps fused basally; M3 trirooted (two labial and one lingual root) and slightly nar- rower than M2; M3 triangular shaped, slightly wider than long, corners some- what elevated but without distinct cusps; M,_3 with moderately high crowns with strongly wrinkled, shallow basins, labial and lingual sides rimmed with numerous small cuspules; trigonid blade of M, better developed than in species of Polydolops, but less than in species of Eudolops. Known Range. — Riochican and Casamayoran of Patagonia, southern Argen- tina. Comments. — Simpson (1935a) erected Seumadia yapa on the basis of an iso- lated M3 (his M4) from the lower fossiliferous level (bed "h" of profile in fig. 3 of Simpson, 1935b, p. 8) of the Rio Chico beds at Cerro Redondo. He noted (1948, p. 65) that The proportions and cusp structure, or its absence, sharply distinguish this [tooth] from the homologous tooth of Polydolops. I agree. The M3 of Polydolops thomasi (AMNH 28440), for example, is notably longer than wide, the basin is slightly deeper, and the enamel less wrinkled than in S. yapa. The M3s in these species are similar in their general overall size, low crowns, shallow basins with wrinkled enamel, and absence of distinct cusps around the rim. These teeth differ from those of species of Eudolops, which are of much larger size, distinctly elongated anteroposteriorly, and have distinct cusps along the rims, especially labially and anterolingually. As Nfs of Amphidolops were unknown, Simpson was not able to compare S. yapa directly with the Casamayoran species A. serrula. He did note, however (1948, p. 65), that Amphidolops has similarly wrinkled enamel and small cusps, but the cusps (also in the MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 55 upper teeth), while small, are numerous, sharp, and distinct, and the crowns are rather high although the apical pattern is shallow. In the process of comparing S. yapa with A. serrula, I have been unable to isolate a single character which would permit distinction of these species at the generic level. In view of this, I formally recognize Seumadia as a junior subjective synonym of Amphidolops. My specific reasons for this move are as follows: 1) the M2 of A. serrula and the M3 of S. yapa have a low crown and a shallow basin in which the enamel is strongly wrinkled; 2) in the M2 of A. serrula (AMNH 28923, 28929), the labial and lingual cusps are low and difficult to differentiate; 3) in S. yapa, the comers are elevated but there are no distinct cusps; 4) in specimens of Polydolops thomasi (e.g., AMNH 28440), the structural differences from M2 to M1 are of similar magnitude to those existing between the M2 of A. serrula and the M' of S. yapa; and 5) in addition, the roots of the M' preserved in AMNH 28929 (A. serrula) show that this tooth was triangular in shape; their placement suggests that the tooth was, as in S. yapa, about equidimensional. Ameghino (1902b, p. 42) proposed a second species for Amphidolops, A. ser- rifer. He later (1903, p. 148) made this the type of a new genus, Anissodolops. As shown here (see p. 25), Anissodolops serrifer is a junior synonym of Polydolops serra Ameghino (1902b, p. 39) and thus has no special affinity with Amphidolops serrula. Amphidolops yapa (Simpson, 1935a). Figures 42-43. Seumadia yapa Simpson, 1935a, p. 6, fig. 3; 1948, p. 65, pi. 6, fig. 6. Type. — AMNH 28431, an isolated M3 (figured by Simpson, 1935a, fig. 3; 1948, pi. 6, fig. 6). Hypodigm. — Type only. Horizon and Locality. — Rio Chico Formation, 37 meters above "Banco Verde" of the Salamanca Formation, Cerro Redondo, west of Puerto Visser, Chubut Province, Argentina. Age. — Riochican. Revised Diagnosis. — Only known Riochican species of Amphidolops; similar in size to A. serrula; M} L = 3.4 mm, W = 3.5 mm. Fig. 42. Amphidolops yapa (Simpson, 1935a, p. 6) (Riochican). AMNH 28431 (type), an isolated right M3: a, occlusal; b, labial; c, lingual views. Scale = 3 mm. 56 FIELDIANA: GEOLOGY Fig. 43. Amphidolops yapa (Simpson, 1935a, p. 6) (Riochican). Stereopairs of AMNH 28431 (type), an isolated M3: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. Comments. — Amphidolops yapa is known only by an isolated M3 and cannot be compared directly with the Casamayoran species A. serrula, as no IVPs are known for that species. In view of this and the established fact that these species come from beds of different ages, as documented by study of associated fauna, their specific distinction is retained. In spite of its incompleteness, A. yapa is important, as it establishes the occurrence of this genus in the Riochican. In view of the similarity in size of A. yapa and A serrula and their occurrence in beds of successive ages, I tentatively recognize the former as the possible direct Riochican ancestor of the latter. Amphidolops serrula Ameghino, 1902b. Figures 44-49; Appendix 8. Amphidolops serrula Ameghino, 1902b, p. 42; 1903, p. 148, fig. 71; Simpson 1948, p. 65, pi. 6, fig. 5, pi. 7, figs. 1, 2. Anadolops thylacoleoides Ameghino, 1903, p. 186, fig. 120; 1904b, p. 258 (said to be new in 1904, but publication in 1903 is prior and valid); Simpson, 1948, p. 65 (as a junior synonym of Amphidolops serrula). Type of A. serrula. — MACN 10357, an isolated left M2 (figured by Ameghino, 1903, fig. 71). Type of A. thylacoleoides. — MACN 10339, a fragment of a left mandibular ramus with M, missing anterior edge, M2 complete, M;! missing posterolingual corner (figured by Ameghino, 1903, fig. 120). Fig. 44. Amphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). MACN 10357 (type), an isolated left M2: a, labial; b, occlusal; c, lingual views. Scale = 4 mm. J J I I [I Fig. 45. Amphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). Stereopairs of MACN 10357 (type), an isolated left M2: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. Fig. 46. Amphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). MACN 10339 (type of Anadolops thylacoleoides), a fragment of a left mandibular ramus with M, missing anterior edge, IvL complete, M-, missing posterolingual corner: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 57 58 FIELDIANA: GEOLOGY Hypodigm. — The two types and AMNH 28438, an isolated M1 (figured by Simpson, 1948, pi. 7, fig. 2); AMNH 28922, an isolated right M, (figured by Simpson, 1948, pi. 7, fig. 1); AMNH 28923, an isolated right M2; AMNH 28929, a fragment of a left maxillary with roots of P3, M1"2 complete and alveoli of M3 (figured by Simpson, 1948, pi. 6, fig. 5); AMNH 28933, an isolated M,; MLP 59-11-28-86, a fragment of a right maxillary with M1"2; and MLP 66-V-4-25, a fragment of a left mandibular ramus with posterior half of M, . Horizon and Locality. — All specimens are from Casamayoran beds in Chubut Province, Patagonia, southern Argentina. MACN 10357 is from the Barranca south of Lago Colhue-Huapi and MACN 10339 is without specific locality data; both specimens were collected by Carlos Ameghino. AMNH 28438 is from the Barranca south of Lago Colhue-Huapi and was collected by members of the Scarritt Expedition in 1930. AMNH 28922, 28923, 28929, and 28933 are from the northeastern part of Rinconada de Los Lopez and were collected by members of the Scarritt Expedition in 1933. MLP 59-11-28-86 and 66-V-4-25 are from Aguada de Batistin, Paso de Indios, and were collected by MLP personnel (these specimens are from the same level as Oxybunotherium praecursor, see Pascual, 1965, and are thus of Upper Casamayoran Age). Age. — Casamayoran. Revised Diagnosis. — Only known Casamayoran species of Amphidolops; similar in size to Riochican species A. yapa. Comments. — Amphidolops serrula was erected by Ameghino (1902b, p. 42) on an unworn isolated left M2 (MACN 10357) in which the crown morphology was perfectly preserved. The noteworthy features of this tooth include the high crown, shallow basin covered with strongly wrinkled enamel, and numerous low cusps along the labial and lingual edges (seven labially, five lingually). A year later Ameghino (1903, p. 186) erected Anadolops thylacoleoides on a fragment of a left mandibular ramus (MACN 10339) with M, missing anterior ■■VHH HI im- 5^— A i hh Is—tyi l! Fig. 47. Amphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). Stereopairs of MACN 10339 (type of Anadolops thylacoleoides), a fragment of a left mandibular ramus with M, missing anterior edge, M2 complete, M;1 missing posterolingual corner: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. Fig. 48. Amphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). AMNH 28929, a fragment of a left maxillary with roots of P, M1'- complete, and alveoli of M3 : a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 59 60 FIELDIANA: GEOLOGY Fig. 49. Atnphidolops serrula Ameghino, 1902b, p. 42 (Casamayoran). Stereopairs of AMNH 28929, a fragment of a left maxillary with roots of P!, M1"2 complete, and alveoli of M3: a, labial; b, occlusal; c, lingual views. Scale = 10 mm (for a and c). edge, M2 complete, and M3 missing posterolingual corner, all deeply worn. Despite the extensive occlusal wear, these teeth are characterized by shallow basins with remnants of strongly wrinkled enamel and with numerous cuspules along the labial and lingual edges. In size, shape, and structure, the M2s of A. serrula and A. thylacoleoides are virtually identical and clearly represent one and the same species. In all cases, the differences used by Ameghino to characterize these species are related to the differences in occlusal wear. This synonymy was earlier recognized by Simpson (1948, p. 65). In that same study Simpson (1948, p. 65) referred some upper teeth to this species; I totally support this move and his reasons for doing so. Eudolops Ameghino, 1897 Eudolops Ameghino, 1897, p. 498. Promysops Ameghino, 1902b, p. 36. Propolymastodon Ameghino, 1903, p. 100. Type of Eudolops. — E. tetragonus Ameghino, 1897, p. 498. Type of Promysops. — P. acuminatus Ameghino, 1902b, p. 36. Type of Propolymastodon. — P. caroli-ameghinoi Ameghino, 1903, p. 100. Revised Diagnosis. — Polydolopids of very large size; known dental formula Ij, Ct, Pf, M|; sectorials include P2-F/P;! and trigonid of M,; P2 and F well- developed, two-rooted, anteroposteriorly elongated blades with sharp, un- MARSHALL: SYSTEM ATICS OF POLYDOLOP1DAE 61 serrated cutting edges, a broad but weak medial rib along labial and lingual surface extending from base of