; UNIVERSITY OF KANSAS miscellaneous TURAL HISTORY MUSEUM ^^^^Na^'sT Systematic Revision of Fossil Prairie Dogs with Descriptions of Two New Species H. Thomas Goodwin LAWRENCE 28 April 1995 The Library Museum of Comparative Zooiogf Harvard University J r a. THE UNIVERSITY OF KANSAS, NATURAL HISTORY MUSEUM PUBLICATIONS ' The University of Kansas Publications, Natural History Museum, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly published in the above series are now published as The University of Kansas Natural History Museum Occasional Papers. The University of Kansas Natural History Museum Miscellaneous Publications began with number 1 in 1946. Longer research papers are published in that series. Monographs of the Natural History Museum were initiated in 1970. Authors should contact the managing editor regarding style and submission procedures before manuscript submission. All manuscripts are subjected to critical review by intra- and extramural specialists; final acceptance is at the discretion of the Director. This publication is printed on acid-free paper. Occasional Papers and Miscellaneous Publications are typeset using Microsoft® Word and Aldus PageMaker® on a Macintosh computer. Institutional libraries interested in exchanging publications may obtain the Occasional Papers and Miscellaneous Publications by addressing the Exchange Librarian, The University of Kansas Library, Lawrence, Kansas 66045-2800, USA. Individuals may purchase separate numbers from the Office of Publications, Natural History Museum, The University of Kansas, Lawrence, Kansas 66045-2454, USA. MCZ The University OF Kansas i.iAi ^9 \^^^ Natural History Museum , ^_^ HARVARD Miscellaneous Publication No. 86 28 April 1995 Systematic Revision of Fossil Prairie Dogs with Descriptions of Two New Species t ' n H. Thomas Goodwin Department of Natural Sciences Loma Linda University Loma Linda, California 92350 USA Present Address: Department of Biology Andrews University Berrien Springs, Michigan 49104 USA Natural History Museum Dyche Hall The University of Kansas Lawrence, Kansas MISCELLANEOUS PUBLICATIONS Editor for this issue: Larry D. Martin Managing Editor. Joseph T. Collins Miscellaneous Publication No. 86 Pp. 1-38; 12 figures; 13 tables; 1 appendix Published 28 April 1995 ISBN: 0-89338-050-4 © 1995 BY THE Natural History Museum Dyche Hall The University of Kansas Lawrence, Kansas 66045-2454 USA Printed by University of Kansas Printing Services Lawrence, Kansas CONTENTS INTRODUCTION 1 ACKNOWLEDGMENTS 1 MATERIALS AND METHODS 2 Age Determination 2 Terminology and Measurement Protocol 2 Morphometric Methods 3 Phylogenetic Analysis 4 Abbreviations 5 SYSTEMATIC ACCOUNTS 5 Genus Cyuomys 5 Description 6 ICynomys vetiis Hibbard, 1942 7 Description 7 Comments 9 Cynomys hihhanU Eshelman. 1975 10 Description 10 Comments 12 Subgenus Cynomys 12 Description 12 Morphometric Relationships Among Fossil and Recent Black-tailed Prairie Dogs 14 Cynomys (ICynomys) sappacnsis new species 16 Description 17 Comments 17 Cynomys (Cynomys) spenceri new species 18 Description 18 Comments 19 Cynomys (Cynomys) ludovicianiis (Ord, 1815) 20 Description 20 Comments 20 Cynomys (Cynomys) cf. Cynomys mexicanus Merriam, 1892 22 Comments 22 Subgenus Leucocrossuromys 22 Description 23 Morphometric Relationships Among Fossil and Recent White-tailed Prairie Dogs 23 Cynomys (Leucocrossuromys) guunisoni (Baird, 1855) 26 Description 26 Comments 28 Cynomys (Leucocrossuromys) niohrarius Hay, 1921 28 Description 28 Comments 28 Cynomys niohrarius niohrarius 29 Cynomys niohrarius churcherii Burns and McGillivray. 1989 29 Comments 29 Cynomys sp 29 PHYLOGENETIC RELATIONSHIPS 30 SUMMARY 31 LITERATURE CITED 32 APPENDIX 35 INTRODUCTION Prairie dogs are large North American ground- dwelling squirrels of the genus Cynomys, closely related to ground squirrels of the subgenus Spermophilus{Spermophlli4s)(Bvy'dnt, 1945; Nadler et al., 1971). Two subgenera and five extant species currently are recognized (Pizzimenti, 1975; Hall, 1981). The subgenus Cynomys, referred to as black- tailed prairie dogs, includes C. mexicanus, today restricted to a small area in northeastern Mexico, and C. ludoviciamis, which ranges widely across the Great Plains from southern Canada to Texas. The subgenus Leucocrossuromys includes three species, collectively termed white-tailed prairie dogs. All three occur on high-elevation basins or plateaus associated with the Rocky Mountains. Cynomys gunnisoni inhabits the southern Rockies; C. parvidens. in southwestern Utah, and C. leucuriis, in the central Rockies and Wyoming Basin, form a closely related pair of allospecies (Pizzimenti, 1975). The terms "black-tailed," "black tails," "white- tailed." and "white tails" will be used frequently in this paper to refer to the respective subgenera. Cynomys is known in the fossil record from the Late Pliocene (Late Blancan) to Holocene. The first fossil species to be described was C. niohrarius (Hay, 1921). Five additional fossil species have since been named (C. vetus Hibbard, 1942; C. meadensis Hibbard, 1956; C. spispiza Green, 1960; C. hihhardi Eshelman, 1975; C. churcherii Bums and McGillivray, 1989). Fossils have also been referred to three extant species (C. ludoriclanus, C. leiicurus. C. gunnisoni). Thus, nine prairie dog species have been recognized in the fossil record. Lacking a comprehensive review, the systematics of fossil Cynomys has become confused. The primary purpose of this paper is to revise the systematics of fossil prairie dogs. In doing so, it is necessary to revise the diagnoses of the genus and two subgenera, and for these purposes I examined samples of all extant species. Extant species, how- ever, are only treated in species accounts if pre- served as fossils. I also consider the phylogenetic relationships among fossil and Recent species of Cynomys. ACKNOWLEDGMENTS This paper was originally part of a dissertation submitted to the faculty of the University of Kansas. I am especially indebted to L. D. Martin, my gradu- ate adviser, for initiating and encouraging my inter- est in fossil prairie dogs. R. M. Timm, R. W. Wilson, L. D. Martin, and two helpful anonymous reviewers carefully read and improved the manuscript. L. Dryden drafted several of the figures. Many indi- viduals and institutions helped me with access to fossil or Recent materials necessary for this study. These include (organized alphabetically by institu- tion): K. L. Rogers, Adams State College; R. H. Tedford, American Museum of Natural History; R. K. Stucky, Denver Museum of Natural History; R. J. Zakrzewski, Fort Hays State University; W. A. Akersten, J. M. Hearst, and J. D. Pinsof, Idaho Museum of Natural History; L. J. Barkley, Los Angeles County Museum of Natural History; W. W. Dalquest, Midwestern University; J. A. Bums, Pro- vincial Museum of Alberta; C. S. Churcher, Royal Ontario Museum; D. Winkler, Southern Methodist University; J. E. Martin and M. Bowman, South Dakota School of Mines; E. L. Lundelius, Jr., and M. C. Winans, Texas Memorial Museum; E. Johnson and M. W. Campbell, Texas Tech University; P. Robinson and S. M. Wallace, University of Colo- rado Museum; H. A. Semken, Jr., University of Iowa; R. M. Timm, University of Kansas Natural History Museum; G. F. Gunnell and P. D. Gingerich, University of Michigan Museum of Paleontology; M. R. Voorhies and R. G. Comer, University of Nebraska State Museum; W. O. Kupsch, University of Saskatchewan; R. Purdy and F. Grady, United States National Museum; A. H. Harris, University of Texas at El Paso; B. H. Breithauptand D. N. Walker, University of Wyoming. Financial support was pro- vided by an Honor's Fellowship from the Graduate School of the University of Kansas; by grants from the Panorama Society of the University of Kansas Natural History Museum and the University of Kansas Dissertation Research Travel fund; and by the Department of Systematics and Ecology, Uni- versity of Kansas, and the Department of Natural Sciences, Loma Linda University. Finally, I am indebted to my very patient wife, Joanie, whose support and encouragement were especially appre- ciated during the final stages of my dissertation work. UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 MATERIALS AND METHODS For this revision, I examined over 400 fossil (pre- Holocene) prairie dog specimens and Recent samples of five extant species. My study focussed on lower jaws with teeth because they are abundant and diagnostic as fossils. Relatively complete skulls, uncommon as fossils, were considered when avail- able. Isolated teeth were only utilized when other, more complete specimens were rare or absent at a locality. Post-cranial elements were not studied. Cranial and dental descriptions and comparisons of Recent prairie dogs have been provided by Hollister (1916) and Bryant (1945). Using charac- ters listed in these sources as a starting point, I made detailed comparisons among fossil and Recent fonns. Special attention was given to qualitative moipho- logical descriptions in accounts of the genus, spe- cies of uncertain subgenus, and the subgenera. Dif- ferentiation among species within each subgenus was based primarily on quantitative features. Fea- tures of external and soft anatomy, biochemistry, genetics, and ecology are taxonomically useful for Recent species (Hollister, 1916; Tileston and Lechleitner, 1966; Nadler et al., 1971; Pizzimenti, 1975 and references therein; Hoogland, 1981; McCullough et al., 1987). These features, however, were not treated in this revision because they cannot be evaluated on fossil forms. Age Determination The chronologic scheme used herein is the Land Mammal Ages proposed by Wood et al. ( 1941 ) and subdivided by Schultz et al. ( 1 978 ). A recent review of Plio-Pleistocene Land Mammal Ages is provided by Lundelius et al. (1987). Fossil prairie dogs are known from the following intervals: Senecan (Late Blancan; ca. 2.5-2.0 my B.P); Sappan (Early Irv- ingtonian; ca. 2.0-0.75 my B.P); Cudahyan (Medial Irvingtonian; ca. 0.75-0.50 my B.P); Sheridanian (Late Irvingtonian; ca. 0.50-0.20 my B.P.); and Rancholabrean (0.20-0.01 my B.P). The last was further subdivided herein into Early (Late Illinoian and Sangamonian; ca. 0.20-0. 10 my B.P.) and Late Rancholabrean (Wisconsinan; ca. 0.10-0.01 my B.P). Some fossils could not be placed more pre- cisely than Sheridanian or Early Rancholabrean; I use the term Illinoian to refer to this interval. Abso- lute age estimates given above are based on previous correlations of faunas with dated stratigraphic se- quences. In general, I considered fossil Cynomys to have approximately the same age as associated fauna and sediments. Dating a fossil usually was based on the age of the associated fauna or sediment as estimated by some combination of biostratigraphic, lithostrati- graphic, paleomagnetic and radiometric evidence. In some cases, I estimated age based on the prairie dog fossils themselves (Goodwin, 1993). Details about specific lines of evidence used in the chrono- logical placement of individual localities are pro- vided elsewhere (Goodwin, 1990a:281 and refer- ences therein; modified as in Goodwin, in prep.). Because prairie dogs are burrowing rodents, their fossils may be younger than the sediments and faunas in which they occur. If the age of a fossil was suspected by the original collector, or if state of preservation or stage of evolution was out of charac- ter with associated fauna, I eliminated the specimen from my analysis. Nonetheless, some fossils consid- ered herein probably are unrecognized intrusives. Terminology and Measurement Protocol Cranial and mandibular terminology generally follow Bryant (1945); dental terminology follows Wood and Wilson (1936). Table 1 lists and briefly describes the cranial and mandibular variables that I measured. Many are not standard measurements taken in studies of Recent skulls. I selected variables based on the frequency of preservation on fossils and probability of systematic usefulness (based on previous studies or my own observations). I took measurements with a dial cali- pers. Table 2 lists and briefly describes the dental variables that I measured. All dental measurements were taken with a Daedal combination linear/rotary measuring stage under an Olympus binocular dis- secting scope. Dental measurements represent mea- sures of the tooth in occlusal view. Because of the shape of Cynomys cheek teeth, slight buccal or lingual rotation may result in somewhat different measures of width. I was unable to find absolute criteria for orientation, but the following protocol minimizes variability due to orientation. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS Table 1. Description of cranial and mandibular measurements taken in this study. Variable Description PALM3 PALP-^ MXALV PALLN SKLN INTOR PSTOR OCCPW OCCPH FORMW FORMH EAML MDALV ANTJW Width of palate between lingual borders of M^ alveoli Width of palate between lingual borders of P-^ alveoli Length of maxillary alveolar row parallel to its long axis Length of palate from notch in premaxillae above the l' alveoli to the anterior margin of the left or right internal nares Length of skull from notch in premaxillae above the l' alveoli to the anteroventral border of the foramen magnum Least width of the interorbital region, including the supraorbital notches (if developed) Post orbital width immediately posterior to the postorbital processes Width of occiput across the paroccipital processes Height of occiput from base of the medioventral wall of the foramen magnum to the dorsalmost point of the occiput Greatest width of the foramen magnum Greatest height of the foramen magnum Greatest anteroposterior dimension of the external auditory meatus Length of mandibular alveolar row parallel to its long axis Length of anterior part of jaw from posterior border of alveolar row to dorsoposterior margin of the I J alveolus P^^ — Tooth oriented with the single root pro- jecting directly away from the viewer. P4-M3 — Tooth oriented with the three roots projecting away from the viewer at approxi- mately equal angles. This was usually easy to determine, less so on M-^, thus measures of width on the latter may be less repeatable. P4 — Proper orientation usually about halfway between two limits defined in the rotation of P4 around the anteroposterior axis of the jaw, namely that point in lingual rotation where the tip of the metaconid extends be- yond the lingual margin of the trigonid, and that point in buccal rotation where the posterolingual root becomes visible below the lingual margin of the talonid. M1-M3 — Tooth oriented such that the two an- terior roots project away from the viewer at approximately equal angles relative to the line of sight. The direction of root projection usually could be estimated from the exposed proximal portion. In some cases, roots were not readily visible, and orientation was esti- mated based on experience gained from measuring numerous teeth. MORPHOMETRIC METHODS It usually was difficult to distinguish among fossil and Recent prairie dogs within a subgenus based on qualitative dental features. Taxonomic decisions at this level were dependent heavily on morphometric evidence. Selection of operational taxonomic units and samples. — For each subgeneric analysis, opera- tional taxonomic units (OTUs) included all Recent species and putative fossil forms (delimited tempo- rally and/or geographically) assignable to that sub- UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 2. Description of dental measurements talcen in this study. See text for orientation criteria. Variable Description LP3 wp3 LP4, LM^ WP4, WM'' LP4. LM^ WTRP4 WTLP4 WM. Greatest length of P-^ perpendicular to long axis of protoloph. Greatest width of P-^ parallel to long axis of protoloph. Greatest length of indicated upper tooth perpendicular to long axis through trigon. Greatest transverse width of indicated upper tooth parallel to long axis through trigon. Greatest length of indicated lower tooth perpendicular to long axis through trigonid. Greatest transverse width across the trigonid of P4 parallel to the long axis of trigonid. Greatest transverse width across the talonid of P4 parallel to the long axis of trigonid. Greatest transverse width across trigonid of indicated lower molar parallel to the long axis of the trigonid. genus. Fossil OTUs do not coirespond with named fossil species for black tails but do for white tails. Samples of fossil OTUs are the pooled minimum number of individuals (MNI) from relevant (appro- priate age and/or geographic location) fossil locali- ties. To avoid circularity, fossil localities dated on the prairie dogs themselves (Goodwin, 1 993 ) are not included in the samples of temporally-defined OTUs. Principal component analysis. — I used corre- lation-based principal component analysis (PCA) to summarize size and shape variation among fossil and Recent OTUs within each subgenus. PCA is a multivariate technique designed to reduce the num- ber of variables that need to be considered from many correlated variables to a few uncorrelated ones (called the principal components) that are linear combinations ofthe original variables (Manly, 1986). I used the SAS statistical package (SAS, 1985) to perform PCA. My interpretation of a principal component was based on the correlations (termed loadings) of the original variables with that component. If all load- ings were high and positive, I interpreted that com- ponent as a general size axis (frequently the first principal component, PCI). If some loadings were high and positive whereas others were high but negative, I considered that component as a shape axis [contrast between variables with positive and negative loadings; frequently principal component two (PC2) and following]. Separate PCAs were performed on cranial data from each subgenus. Fossil skulls usually are dam- aged, thus only a subset ofthe original variables was used in each cranial PCA in order to increase sample size. PCAs also were performed on lower dental data. For the subgenus Leucocrossuromys, a single analysis was done using a set of 7 dental variables (measurements of P4-M3). For the subgenus Cynomys, separate analyses were done using ante- rior (P4-M1) and posterior (M2-M3) dentition, re- spectively. Statistical comparisons. — I made comparisons among fossil and Recent OTUs within each subge- nus using as variables the first two axes (PC 1 , PC2) of the cranial PCA (for Leucocrossuromys only); alveolar length (MDALV); and PCI and PC2 of each dental PCA. Principal components were used as variables because they generally suinmarize varia- tion in size (PCI ) and aspects of shape (PC2). Each variable was tested for normality within each sample using the Kolmogorov-Smirnov test (test statistic derived by the BASTAT routine ofthe BIOM statis- tical package; Rohlf, 1985) and for homogeneity of variances among samples using the F-max test. Unless noted otherwise, these assumptions were found to be valid. I used the MCPAIR routine of BIOM (Rohlf, 1985) to make statistical compari- sons among sample means. Phylogenetic Analysis I investigated phylogenetic relationships among Recent and fossil prairie dogs using the computer program PAUP (phylogenetic analysis using parsi- SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS mony; Swofford, 1985). I treated characters as or- dered and employed the branch and bound algo- rithm to find all equally parsimonious phylogenetic hypotheses. These hypotheses were rooted using a hypothetical ancestor based on the shared morphol- ogy of Spermophilus richardsonii and S. parryii. one or both of which usually were placed as or within the sister group of Cynomys in phylogenetic hypotheses generated in a preliminary study of relationships among species of the subgenus Spermophilus — genus Cynoinys clade (Goodwin, 1990a). For each hypothesis, PAUP generated a consistency index which is the theoretical minimum number of evolutionary steps divided by the actual number of steps. Abbreviations Variables. — Abbreviations of cranial and dental variables are given in Tables 1 and 2. Institutions and collections. — Material perti- nent to this study was obtained from numerous institutions and collections. The following abbre- viations are found in the text, figures, tables, and appendix: ADAM — Adams State College, Alamosa, Colo- rado AMNH — American Museum of Natural History, New York DMNH — Denver Museum of Natural History, Denver FHSU — Sternberg Memorial Museum, Fort Hays State University, Hays, Kansas IMNH — IdahoStateUniversity, Museum of Natu- ral History, Pocatello IOWA — University of Iowa, Iowa City KUM — Natural History Museum, University of Kansas, mammalogy collection. Lawrence KUVP — Natural History Museum, University of Kansas, vertebrate paleontology collection, Lawrence MWU — Midwestern University, Wichita Falls, Texas PMA — Provincial Museum of Alberta, Edmonton ROM — Royal Ontario Museum, Toronto SMU — Southern Methodist University, Dallas SDSM — South Dakota School of Mines, Geology Museum, Rapid City TMM — Texas Memorial Museum, University of Texas, Austin TTU — The Museum, Texas Tech University, Lub- bock UCM — University of Colorado Museum, Boulder UMMP — University of Michigan Museum of Pa- leontology, Ann Arbor UMTG — University of Montana, Geology Mu- seum, Missoula UNSM — University of Nebraska State Museum, Lincoln USG — University of Saskatchewan, Department of Geology, Saskatoon USNM — United States National Museum, Wash- ington UTEP — University of Texas. El Paso UWYA — University of Wyoming, Anthropology Museum, Laramie UWYG — University of Wyoming, Geology Mu- seum, Laramie SYSTEMATIC ACCOUNTS ORDER RODENTIA Family Sciuridae Genus Cynomys Figures lA, IC, IE, IG Synonomy. — Summarized by Hollister, 1916:10, and Hall, 1981:410, for Recent forms. Type Species. — Cynomys socialis Rafinesque, 1817 {=Cynomys ludoviciamis). Geologic and Geographic Range. — Late Blancan (Senecan) to Recent; restricted to a zone from northern Mexico to southern Canada across mid-continental North America. Emended Diagnosis. — Large ground squirrels resembling the subgenus Spermophilus, but with relatively larger cheek teeth; maxillary tooth rows strongly convergent posteriorly; protolophid of P4 complete or nearly so and well developed; metalophid of M3 complete, merges lingually with posterior wall of trigonid well up from floor of talonid. UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 H Fig. 1 . Comparisons of the genus Cynomys (A, C, E, G) with the large ground squirrel. Spermophihts parryii (B, D, F, H). (A-B) Anterior view of L maxilla; (C-D) occlusal view of LM2-M3; (E-F) occlusal view of LP4; (G-H) posterior view of LM3. Arrows point to generic characters noted in text. Description Skull. — The skull oi Cynomys is robust, more so than is typical for other ground squiirels. In dorsal view, the skull resembles the subgenus Spermophi Ins in that the rostrum has subparallel sides, the