tooth to tip of crown, posterolingual corners swollen internally, more so in P than in P; P narrower than P but both of similar length; P with small but distinct anterobasal cuspule, a feature not present on P; molar basins deep with moderately wrinkled enamel; M' slightly broader posteriorly than anteriorly, four primary cusps along labial edge, a fifth cusp between and immediately linguad of posterior two of primary cusps (a small cuspule may occur between primary labial cusp two and three); one or two large cusps on crown surface above anterolingual root and three cusps aligned in a row on crown surface above posterolingual root of M1 (more pos- terior of these smaller than two anterior, which are subequal in size); Nf- slightly broader anteriorly than posteriorly, four primary cusps along labial edge, a fifth cusp between and immediately linguad of posterior two of primary cusps, one large cusp over crown surface above anterolingual root and three cusps aligned in a row on crown surface above posterolingual root (more post- erior of these smaller than two anterior, which are subequal in size); Nf' elon- gated, triangular shaped, with apex pointing posteriorly (anterior part thus broader than posterior part), four or five distinct cusps along labial surface (first and second largest) and two distinct cusps along anterolingual side of crown; lower antepremolar teeth include two very reduced incisors and a large, single- rooted procumbent canine; P, and P2 absent; P;i well-developed (but not large compared with other polydolopids), two-rooted blade with distinct rib on mediolabial and mediolingual surfaces, cutting edge sharp, only weakly ser- rated; M,..i subequal in size; M, trigonid well elevated above rest of molar series, bladelike with weakly serrated cutting edge as in P;1; IvL talonid with four distinct cusps labially (first very small, third larger, second and fourth largest and subequal in size) and three lingually (first and second very large, fused basally, giving appearance of single large cusp); M-> and M:t have four large cusps labially and three lingually. Comments. — An excellent discussion and justification of the generic synonymies recognized here is given by Simpson (1948, pp. 65-67). Eudolops tetragonus Ameghino, 1897. Figures 50-59; Appendix 9. Eudolops tetragonus Ameghino, 1897, p. 498, fig. 74; 1898, p. 185; 1902a, p. 213; Simpson, 1948, p. 67 (partim). Promysops acuminatus Ameghino, 1902b, p. 36; 1903, p. 84, figs. 3, 7, 8, 44; 1906, p. 363, fig. 208. Propolymastodon acuminatus Schlosser, 1925, p. 26, fig. 38B. Eudolops acuminatus Simpson, 1948, p. 67. Propolymastodon caroli-ameghinoi Ameghino, 1903, p. 100, figs. 18-22; 1904a, p. 43, fig. 27; 1906, p. 362, fig. 206; Simpson, 1928, p. 8, fig. 4; Schlosser, 1925, p. 26, fig. 38A. Eudolops caroli-ameghinoi Simpson, 1948, p. 67, pi. 2, fig. 10, pi. 7, figs. 4, 5; Paula Couto, 1952, p. 21, fig. 7C. Propolymastodon cardatus Ameghino, 1903, p. 105, fig. 23; Schlosser, 1925, p. 26, fig. 38C; Simpson, 1948, p. 67 (as a junior synonym of E. acuminatus). Eudolops sp. Simpson, 1948, p. 68. Type of E. tetragonus. — MACN 10358, an isolated right M1 missing postero- labial corner (figured by Ameghino, 1897, fig. 74). Type of P. acuminatus. — MACN 10340a, anterior part of an edentulous right mandibular ramus with alveoli of two small incisors and C, part of diastema and alveoli of P8, and anterior alveolus of M, (figured by Ameghino, 1903, figs. 3, 8). 62 FIELDIANA: GEOLOGY Fig. 50. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). MACN 10334 (type of Propolymastodon caroli-ameghinoi), greater part of a left mandibular ramus with P:i-M:! complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. Type of P. caroli-ameghinoi. — MACN 10334, greater part of a left mandibular ramus with P3-M3 complete and an isolated right C, supposedly of same indi- vidual (figured by Ameghino, 1903, figs. 18-22; 1904a, fig. 27; 1906, fig. 206; Schlosser, 1925, fig. 38A; Simpson, 1928, fig. 4; 1948, pi. 2, fig. 10; Paula Couto, 1952, fig. 7C). Type of P. cardatus. — MACN 10333, part of a left mandibular ramus with M,.2 (figured by Ameghino, 1903, fig. 23; Schlosser, 1925, fig. 38C; P3 is shown in Ameghino's figures, but is not now present). Hypodigm. — The four types and AMNH 28430, an isolated right M1; AMNH 28435, a fragment of a right mandibular ramus with M2_3 complete (figured by Simpson, 1948, pi. 7, figs. 4, 5); AMNH 28437, an isolated right M3; MACN 10340b, a fragment of a left maxillary with a complete M;! (figured by Ameghino, 1903, figs. 7, 44; 1906, fig. 208; Schlosser, 1925, fig. 38B); MLP 59-11-28-79, a fragment of a right mandibular ramus with posterior root of M,, M2 missing only anterolingual corner, and alveoli of M3; MLP 59-11-28-80, a fragment of a left mandibular ramus with M2.3 complete (apparently same individual as MLP MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 63 fcjrv'p L -L Fig. 51. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MACN 10334 (type of Propolymastodon caroli-ameghinoi), greater part of a left mandibular ramus with P,-M:, complete: a, labial; b, occlusal; c, lingual views. Scale = 20 mm. 59-11-28-79); MLP 77-VI-14-5, a fragment of a right maxillary with P'-M2 com- plete. Horizon and Locality. — All specimens are from Casamayoran beds in Chubut Province, Patagonia, southern Argentina; AMNH 28430 is from Bahia Solano and was collected by members of the Scarritt Expedition in 1931; AMNH 28435 is from Cerro Blanco and was collected by members of the Scarritt Expedition in 1930; AMNH 28437 is from the Barranca south of Lago Colhue-Huapi and was collected by members of the Scarritt Expedition in 1930; MACN 10333, 10334, 10340a, 10340b, and 10358 are without specific locality data, but are probably 64 FIELDIANA: GEOLOGY Fig. 52. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MACN 10333 (type of Propolymastodon cardatus), part of a left mandibular ramus with M,.2: a, labial; b, lingual views. Scale = 10 mm. from the Barranca south of Lago Colhue-Huapi and were collected by Carlos Ameghino; MLP 59-11-28-79 and 59-11-28-80 were collected by MLP personnel from Aguada de Batistin, Laguna de La Bombilla, Paso de Indios, Proximo a casa de Porfirio Calfo (they are from the "tobas rosadas," from the same level as Oxybunotherium praecursor and are thus of upper Casamayoran age, see Pascual, 1965, p. 59); MLP 77-VI-14-5 was collected by H. Herrera from the Barranca south of Lago Colhue-Huapi. Age. — Casamayoran. Revised Diagnosis. — Differs from Eudolops hernandezi in that known compa- rable (in some cases average) linear tooth dimensions are smaller by the fol- lowing percentages: M1 L 14%, M2 L 27%, M2 W 12%, M3 L 30%, M3 W 12% (M1 W about same in both species); anterior end of M1 proportionately smaller in occlusal view, compared with posterior end, than occurs in E. hernandezi; M3 with four distinct labial cusps (compared with five in £. hernandezi) and pro- portionately shorter (L/W = 1.16) than in E. hernandezi (L/W = 1.47). Comments. — Eudolops tetragonus was founded by Ameghino (1897, p. 498) on an isolated right M1 (MACN 10358). The lingual side of the tooth is deeply eroded by occlusal wear and the posterolabial corner is now missing. This breakage almost certainly resulted after this tooth was figured in occlusal and labial views by Ameghino (1897, fig. 74), since there is no indication of break- age in that illustration; the other coronal features, including the deeply eroded lingual wear basin, are precisely indicated. As pointed out by Ameghino, this tooth is easily distinguished from those of Polydolops by its greater size, smaller crown area over the anterolingual root than over the posterolingual root, only four distinct cusps labial to the basin (instead of six), and a lack of accessory stylar cuspules labial to these. Ameghino (ibid.) believed the type to be an M1 (his Mr>). Simpson (1948, p. 67) compared the type with a left maxillary fragment with M1"3 (AMNH 28932), which he tentatively referred to E. tetragonus. He concluded that the type com- pared most closely with the M2 (his M3) of AMNH 28932 and regarded it as such. In this study (see p. 70), I recognize AMNH 28932 as the type of a new species of Eudolops and the type of E. tetragonus as an M1. The identity of the type of E. tetragonus as an M1 or M2 became clear after it was compared with MLP 77-VI-14-5, a right maxillary fragment with P2-M2, MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 65 collected in 1977 by Senorita Hebe Herrera from Casamayoran beds at the Bar- ranca south of Lago Colhue-Huapi. The type of E. tetragonus is slightly wider than the M1 in MLP 77-VI-14-5 (see Appendix 9), but these teeth are otherwise identical in length and in all comparable aspects of crown morphology. The M1 of MLP 77- VI-14-5 is less worn and retains more details of crown morphology than the type. In MLP 77-VI-14-5 and the type (see Ameghino, 1897, fig. 74), there are four primary cusps along the labial edge of the tooth and a fifth cusp between and immediately lingual to the posterior two of these. Along the lin- b Fig. 53. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MACN 10340a (type of Promysops acuminatus), anterior part of an edentulous right man- dibular ramus with alveoli of two small incisors and C, part of diastema and alveoli of P:t, and anterior alveolus of M,: a, labial; b, occlusal views. Scale = 15 mm. 66 FIELDIANA: GEOLOGY Fig. 54. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). MACN 10358 (type), an isolated right M' missing posterolabial corner: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. gual side of the M1 in MLP 77- VI-14 -5 there is one large cusp on the crown surface above the anterolingual root and three cusps in a row on the crown surface above the posterolingual root (the most posterior of these is smaller than the two anterior, which are subequal in size). The type of E. tetragonus differs from the M2 of MLP 77- VI-14 -5 in being distinctly shorter (see Appendix 9) and in its anterior half being distinctly wider than the posterior half (in the M's the anterior half is narrower than the posterior). Simpson (1948, p. 68) referred an isolated, unworn right M1 (AMNH 28430) to Eudolops sp. This specimen is virtually identical in size to MACN 10358 and the M1 of MLP 77- VI-14 -5 (see Appendix 9). In view of the size similarity and agreement in general coronal morphology with these specimens, I here refer AMNH 28430 to E. tetragonus. There are, however, two minor differences in cusp structure between MLP 77-VI-14-5 and AMNH 28430 which require men- tion, although I choose to attribute these differences to individual variation. In AMNH 28430, an additional small cuspule occurs between the second and third large labial cusp and two distinct cusps occur on the crown above the anterolin- gual root, rather than one as in MLP 77- VI-14 -5. The latter feature may be attributed, at least in part, to the differences in occlusal wear between these specimens, as MLP 77- VI-14 -5 shows some wear while AMNH 28430 is un- worn. Along the lingualmost surface of the crown above the anterolingual root of the MLP specimen, there occurs a slight vertical depression which suggests i_j i i i i Fig. 55. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MACN 10358 (type), an isolated right M1 missing posterolabial corner: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 67 Fig. 56. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). MACN 10340b, a fragment of a left maxillary with a complete M1: a, labial; b, occlusal; c, lingual views. Scale = 4 mm. that two cusps were originally present but have been lost as a result of occlusal wear. The single anterolingual cusp in MLP 77-VI-14-5 may thus represent the basally fused crowns of two once-distinct cusps. The M1 of £. tetragonus is represented by two referred specimens. MACN 10340b, a fragment of a left maxillary with a complete M', was figured by Ameghino (1903, figs. 7, 44; 1906, fig. 208) as an "[u]ltima muela inferior del lado izquierdo," and was referred to Promysops acuminatus Ameghino (1902b, p. 36), a species founded on the basis of an edentulous lower jaw fragment. MACN 10340b is virtually identical in size and structure to the second referred specimen, AMNH 28437, an isolated right M:l. Both specimens are distin- guished by two large cusps along the anterolabial and anterolingual sides of the crown and the presence of two small cusps along the anterolabial edge. The anterior part of the M' is much broader than the posterior. My criteria for assigning these isolated M:is to E. tetragonus are as follows: 1) they are of the size expected for the M1 of this species, as evidenced from structure and size of the M;) and M2; and 2) they have a well-developed basin with moderately wrinkled enamel and well-developed marginal cusps (at least anterolabially and anterolingually), as do M1"2. Of less yet notable significance is the fact that the species to which MACN 10340b was originally referred by Ameghino, P. acuminatus, is for reasons given below also referred to £. tet- ragonus. Fig. 57. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MACN 10340b, a fragment of a left maxillary with a complete M1: a, labial; b, occlusal; c, lingual views. Scale = 5 mm. 68 FIELDIANA: GEOLOGY The P8 and P8 of £. tetragonus are known only in MLP 77- VI-14 -5. Both teeth are well-developed, two-rooted anteroposteriorly elongated blades in which the cutting edges are sharp but unserrated. Both have broad but weak medial ribs along the labial and lingual surfaces, which extend from the base of the tooth to the tip of the crown. The posterolingual corners of both are swollen internally, more so in P! than in F2. The P2 is narrower than the P5 and has a small but distinct anterobasal cuspule not present in the P1. The lower dentitions of E. tetragonus are represented by the types of Pro- mysops acuminatus Ameghino, 1902b, p. 36 (MACN 10340a), Propolymastodon caroli-ameghinoi Ameghino, 1903, p. 100 (MACN 10334), Propolymastodon car- datus Ameghino, 1903, p. 105 (MACN 10333), and by three additional referred specimens (AMNH 28435, MLP 59-11-28-79, MLP 59-11-28-80). I will first dis- cuss reasons for recognizing P. acuminatus, P. caroli-ameghinoi, and P. cardatus as synonymous, and will then present justification for regarding the types of these species and the referred specimens as the lower dentition of E. tetragonus. Promysops acuminatus Ameghino (1902b, p. 36) was founded on an edentu- lous fragment of a right mandibular ramus (MACN 10340a) with the alveoli of three antepremolar teeth, alveoli of P2, and anterior alveolus of M,. The labial- Fig. 58. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). MLP 77-VI-14-5, a fragment of a right maxillary with P^-M2 complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 69 Fig. 59. Eudolops tetragonus Ameghino, 1897, p. 498 (Casamayoran). Stereopairs of MLP 77-VI-14-5, a fragment of a right maxillary with F-M2 complete: a, labial; b, lingual views. Scale = 10 mm. most of the three antepremolar teeth is very large; the two lingual ones are very small and are set one above the other. This arrangement in relative size and placement of the antepremolar teeth is identical to that seen in Epidolops ameghinoi (see p. 76). Assuming these teeth to be homologous in the two species, then the large labial alveolus in MACN 10340a can be regarded as a canine and the two smaller lingual ones as incisors. The diastema is sufficiently well preserved to demonstrate that P, and Pj were absent. The posterior edge of the mandibular symphysis extends to below the anteriormost edge of the P.,. So far as comparable, the type of P. acuminatus agrees well with that of P. caroli- ameghinoi (see below); these species are regarded as synonymous. Ameghino (1903, figs. 7, 44; 1906, fig. 208) referred a second specimen (MACN 10340b) to P. acuminatus; he discussed and figured it as an "[u]ltima muela inferior del lado izquierdo." This specimen is, as noted above, a frag- ment of a left maxillary with a complete VP, referrable to E. tetragonus. Propolymastodon caroli-ameghinoi was erected by Ameghino (1903, p. 100) on the greater part of a left mandibular ramus (MACN 10334) with the roots of two antepremolar teeth, P.j-M.{ complete, and a supposedly associated but isolated right lower "incisor". As seen in Ameghino's illustration (1903, fig. 19), the broken anterior edge of the mandibular ramus preserves the root of a single- rooted tooth lingually and the alveolus of another single-rooted tooth im- mediately labial to it. The symphysis of MACN 10334 is broken further pos- teriorly than occurs in the type of P. acuminatus (MACN 10340a). By extrapola- tion posterior to the alveoli in MACN 10340a, the two teeth in MACN 10334 can be interpreted to represent the posterior edge of the canine on the labial side and the lower of the two incisors on the lingual side. The upper of the two incisors as preserved in MACN 10340a is thus not preserved in MACN 10334. 70 FIELDIANA: GEOLOGY The isolated lower right procumbent tooth was interpreted to be an incisor by Ameghino and was figured by him (e.g., 1903, fig. 22) in four views. I could not locate this tooth in the Ameghino Collection, although Ameghino's figures are adequate for comparative purposes. The crown of the tooth is slightly concave lingually, with a distinct wear surface dorsolingually; enamel occurs only along the labial and ventral surfaces. The tooth is larger than but structurally similar to the canine of £. ameghinoi and, assuming that this tooth is indeed referrable to the same species as the left jaw of P. caroli- ameghinoi , it may be regarded as a lower canine of £. tetragonus. As seen in MACN 10334, the P( is a well-developed (but not large compared with other polydolopids), two-rooted blade with a distinct rib on the mediola- bial and mediolingual surface. The cutting edge is sharp and weakly serrated. The molars are subequal in size, both in length and width, and the enamel in the basins is moderately wrinkled. The trigonid of the M, is well elevated above the rest of the molar row and bladelike, with a weakly serrated cutting edge as in the P;!. The M, talonid has four distinct cusps labially (the first is very small, the third is larger, and the second and fourth are largest and subequal in size) and three cusps lingually (the first and second are very large and fused basally, giving the appearance of a single large cusp). The M2 and M:3 each has four large cusps labially and three lingually. Propolymastodon cardatus was erected by Ameghino (1903, p. 105) on a frag- ment of a left mandibular ramus (MACN 10333) with P;i-M2. This specimen was described and figured by Ameghino (e.g., 1903, fig. 23) as having the P3, although this tooth is now missing, as I could not locate it in the Ameghino Collection. The description of the P:i given by Ameghino and my personal study of the Mj.2, which are still preserved, reveal that in size (Appendix 9) and structure these teeth are virtually indistinguishable from the type of P. caroli- ameghinoi (MACN 10334). The only discernable differences are the slightly smaller M2 in MACN 10333 and the slightly larger third cusp on the labial side of the M, talonid in MACN 10334. Thus, the types of P. acuminatus, P. caroli- ameghinoi, and P. cardatus are of one and the same species and are here regarded as synonymous. Association of the lower dentitions for which the name of P. acuminatus has priority and the upper dentitions for which the name E. tetragonus has priority is based on their similarity in size and complementarity in structure. These similarities are readily seen on comparing MACN 10334 (Pj-M3) with MLP 77- VI- 14-5 (P2-M2). The premolars are similar in structure and, although well developed, are not large in comparison with other polydolopids. The upper and lower molars are characterized by being deeply basined with moderately wrin- kled enamel, with large cusps along the labial and lingual sides of the basins. Lastly and most convincingly, when comparison is made of complementary structures and wear facets, these dentitions can be demonstrated to occlude almost perfectly. For these reasons I recognize Promysops acuminatus, Pro- polymastodon caroli- ameghinoi , and Propolymastodon cardatus as subjective junior synonyms of Eudolops tetragonus. Eudolops hernandezi sp. nov. Figures 60-61; Appendix 9. ?Eudolops tetragonus Simpson, 1948, p. 67, pi. 7, fig. 3. Etymology. — hernandezi, named in honor of Justino Hernandez who, as a MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 71 Fig. 60. Eudolops hernandezi sp. nov. (Casamayoran). AMNH 28932 (type), a fragment of a left maxillary with worn M13: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. member of the Scarritt Expedition to Patagonia in 1933, collected the specimen. Type. — AMNH 28932, a fragment of a left maxillary with worn M1'3 (figured by Simpson, 1948, pi. 7, fig. 3). Hypodigm. — Type only. Horizon and Locality. — Rinconada de Los Lopez, north-northeast of Tapera de Lopez, Chubut Province, Patagonia, southern Argentina. 