interorbital and postorbital widths are subequal, and the zygomatic arches are expanded, especially at the squamosal roots. The maxilla and palate exhibit several distinctive features. The infraorbital foramen is strongly trian- gular, the lateral wall sloping ventrolaterad, resem- bling advanced members of the subgenus Spermo- philus. The basal (ventral) wall of the foramen usually is robust and inclined, sloping lateroventrad from its medial end (Fig. lA); in the subgenus Spermophilus this margin is more slender and usu- ally horizontal in orientation (Fig. IB). The masse- teric tubercle typically is massive, positioned at the ventrolateral comer of the foramen, and laterally extended. The zygomatic plate of the maxilla, in anterior view, is deeply concave along its ventral margin in advanced species oi Cynomys (Fig. lA); in ground squiiTcls this margin is less deeply concave (Fig. IB). The alveolar rows on each side of the palate are strongly convergent posteriorly in most specimens. In a few specimens of ground squirrels, I have noted weak posterior convergence but never as strongly as is typical for Cynomys. Upper dentition. — The upper incisors are ro- bust and procumbent. The upper cheek teeth, as well as the lowers, are large relative to skull size, and extremely hypsodont in advanced fossil and extant species. In this respect, prairie dogs differ from all known fossil and living ground squirrels. Extreme hypsodonty is especially evident lingually at the protocone. Several early prairie dogs have less hypsodont teeth, resembling large advanced ground squirrels; the extreme hypsodonty evident today developed since the origin of the clade. P-^ is large; is usually somewhat flattened anteri- orly; and bears a high, functional protoloph. P-+-M- are triangular in occlusal view, as in the subgenus Spermophilus, but are relatively wider. On Mi-M^, the buccal one-half of the protoloph frequently is expanded along its posterior margin, approaching an accessory lophule on M^ (Fig. IC; contrast with Fig. ID). The expanded section of the protoloph is terminated lingually by a sharp indentation from the posterior direction (Fig. IC). In morphologically derived prairie dogs, M-^ is long relative to M^; early prairie dogs appear to have a much shorter M-^. The metaloph on M3 is well developed and extends all the way across the tooth. Lower jaw and dentition. — The portion of the lower jaw beneath the diastema is robust and short relative to the length of the jaw. The mental foramen typically is positioned somewhat anterior to the plane of the anterior root of P4, more anteriorly than usual for many members of the subgenus Spermo- philus. The coronoid process projects strongly dor- sad, and the angular process turns inward at an angle of about 90° to the plane of the posterior part of the ramus, in both respects resembling the subgenus Spermophilus. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS ^ B A ' Fig. 2. (A-D) Holotype of Cynomys vetus (KUVP 6187): (A) dorsal view of preserved and partially recon- structed skull roof (reconstruction shown as dotted line): (B) lateral view of R jugal angle; (C) anterodorsal view of squamosal root of zygomatic arch; (D) anterior view L maxilla. (E) Anterior view of L maxilla of C. sappaensis (UNSM 1 1761 ). Scale bar represents 5 mm. P4 usually exhibits a complete, well developed protolophid (Fig. IE); the protolophid is variably developed in other ground squirrels, but when promi- nent is separated from the metaconid by a distinct notch (Fig. 1 F). M 1-M2 bear a complete metalophid, particularly developed in advanced species. The talonid of these two teeth, when unworn, typically bears a lophid or one or more developed cuspulids, but these structures frequently disappear with mod- erate tooth wear. Similar structures seldom are found in the subgenus Spermophilus. M3 exhibits a com- plete metalophid on all prairie dogs (Fig. IG; con- trast with Fig. 1 H), but it is not as strongly developed in early forms. The talonid bears a deep to very deep basin trench along the lingual border of the ectolophid, much more developed than in any extant species of Spermophilus. ICynomys vetus Hibbard, 1942 Figures 2A-D, 3A Cynomys vetus Hibbard. 1942:268. Holotype and Only Specimen.— KUVP 6187, fragmentary skull including palate with right and left P3-M3. isolated right and left II, and parts of skull roof, squamosal, and jugal of a mature to old adult. Horizon and Type Locality. — Probably Late Blancan (Senecan); reported from the "Early phase of the Loveland loess (brown zone, whitened by calcareous matter and containing large limestone concretions, occurring below the typical red phase)" (Hibbard, 1942:268); unnamed locality. Sec. 3, T. 1 S., R. low, Jewell County, Kansas. It should be noted that these deposits are clearly not equivalent with the Loveland as the term is usually used (Illinoian complex of loesses and paleosols; Schultz and Mar- tin. 1970). Geologic and Geographic Range. — Known only from the type locality. Emended Diagnosis. — Smaller than all extant Cynomys; distinguished from all extant and fossil species known from appropriate material by rela- tively shorter M^, more circular P3, more concave anterior margin of the squamosal arm of zygomatic arch (anterodorsal view), less concave ventral mar- gin of zygomatic plate (anterior view). Description The following cranial measurements were ob- tained or estimated from the type: INTOR, 9.4 mm; PSTOR, 12.7 mm (both are estimates made on reconstructed skull roof); MXALV: left, 14.55 mm; right, 13.65 mm. Dental measurements are given in Table 3. Skull roof. — The skull roof is incompletely pre- served, but it was possible to reconstruct the outline of a portion thereof by projecting a mirror image of the preserved fragment (Fig. 2A). The suture be- tween the frontals and the nasals and right premax- illa is preserved (Fig. 2A). The premaxilla does not extend posteriad beyond the nasals as it does in many specimens of Cynomys gunnisoni. The interorbital width of the frontals, as reconstructed in Fig. 2A, is less than the postorbital width. The supraorbital notches are well developed and deep. 8 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 3. Measurements of upper dentition for three fossil prairie dogs. ?C. vetus C. sappaensis C. spenceri Variable KU6187 UNSM 11761 UNSM 33798 LP3 2.83 3.02 Wp3 3.02 3.50 LP4 3.04 — 3.20 Wp4 4.29 4.52 lm' 2.64 2.87 3.08 wmi 4.68 4.61 4.75 LM- 2.78 2.84 3.06 WM2 4.55 4.43 4.66 LM-3 3.16 4.04 WM-"* 4.27 4.25 more so than on most prairie dogs but within the range of variation exhibited by the genus. The base of the left postorbital process suggests that it was moderately robust. Jugal. — The jugal angle of the zygomatic arch, in lateral view, is relatively triangular in shape, capped ventrally by a distinct apical prominence (Fig. 2B). The arm of the jugal extending posteriad from the triangle tapers gradually; it does not thin rapidly into a thin sheet as in the subgenus Leucocrossuromys. In these respects, Cynomys vetus resembles the subgenus Cynomys, although the jugal is even more triangular in the latter. Squamosal. — The anterior margin of the squa- mosal root of the zygomatic arch, viewed anterodorsally, is strongly concave (Fig. 2C), not flattened as in many prairie dogs. The posterolateral portion of the squamosal root is not extended as strongly posteriad as in many prairie dogs. Viewed laterally, the portion of the squamosal ventral and posterior to the root of the zygomatic arch and anterior to the auditory bulla extends more posteriad relative to the position of the root than it does in any other Cynomys specimen that I examined. Maxillae and palate. — The infraorbital foramina are moderately large, triangular. The masseteric tubercles are moderately developed but not strongly directed laterally. Viewed anteriorly, the ventral margin of the zygomatic plate of the maxilla is weakly concave and merges gradually with the border of the alveolar row (Fig. 2D), not strongly concave as in other Cynomys for which this charac- ter is known (Fig. 1 A). Viewed ventrally, the zygo- matic notches terminate opposite Ml. not M2 as appears to be suggested in the original description (Hibbard. 1942:268). The palate appears narrow (Fig. 3A), at least in part because the tooth rows were rotated ventro- mediad around the palatal midline when the fossil was originally prepared and glued together. Allow- ing for this distortion, the palate still would probably be narrower than typical for ground squirrels of the subgenus Spermop/iilus and most specimens of prai- Fig. 3. Occlusal views of: (A) holotype oflCynomys vetus (KUVP 6187); (B) L maxilla with M1-M2 referred to C. sappaensis (UNSM 1 1761 ). Scale bars represent 5 mm. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 9 rie dogs. The tooth rows converge posteriorly, more gests that some of the advanced features character- strongly than in Spermophilus. Lengths of right and istic of both Cynomys and Spermophilus arose in left alveolar rows differ significantly due to the parallel, unless moiphologically derived species of strong anterodorsad direction of the root on left P3. the subgenus Spermophilus are more closely related Upper dentition. — The incisors are strongly to Cynomys than is ?C. vetus. curved and relatively deep for the size of the speci- I follow Hibbard (1942) in assigning vetus to the men. The cheek teeth are heavily worn, and details genus Cynomys. but I query the assignment because ofcusp and loph morphology are largely obliterated, of the uncertainty noted previously. The fossil is P3 is large and rounded, lacking the anterior either a primitive prairie dog or a ground squirrel flattening typical for advanced prairie dogs. The convergent on the prairie dog morphotype. I doubt ridge bounding the anterior cingulum appears low. that it has special relationship with the subgenus not as well developed as in the subgenus Cynomys. Cynomys. thus I regard the similarity of jugal con- P4-M2 are triangular in occlusal outline. All three of formation as convergence. Alternatively, the trian- these teeth exhibit a developed, buccally extended gular jugal may be a primitive state retained by the parastyle. The parastyle and metacone both extend subgenus Cynomys. However, this interpretation is more buccad than does the paracone, thus the buccal not supported by outgroup comparisons — ground margins appear indented at the paracone. especially squirrels lack a strongly triangular jugal. on M'. M3 is much shorter relative to the length of The relationship between ICynomys vetus and M- than in other prairie dogs or advanced members other fossil prairie dogs deserves attention. It is of the subgenus Spermophilus. Although worn, M3 small and primitive as is C. hihbardi, but the latter is appears to have had a developed metaloph. known only from a lower jaw with teeth. Relative sizes of M-^ and M3 are usually correlated in ground- CoMMFNTS dwelling squirrels. Assuming this correlation to be true for these forms, ?C. vetus had shorter posterior The large P3 and triangular upper cheek teeth molars than did C. hihbardi. This suggests that ?C. exhibited by ICynomys vetus support a relationship vetus was the more primitive of the two. with the advanced ground squirrel clade which M1-M2 of the fossil resemble in general shape includes Cynomys and the subgenus Spermophilus. and size those of a small prairie dog from the Sappa The posteriorly convergent tooth rows support a Local Fauna, described herein as a new species, relationship with Cy«c»wy5, and the conformation of Detailed dental comparisons are not possible be- the jugal angle exhibited by the fossil is similar to cause of the extreme wear on ICynomys vetus. but (but not as well developed as ) the subgenus Cynomys. the conformation of the ventral margin of the zygo- This is consistent with Hibbard 's original sugges- matic plate on the Sappa form is clearly prairie dog- tion that ?C. vetus was a member of the latter like and differs from the condition seen in the type (Hibbard, 1942). of IC. vetus. However, ICynomys vetus lacks a number of Dalquest (1967) referred prairie dogs from the derived features characteristic of the subgenus Slaton Local Fauna of Texas to Cy^cwy^ I't-m^ and Cynomys or even of the genus. In contrast with the suggested that they represented a white tail. The subgenus Cynomys. the fossil lacks the extreme Slaton prairie dogs are relatively small and P4-M2 development of the anterior cingular ridge on P3. In are roughly comparable in size with the type of ?C. contrast with all other known prairie dogs, the fossil vetus. However, subsequent work showed that the lacks apparently derived features of the P3, M^, Slaton specimens represent an advanced black tail conformation of the squamosal, and conformation (Dalquest, 1988) with no relationship to ?C. vetus. of the zygomatic plate (see the description for de- Eshelman (1975) referred associated left and tails). In these respects, ?C. vetus is more similar to right lower jaws from the White Rock Local Fauna ground squirrels of the subgenus Spermophilus. In to ICynomys vetus based on size. The absence of at least one feature — the short M^ — the fossil ap- cranial and upper dental elements makes this assign- pears to be even less derived than many extant ment uncertain. The jaws are robust; the mental species of the subgenus Spermophilus. This sug- foramen is positioned far forward; and P4 appears to UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 exhibit a complete protolophid (descriptions and fig. 4C-D in Eshelman, 1975). These characters are consistent with assignment to Cynomys. However, the metalophid on M2-M3 is incomplete, removing the specimens from Cynomys as diagnosed herein. Cynomys hihhardi Eshelman, 1975 Figures 4A. 4C-D Cynomys hibhardi Eshelman, 1975:27. Holotype.— UMMP V6 1 648. left lower jaw with M-M,. Horizon and Type Locality. — Late Blancan (Senecan); collected "at UM-K9-72 in the silty clay lithosome, approximately 2.2 m below the contact with the sand lithosome, Belleville Fomiation, SE I / 4, SEl/4, SWl/4, Sec. 34, T. 1 S, R. 5 W, Republic County, Kansas" (Eshelman, 1975:27). Referred Specimen. — UMMP V745 10, unworn right M„ from Nash Local Fauna, Meade County, Kansas. Geologic and Geographic Range. — Known from the Late Blancan (Senecan) and Early Irvingtonian (Sappan) of Kansas. Emended Diagnosis. — Cynomys, but smaller than all extant species; teeth less high crowned than other fossil and Recent species, possibly excepting ICynomys vetus: metalophid complete on M1-M2 but deeply notched; metalophid complete on M3 but less well developed than in other species. Description Lower jaw. — Selected measurements of the lower jaw and dentition are provided in Table 4. A lateral view of the holotype is given in Figs. 4A. The portion of the jaw beneath the diastema is similar in proportion to other prairie dogs, but is shorter and more robust than typical for ground squirrels of the subgenus Spermophilus. The position of the mental foramen — somewhat anterior to the plane of the anterior root of P4 — also resembles other Cynomys rather than Spermophilus. The symphy seal region of the jaw, viewed dorsally, is shorter anteroposteriorly than in any other prairie dog specimen that I have examined for this character. The masseteric ridge is well developed ventrally; the anterior limit of the masseteric fossa lies at the level of the posterior half of P4. The condyloid process is relatively longer and less robust, and bears a relatively deeper fossa on its lateral surface, than is typical for extant C. ludovicianus and C. leuciirus, but these features resemble some specimens of C. gunnisoni. Lower dentition. — An occlusal view of the cheek teeth of the holotype is given in Fig. 4C. 1 1 and P4 are missing, but the preserved alveoli suggest that the former was rather robust and that the latter was relatively long with a single, transversely expanded anterior root. MJ-M3 are relatively narrow trans- versely and exhibit an anterior bulge of the protolophid, but in both respects are within the range of variation exhibited by extant prairie dogs. Mj- Mo both exhibit a complete, but deeply notched, metalophid which forms an incomplete posterior boundary to the anterobuccally-posterolingually oriented trigonid basin. The talonid basin of these teeth is rugose, especially on Mo where the rugosity forms a distinct, transversely oriented lophid run- ning from the base of the entoconid toward the base of the hypoconid. This lophid is particularly high in an unworn specimen from the Nash Local Fauna (Eshelman and Hibbard, 1981) referred here to C. hihhardi (UMMP V74510; Fig. 4D). On both M, and Mt, the rugose portion of the talonid is separated from the well developed ectolophid by a deep trench. M3 is elongate relative to the Mt, much as in other prairie dogs. Its trigonid resembles Mj-Mt in most respects. The metalophid is slightly less devel- oped but is complete, merging with the posterolingual wall of the trigonid about half-way up from its base. In most ground squirrels, even advanced forms of the subgenus Spermophilus, the metalophid projects towards the floor of the talonid basin and only joins the posterolingual wall of the trigonid, if at all, near its base. The talonid of M3 bears a low, rugose ridge which, from a point just posterior to the metaconid, arches buccad and slightly posteriad toward the ectolophid, and then posteriad and slightly linguad until it merges with the posterolophid. This ridge is separated from the ectolophid by a well developed trench, but the trench becomes very narrow near its midpoint. There is no bridge connecting the ectolophid and talonid. The hypoconid is strongly deflected anteriorly, and there is no ectostylid in the hypoflexid, in both respects resembling black-tailed rather than white-tailed prairie dogs. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 11 tmf4^ Fig. 4. Lateral view of hoiotype of (A) Cynomys hihhardi (UMMP V61648). (B) C. sappaensis (UNSM 1 1 760). (C-F) Occlusal views of dentition of (C) hoiotype of C. hihhardi and (D) RM2 referred to C. hiliJnvdi (UMMP V74510), (E) hoiotype of C. sappaensis, and (F) lower dentition referred to C. sappaensis (UNSM 1 1759). All occlusal views but (D) are stereophotos. Scale bars represent 5 mm. Scale bar under (A) also applies to (B); scale bar under (F) also applies to (C) and (E). 12 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 4. Measurements of lower jaws/dentition for several fossil prairie dogs. C. hibbardi C. sappaensis UMMP UMMP UNSM UNSM Variable V61648 V74510 11759 11760 ANTJW 21.10 — MDALV 13.65 13.80 — LP4 3.31 3.05 WTRP4 3.64 WTLP4 3.88 LMi 2.91 — 3.03 2.88 WMi 3.80 — 3.92 LMt 3.06 3.21 3.13 3.10 WM^ 4.04 4.09 4.20 LM3 4.11 4.08 WM3 4.09 3.89 Comments Cynomys hibbardi resembles other prairie dogs in a number of respects — the proportions of the diastema, position of the mental foramen, develop- ment of the metalophid on M3, and several other features. 1 agree with Eshelman (1975) that the species should be included within the genus Cynomys instead of the genus Spermophilus. However, it is possible that many of these prairie dog-like features evolved in parallel, and more complete material is needed in order to more fully assess the affinities of this species. Eshelman (1975) further suggested that the spe- cies be assigned to the subgenus Cynomys. It re- sembles that subgenus in several features, but most of these similarities probably are primitive for the genus (e.g., absence of a bridge connecting the ectolophid and talonid on M3; anteriorly deflected hypoconid on M3; transversely narrow cheek teeth). Indeed, the species appears to be primitive with regards to the morphology of prairie dogs in most respects, and I believe it is best treated as a species with uncertain subgeneric affinities. It may lie close to the ancestry of all later taxa. Cynomys hibbardi is clearly distinguishable from later fossil prairie dogs on a number of features, notably its lower crowned teeth and less developed metalophid on M3. It appears to be more advanced than ?C. vetus in the more elongate posterior molars (see the account of that species). The type specimen was recovered from the same horizon as UMMP V61649, a relatively complete right lower jaw with heavily worn P4-M3 that re- sembles Cynomys hibbardi in overall size (Eshelman, 1975). However, this specimen differs in several respects from the type, and it is not clear that it represents the same species. Subgenus Cynomys Rafinesque, 1817 Figures 5A, 5C, 5E, 6A, 6C, 6E Type Species. — As for the genus. Geologic and Geographic Range. — Early Irvingtonian ( Sappan ) to Recent; restricted through- out its history to mid-continental North America. Emended Diagnosis. — Cynomys, distinguished from all other prairie dogs by a large, high ridge bordering anterior cingulum of P3; from ?C. vetus by more elongate M-^; from C. hibbardi by better devel- oped metalophid on Mj and M2; and from the subgenus Leucocrossuromys by a strongly devel- oped jugal triangle, less wide trigonid on P4 (relative to the width across the talonid), absence of a bridge between the talonid and ectolophid of M3, and the strong anterior deflection of the hypoconid on M3. Description Skull and upper dentition. — Distinctive as- pects of the skull and upper dentition are shown in Fig. 5A, C, and E and can be compared with equiva- lent features in Leucocrossuromys (Fig. 5B, D, F). Viewed dorsally, the angle between the lateral mar- gins of the rostrum and the anterodorsal margin of the zygomatic plate is typically more abrupt than in other prairie dogs, but this character exhibits consid- erable variation. The well developed jugal triangle is one of the most distinctive features of the subgenus Cynomys, distinguishing it from the subgenus Leucocrossu- romys (Fig. 5 A). Astrongaupraiugalprocessprojects, dorsad and slightly anteriad from the triangle and forms the posterior margin of a robust lateral margin to the zygomatic plate. Posteriorly, the jugal triangle tapers gradually, thus the lateral surface of the zygo- matic arch remains thickened along about one-half of its length. In posterior view, the occipital plate exhibits a domed dorsal margin which follows a more or less SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 13 follows a more or less continuous slope from its inception near the anterolingual margin of the cin- gulum to its termination near the apex of the paracone. No other known prairie dog taxon, fossil or Recent, bears this confoiTnation of the cingular ridge. Other- wise, the upper dentition of the subgenus is similar to that of Leucocrossuromys. D F Fig. 5. Comparisons of the subgenera Cynomys (A, C, E) and Leucocrossuromys (B, D, F). (A-B) Lateral view of L jugal angle; (C-D) posterior view of occipital plate; (E-F) buccal view of the RP-\ Arrows point to subgeneric characters noted in text. continuous arc from the level of the external audi- tory meatus to the dorsalmost point of the occiput (Fig. 5C), in this respect resembling most species of ground squirrels but differing from some species of Leucocrossuromys. P3 bears a well developed ridge bounding the anterior cingulum externally (Fig. 5E). This ridge Fig. 6. Comparisons of the subgenera Cynomys (A, C, E) and Leucocrossuromys (B, D. F). (A-B) Lateral view of L lower jaw; (C-D) occlusal view of RP4; (E-F) occlusal view of LM3. 