72 FIELDIANA: GEOLOGY Age. — Casamayoran. Diagnosis. — Largest known species of Casamayoran Polydolopidae; differs from E. tetragonus in comparable (in some cases average) linear tooth di- mensions, being larger by the following percentages: M1 L 14%, M2 L 27%, M2 W 12%, M3 L 30%, M3 W 12% (M1 W is about the same in both species); anterior end of M1 proportionately larger in occlusal view relative to posterior end than in E. tetragonus; M3 with five distinct labial cusps (four in E. tetragonus) and proportionately longer (L/W = 1.47) than in E. tetragonus (L/W = 1.16). Comments. — The type of Eudolops hernandezi (AMNH 28932) was tentatively referred by Simpson (1948, p. 67) to Eudolops tetragonus. Based on comparison with AMNH 28932, Simpson regarded the type of E. tetragonus (MACN 10358) as an M2 (his M3), although I have attempted to demonstrate elsewhere (see p. 64) that the type is an M1. The type of E. hernandezi differs from the type and referred specimens of E. tetragonus as follows: 1) most comparable linear tooth dimensions are larger (see diagnosis above); 2) the anterior end of the M1 is proportionately larger in occlusal view than the posterior relative to E. tetragonus; 3) the M3 has five distinct labial cusps compared with four in E. tetragonus; and 4) the M3 is absolutely and relatively longer than in E. tetragonus. The M1 of E. hernandezi Fig. 61. Eudolops hernandezi sp. nov. (Casamayoran). Stereopairs of AMNH 28932 (type), a fragment of a left maxillary with worn M1'3: a, labial, b, occlusal; c, lingual views. Scale = 20 mm. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 73 also has a cuspule between the second and third large cusps along the labial side of the tooth, while in E. tetragonus a cusp in this position is known only in a single specimen (AMNH 28430). The phylogenetic relationship of £. tetragonus and E. hernandezi is subject to debate. The type of E. hernandezi is the only specimen of Eudolops known from the Casamayoran beds at Rinconada de Los Lopez. Although all specimens of E. tetragonus are also from beds of Casamayoran age, they are from different lo- calities than that of E. hernandezi. Specimens referred to Polydolops thomasi are known from the same localities as E. tetragonus as well as from the Rinconada de Los Lopez, suggesting a similar age for each of these local faunas. The possibil- ity nevertheless exists that the Casamayoran local fauna from the Rinconada de Los Lopez is slightly later than local faunas from the other Casamayoran beds which have yielded specimens of E. tetragonus. This being the case, E. hernan- dezi could represent a slightly larger descendant of a population of E. tetragonus. This suggested relationship, however, is only speculative; it has not been, and based on present evidence cannot be, demonstrated that such a time difference exists between these local faunas. It is more appropriate to recognize the more likely existence of two contemporaneous and possibly competitively exclusive species of Eudolops during Casamayoran time in Patagonia. E. hernandezi is apparently the more derived of the two and may be regarded as having evolved from an ancestral form morphologically similar or even identical to E. te- tragonus. Subfamily EPIDOLOPINAE Pascual & Bond, 1981, p. 483. Diagnosis. — Polydolopids of medium size; dental formula %, C}, Pj^, Mj; molars quadritubercular, sectorial blades P/P:{; I1'3 present but very reduced, single rooted, apparently proodont; I, large, I2 very reduced (both procumbent); upper and lower C large, single rooted, semiprocumbent; PI sometimes absent, when present small, vestigial, single (rarely double) rooted [P1 and V1 are dou- ble rooted in MNRJ 2879-V] with crown tilting toward C, separated from C by short diastema and from P2 by long diastema; P2 very small, single or double rooted, crown without cusps, flattened, irregularly oval in outline; P3 with axis of tooth set at about 30° angle relative to axis of molar series; P'1 very large blade with one or two narrow ribs along anterolabial and anterolingualmost edges, dorsal and dorsoposterior edge of blades strongly serrated; P;1 very large blade with distinct but narrow rib extending from anterior edge of posterior root to tip of crown on labial side, a very narrow, faint rib occurring along anterolin- gual edge, blade edge strongly serrated; upper and lower molars bunolopho- dont with weak lophs connecting primary labial and lingual cusps, basins shal- low with very lightly wrinkled enamel, molars decreasing rapidly in size from M1 to M4 and from M2 to M4; M'"3 with two primary lingual cusps, two primary labial cusps, and two somewhat smaller cusps between these but set closer to labial primary cusps; M1*8 with three distinct roots, two labial ones unequal in size, lingual one enormous and as wide as combined width of two labial roots; M1 with distinct but narrow stylar shelf; M2 similar in structure to M1 but stylar shelf either absent or only a hint of its presence; M2 smaller than M1 , posterior edge narrower than anterior (posterolabial primary cusp more reduced than in M1 but still well developed); M3 similar to M1"2 but much smaller, posterior edge very narrow, posterolabial and posterolingual cusps very reduced, no trace of 74 FIELDIANA: GEOLOGY stylar shelf; M4 very reduced, oval in occlusal view, single rooted; lingual edges of M,.3 trigonids weakly serrated; M, trigonid elevated above rest of molar series but not modified into blade (labial and lingual trigonid cusps distinct), antero- labial cingular shelf weakly developed; M2 trigonid somewhat elevated above talonid but not to degree seen in M,; M, greatly reduced in size, single rooted, oval in occlusal view, only hint of distinction between two primary labial cusps, lingual cusps indistinct. Known Range. — Riochican of Patagonia, southern Argentina, and Brazil. Epidolops Paula Couto, 1952 Epidolops Paula Couto, 1952, p. 7. Type. — E. ameghinoi Paula Couto, 1952, p. 7. Revised Diagnosis. — As for the subfamily Epidolopinae. Known Range. — Riochican of Argentina and Brazil. Comments. — Undoubted epidolopines, apparently representatives of the genus Epidolops, have been found in a new Riochican locality in the "Banco Negro Inferior" at the east end of the Barranca south of Lago Colhue-Huapi, Chubut Province, Patagonia, southern Argentina (Pascual & Bond, 1981, p. 483). These specimens, now being catalogued into the MLP collections, were collected by use of screen washing techniques by Robert F. Butler, Malcolm C. McKenna, William D. Turnbull, and L. G. Marshall during January and Febru- ary 1979. Epidolops ameghinoi Paula Couto, 1952. Figures 62-68; Appendices 10-12. Epidolops ameghinoi Paula Couto, 1952, p. 7, figs. 1-4, 5A, 6A, 7A; 1961, p. 324, figs. 1-6; 1970, p. 19. Epidolops gracilis Paula Couto, 1952, p. 10; 1961, p. 328, fig. 7. Type of E. ameghinoi. — DGM 321-M, a nearly complete but much crushed and flattened skull, lacking posterior part of braincase, part of basisphenoid, and anterior end of premaxillae, with left P2-M2 and right P'-M2 complete and alveoli of left and right C and M5 "* and left P2; an associated mandible lacking only anterior end of incisor border, with right I,^, left C, and left and right P2-M3 complete and alveoli of Mj (skull and/or upper dentition are figured by Paula Couto, 1952, figs. 1, 2; 1961, figs. 1, 2; mandible and/or lower dentition are figured by Paula Couto, 1952, figs. 2, 3, 5A, 6A, 7A; 1961, figs. 3, 4, 5). Type of E. gracilis. — DGM 188-M, a fragment of a right mandibular ramus with P2-M3 and alveoli of M4 (figured by Paula Couto, 1961, fig. 7). Hypodigm. — The two types and AMNH 49806 ( = MNRJ 1346-V), a frag- ment of a right mandibular ramus with roots of P2 and with P;i complete; AMNH 49807 ( = MNRJ 1375 -V), a fragment of a left mandibular ramus with M,.2 complete and alveoli of M3.4; AMNH 49808 ( = MNRJ 1379-V), a fragment of a right mandibular ramus with root of C, P2-M2 complete and alveoli of M3.4; AMNH 49809 ( = MNRJ 1380 -V), a fragment of a right mandibular ramus with alveoli of C and M3.4, and P2-M2 complete; AMNH 49810 ( = MNRJ 1391-V), a fragment of a left mandibular ramus with alveoli of C, and P2.3 complete; AMNH 49811 ( = MNRJ 1393 -V), a fragment of a right mandibular ramus with P2 and M,.2 complete and P3 missing tip of crown; AMNH 49812 ( = MNRJ 1412-V), a fragment of a right maxillary with F-M2 complete; AMNH 49813 ( = MNRJ 1406-V), a fragment of a left maxillary with F-M1 complete; AMNH MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 75 49814 ( = MNRJ 1404-V), a fragment of a left mandibular ramus with base of M,, M2 complete, and alveoli of M3; AMNH 49860 (exchange from DGM in 1951), a right mandibular ramus with alveoli of C-P;i and M3.4, and M,.2 com- plete; DGM 170-M, a right mandibular ramus with alveoli of C-M, and M», and M2.3 complete; DGM 171-M, a left mandibular ramus with alveoli of I14, C, and Mi and with P2-M3 complete; DGM 172-M, an edentulous right mandibular ramus; DGM 173-M, a fragment of a right mandibular ramus with M2 complete and alveoli of P3-M, and M^_4; DGM 174 -M, a fragment of a right mandibular ramus with P3-M3 complete and alveoli of M4; DGM 175-M, a fragment of a right mandibular ramus with P3-M2 complete, and alveoli of M3.4; DGM 176-M, a right mandibular ramus with roots of P2, and with P3-M3 complete; DGM 177-M, a fragment of a right mandibular ramus with roots of P2, P3-M3 com- plete, and roots of M4; DGM 178-M, a left mandibular ramus with roots of Ii-C and M^, and with P2-M2 complete (figured by Paula Couto, 1952, fig. 4); DGM 179 -M, a fragment of a right mandibular ramus with P2-M3 complete and alveoli of M4; DGM 180-M, a fragment of a left mandibular ramus with M,.2 complete and roots of M,3.4; DGM 181-M, a right mandibular ramus with alveoli of I,-C and M3.4, and with P2-M2 complete; DGM 182-M, a fragment of a left mandibular ramus with roots of P2 and M,.4/ and P3 complete; DGM 183-M, a fragment of a left mandibular ramus with P2-M2 complete and alveoli of M3.4; DGM 184-M, a right mandibular ramus with P2-M2 complete and alveoli of Fig. 62. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). Stereopairs of DGM 321-M (type), a nearly complete but much flattened skull, lacking posterior part of brain- case, part of basisphenoid and anterior end of premaxillae, with left F-M2 and right F-M2 complete, and alveoli of left and right C, M3"4, and left P2: occlusal view. Scale = 30 mm. Fig. 63. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). Stereopairs of DGM 321-M (type), a nearly complete mandible lacking only anterior end of incisor region, with right I,.2/ left C, and left and right P2-M;i present, and with alveoli of IVL.: a, occlusal; b, right lateral views. Scale = 20 mm. 76 MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 77 M3.4; DGM 185-M, a fragment of a left mandibular ramus with roots of M,.4; DGM 187-M, a fragment of a left mandibular ramus with alveoli of C-P3 and M^, and M,.2 complete; DGM 189-M, an edentulous right mandibular ramus; DGM 190-M, a fragment of a left mandibular ramus with roots of I,— P2 and M3.4, and with P3-M2 complete; DGM 191-M, a fragment of a left mandibular ramus with P2-M, complete and roots of M2.4; DGM 192-M, a fragment of a right mandibular ramus with alveoli of M,.4; DGM 193-M, a fragment of a right mandibular ramus with M,.2 complete and alveoli of M3.4; DGM 194-M, an edentulous right mandibular ramus; DGM 195-M, a fragment of a left man- dibular ramus with alveoli of M,.4; DGM 196-M, a fragment of a left mandibular ramus with P3-M2 complete and alveoli of M:{.4; DGM 197-M, a fragment of a left mandibular ramus with M2.3 complete and alveoli of Mj; DGM 198-M, a frag- ment of a left maxillary with PMvF complete; DGM 199-M, a fragment of a right maxillary with P-M2 complete; DGM 200-M, a fragment of a right maxillary with P complete and alveoli of P2 and M1"2; DGM 201-M, a fragment of a right maxillary with P; DGM 202-M, a fragment of a left maxillary with part of P; DGM 203-M, a fragment of a left maxillary with P^-M3 complete; DGM 204-M, a fragment of a left maxillary with P'-M3 complete; DGM 205-M, a fragment of a left maxillary with P-M1 complete and alveoli of M2; DGM 206-M, a fragment of a left maxillary with M1"3 complete and alveoli of M4; DGM 207-M, a fragment of a left mandibular ramus with alveoli of I,— P, , P2 present, base of P3, and alveoli of M,.2; DGM 208-M, a fragment of a right mandibular ramus with M, and M3 complete and roots or alveoli of P2.3, M2 and M*; DGM 209-M, an edentulous fragment of a right mandibular ramus with alveoli of P2-M»; DGM 635-M, a left maxillary fragment with P2'3 and alveoli of M1; DGM 636-M, a nearly complete left mandibular ramus with P2-M, present, and alveoli of M2.4; DGM 637-M, a partial right mandibular ramus with P3-M3 present, and alveoli of P2 and M,; DGM 638 -M, a fragment of a right mandibular ramus with P3-M2 Fig. 64. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). Stereopairs of DGM 898-M, a fragment of a right maxillary with alveoli of C, P1 and P-M4, and with P2 complete: occlusal view. Scale = 15 mm. 'o ^^ * § c *J -C "f E v JA b.2 a ».w o •S ■ ^ - a.* in 2 « 3 CM 3 <* r 1 "* U ft 'Si a g « o 3 o to ■s ■a" ? 3 . 3 x S 3 .2 jy ja |S "<3 78 MARSHALL: SYSTEM ATICS OF POLYDOLOPIDAE 79 present, and alveoli of M3.4; DGM 639-M, a fragment of a left mandibular ramus with M,.;, present and alveoli of M»; DGM 640-M, a partial left mandibular ramus with M2 present and alveoli of P2-M, and M3.4; DGM 641-M, a partial right mandibular ramus with M,.2 present and alveoli of M;1.4; DGM 800-M, a right maxillary fragment with F-M3 complete; DGM 892-M, a right mandibular ramus with alveoli of I,-C and M2.4, and P2-M, complete; DGM 893-M, a frag- ment of a left maxillary with F2-M2 complete; DGM 897-M, a fragment of a right mandibular ramus with M2.4 complete; DGM 898-M, a fragment of a right maxillary with alveoli of C and P', P2 complete, and alveoli of P'-M4; DGM 899-M, a fragment of a left mandibular ramus with M2 complete and alveoli of C-M, and M^; DGM 900-M, a left mandibular ramus with alveoli of C and P, , roots of P2, P3-M2 complete, and alveoli of M3.4; DGM 901-M, a right mandibu- lar ramus with alveoli of P,.2/ P3-M2 complete, and alveoli of M3.4; DGM 902-M, a left mandibular ramus with P3 complete and roots of P,.2 and M,.4; DGM 903-M, a right mandibular ramus with alveoli of C-P2 and M2.4, and P3-M, complete; DGM 904 -M, a fragment of a left mandibular ramus with alveoli of C-P3 and with M,.2 complete; DGM 905-M, a fragment of a right mandibular ramus with P2-M2 complete; DGM 906-M, a fragment of a right mandibular ramus with P2-M2 complete; DGM 907-M, a fragment of a left mandibular ramus with M,.3 complete and roots of M^ DGM 908-M, a fragment of a left mandibular ramus with alveoli of C-Pt, roots of P2, and with P3 complete; DGM 909-M, a right maxillary fragment with M1'3 complete and alveoli of M4; DGM 910-M, a right maxillary fragment with P'-M1 complete and roots of M2"4; DGM 911-M, a right maxillary fragment with P-M2 complete; DGM 912-M, a right maxillary fragment with P2-MJ complete and alveoli of M4; DGM 913-M, a right maxillary fragment with P2-M2 complete; DGM 914-M, a right maxillary frag- ment with alveoli of P2, P3-M2 complete; DGM 915-M, a left maxillary fragment with P3-M' complete and alveoli of M2; DGM 916-M, a fragment of a left maxil- Fig. 66. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). Stereopairs of MNRJ 2492-V, a nearly complete right mandibular ramus with P2-M| complete: a, labial; b, occlusal; c, lingual views. Scale = 15 mm. Fig. 67. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). DGM 800-M, a righ maxillary fragment with F2-M3 complete: a, labial; b, occlusal; c, lingual views. Scale = !l 80 MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 81 lary with M2 "3 complete and alveoli of M4; DGM 917-M, a fragment of a left maxillary with roots of P1'2 and alveoli of P3-M2; DGM 918-M, a fragment of a left maxillary with P2 and M1"2 complete and P3 very broken; MNRJ 1373-V, a fragment of a right mandibular ramus with M,.3 complete and alveoli of M4; MNRJ 1374-V, a fragment of a right mandibular ramus with M,_2 complete and alveoli of M3.4; MNRJ 1376-V, a fragment of a right edentulous ramus with alveoli of M,_4; MNRJ 1378-V, a fragment of a right mandibular ramus with M,.;,; MNRJ 1381 -V, a right mandibular ramus with P2-M2 complete and roots of M:».4; MNRJ 1383-V, a fragment of a left mandibular ramus with M,.;, complete and alveoli of M4; MNRJ 1386-V, a fragment of an edentulous left mandibular ramus with alveoli of M,.4; MNRJ 1387-V, a fragment of a left mandibular ramus with roots of C-P3 and M,_2 complete; MNRJ 1388-V, a right ramus with roots and/or alveoli of C-M»; MNRJ 1390-V, a fragment of a right mandibular ramus with alveoli of 1,-C, roots of P2 and P:t complete; MNRJ 1392-V, a fragment of a left mandibular ramus with M,.2 complete and alveoli of M;,_4; MNRJ 1394-V, a fragment of a right mandibular ramus with M,.:, complete and alveoli of M4; MNRJ 1400-V, a fragment of a left mandibular ramus with P2-M2 complete; MNRJ 1401-V, a fragment of a left mandibular ramus with P2.:i complete; MNRJ 1402-V, a fragment of an edentulous left mandibular ramus with alveoli of M,.4; MNRJ 1403-V, a fragment of an edentulous left mandibular ramus with alveoli of M,.4; MNRJ 1405-V, a fragment of an edentulous left mandibular ramus with alveoli of M,.4 and roots of P3; MNRJ 1407-V, a fragment of a right maxillary with P2-M2 complete; MNRJ 1409-V, a left maxillary fragment with P2 and M1'2 com- plete, and roots of P3; MNRJ 1410-V, a fragment of a right maxillary with P3-M' Fig. 68. Epidolops ameghinoi Paula Couto, 1952, p. 7 (Riochican). Stereopairs of DGM 800-M, a right maxillary fragment with F-M3 complete: a, labial; b, occlusal; c, lingual views. Scale = 10 mm. 82 FIELDIANA: GEOLOGY complete; MNRJ 1411-V, a fragment of a right maxillary with P! complete and alveoli of M1"2; MNRJ 1415-V, a fragment of a left maxillary with M2; MNRJ 1416-V, a fragment of a right maxillary with M1"2; MNRJ 1417-V, a fragment of a left maxillary with P3 and M2 (very broken); MNRJ 1418-V, an isolated P3; MNRJ 1419 -V, an isolated P3; MNRJ 1420 -V, an isolated P3; MNRJ 1421-V, an isolated upper M' or M2; MNRJ 1422-V, an isolated broken upper molar; MNRJ 1423-V, two isolated molars; MNRJ 2492-V, a nearly complete right mandibular ramus with P2-M4 complete (figured by Paula Couto, 1961, fig. 6); MNRJ 2879-V, a fragment of a right maxillary with alveoli of C-P2 and M1"2, and with P3 com- plete; MNRJ 2880-V, a right mandibular ramus with alveoli of Ii-Pi, roots of P2/ P3-M3 complete and alveoli of M^; MNRJ 2881-V, a left mandibular ramus with alveoli of C-P], roots of P2, Pa-Ma complete, and alveoli of M,. Horizon and Locality. — All specimens are from fissure deposits in the Itaborai Formation, in the limestone quarry at Sao Jose de Itaborai, northwest of Niteroi, the capital of the state of Rio de Janeiro, and on approximately the same latitude as the city of Rio de Janeiro, Brazil (Rodrigues Francisco & de Souza Cunha, 1978). Age. — Riochican. Revised Diagnosis. — As for the subfamily Epidolopinae. Known Range. — Riochican of Brazil. Comments. — The larger specimens assigned by Paula Couto to £. ameghinoi generally have the P, present and the P2 double rooted. In contrast, the smaller forms which he assigned to E. gracilis generally have the P, absent and the P> single rooted. The largest variation seen in the sample from Itaborai occurs in length and width of P3 (Appendix 12), with CVs ranging between 10 and 13. In linear measurements of the molars, CVs range between 5 and 7, and in some cases the sample sizes are as high as 46. The characters used by Paula Couto (1952, p. 11) to distinguish E. gracilis from E. ameghinoi lose their uniqueness upon study of the large sample of Epidolops. Intergrades exist for all the "di- agnostic" features used to separate these species. All evidence thus suggests that the sample of Epidolops from Itaborai is referrable to one rather homoge- neous species, E. ameghinoi. EVOLUTION OF POLYDOLOPIDAE Members of the extinct South American marsupial family Polydolopidae are known in South America from beds of Riochican (middle to late Paleocene), Casamayoran (early Eocene), and Mustersan (middle Eocene) age in Patagonia, southern Argentina; in beds of Riochican age in Brazil; and apparently in beds of Deseadan (early Oligocene) age in Bolivia. Five genera and 13 species, placed in two subfamilies, are recognized: Epidolops ameghinoi in the Epidolopinae; and Polydolops clavulus, P. kamektsen, P. rothi, P. winecage, P. serra, P. thomasi, P. mayoi, Pseudolops princeps, Amphidolops yapa, A. serrula, Eudolops tetragonus, and E. hernandezi in the Polydolopinae. A specimen from beds of Deseadan age in Bolivia is tentatively referred to Polydolops? sp. indet. and specimens from beds of Eocene age on Seymour Island, Antarctica, are tentatively referred to Polydolops (Simon, 1982). Lower dentitions of the various species of Polydolopidae showing relative size and proportions of teeth are compared in Figure 69, and upper dentitions are compared in Figure 70. The size distribution of polydolopid species as MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE A B C P3 M, M2 M3 83 Fig. 69. Comparison of lower dentitions of various species of Polydolopidae showing relative size and proportions of teeth. All illustrations are drawn to same scale. A, labial; B, occlusal; C, lingual views. 