14 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Lower jaw and dentition. — Aspects of the lower jaw and dentition are shown in Fig. 6 (A, C, E) and contrasted with Leiicocrossuwmys (B, D. F). On average, the portion of the lower jaw beneath the diastema is more robust than in other prairie dogs (Fig. 6A). As a result, the dorsal margin of this region, viewed laterally, tends to merge gradually with the anterior margin of the tooth row. The lower dentition is distinctive in several ways. On average, P4-M3 are relatively narrow buccolingually, espe- cially when compared to Leiicocrossuromys. P4 bears a buccolingually expanded trigonid, typical for prairie dogs generally, but the trigonid width is usually subequal to or less than the talonid width. This results from an enlarged, buccally expanded hypoconid (Fig. 6C). M 1-M2 usually lack a developed mesolophid on the floor of the talonid, but they frequently exhibit one or more conulids, frequently oriented in a line to present an incipient mesolophid. M3 bears several subgeneric characters, noted in the diagnosis (Fig. 6E). It usually lacks an ectostylid in the hypoflexid. The talonid platform is heavily rugose, especially along a curving ridge delimiting the anterior and buccal margins of the platfomi. MORPHOMETRIC RELATIONSHIPS AMONG FoSSIL AND Recent Black-tailed Prairie Dogs 1 considered morphometric relationships among Recent and putative fossil black tails in order to clarify relationships and aid in making taxonomic decisions. Initial inspection of fossil material sug- gested the presence of three temporally successive black tails with the most recent ( Late Rancholabrean ) possibly divisible into two geographic morphs. The earliest, from the Early Irvingtonian (Sappan), is easy to distinguish from later members of the subge- nus on qualitative features; therefore, it was not considered in this morphometric analysis. The other fossil OTUs — one from the Sheridanian and Early Rancholabrean (Illinoian and Sangamonian) and two from the Late Rancholabrean ( Wisconsinan) — are similar to each other and to two extant species {Cynomys ludoviciamis, C. mexicanus) in general cranial and dental morphology. These five are the initial OTUs in the following analyses. Unfortu- nately, the smaller of the two Late Rancholabrean morphs is only known from a few specimens recov- ered from one locality in southeastern New Mexico. It had to be removed from statistical comparisons, although descriptive comparisons were made where possible. Cranial analysis. — Mean values for 12 cranial variables are given for black tails in Table 5. Sample sizes for two represented fossil fornis (cranial mate- rial was unavailable for the small, Wisconsinan moiph) are too small to allow for meaningful statis- tical comparisons. However, the large Wisconsinan form resembles Cynomys ludovicianus in being generally larger than the Illinoian form and C. mexicanus. Nine of the cranial variables were used in a PC A, and the congelations of these variables with PCI and PC2 are given in Table 6. The pattern of con'elations suggests that PC 1 represents skull size, especially as reflected in variables PALP3, MXALV, PALLN, and INTOR, but with a contrast to variable PSTOR. Specimens with high scores on PC 1 tend to be large, but with a relatively constricted postorbital region (PSTOR). PC2 appears to reflect a contrast between palatal width (especially as represented by PALM-^) and EAML on the one hand, and MXALV and FORMH on the other. Specimens with high scores on PC2 exhibit a relatively broad palate and large external auditory meatus but a short alveolar row and low foramen magnum. Figure 7 plots bivariate means of fossil and Recent samples on PCI and PC2. The extant forms (Cynomys mexicanus, C. ludovicianus) are clearly distinguishable on both axes, suggesting differences in size and shape. The small sample of large Wisconsinan fossils differs from both Recent taxa, especially on PC2, but falls out closer to C. ludovicianus than to C. mexicanus on both axes. The one Illinoian specimen resembles C. mexicanus on PC 1 but exhibits a lower score on PC2 than is typical for that species, suggesting differences in shape. Mandibular/lower dental analysis. — Mean values for 10 mandibular/lower dental variables for black tails are shown in Table 5. Separate PCAs were performed on the anterior and posterior dental variables, and the correlations of the original vari- ables with PCI and PC2 are shown in Table 7. In both analy.ses, 78% of total variation in the original data is accounted for by PCI; this axis clearly represents general size. PC2 represents a shape axis, in both analyses reflecting a contrast between mea- SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 15 Table 5. Summary statistics for cranial and lower dental variables of two Recent and three putative fossil forms of the subgenus Cynomys. Sample size is given at the head of each column and is only repeated when it changes. Large Small Cy nomys Illinoian Cy nomys Wisconsinan Wisconsinan mexicanus black tail ludo} 'icianus black tail black tail Variable X (sd)« X (sd)« X [sd)« X (sd) n X (sd)« Cranial PALM-'' 5.42 (0.42) 13 4.95 (0.28) 2 4.93 (0.39) 25 4.74 (0.65) 6 _ PALP-^ 10.57 (0.56) 9.85 (0.57) 10.76 (0.52) 10.47 (0.39) MXALV 15.45 (0.50) 15.72 (0.04) 16.86 (0.58) 17.65 (0.92) PALLN 32.98 (1.60) 33.45 (— ) 1 34.87 (1.40) 34.60 (0.54) 4 SKLN 53.74 (2.60) 56.37 (2.70) 56.45 (1.00) 3 INTOR 12.59 (0.97) 12.70 (— ) 12.99 (0.84) 12.92 (1.10) 5 PSTOR 13.90 (0.81) 14.25 (— ) 13.31 (1.10) 13.14 (0.78) — OCCPW 27.15 (0.95) — 28.08 (0.88) 28.22 (0.66) 4 — OCCPH 17.07 (0.75) — 17.66 (0.81) 17.88 (0.28) FORMW 8.85 (0.33) 8.75 (— ) 8.84 (0.26) 8.68 (0.21) FORMH 6.32 (0.60) 6.55 (— ) 6.93 (0.46) 7.24 (0.61) — EAML 3.80 (0.30) 3.50 (— ) 3.57 (0.30) 3.75 (0.09) 5 Mandibular/ lower dental MDALV 13.71 (0.30) 10 14.33 (0.54) 23 15.20 (0.54) 28 15.29 (0.57) 49 14.10 (0.40) 4 LP4 2.98 (0.10) 3.18 (0.16) 19 3.33 (0.16) 3.38 (0.16) 34 3.15 (0.10) 3 WTRP4 3.77 (0.16) 3.97 (0.31) 4.30 (0.20) 4.32 (0.28) 3.94 (0.10) WTLP4 3.80 (0.20) 4.00 (0.33) 4.23 (0.23) 4.36 (0 .27) 3.96 (0.23) LMj 2.89 (0.17) 2.92 (0.15) 3.16 (0.14) 3.18 (0.16) 33 2.86 (0.01) 2 WM, 4.22 (0.12) 4.33 (0.28) 4.69 (0.21) 4.69 (0.24) 4.26 (0.04) LM2 3.07 (0.12) 3.12 (0.14) 18 3.32 (0.15) 3.36 (0.16) 34 3.19 (0.10) WM2 4.45 (0.11) 4.53 (0.32) 4.86 (0.19) 4.90 (0.24) 4.60 (0.05) LM3^ 4.36 (0.19) 4.45 (0.24) 22 4.83 (0.26) 4.86 (0 27) 48 4.52 (0.13) WM3 4.39 (0.14) 4.38 (0.21) 4.74 (0.18) 4.72 (0.20) 4.43 (0.07) sures of tooth length and width (length of Mj and Mt had particularly high correlations with PC2 in the respective analyses). Thus, specimens with high scores on this axis tend to have relatively long teeth (especially Mj or M2) compared to tooth width. Statistical comparisons were made among black tail OTUs using mean values of MDALV, PCI derived from the anterior and posterior dentition (general measures of size), and PC2 derived from the anterior and posterior dentition (general mea- sures of shape). Within each sample, all variables were found to be normally distributed; variances of compared samples were homogeneous except for PCI derived from M2-M3. These comparisons are displayed in Table 8; the means connected by a line are not significantly different. The large Wisconsinan form does not differ from extant Cynomys ludovicianus (p > 0. 1 ) in any comparison. However, these two show highly significant differences (p < 0.01) with both the Illinoian form and extant C. mexicanus in all com- parisons of size (MDALV, both PC 1 s). The Illinoian form and C. mexicanus are significantly different (p < 0.05) only in MDALV. There are no differences among groups in mean values of PC2, thus no differentiation in dental shape is evident. Because of small sample size, the small Wisconsinan form is not included in Table 8. How- 16 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 6. Correlations (loadings) of nine original cra- nial variables with principal component axes 1 and 2. Sample includes Recent and fossil black tails. Variable PCI PC2 PALM-"* PALP-"* MXALV PALLN INTOR PSTOR FORMW FORMH EAML % total variance explained -0.15 0.49 0.71 0.80 0.67 -0.54 0.35 0.13 0.29 26 0.81 0.39 -0.52 -0.02 0.36 0.15 0.16 -0.62 0.64 23 ever, descriptive comparisons of original lower den- tal variables (Table 5) indicate close size resem- blance to the Illinoian morph. There is also general resemblance to Cynomys mexicanus. Morphometric variation in jaw proportions was examined with a bivariate plot of MDALV versus ANTJW (Fig. 8). Cynomys mexicanus differs from C. ludovicanus and the Illinoian and large Wisconsinan black tails in having a relatively low MDALV/ANTJW ratio. However, C. mexicanus resembles the small Wisconsinan form (marked by arrows. Fig. 8) in this character. Taxonomic implications. — Three taxonomic judgments seem wananted. First, large Wisconsinan black tails should be referred to the extant species, Cynomys ludovicianus. The two fonns do not differ significantly in any of the statistical comparisons. The sample of fossil skulls appears to differ from C. ludovicianus in aspects of shape (see the preceding cranial analysis), but two of three individuals in this sample came from essentially the same locality. The distinctive "shape" likely represents a local geo- graphic morphotype. Second, the Illinoian fonn should be distinguished taxonomically from both Cynomys ludovicianus and C mexicanus. It differs significantly from the former in size and from the latter in mandibular proportions. Third, the small Wisconsinan form from southeastern New Mexico likely represents C. mexicanus. The two resemble each other in size and MDALV/ANTJW ratio. Thus, there are four recognizable black tails: a 1.0 CM O -1.0 - CM IL CL LW "I" -1.5 T -1.0 -.5 T r 0 .5 PCI Fig. 7. Scatter plot of means of Recent and fossil black tails on PCI and PC2 derived from analysis of cranial variables. Bars represent standard errors of means. OTU abbreviations and sample sizes are: CL, Cynomys ludovicianus. n = 25; CM, C. me.xicanus. n = 13; IL, Illinoian black tail, n = 1 ; LW, large Wisconsin black tail, n = 3. primitive, probable black tail from the Early Irvingtonian, and a small but advanced fonn from the Late Irvingtonian and Early Rancholabrean, both described herein as new species; and the two extant forms, Cynomys ludovicianus and C. mexi- canus. Cynomys (ICynomys) sappaensis new species Figures 2E, 3B, 4B, 4E-F Holotype.— UNSM 11760, left lower jaw with P4-M3. Horizon and Type Locality. — Sappa Local Fauna, Early Irvingtonian (Sappan); "UNSM col- lection locality Hn-102 in the NWl/4 SWl/4 SEl/4 NEl/4, Sec. 1 1, T. 2 N., R. 20 W., Harlan County, Nebraska" (Martin and Schultz. 1985). Referred Specimen. — UNSM 1 1759, right lower jaw with fragments of P4-M2, and UNSM 1 1761, left maxillary fragment with M 1-M2, both from type locality. Geologic and Geographic Range. — Known only from the type locality. Diagnosis. — Small Cynomys with greater hypsodonty and higher metalophid on MJ-M2 than C. hihhardi: distinguished from all later species by lesser hypsodonty and slightly less developed metalophid on M3. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 17 Table 7. Correlations (loadings) of original lower dental variables with principal component axes 1 and 2. Two analyses are represented, one on P4-M j and another on M2-IVI3. Samples include Recent and fossil black tails. Variable PCI LP4 WTRP4 WTLP4 LMi WM, % total variance explained LM2 wm't LM3" WM3 % total variance explained 0.90 0.94 0.90 0.79 0.87 78 0.80 0.92 0.92 0.90 78 PC2 0.22 -0.20 -0.32 0.57 -0.20 11 0.58 -0.20 0.06 -0.37 Table 8. Comparisons of means among Recent and putative fossil black-tailed prairie dogs. Means con- nected by a line are not significantly different. Taxon abbreviations are identified in the legend of Fig. 7. Variable MDALV PCl:P4-Mi PC2:P4-Mi PC1:M2-M3 13 PC2:M2-M3 Taxon and mean CM IL CL LW 13.71 14.33 15.20 15.29 CM IL CL LW -2.98 -1.77 0.56 1.21 LW CM IL CL -0.13 -0.09 0.04 0.11 CM IL CL LW -2.24 -2.02 0.57 0.88 CL CM LW IL -0.63 -0.15 0.06 0.07 Etymology. — Named after the Sappa Formation type locality, from which the holotype was recov- ered. Description Measurements are given in Tables 3 and 4. Maxilla and upper dentition. — Only a frag- ment of the left maxilla is preserved in one referred specimen. It is shown in anterior view in Fig. 2E. The ventral portion of the zygomatic plate is pre- served, and in anterior view the ventral margin appears to exhibit the deep concavity typical for Cynomys. The preserved portion of the palate ap- pears to indicate posterior convergence of the tooth rows. M1-M2 are triangular in occlusal outline and are much wider than long, especially Ml (Fig. 3B). Ml exhibits a posterior expansion of the buccal portion of the protoloph as in most later Cynomys and a buccally expanded anterior cingulum which ex- tends slightly beyond the buccal margin of the paracone. Both upper teeth are less hypsodont than on later prairie dogs. Lower jaw and dentition. — Preserved features of the lower jaw resemble Cynomys. In lateral view (Fig. 4B), the portion of the lower jaw beneath the diastema is robust and the mental foramen placed anterior to the plane of the anterior root of P4. All cheek teeth appear to be relatively narrow on the holotype (Fig. 4E), but somewhat wider on UNSM II 759 (Fig. 4F). The trigonid on P4 of the holotype is relatively compressed anteroposteriorly, but is more robust in this dimension on UNSM 11 759. In both specimens, the protolophid is high and com- plete and the hypoconid is large and expanded buccally. The Mj and M2 of the holotype exhibit a complete, high metalophid (UNSM 11759 is worn and difficult to evaluate, but probably had a similar metalophid) and a squared off entoconid (damaged on UNSM 11759). M3 resembles the subgenus Cynomys in the absence of a bridge between the talonid and ectolophid and the presence of an ante- riorly deflected hypoconid. The metalophid on M3 is complete, more developed than in C. hihhardi, but less developed than in later prairie dogs. Comments There is variation in the small Sappa sample, and two forms may be present. The holotype differs from both referred specimens in the relatively lesser width of the teeth, and from UNSM 1 1 759 in the less robust trigonid on P4 and smaller size. However, two lines of evidence suggest that all the material repre- sents one species. First, all specimens appear to 18 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 E E < Q 16 15 14 - 13 exhibit a similar level of hypso- 1 7 -\ donty, suggestive of a similar "stage of evolution." If the sample repre- sents two species from significantly different time periods, one might expect that they would differ in this respect as known later forms have greater hypsodonty. Second, both of the lower jaws exhibit char- acteristics of black-tailed prairie dogs, and it is difficult to explain the cooccurrence of two species from the same subgenus, a pattern never documented elsewhere. There are many examples in the fossil record of two prairie dog species from the same locality, but they are always from separate sub- genera. Thus, at present I believe that the within- sample variation evident in the Sappa Local Fauna probably reflects population or small-scale tempo- ral variation in the same species. 1 have tentatively placed Cynomys sappaensis in the subgenus Cynomys, based on shared features of P4 and M3. However, these shared features may be primitive for prairie dogs. Available material does ♦ ♦ O ♦ .0 I . o ♦ B n D[? O O ° °^ □ mexicanus ♦ ludovicianus n lllinoiaii black lail O Wisconsin black tails 21 22 — r~ 23 — r~ 24 — r~ 25 26 ANTJW (mm) Fig. 8. Scatter plot of individual black tails on MDALV versus ANTJW. Arrows mark two Lost Valley specimens. Referred Specimens. — See the appendix. Geologic and Geographic Range. — Late Irvingtonian(Sheridanian) and Early Rancholabrean (Late lllinoian and Sangamonian) of the central and southern Great Plains. Diagnosis. — Subgenus Cynomys; dentition av- erages slightly larger than C. mexicanus but di not preserve the derived^features of P3 and the jugal astema relatively shorter; averages much smaller evident in other black tails, thus subgeneric place- ment is made tentatively. Morphologically, this spe- cies is intemiediate between C hibhairli and Late Irvingtonian black tails, but it is not possible with present evidence to determine if this represents an evolutionary lineage. Cynomys (Cynomys) spenceri new species Figure 9 Cynomys veins: Dalquest, 1967:5. Cynomys niohrarins: Martin, 1969:30. Holotype. — UNSM 33798, skull preserving ros- trum with right and left II; palate with left P3-M3 and right P4-M3; most of skull roof; complete right zygomatic plate with jugal; much of the occiput; and right auditory bulla. Horizon and Type Locality. — Angus Local Fauna, Late Irvingtonian (Sheridanian); UNSM collection locality NO-101, 1 1/2 miles SW of An- gus, SWl/4, NEl/4, Sec. 33, T4N, R6W. Nuckolls County, Nebraska (Schultz and Tanner, 1957). than C. ludovicianus. Etymology. — Named in honor of Dr. Lee A. Spencer, whose enthusiasm for fossil mammals and earth history sparked my interests in the same. Description Skull and upper dentition. — The holotype of Cynomys spenceri is shown in lateral and ventral views in Fig. 9. Upper dental measurements are provided in Table 3; average cranial measurements for the holotype and one referred specimen are provided in Table 5 (lllinoian black tail). The holo- type skull exhibits two diagnostic black-tailed fea- tures— the well developed jugal angle and large, high ridge bordering the anterior cingulum of P3. The lllinoian black tail plotted in Fig. 7 is the holotype of Cynomys spenceri. This skull resembles C mexicanus in small size and broad postorbital region (low score on PCI) but differs from that taxon in exhibiting a relatively narrower palate, higher foramen magnum, and smaller external audi- tory meatus (low score on PC2). In these features, C. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 19 Fig. 9. Holotype of Cynomys spenceri (UNSM 33798 ) in ( A) lateral, (B) ventral views. Scale bar represents 5 mm. spenceri resembles C. ludovicianus. However, sum- mary statistics for cranial variables (Table 5) indi- cate that C. spenceri (a small sample, // = 2) differs from C. ludovicianus in that the palate is relatively less constricted posteriorly (PALM-^/PALP^ ratio much greater for C. spenceri). Upper dentitions are known from the type and several referred specimens but do not appear to differ from Cynomys mexicanus and C. ludovicianus except in size. Lower jaw and dentition. — Measurements of the lower jaw and dentition are given in Table 5 (Illinoian black tail). Cynomys spenceri averages smaller than C. ludovicianus in all variables and slightly larger than C. mexicanus in all variables except WM3 . In most respects other than size, characteristics of the lower jaw and teeth are shared with extant black tails. However, the diastema is relatively short as in Cynomys ludovicianus as opposed to the long diastema seen in C. mexicanus. This shortened di- astema causes the relatively low values for ANTJW, and thus the higher MDALV/ANTJW ratios seen in Fig. 8. Some of the "outlying" points for C. spenceri on this figure may result, in part, from slight damage to the anterior end of the lower jaw, and thus to an artificially shortened diastema. Comments Cynomys spenceri is an advanced black tail which exhibits greater morphologic similarity to C. ludovicianus than to C. mexicanus. Its temporal range probably extends from the Late Irvingtonian (Sheridanian) into the Sangamonian. Single speci- mens from the Sangamonian Mesa de Maya Local Fauna (UWYG 6032; Hager, 1975) and the Sangamon soil in Harlan County, Nebraska (UNSM 50778) are small, and probably referable to C. spenceri. However, several isolated specimens of 20 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Early Rancholabrean (Late Illinoian) age are larger than typical of C. spenceri, approximating the aver- age size of C. ludoviciamis. Because fossil data document a continued presence of small-sized C. spenceri during this interval, it is possible that at least some of these specimens are intrusives from the Late Rancholabrean. A single, isolated tooth (USNM 304236) from the Middle Irvingtonian (Cudahyan) Hall Ash Local Fauna (Eshelman and Hager, 1984) likewise seems anomalous. The specimen is probably a P4 (the roots are not preserved so this is not certain) and morpho- logically resembles black tails. It resembles Cynomys ludovicianus in being larger than typical for C. spenceri. It is not clear whether this represents a large black tail preceding C. spenceri; chance sam- pling of an extreme individual from the local popu- lation (one of seven measured P4S from the type locality of C. spenceri approximates the Hall Ash specimen in size); or a Late Rancholabrean intru- sive. Further work is needed to characterize this poorly known black tail from the Cudahyan interval. Wisconsinan bl :k tail) and Recent Cynomys ludovicianus are provided in Table 5. Most fea- tures of the skull and upper dentition are shared with other black tails and are described in the account of the subgenus. However, the species differs morphometrically from C. mexicianus in a number of respects. Fig. 7 and Table 5 suggest that C. ludovicianus is larger, but with a relatively more constricted postorbital region; exhibits a relatively narrower palate, especially between the M^s; and has a smaller external auditory meatus but higher foramen magnum (all differences con- tributing to low scores on PC2 in Fig. 7). Lower jaw and dentition. — Average mandibu- lar and lower dental measurements for fossil and Recent samples are given in Table 5, and an occlusal view of a lower dentition is shown in Fig. lOA. Again, most morphological features are shared with other black tails as described in the account of the subgenus. Like Cynomys spenceri, C. ludovici- anus differs from C. mexicanus in its relatively shorter diastema (Fig. 8). Cynomys (Cynomys) ludovicianus (Ord, 1815) Figure lOA Arctomys ludoviciana Ord, 1815. Cynomys meadensis Hibbard, 1956:172. (For a listing of synonyms in the literature of modern Cynomys ludovicianus, see Hollister, 1916:14, and Hall, 1981:411). Referred Specimens. — See the appendix. Geologic and Geographic Range. — Late Rancholabrean (Wisconsinan) of the central and southern Great Plains and across the Southwest (south of the Colorado Plateau) to southeastern Arizona; Recent of same general region and ex- tending onto the northern Great Plains. Emended Diagnosis. — Subgenus Cynomys, averaging larger than all other members of the subgenus; palate relatively more constricted pos- teriorly than in C. mexicanus. Description Skull and upper dentition. — Average cranial measurements for samples of fossil (large Comments Cynomys ludovicianus apparently arose near the beginning of the Late Rancholabrean through anagenetic change of ancestral C. spenceri. How- ever, the variation within each of these chrono- species makes it difficult to precisely delimit the temporal boundary between them. The type (UMMP 31963; Hibbard, 1956) and one referred specimen (UMMP V60532) of Cynomys meadensis from the Blancan Deer Park Local Fauna probably represent C. ludovicianus. Both specimens are high crowned and more ad- vanced than C. hibhardi from the slightly younger White Rock Local Fauna, indicating that they are almost certainly intrusive. Hibbard came to this same conclusion (written communication to L. D. Martin). The type preserves only Mj-Mt, but UMMP V60532 preserves M3 which exhibits a strongly deflected hypoconid, typical of black tails. The talonid platform lies adjacent to the ectolophid on this tooth, blocking the basin trench as in white tails. However, no distinct bridge is formed, and a similar conformation is sometimes seen in black tails. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 21 Fig. 10. Occlusal views of lower dentition, all stereophotos, for (A) Cynomys ludovicianus (UMMP 3 1 759), (B) C. gunnisoni (TMM 4 1 228-1 69 1 ), and (C) C. niobrarius (KUVP 55595). Scale bars represent 5 mm; bar under (C) also applies to (A). Hibbard (1956) diagnosed this fossil species specimens approximate Cynomys ludovicianus based on the presence of a "small, round conulid, instead of a transverse mesolophid, on the lingual side of the talonid basin" of M i and M2. However, a similarly reduced "mesolophid" is encountered in other black tails. In size, both of the Deer Park more closely than C. spenceri, thus the assignment to the former taxon. A single M3 from the mid-Wisconsinan Craigmile locality. Mills County, Iowa deserves comment. The locality is dated at 23,000 yr B.P. 22 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 (Rhodes, 1984) and thus represents the mid- Wisconsinan. This specimen clearly exhibits black tail morphology but is extremely small (LM3 = 4. 16, WM3 = 4.32; compare with Table 5 ), smaller than any other fossil or Recent specimen of Cynomys ludovicianus that I measured, approach- ing the size of some specimens of C. mexicanus. Fossils from other horizons or localities from the Great Plains thought to be of similar age (sites in Wilson County, Kansas; Citellus zone of southern Nebraska; Burnham Site, Wood County, Okla- homa; several sites in Denton County, Texas; see the appendix) approach average size for fossil C. ludovicianus, thus general size decrease during the mid-Wisconsinan is not indicated. This local- ity probably approximated the eastern mid- Wisconsinan range boundary for the species, and the Craigmile population may have been coloniz- ing marginal habitat. Cynomys (Cynomys) cf. Cynomys mexicanus Merriam, 1892. Referred Specimens. — See the appendix. Comments As noted in the morphometric analysis of black tails. Recent Cynomys mexicanus differs from fos- sil and Recent C. ludovicianus in skull size and shape (Fig. 7), much smaller size of the lower alveolar row and dentition (Table 5), and a rela- tively low MDALV/ANTJW ratio (Fig. 8) result- ing from the elongate diastema. Cynomys mexicanus has not been reported as a fossil from its present range in northeastern Mexico (Ceballos- G. and Wilson, 1985). Alvarez (1983) recently reported fossil C. ludovicianus from Mexico, and the possibility that these fossils are related to C. mexicanus needs investigation. I tentatively refer the mid-Wisconsinan fossil black tails from Lost Valley, Eddy County, New Mexico (Harris, 1987) to this species based on overall similarity in size (much smaller than Cynomys ludovicianus) and on a similarly elon- gate diastema resulting in a low MDALV/ANTJW ratio (Fig. 8, marked by arrows). The fossils ap- pear to be more robust in general proportions, with relatively deeper lower jaws, than is typical of C. mexicanus, but in this character they fall within the range of variation exhibited by Recent speci- mens. If these fossils do represent Cynomys mexicanus, the record is of considerable interest. Southeastern New Mexico is well north of the present range of the species, indicating a more extensive range than that of today. Additionally, the record supports an origin of C mexicanus prior to the mid-Wiscon- sinan (about 30,000 yr B.R). Other lines of evi- dence support a close relationship between C mexicanus and C. ludovicanus (Pizzimenti, 1975; McCullough and Chesser, 1987; McCullough et al., 1987). Cynomys mexicanus may be a Pleis- tocene "relict," separated from the main range of black tails by events during the Pleistocene (Hoffmann and Jones, 1970). Genetic distance has been interpreted as suggesting separation of these species about 42,000 yr B.P. (McCullough and Chesser, 1987). Thus, the Lost Valley record may represent the early history of C mexicanus. This might explain the more robust lower jaws (resem- bling C. ludovicianus) than typical of C. mexi- canus— the latter has subsequently diverged in this character from the ancestral morphotype. Subgenus Leucocwssuwmys Hollister, 1916 Figures 5B, 5D, 5F, 6B, 6D, 6F Type Species. — Cynomys gunnisoni (Baird, 1855). Emended Diagnosis. — Cynomys, distinguished from all other prairie dogs by presence of a mod- erately to well developed bridge connecting the ectolophid and talonid on M3 of most specimens, reduction or loss of the strong anterior deflection of the hypoconid on M3, broadened trigonid and reduced hypoconid on P4; further distinguished from subgenus Cynomys by lesser development of jugal angle and of ridge bordering anterior cingu- lum on P-^. Geologic and Geographic Range. — Probably Early Irvingtonian (Sappan) to Recent; fossil forms distributed over the northern and central Great Plains and Rocky Mountain region; Recent forms restricted to the latter. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 23 Description Skull and upper dentition. — Some distinctive aspects of the skull and upper dentition are illus- trated in Fig. 5 (B, D, F) and may be compared with equivalent features in the subgenus Cynomys (Fig. 5A, C, E). Viewed dorsally, the angle be- tween the lateral margins of the rostrum and the anterodorsal margin of the zygomatic plate is typi- cally less abrupt than in the subgenus Cynomys, but this character is variable. Viewed laterally, the jugal lacks the prominent, downward-pointing angle (Fig. 5B) exhibited by black tails (Fig. 5A). In posterior view, the occipital plate frequently exhibits a relatively flattened dorsal margin (Fig. 5D). The ridge bounding the anterior cingulum on P3 is typically small. In buccal view, this structure usually is terminated by a distinct notch well below the apex of the paracone (Fig. 5F). In this respect, Leucocrossuromys resembles advanced ground squirrels such as Spermophilus richardsonii but differs from the subgenus Cynomys which exhibits a large, well developed cingular ridge (Fig. 5E). In other respects the upper dentition is similar to that of the subgenus Cynomys. Lower jaw and dentition. — Several features of the lower jaw and dentition are shown in Fig. 6 (B, D, F) and contrasted with black tails (A, C, E). Typically, the portion of the lower jaw beneath the diastema is less robust than in black tails. As a result, the dorsal margin of this region of the jaw, in lateral view, tends to drop abruptly from the anterior margin of the tooth row. The lower denti- tion exhibits several distinctive characters. On average, P4-M3 are relatively wider buccolingually across their trigonids than on black tails. P4 is especially distinctive because the wide trigonid frequently contrasts with a narrow talonid, the latter resulting from a reduced hypoconid (Fig. 6D). This conformation of P4 is not typical for other prairie dogs. M1-M2, especially M2, usually exhibit a mod- erate to large mesolophid which traverses the talonid basin buccolingually, sometimes com- pletely dividing it into anterior and posterior por- tions (Dalquest, 1988). However, this structure wears rapidly and is not visible on many speci- mens. M3 bears several subgeneric characters (Fig. 6F) as noted in the diagnosis. The bridge between the ectolophid and talonid, and the reduction or loss of the anterior deflection of the hypoconid, are derived characters not found in other prairie dogs. The hypoflexid, between the protoconid and hypoconid, typically bears one or two ectostylids positioned externally, internally, or sometimes in both positions. These structures are only occasion- ally present in black tails. The talonid platform is often bounded anteriorly by a ridge coursing linguad from the point of contact between the ectolophid and talonid. MORPHOMETRIC RELATIONSHIPS AmONG FoSSIL AND Recent White-tailed Prairie Dogs Middle Irvingtonian (Cudahyan) and Late Rancholabrean fossils from southern Colorado and New Mexico resemble Cynomys gunnisoni, present in that region today, in small size and in several qualitative features. Late Irvingtonian (Sheridanian ) through Late Rancholabrean fossils from the central and northern Great Plains are distinctly larger than other fossil or Recent white tails, indicating the presence of at least one extinct fossil species. How- ever, three species of large, white tail fossils have been described, and taxonomic relationships among them need investigation. Cynomys niohrarius, de- scribed by Hay (1921) from a damaged, somewhat distorted skull (AMNH 2715), was recovered from Pleistocene beds in northwestern Nebraska. The fossil was collected from the vicinity of the type section for the Sheridanian (Late Irvingtonian), and the fossil may be of equivalent age; however, this is not certain. Cynomys spispiza, described by Green (1960) from a partial lower jaw with P4-M3 (SDSM 57100), was recovered from Late Rancholabrean deposits in southcentral South Dakota. Because of the geographic proximity of type localities, the two named forms probably do not represent contempo- rary species. However, they might represent sepa- rate chronospecies. Cynomys churcherii was de- scribed from a large sample of cranial and post cranial fossils (the type, PMA P85.9.12, includes most of an entire skeleton) from the northwestern Plains of southern Alberta (Bums and McGillivray, 1989). The fossil sample is Late Rancholabrean 24 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 9. Summary statistics for cranial variables of three Recent and two putative fossil forms of the subgenus Leucocwssuromys. Sample size is given at the head of each column and is only repeated when it changes. Names of fossil OTUs correspond to putative taxa recognixed in previous studies. Cynomys Cynomys Cynomys Cynomys Cynomys niobrariusi gunnisoni par videns leucurus churcherii spispiza Variable X (sd) n X (sd) /; X (sd) n X (sd) n X (sd) n PALM-'' 5.45 (0.36) 11 5.51 (0.33) 5 5.47 (0.35) 15 5.06 (0.39) 15 5.17 (0.54) 9 PALP-"* 9.78 (0.51) 9.67 (0.87) 9.65 (0.81) 10.90 (0.32) 14 10.41 (0.56) 7 MXALV 