1, Polydohps mayoi; 2, Eudolops tetragonus; 3, Polydolops thomasi; 4, Amphidolops serrula; 5, Polydolops winecage; 6 and 7, Polydolops serra; 8, Polydolops rothi; 9, Polydolops kamektsen; 10 and 11, Polydolops clavulus; and 12, Epidolops indicated by the relationship of length versus width of P:t is compared in Figure 71, length of P3 versus length of M, in Figure 72, and length versus width of M2 in Figure 73. These plots demonstrate that absolute and/or relative size is sufficient to differentiate known species, especially those of a given age from the same geographic area. A summary of distribution of some character states of better-known poly- dolopid genera is presented in Appendix 13; based on these characters, a cladogram showing the probable phylogenetic relationships of Epidolops, Eudolops, Amphidolops, and Polydolops is presented in Figure 74. roS.2 3 in Oi S > .2 .£ "- "^ w v. 00 rH • -3 c « O o <« 3 « O UJ V- o y « 3 Oh O o « u > — < J2j £ ID.JjTj t/1 -3 5 S 35 -« o | re 'S. w p = :, s- a v«n o T3 2§ « o > ^ g-n "5 £ £ ■a a« •^ Hi N ^3 s = ft. 35 "S o o o — — * r- O O -~ TJ ts -3 O -s: ** £ &* Q. a r w n ^ Ifl <0 84 MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 85 The criteria employed in selection of the morphocline series for the characters used in constructing the cladogram in Figure 74 are as follows (see Appendix 13 and summary of character state distribution in caption to Figure 74); in all cases, the ancestral states most closely approximate those inferred to have existed in ancestral didelphoids as summarized by Reig, Kirsch & Marshall (in prep): Character 1 — The general trend in polydolopid evolution has been from very small to very large size. Characters 2 (4, 5, 6, 7, 8, 15, 17, 24) — The highest dental formula known in any polydolopid is I§, C\, F^, M$, and deviation from this formula has involved loss or reduction in the number of teeth, a condition regarded as derived. Character 3 — In the ancestral polydolopid the sectorials were P'/P.,, and variations from this condition to include V1 and/or the trigonid of M, are regarded as derived. Character 9 — P2 in ancestral polydolopids was well developed, two rooted, bladelike, cutting edge not serrated. Derived conditions include development of ser- rated edges (Polydolops) or loss (Amphidolops) or reduction (Epidolops) of this tooth. Character 10 — P2 in ancestral polydolopid is a well-developed, two-rooted blade, cutting edge not serrated. Variations from this condition are regarded as derived. Character 11 — P:1 in ancestral polydolopid is aligned in same anteroposterior axis as molar series. Characters 12, 13 — M1 in ancestral polydolopid is quadritubercular with two pri- mary labial and two primary lingual cusps and two smaller cuspules between them, which are set closer to primary labial cusps; weak loph connects primary labial and primary lingual cusps; distinct but narrow stylar shelf. Proliferation of cusps in some taxa, especially along stylar shelf, is regarded as derived. 6H mm 5- PoW 4 - 3- Polydolops mayoi Polydolops thomasi (•) Polydolops serra Polydolops rothi Polydolops winecage Epidolops ameghinoi (♦) Eudolops tetragonus & Polydolops ~ clavulus "" 1 ■ 1 6mm 7 P^L Fig. 71. Size distribution of various species of Polydolopidae as indicated by re- lationship of length vs. width of P3. 86 FIELDIANA: GEOLOGY mm 6 - 5 - 4 - M.,L 3 - 2 - & Eudolops tetragonus 0 Polydolops mayoi Polydolops thomasi Polydolops serra Polydolops rothi Polydolops clavulus Epidolops ameghinoi P3L Fig. 72. Size distribution of various species of Polydolopidae as indicated by re- lationship of length of P3 vs. length of M, . Character 14 — Ancestral condition of M3 is regarded as best approximated by that seen in Epidolops. Character 18 — P2 in ancestral polydolopid very small, double rooted, crown without cusps, flattened, irregularly oval in outline. Character 19 — P3 primitively well developed, bladelike, cutting edge not serrated. Character 20 — P, primitively aligned in same anteroposterior axis as molar series. Character 21 — M, primitively anterior edge of trigonid elevated above talonid but not modified into blade; labial and lingual edges of talonid without cusps. Character 22 — M2 primitively quadritubercular with weak lophs connecting primary labial and lingual cusps; trigonid somewhat elevated above talonid. Proliferation of cusps along labial and lingual surfaces of tooth are regarded as derived conditions. Character 23 — M:! primitively quadritubercular with weak lophs connecting primary labial and lingual cusps. Proliferation of cusps along labial and lingual surfaces of tooth is regarded as derived. Character 26 — Molar basins primitively shallow with lightly wrinkled enamel. In- crease in depth of basins and of enamel wrinkling are regarded as derived conditions. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 87 Utilization of these morphocline series are complementary, one with the other, and the final result as shown in the cladogram in Figure 74 is judged to be a reasonable approximation of the phylogenetic relationships of these polydolopid genera. Polydolopids are represented by five species in beds of Riochican age, seven in beds of Casamayoran age, one in beds of Mustersan age, and apparently one in beds of Deseadan age. Factors influencing the decline in diversity to extinc- tion of polydolopids are complex, but this process is correlated with the appear- ance and/or diversification of other mammal groups which had similar dental specializations and presumably ecologies as well. The appearance of one group coincides with the disappearance of another, and by such process the fossil record documents the successive replacement through time of (and by) different groups of animals stemming from different ancestors but occupying the same adaptive zone. These "ecological replacements" or "evolutionary relays" may have resulted from active competition between the successive groups filling these roles, or from passive replacement resulting from the disappearance of one group due to chance processes or as a result of concurrent environmental change. Alternatively, such faunal changes were the result of a combination of these or other possibilities (Marshall, 1980a, p. 132). mm 5 - M2W 4 - 3 - 2 - Polydolops mayoi Eudolops tetragonus Epidolops ameghinoi Polydolops kamektsen I X Amphidolops • /i^ \ serrula • • • Polydolops thomasi Polydolops rothi Polydolops clavulus M2L Fig. 73. Size distribution of various species of Polydolopidae as indicated by re- lationship of length vs. width of M2. 88 FIELDIANA: GEOLOGY Epidolops Eudolops Amphidolops Polydolops Fig. 74. Cladogram showing probable phylogenetic relationships of better-known gener; of Polydolopidae. Diagram indicates only relative position of common ancestor. Number; refer to distribution and relative times of acquisition of derived character states. 1, Medium size: dental formula I5, C\, P3, Mf; lower canine well developed, singli rooted, procumbent; sectorial blades P2 and P3/P3; P{ either vestigial and single rooted or absent; P2 well-developed, two-rooted blade; P3 well-developed, two-rooted blade P3? aligned in same anteroposterior axis as molar series; molars quadritubercular; mola basins shallow with very lightly wrinkled enamel. 2, P2 very small, single or double rooted, crown without cusps, flattened, irregularl; oval in outline, loss of sectorial function; axis of P3 set at about 30° relative to axis o molar series. 3, Complete loss of Py, loss of Mf; trigonid of Mj elevated above rest of molar serie and modified into shear blade; molars multituberculate; supernumary cusps develo] along labial and lingual sides of M[ talonid basin, on M2.3, and on labial sides o M1"2. 4, Marked increase in size: loss of P2; secondary size reduction of P3 relative to molars molar basins deep with moderately wrinkled enamel. 5, Development of prominent row of stylar cusps. 6, Apparent loss of P2; molar basins with strongly wrinkled enamel; some size increase 7, Loss of last two lower incisors; size increase of lower canine; proliferation of labia cusps on upper molars (especially on M1"2); secondary size decrease in some species marked size increase in others. It is generally agreed that polydolopids were somewhat rodent- ani primatelike in structure and ecology. This conclusion is based on comparison c the dental specializations of polydolopids with rodent, primate, and livin; (phalangeroid and caenolestoid) marsupial groups (e.g., Simpson, 1933; Rose 1975; Kay & Hylander, 1978; Marshall, 1980a) and the demonstrable similarit in structure and presumably function of the dental apparatus of these groups In view of this, it is of interest that caviomorph rodents and platyrrhin primates first appear in the South American fossil record in beds of Deseadai (early Oligocene) age and are dominant elements in subsequent land mammc faunas on that continent. These groups are believed to have arrived in Soutl America as waif immigrants, either from Africa or North (or Central) America during or just before Deseadan time (see Marshall, Hoffstetter & Pascual [ii press] for a review of pertinent literature). In addition, members of the marsu pial family Caenolestidae having polydolopidlike dental specializations appea in the Deseadan (i.e., Palaeothentinae) and Colhuehuapian (late Oligocene Abderitinae) (Marshall, 1976, 1980a). It thus appears permissible to specular that the ecological roles filled by Polydolopidae in Riochican and Casamayorai faunas came to be filled, at least in part, in the Oligocene and later faunas b; members of the marsupial family Caenolestidae (Abderitinae, Palaeothentinae and by immigrant caviomorph rodents or platyrrhine primates. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 89 LITERATURE CITED Abbie, A. A. 1937. Some observations on the major subdivisions of the Marsupialia with special reference to the position of the Peramelidae and Caenolestidae. Jour. Anat., 71: 429-36. Ameghino, F. 1897. Mammiferes cretaces de l'Argentine. (Deuxieme contribution a la connaissance de la faune mammalogique des couches a Pyrotherium) . Bol. Inst. Geog. Argentine 18: 406-521. . 1898. Sinopsis geologico-paleontologica. Segundo censo de la Repiiblica Argen- tina. Fol., Buenos Aires, 1: 112-255. . 1902a. L'age des formations sedimentaires de Patagonie. An. Soc. Cien. Argen- tina, 50: 109-30, 145-65, 209-29 (1900); 51: 20-39, 65-91 (1900); 52: 189-97, 244-50 (1901); 54: 161^80, 220-49, 283-342 (1902). 1902b. Notice preliminaries sur les mammiferes nouveaus des terrains cretaces de Patagonie. Bol. Acad. Cienc. Cordoba, 17: 5-70 (1902-04). -. 1903. Los diprotodontes del orden de los plagiaulacoideos y el origen de los roedores y de los polimastodontes. An. Mus. Nac. Buenos Aires, (3)1: 81-192. 1904a. Paleontologia argentina. Publicaciones de la Universidad de La Plata. Faculdad de ciencias fisico-matematicas, no. 2: 1-79. -. 1904b. Nuevas especies de mamiferos cretaceos y terciarios de la Repiiblica Argentina. An. Soc. Cien. Argentina, 56: 193-208 (1903); 57: 162-75, 327-41 (1904); 57: 35-71, 182-92, 225-91 (1904). [Also issued as a separate, paged consecutively 1-142, Buenos Aires, 1904.] -. 1906. Les formations sedimentaires du cretace superieur et du tertiare de Patagonie avec un parallele entre leurs faunes mammalogiques et celles de l'ancien continent. An. Mus. Nac. Buenos Aires, (3)8: 1-568. Archer, M. 1978. The nature of the molar-premolar boundary in marsupials and a re- interpretation of the homology of marsupial cheekteeth. Mem. Qd Mus., 18: 157-64. Biggers, J. D. and E. D. DeLamater. 1965. Marsupial spermatozoa pairing in the epididymis of American forms. Nature, 208: 401-14. Clemens, W. A. and L. G. Marshall. 1976. Fossilium catalogus: American and European Marsupialia. W. Junk, The Hague, Paris 123: 1-114. Gregory, W. K. 1910. The orders of mammals. Bull. Amer. Mus. Nat. Hist., 27: 1-524. Hayman, D. L., J. A. W. Kirsch, P. G. Martin and P. F. Waller. 1971. Chromosomal and serological studies of the Caenolestidae and their implications for marsupial evolu- tion. Nature, 231: 194-95. Kay, R. F. and W. L. Hylander. 1978. The dental structure of mammalian folivores with special reference to primates and Phalangeroidea (Marsupialia). hi The ecology of ar- boreal folivores, Montgomery, G. G., ed., pp. 173-91. Smithsonian Inst. Press, Wash- ington, D.C., 574 pp. Kirsch, J. A. W. 1977. The comparative serology of Marsupialia, and a classification of marsupials. Austral. Zool., Suppl. ser. 52: 1-152. Marshall, L. G. 1976. Revision of the South American Fossil Marsupial subfamily Ab- deritinae. Mar del Plata Mus. Munic. Cienc. Natur., Publ. 2(3): 57-90. . 1979. Evolution of metatherian and eutherian (mammalian) characters: a review based on cladistic methodology. Zool. Linn. Soc. London, 66: 369-410. . 1980a. Systematics of the South American marsupial family Caenolestidae. Field- iana: Geology, n.s. no. 5: 1-145. . 1980b. Marsupial Paleobiogeography. In Aspects of Vertebrate History, Jacobs, L. L., ed., pp. 345-86. Museum of Northern Arizona Press, Flagstaff, 407 pp. Marshall, L. G., R. F. Butler, R. D. Drake and G. H. Curtis. 1981. Calibration of the beginning of the Age of Mammals in Patagonia. Science, 212: 43-45. Marshall, L. G., R. Hoffstetter and R. Pascual. In press. Geochronology of the con- tinental mammal-bearing Tertiary of South America. Fieldiana: Geology. Obenchain, J. B. 1925. The brains of the South American marsupials Caenolestes and Orolestes. Field Mus. Nat. Hist., Zool. ser., 14(3): 175-232. 90 FIELDIANA: GEOLOGY Odreman Rivas, O. E. 1978. Sobre la presencia de un Polydolopidae (Mammalia, Mar- supialia) en capas de edad Mustersense (Eoceno medio) de Patagonia. Obra del Cen- tenario del Museo de La Plata, 5: 29-38. Pascual, R. 1965. Un nuevo Condylarthra (Mammalia) de edad Casamayorense de Paso de Los Indios (Chubut, Argentina). Breves consideraciones sobre la edad Casa- mayorense. Ameghiniana, 4(2): 57-65. . 1980a. Nuevos y singulares tipos ecologicos de marsupiales extinguidos de America del Sur (Paleoceno tardio y Eoceno temprano) del Noroeste Argentino: Act. II Congr. Argent. Paleont. y Bioestr., y I Congr. Latinoam. Paleont. Buenos Aires, 1978, 2: 151-73. -. 1980b. Prepidolopidae, nueva familia de Marsupialia Didelphoidea del Eoceno sudamericano. Ameghiniana, 17(3): 216^12. 1981. Adiciones al conocimiento de Bonapartherium hinakusijum (Marsupialia, Bonapartheriidae) del Eoceno temprano del Noroeste Argentino. Anais II Congresso Latino-Americano Paleontologia, Porto Alegre, Abril 1981, 2: 507-20. Pascual, R. and M. Bond. 1981. Epidolopinae subfam. nov. de los Polydolopidae (Mar- supialia, Polydolopoidea). Anais II Congresso Latino-Americano Paleontologia, Porto Alegre, Abril 1981, 2: 479-88. Pascual, R. and H. E. Herrera. 1973. Adiciones al conocimiento de Pliolestes tripotamicus Reig, 1955 (Mammalia, Marsupialia, Caenolestidae) del Plioceno superior de la Argentina. Ameghiniana, 10(1): 36-50. . 1975. Stilotherium Ameghino, 1887, el mas primitivo Caenolestidae conocido. Consideraciones sobre la transicion Didelphidae-Caenolestinae (Marsupialia). Act. 1st Cong. Argent. Paleont. y Bioestr., Tucuman, 1974, 2: 417-30. Pascual, R. and O. E. Odreman Rivas. 1971. Evolucion de las comunidades de los vertebrados del Terciario Argentino. Los aspectos paleozoogeograficos y paleo- climaticos relacionados. Ameghiniana, 8: 372-412. Patterson, B. and L. G. Marshall. 1978. The Deseadan, Early Oligocene, Marsupialia of South America. Fieldiana: Geology, 41: 37-100. Paula Couto, C. de. 1952. Fossil mammals from the beginning of the Cenozoic in Brazil. Marsupialia: Polydolopidae and Borhyaenidae. Amer. Mus. Novit., 1559: 1-27. . 1961. Marsupiais fosseis do Paleoceno do Brazil. An. Acad. Brasil. Cienc, 33: 321-33. — . 1970. News on the fossil marsupials from the Riochican of Brazil. An. Acad. Brasil. Cienc, 42: 19-34. Reig, O. A., John A. W. Kirsch and L. G. Marshall. In preparation. Systematic re- lationships of extant and Neocenozoic opossumlike marsupials, with comments on the classification of these and other Cretaceous and Paleogene New World and European Metatherians. Ride, W. D. L. 1962. On the evolution of Australian marsupials. In The Evolution of Living Organisms, Leeper, G. W., ed., pp. 281-306. Melbourne Univ. Press, 459 pp. Rodrigues Francisco, B. H. and F. L. de Souza Cunha. 1978. Geologia e Estratigrafia da Bacia de Sao Jose, Municipio de Itaborai, R. J. An. Acad. Brasil. Cienc, 50: 381-416. Rose, K. D. 1975. The Carpolestidae-Early Tertiary primates from North America. Bull. Mus. Comp. Zool., 147: 1-74. Schlosser, M. 1923. Grundzuge der Palaontologie (Palaeozoologie) von Karl A. von Zit- tel, II Abteilung-Vertebrata. Neuarbeitet von F. Broili und M. Schlosser. Munich and Berlin, R. Oldenbourg, 706 pp. (Marsupials, pp. 433-42). . 1925. Class V. Mammalia, Vol. III. Textbook of Paleontology, K. A. von Zittel, revised by Max Schlosser. MacMillan & Co., London, 316 pp. (Marsupials, pp. 21-39). Simon, C. 1982. Fossilized jaw first sign of land mammals in Antarctica. Sci. News, vol. 121(13): 213. Simpson, G. G. 1928. Affinities of the Polydolopidae. Amer. Mus. Novit., no. 323: 1-13. . 1930. Post-Mesozoic Marsupialia. In Fossilium Catalogus. 1: Animalia. W. Junk, Berlin, Pt. 47: 1-87. MARSHALL: SYSTEMATICS OF POLYDOLOPIDAE 91 . 1933. The "plagiaulacoid" type of mammalian dentition. A study of convergence. Jour. Mammal., 14: 97-107. -. 1935a. Descriptions of the oldest known South American mammals, from the Rio Chico Formation. Amer. Mus. Novit., no. 793: 1-25. 1935b. Occurrence and relationships of the Rio Chico fauna of Patagonia. Amer. Mus. Novit., no. 818: 1-21. -. 1936. Notas sobre los mamiferos mas antiguos de la coleccion Roth. Inst. Mus. Univ. Nac. La Plata. Obra del Cincuentenario, vol. 2: 63-94. 1948. The beginning of the Age of Mammals in South America. Part I. Bull. Amer. Mus. Nat. Hist., 91(1): 1-232. -. 1964. Los mamiferos casamayorenses de la Coleccion Toumouer. Rev. Mus. Ar- gent. Cien. Nat. Paleont., 1(1): 1-21. 1967a. The Ameghinos' localities for early Cenozoic mammals in Patagonia. Bull. Mus. Comp. Zool., 136(4): 63-76. 1967b. The Beginning of the Age of Mammals in South America. Part II. Bull. Amer. Mus. Nat. Hist., 137: 1-259. -. 1971. The evolution of marsupials in South America. An. Acad, brasil. Cienc, 43 (supp.): 103-18. Tournouer, M. A. 1903. Note sur la geologie et la paleontologie de la Patagonie. Bull. Soc. Geol. France, ser. 4, 3: 463-73. Appendix 1. Measurements of lower cheek teeth of some species of Polydolops. P2 P3 Ml M2 Specimen L W L W L W L W clavulus MACN 10356 1.2 1.0 2.7 2.3 1.8 1.8 MACN 10360 2.8 2.4 1.8 1.7 1.8 1.7 kamektsen AMNH 28525 2.3 2.0 rothi MLP 11-122 4.0 2.9 3.0 2.6 2.8 2.5 winecage AMNH 27893 4.3 3.9 3.1 2.6 Appendix 2. Measurements of cheek teeth of Polydolops serra. Specimen Upper Dentition MACN 10353 AMNH 28409 AMNH 28412 AMNH 28425 AMNH 28426 AMNH 28427 AMNH 28885 Lower Dentition AMNH 28408 AMNH 28429 MACN 10341 MACN 10359 MACN 10361 MACN 10363 P2 L W P3 L W 1.9 1.5 4.0 3.4 Ml L W 3.3 3.4 3.5 3.4 3.6 3.5 3.6 3.6 3.4 3.6 3.7 3.7 M2 L W 2.9 ca 3.1 2.5 4 2.6 . ca2.6 2 2.8 3.0 3.2 2.5 2.7 M3 L W 2.0 92 & -1 -J nO co co • ** ' co co co ' on . co ! "*! ON in \ co ! . oo ! co 5 t 1 to M rn oo in oo co co co " 00 O a . -I ^ I •" - no --1 I , ■ ^ LO CO CO CO CO CO CO CO CO CO O CO CO CO Tt* t}< O NO NO IO CO CO CO CO ■>* co ■* co in ON tN CNj rf CO Tf ^ Tjl ID NO O NO ON a, : :«> : £ . . "* . "-1 ON d , cn 00 _ ; co ; o co ^t1 " ' (N ^ •* • ' tj< ' 10-^ \ ; uS \ in in in ! | ! ! ! | 1 ] in in in \ '. '. 10 i^S ON 1— I c o 2 :-::::: :o :::::::::::::::: I CO j :° ::::: :^ :::::::::::::::: § .CM ri ,£ c 0> 4P NO t^ CN CO On O •n ooooooono 00 i— 1 cm co t*< in in 3333333^^ > > > I I I I I I I On On On On On On On Z*-J >—? »-y .-- .-, >__ .-_ I—, I—, ij> W^ W^ VJ> V-»^ W^ VJ> V*-' N^ ^^ ZZZZZZZZZininmtninininNONONO cm on r-j. op 00 co "41 ■4| --, CM 1— 1 i— 1 1— 1 v ■" B > > > z ON t"^ Cv t-N ^7 NO [^ f^ 1^ <^ CU 0_ C2-, Pu £ J J J JZ 94 Appendix 5. Statistics for some cheek tooth dimensions of Polydolops thomasi. Dimension N OR x s CV 3 3.0-3.4 3.20 3 3.0-3.8 3.30 8 4.3-4.9 4.50 8 3.9-4.7 4.36 6 3.4-3.5 3.43 6 3.6-3.9 3.77 15 4.8-5.4 5.22 14 3.8-4.9 4.41 19 3.9-5.6 4.38 21 2.9-4.5 3.41 24 3.^4.0 3.73 24 2.5-3.3 3.04 8 3.4-3.8 3.69 7 2.5-3.1 2.83 4 11.6-12.3 12.03 2 16.2-17.0 16.61 Appendix 6. Measurements of lower cheek teeth of Polydolops mayoi. P3 Ml M2 Upper Cheek Teeth P3 L W M1 L W M2 L W Lower Cheek Teeth P3 L W M, L W M2 L W M, L W M,_, L P^M3 L .20 6.25 .44 13.33 .21 4.67 .23 5.28 .05 1.46 .12 3.18 .19 3.64 .34 7.71 .34 7.76 .37 10.85 .16 4.29 .20 6.58 .16 4.34 .20 7.07 .31 2.58 .55 3.31 Specimen L W L W L W MLP 52-XI-4-176 5.2 5.0 5.2 3.6 4.0 3.8 MLP 59-11-28-95 5.5 3.9 4.1 3.8 MLP 69-III-24-1 5.3 5.9 5.0 4.1 4.7 4.1 MLP 69-111-24-12 5.1 4.8 95 -2 w c •a , . (Q ra TJ X TO ^_, X 9 O X CD x >- 0) ( Ol C £ X >-* Ol a. 3 6 o 0* o E M c •c 73 -x> a%e re u £ ? n aCT3^ SS &, « m P •SC3C3? 3 «a 5 c •3 >> 6 c c a 3 nju(UXi«i4!!*iCi.4:0(5(X0M) VO bo 55 01 "3 ._ o ^ 03 M>~ ^ C i ^ tj t -s - C m c 03 x .2 v, F P S - r- OJ fcc _ O S F •*> — m ^ ,Z ^ = 8 'x1 >h oi SI P o.. n . (0 C, n -< S a, c 3 <£ o -v x> aJ-c bo c •4= — x 03 3 2 £> J-1 . C C^h-mS-3c 91 V i (0 O 9 ■> 2& «° 41 0) C ^ X _ O O ~ ^ C Oi 03 H ^= X 2 bC ■8 5 ra -a B a <» * 3 TO 5 13 XT 01 o u oi x 03 3 C S 8 2-2 "P o S ^2 * c ** X 03 O O) H C T3 « g a; a « "> X 03 jH *i Cl,*i O 03 ^ -• '2 C I; m in O 3 _ . Dm U 2 -3 03 X? 6 3 w to c <* Xi C -3 „ i_ _ an c ,^ 0 * •-3 3 8 E X 3 M TJ X 11 xi E -c c to § b 42 ra •S . E 5 5 (0 £ E 3 j* ao lo S3 ^^ < "6 ft 03 bO « g > v3 X 03 01 I-H > T— 1 i = SP x *; PH (l<° 3 « ,. * a,73 5 ^ w o C3 J5 oi o *-. ai 23x g x c 3 c^a 03 O 01 T3 >-. 3 X *! » ?, C 3 r "ra E JS "P o x S oi « u 00 5 C i! C 03 S 03 X •43 x a, c -c c 3 _ > _. X ° 2 c " M ij 03 !2 C -g 03 5 .&-S 3 rx .a oo i to o ... o Si tSu*; c*. o 03 O Q 01 o o o^ g-oo° ^2.5 O X X 3 bO P P o X 01 3 ^2 03 , x fc P fj oi oi 3 o 1/5 c c i ^1-2 I c o .H ° >M 41 > "3 u ^5 03 S > 03 Dtf c X U IH o > oi ^ a "43 03 Ml'P 3 03 " — ' 03 33 u tC J 03 ^ U^ D? f^Ti 11* c ■3 o 3 u 0) 96 *-f § E J^ S.-S >* J2, C c o> « 5 ~K .5 ai « C- C o -^ . E >^ jU 0) rH >- v- E-£ c Oi w— ; « o " E <2/C .2 S di js — 60 "3 •- 2 « E ~~ ^1- 3 J2 v- in g 3 o o c -c c c * * o o c c M -* c c 3 3 5^-^ jL o ~ C •6 2 ■ Ed ai 'C a 01 •*-* o> c in l« in >> . o . »e tu *; 3 — in 60—. jjT -c e-^^E is g-^fc >*m(n oi ** 5 IS 60 g. S 1£ g.js > 2°>vc 2 >vc — o c o c 5 01 0< O c c o c «m .2 '1 c ^h 3 — i— i C T^ S S o S?-c E B X, E t3 2 CL._q -t-> w *t: u 3 E C y E c Cl- in 3 E tj c u = <-» 3 S3 3 a C 3 -3 u 3 ■a H u 0 c 2 — C <" i« c <" in ^ ^> 3S^ "2 c 0 E c 3 E c 3 c ^ 3 s -3 „ »-3-2 ,-a 60&" S^ jsS o^ 3 97 &T3 > TS V O =3 o *- T5 -,*■ M> C ft i * S 6 * w 73 o Oh C 5 3> o -a g •3 ai ex a> rQ 5 ° .H o> « M £ o 01 0> -Q c *C "J o g <° • . O -R H pfi U (0 3 O JC o (X o Oh '3 i5 0 , 1 J ■3 -a *3 oi" T3 01 "ffi > "3 0) * 2 "V o 3 -i o k.b0 jd 2 « S^ | m ^, (4 e w S « S S 2 VC 01 nj Cu C C 3 5 ft >H oj C T3 • c *• <0 en • — I o> •- .2 "3«# .£> '35 (0 J u 5fl 01 01 3 (- ro ;Bu I g3.fi ^i O u 3 98 J J J <: ~i J j S \»' S •: ••• % OOhN " in in 1s . fn ^* ^ . m m s n ts s £ -J ' v© in in t-^ ^ tn J JO - ~ 3 — - O-S > X frl 3 | »■* ._ rr* H "T"1 w So 3 Ji-SS-o 32S15 3 3-8 99 rH J H00ON00 ■* co ** co co vn nJ 00 vO 00 \0 K iri iri in in iri h " " O\ <<<<<<< oo oo ! 2 2 °^ • • io^osoN ! o> [oooo ! 2 2 uuO 8 8° uuuu uu 33 * :::::":::::::::::::::::::::: - ; : ; ; ;s ;;;;;;;;;;;;;;;;;;;;; : ^, o Tf o »-i ■ «-■ ; ; i-< ; i-< ; o ; o ; ; o ; ; co cn i-< h o> oq ; o r\i c\i 1-5 i-5 ' oi i-j ; o ' o; ; oq ] ; o j ! ''i °. C vO 'tvt^ONOO " IO ^ O N ' N MT If) N >C '101010 On CJ\ rj [(Ntsi [rsifNrsioi ] oi csi r\i csi ] «n cn cn oi 't-^t^oooo "looiovoioso 'tot^vo ' oo vo i i oi ooom»no> o o oo oq o oq oooqi-; , *"• *-| oi co ro oi c\i co . oq to in in in ! ■»* io 'in ! \ <& ! ** j^TjiiniiS rlooJ-^r^^dlfi^fA4iAvA^o0^^^^ll5l§8 (NfOfO0000X^(^<^^^CT\Ov^ ^-* ^-* ^-" ^-* ^-i «-l ^-* ^-t •— • ^-« *— * *-« ^-H ^-* ^-t __ *—m t— — .— » •— « t-^ f^ t—m —^ *—m >— » »-» *—m UUOUUUUUOUUUUygZ2Z2ZZ2ZZZZ2Z 103 Appendix 12. Statistics for some cheek tooth dimensions of Epidolops ameghinoi. Dimension N OR X s PPEI i Cheek Teeth p3 L 22 4.1-5.8 4.97 0.55 W 21 3.4-5.3 4.40 0.57 M1 L 24 2.1-2.6 2.32 0.17 W 24 3.0-3.9 3.39 0.22 M2 L 20 1.9-2.4 2.13 0.13 W 20 2.8-3.5 3.16 0.17 M3 L 7 1.4-1.6 1.53 0.10 W 7 2.1-2.6 2.37 0.17 3WE r Cheek Teeth P3 L 35 4.5-6.1 5.17 0.52 W 35 3.1-5.5 4.18 0.55 Mj L 44 2.7-3.3 2.93 0.15 W 43 2.3-3.0 2.61 0.16 M2 L 46 2.5-3.0 2.69 0.14 W 46 2.3-3.1 2.64 0.16 M3 L 19 1.8-2.3 2.02 0.14 W 19 2.0-2.4 2.15 0.13 M,. -4 L 28 8.6-11.0 9.37 0.62 Py- M4 L 20 12.7-17.0 14.89 1.07 104 o> ra. CO Uj 'S. o "o "o 60 5 8- o o c ^ i a E -a* 2 o "S c o 60 „ 73 O C C - -£ « C Cu 3 clcz- — 2 _ 73 73 U 1 1 8 •- •- tu a* f= -a Qu 5 a. 5 .2 o 3 o 3.5>- S 2 8» 2 5 S "° "a "° 73 •_ a. ^ -^ co •§! § „ ■£ co 60 £ 5o 'in £ fc » P 0 u E "^ c „ <" £ J2 s >- E C G t_ X> JO '"■ <£ " ai c Jj uo.s c rx'-E o £ a, « -q ST x s- o u x: £ o — X> "3 m — .£ 71 73 .» " t: «- co > ,«X» 3 I t oi cj 73 60 C C 5 c MS V) n X) 1 o> E 0 •5 a. 5 0) 60 c ^' ^ JO £il o E ffi c 3 o o •3 u T3 C * I "co E en b 9 > _,-_ Ou E ** o 73 c o 60 „ "E C ~ 5 73 O C C co .* « C O- 3 2 "0" c o 60 _ 'C c 73 O C C CO -^ m C &H 3 E -£ ¥ > — ' - _2 Z it > C4 G in (O "co 6 ^ C _ -S .j3 C ty g a, U. 73 v2 tn rM ci en J; i- oi i- O 01 c Oi rx c - M ** M 01 O 01 o.'go"gu.5o5s _j - _ .= — i ^ — ^ — in so BO 105 £ a; bO ■ •a c .5 S * 2 32 jd 3 8'c S 3 2 c bo « * c « -a n c 3^ nj re (8 _ S2^35 « ;■ x a) ■2P° x- 2 <3 r- OJ 5b c o2 b'l •si 5 C 3 55 & O T3 £_£ 2 * T3 01 C bO (13 T3 _, * — - 3 "re bo ^■^^ .S X*T3 .2 ±3 o 2 s T. P bO u C -v >^ --" .2 "o3 -q _re .O re 03 73 3 (8 CO 3 O 5 -a X re to e -c £ n a CM $ 3 * -^ _ o> » £ I J > s 1 &^ ,2 3 « 2 oi oj g T3 « 2 a "bb 3> 2 6 5b 6 JS c H +, o> 6^ G 'E u J" 15 u S3 S. I x "2 -2 a o — "^ _v .2 Cu l- * «•! s-i ° ^~ > !> H (8 S » (5 3 ^ *\ J3 H 5 - OJ to bO « 3 C CO C- ' ' ' I fi-E £ Si s •M •> 55 ra S OJ ^ «J -I A ■* ££ §>§ a, to .3 Xr >" ns cj Dh-3 3 X=3 O C bO - « « C O .82 £ *> S T3 « h -* JS * .S | 2s-3 y^ e -3 3 q> c 3 oj (is y ^ o ° oj > cin I-, m u 3 3 oj u o bO o 15 J2 >^ ro >-< IQ — ' O ft 3 > Ui .» i- ■a i2 3 j= c g be « O 2 r "3 TS •*-» (1J ^- 3 c -m o C l_ OJ 2 0/ -= £ oi JX Cl.* "3 fi- ts in T3 3 3 o> bO u .Sco tS a 2xs 6C OJ C 3 "3 "5 c IH C 2 en O 5, u o o c 2 ° 3 h h " O >^ C oj C 5 ts a £ g c u 3 ^ 2 * u b in 01 01 c « * u Ih £ ro (0 0 ta E C M -a ra OJ U5 bO nj ■'-' tn vo »h •C I •- 82 .3 O O 3 5 j- o ?- ^ jC ^ oj _ ±! oi a •£ -e S 'i > ^ re > -a (8 3 p03^ B 3 u o 01 j- Vh .2 -C rti 3 —i to re !X 3 en £>3 ^3 v£) ro l"bb§ S| 2 § bo.£ Si 2 i C S x ^Oo" o « re m j2 > OJ £X^5 "9 s? ;r "3 ~r- -rj bo tj JJ Sts S2 2 •S-Sa'gf — ' -z oi oi 5 Oi X> « j2 Xi &. ^ £ ^ .2 « 2 3 .?2 x O o 106 ™ £e * <0 m _ -C E 2 « « £ O1 "S r> •a-s <« £ 3 JS 3 — •- >. C 3 C C— ' Q. 73 3 3 a § 6 a 3 or ^ 1 1 s s *§■§.!.§ 2 w b n * o 5 3 O £ 3 C C m AS o 0 3. a. JS _o i^"o5 in > •3 -« S ^3 * (0 v — a. « «5 JS o .22 * in O 8.8 a-£ P 3 ai- JS — 2- r3 .2 * a* E £ S £ «tj s y ° f^ 01 « vim 3. 00 C O E o> & T3 5 * in 0J 36 _ o 5-2 £ 3 ■* _2 X> C tfl 3 •^g 3 o J* C «» yS o 2 60 Q- C u C'C re T3 60 C T3 fB 01 Si c «3 re o re in re nj 3 O 3 x 60 Cu 60 w S o o> a-'G-a f c 2 ■£ « .o >- -C 73 O- -_ - / Q._CU — 5 «£ -CO*; 1/1 f-§ i- 60-" 2/1 ~ &0.3 t: com «3 0u O. 3 3 (9 ~ U ., g o _ w u J" ^2 C O M>g TO;/) >- c 3 « u o ■_ 'C Cl.^ B « Oi JS 3 u « n B !X^ 3 C tn 5 '3 — «1 9) 15 i c° T3 W J2-o ^o" -? 60 SI o a.«« og.A « u A c c N (/I 4 8 2 _2 OB 1 — 1 01 3 3 be c 01 U G ^_ p £ '-J fi c B c S o -t- (0 on 3 cr 3 "g -2 £ 3!JS^ S o "o £ ° ^ S" o JU Js « E « > m *3 > u 3 ^ 3 CT in T3 J= tJ >> 0< .ti c — in > .3 « * "S .S "g .y 2 3 1 c°3 a en — ^> ™ &"g "•§ 3 w'o^l C =2 > I" •S <; o ft. 3 J8 ** S *; O. ."3 60 £ ^ ^ S 3 T3 ^ C tiOT3 *ca T3 O « «j — > m c 01 J= ^ *- m J3 to to u J "3 01 s> «5^ E c > o m > m O. tn 60.5 E *- c o 5 3 (8 D. (A O 'C Oi o > j= j- S2 13 60&J2 O) c a. 2 DC JH 3 E 107 0 o E 01 M> Wl j &. t/i 73 c TS (0 a. <3 uT ■5b- ■J3 c in 0) 01 !/J > X M « a u •2 o .■tS !• JJ £.5 Xi ex O 3 > S §£ •73 to "«>£ 5o C C o> "S £ C o — i" >-i 0) J3 3 £ _ ° 3 c o '5b ^ to u 01 <« 60 tj 60 O "9 ^ 01 CX i-i 'Jj o ■c o II: 2 - ' x: o . C S 1 | o> (0 > _ "> 2 &S ^ >^ J2 a, *. 60 0 73 C 0> h.Stj t~ '£ oi 73 3 to 01 a x> 01 cu ^ 73 "O (0 (0 a; 13 3 £ C ^3 5 « J3 >— , X O <0 ._ CO 5 ° I ?J I i? a.T3 " 7 O "5 -w -.OH c e- so o "S 1/5 £ M-, 2 s 2 o S-6 3 .a x to ! in O Oh O o> ~ '2-° bo.2 ^ ni (0 01 01 73 > ^3t3 01 c/) *-• fli 73 o 11 - .2 c c £ ^c- ° to — * 2 oi •2 ^ « a C t/> ~ 73 °-3 2 sL c « c « •43 3 « - 1 ^ 5 2 01 , box > 73 ffl 73 0_ C to a 3 d c <-. ■2 =" 04 13 .2 3 < 5 "73 .13 « JS O 3 >- 60 Oi O v-. III a u J (0 c c £ 01 ;3 ^ TO 3 Cca o 2 oi JH 60 S 3 TJ fi O) >-i o .2 -a U >-i 01 3 "S S 5 P r •3 CX "j 60_ in « 73 o- 8 01 111 "« 2 2 c * O ji ° d7 OI ^ !^ iP O -3 73 m w CX?o I* TS bO 0 « -2 -co — Jj 73 c o 2 « o M _ "-' — ' C . 3 ^ ^ c« 5o£-^ •-1 C T" 3 01 ^ a; 73 _ 73 •35 2 -55 „ en ■2 x ^2 ^ 2 II XI 73 3 _ _■- U O 73 C 4, ■*-• C O o fN J3 O 73 ^ 2 * •J3 •— . « C C " « O o 108 ■* SS-* m 01 <2 c "T 3 5 ^ £ £ <« "a! -tr £ £ 2? (A S gpri. cusp Dnid 1 ut nc 5 •c &• a 3 I" Mu U 1* nnectin lingual th; trigi lonid b 2 n 9 u la 11 £ — 0) c 0 01 c — X> lophs co bial and worn tee above ta X, 9 3 c ."t^ V) B c E T3 0 ■O Q.'ra 0 <0 Gb s °2 * h a z k. n > 0 0 12 _ v 3 5 0 A u 3 ta z 2 0 bb c A 1 01 S 6 O to =5 C m a» x; ■a z £ £ 0 | £ (0 a. C 01 0 C 3 « p St. s E 0 a ~ c s o> X m Tj "33 01 E z S > c o> 01 TJ ■a O >. 3 _^ ■2 a 5 * >. 2 ^ bo ^ £ 0 "3 "S jC ^ c > o — . .5 p^ _0J Cj £ «5 (/> 109 Field Museum of Natural History Roosevelt Road at Lake Shore Drive Chicago, Illinois 60605-2496 Telephone: (312) 922-9410