15.03 (0.32) 15.35 (0.74) 15.67 (0.68) 16.41 (0.34) 15 16.28 (0.81) 9 PALLN 32.53 (0.82) 33.03 (1.10) 33.66 (0.97) 14 36.00 (0.88) 13 34.62 (1.60) 5 SKLN 52.34 (1.70) 53.06 (2.20) 54.46 (1.70) 57.63 (1.40) 12 56.03 (2.50) 3 INTOR 11.65 (0.67) 12.60 (1.30) 12.32 (0.54) 12.84 (0.60) 13 12.52 (0.68) 7 PSTOR 13.84 (0.91) 14.53 (0.42) 13.62 (0.72) 15 13.30 (0.77) 12 12.53 (0.99) 6 OCCPW 27.17 (0.55) 28.37 (1.20) 28.08 (0.80) 14 29.46 (0.74) 9 28.43 (1.50) 5 OCCPH 16.37 (0.55) 16.08 (0.58) 16.75 (0.60) 15 17.41 (0.82) 11 16.91 (0.54) FORMW 8.05 (0.32) 8.60 (0.16) 8.73 (0.28) 8.86 (0.53) 8.27 (0.32) 3 FORMH 6.20 (0.60) 5.91 (0.38) 7.17 (0.34) 6.53 (0.40) 6.55 (0.61) EAML 3.49 (0.28) 10 3.56 (0.51) 3.96 (0.38) 3.83 (0.40) 7 3.79 (0.37) 5 (22,000 to 33,000 yr B.P.) and might represent a geographically distinct contemporary of C spispiza. To assess taxonomic relationships, I investigated morphometric relationships among Recent species {Cynomys gunnisoni , C. leucurus, C.parvidens) and the following putative fossil taxa: C. cf. gunnisoni (small white tails from the southern Rockies), C. niohrarius (herein including large Sheridanian and Early Rancholabrean white tails), C. spispiza (large Late Rancholabrean white tails south and east of southern Alberta), and C. churcherii (large Late Rancholabrean white tails from southern Alberta). Cranial analysis. — Mean values for 12 cranial variables are given in Table 9. Cynomys niobrarius and C. spispiza are combined because of small sample sizes, and fossil C cf. gunnisoni is not included because appropriate material was unavail- able. Five cranial variables are used in a PCA, and the correlations of these variables with PC 1 and PC2 are given in Table 10. The pattern of correlations suggests that PCI primarily represents skull size as reflected by variables PALP3, MXALV, and INTOR, and secondarily reflects a contrast between these variables and PALM3 and PSTOR. Specimens with high scores on PCI tend to be large, but with relatively constricted posterior palates (PALM^) and postorbital regions (PSTOR). PC2 reflects width of the skull roof (INTOR, PSTOR) and of the posterior palate (PALM-"*). Figure 1 1 A plots bivariate means of fossil and Recent skulls on PCI and PC2. Statistical compari- sons of mean values of PCI and PC2 are given in Table 1 1 . There are no differences among Recent or between represented fossil OTUs in PCI (p > .1 in all cases), but both fossil samples differ significantly (p < .0 1 ) from the three Recent forms in this variable. High values of PC 1 for fossil forms reflect large size contrasting with relatively narrow posterior palates and postorbital regions, a pattern also detectable from inspection of Table 9. There are no significant differences among white tails on PC2, but the differ- ence between Cynomys niobrariusi spispiza and C. parvidens approached statistical significance (.05 < p<.l). Mandibular/lower dental analysis. — Mean values for 10 mandibular/lower dental variables are shown in Table 1 2. Seven variables (measurements of P4-M2) were used in a PCA, and the correlations of original variables with PCI and PC2 are given in Table 10. Of the total variation in the original set of seven variables, 78% was accounted for by PC 1 , and this clearly represents general size. PC2 represents a shape axis, reflecting contrast between two mea- sures of length (LMj, LM2; LP4 was not highly SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 25 Table 10. Correlations (loadings) of five original cra- nial and seven original lower dental variables with PC 1 and PC2. Two analyses are represented, one on cranial and the other on lower dental variables. Samples include Recent and fossil white tails. Variable PCI PALM^ -0.39 PALP^^ 0.81 MXALV 0.73 INTOR 0.63 PSTOR -0.49 % total variance explained 39 LP4 0.90 WTRP4 0.91 WTLP4 0.87 LM, 0.84 WM, 0.90 LMt 0.84 WM2 0.90 % total variance explained 78 PC2 0.76 0.26 -0.03 0.68 0.64 30 0.17 -0.33 -0.29 0.49 -0.23 0.50 -0.25 12 Taxonomic implications. — Thiee taxonomic judgments seem warranted. First, fossils referred to Cynomys gunnisoni do represent that species. Sec- ond, large fossils from the northern and central Plains are distinct from all extant white tails, and this 1.5 1.0 CM O Q. 0 - -1.0 -1.5 PV LC CH GN NB/SP T -1.0 T -.5 0 PCI 1.0 1.5 correlated with this axis) and all four measures of width. Specimens with high scores on PC2 tend to have relatively long teeth (especially Mj-Mt) com- pared to tooth width. Figure IIB plots bivariate means of fossil and Recent samples on PCI and PC2. Statistical com- parisons of mean values for MDALV, PC 1 , and PC2 are given in Table 1 1 . Variation among three Recent white tails is not significant in any comparison. This reflects morphologic similarity among Recent spe- cies, but with larger sample sizes some significant differences probably would be detected. Fossils referred to Cynomys gunnisoni do not differ from extant C gunnisoni in any comparison. The three large fossil forms differ significantly (p < .01 ) from all Recent samples and fossil C. cf. gunnisoni in comparisons of size (MDALV, PCI). There are no differences among the three, large fossil forms in comparisons of size (p > . 1 ). However, C. churchehi is significantly (p < .01 ) different from C. niohrarius and C spispiza in comparisons of shape (PC2). No significant differences are evident between C. niohrarius and C. spispiza. .8 - CM O 0 Q. -.4 -.8 PV FG GN NB + LC sp I CH I B -1 0 Fig. 1 1 . Scatter plots of means of Recent and fossil white tail OTUs on PCI and PC2. (A) PCs derived from analysis of five cranial variables; (B) PCs derived from analysis of seven lower dental variables. Bars represent standard errors of means. OTU abbreviations are identi- fied in Table 1 1 . Sample sizes for: (A) CH, n = 1 1; GN, n = ll;LC,n= 14;NB/SRn = 6;PV,n = 6.(B)CH,n = 21; FG,n-6;GN, n= 15;LC, n= 19: NB, n = 24; PV, n - 6; SPn = 21. 26 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 1 1 . Comparisons of means among Recent and putative fossil white-tailed prairie dogs. Means connected by a line are not statistically different. OTU abbreviations used are: CH, C. chuirherii; FG, fossils referred to C. gunnisoni: FW. fossil white tails (combined C. niohrarius and C . spispiza); GN. C. gunuisoni: LC, C. leuciinis: NB. C. niohrarius: PV, C.parvideiis: SP, C. spispiza. Names of fossil OTUs correspond to putative taxa recognized in previous studies. Variable Taxon and mean GN PV LC FW CH PCI: Cranial -1.30 -0.89 -0.42 1.20 1.50 FW GN CH LC PV PC2: Cranial FG GN -1.00 -0.15 0.06 0.18 0.94 PV LC CH NB SP MDALV 13.30 GN 13.50 13.70 13.90 14.60 CH 14.80 NB 14.90 FG PV LC SP PCI: P4-M2 -2.90 -2.60 -2.20 -2.00 1.10 1.20 2.00 CH GN FG LC SP PV NB PC2: P4-M2 -0.78 -0.18 0.11 0.11 0.29 0.44 0.47 should be recognized taxonomically. Third, the simi- larity among all three samples of large fossil white tails probably indicates that they are conspecific (C. niohrarius). Shape differences between the north- em C. churcherii and southern C. niohrarius and C. spispiza appear to be meaningful, but can be inter- preted as geographic variation within a single, wide- spread species, perhaps representing subspecific differentiation. Cynomys {Leucocrossuromys) gunnisoni (Baird, 1855) Figure lOB Spermophilus gunnisoni Baird, 1855. (For a listing of synonyms in the literature of modern Cynomys gunnisoni, see HoUister, 1916:29 and Hail, 1981:414-415). Referred Specimen.s. — See the appendix. Geologic and Geographic Range. — ?Middle Irvingtonian (Cudahyan) to Recent of the southern Rocky Mountains and adjacent highlands. Emended Diagnosis. — Subgenus Leucocrossur- omys, averaging smaller than other white tails; distin- guished by less consistent development of bridge con- necting ectolophid and talonid on M3 and lesser reduction of the anterior deflection of hypoconid on M3. Description Skull and upper dentition. — Average cranial measurements for a sample of Recent Cynomys gunnisoni are given in Table 9. Most features of the skull and upper dentition are common to Leuco- crossuromys. However, skulls of C gunnisoni differ in three respects. First, the jugal angle, although less developed than in black tails, is often more devel- oped than in other white tails. Second, the dorsal margin of the occiput usually is less flattened than in other white tails (Fig. 5D), although it is not typi- cally as domed as in black tails (Fig. 5C). Third, the posterodorsal margins of the premaxillae frequently extend posteriorly beyond the nasals on the skull roof, more so than is typical for other white tails. The upper dentition is characteristic of white tails. Lower jaw and dentition. — Average mandibular and lower dental measurements for fossil and Recent samples are given in Table 12. Most morphological features are shared with other white tails as described in the account of the subgenus. However, Cynomys gunnisoni Q\\\\h\is less consistent development of the bridge between the ectolophid and talonid and lesser reduction of the anterior deflection of the hypoconid on M3. The latter feature can be seen in Fig. lOB, in contrast to Fig. IOC. Some specimens approach the state exhibited by black tails. SYSTEMATIC REVISION OF EOSSIL PRAIRIE DOGS 27 it 1 Dh c ! 1 73 o ^ C ^ OX) >-J o y5 o 3 C I) 03 on X Xi 03 3 > (1) J= o3 ■*— • ^l-H 3 o D- C/0 O 13 O C t^ O 1 D- on O c/3 ^1-. O aj o > C/5 D c^ H 3 O i^ 3 <2 •n ^M c o r3 «) 4—* (U » ^ s r- c r s: :>r 0 So T3 T3 1/5 C/5 l\ GO C/5 1^ X .2 00 ON in ^ Oi 3 ^ ON — (-<~i — r<~) o o o o o o o o O o -^Ol'^OIt^ONOOiriONON ON r-J lO (N — O r<-^ r<-5 ON — ; (N .^ On 00 ? rT n o o o O O o o o o o o o VOOOO — OOOOONONOOO lO o \q — ; ON — > — _ en ON — < ■rt r«-i Tt Tt (N in ro iri Tt lo CO m in n-1 m — ' o fN ■? 00 ^O o ^ in o o o O o O o o o o CNOVOmOOOOc^iONC^I 00 (N rt — — ON (N — ; ON O ■^rn— 'ooinoq r<^ (N r<-i r<-i fN ^' r<-] Tt '^' '^' vo in o o^ o t^ o ^ ON r-1 — ^ fN o — — -- -^ o o o o o o o o o o ooono— '^ONor--ONoo '*t~-;Or-~;r-;^ON\oinin r<-i (N ■^ rn (N ■<+' J _, <"1 ^ l"^ 28 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Comments Cynomys gimnisoni clearly is a member of the subgenus Leiicocrossiiromys as indicated by several shared, derived characters. However, the species differs in the degree to which some characters are developed. In several cases (shape of the occiput, states of the bridge and hypoconid on M3). C. gimnisoni probably is primitive for white tails. This is consistent with previous interpretations that the species is primitive in aspects of its biology (Nadler etal., 1971;Pizzimenti. 1975). Fossils referred to Cynomys gimnisoni resemble Recent specimens in most respects, indicating little change in this lineage since the Pleistocene. How- ever, fossils appear to differ in the relatively smaller (especially shorter) M3 (Table 12). The meaning of this difference is not clear, but it does not in itself warrant the erection of a new taxon. The record of Cynomys gimnisoni from the Hansen Bluff Local Fauna, Alamosa County, Colo- rado (Rogers et al., 1985), suggests a history back to the Cudahyan. The Hansen Bluff fossils are indistin- guishable from Late Rancholabrean C. gimnisoni. and possible intrusiveness must be considered. How- ever, there was no field evidence that the prairie dog fossils were out of context (Rogers, pers. comm.). Cynomys {Leucocrossuwmys) niohrarius Hay, 1921 Figure IOC Cynomys niohrarius Hay, 1921:615. Cynomys spispiza Green, 1960:545. Cynomys cf. leucurus: McDonald and Anderson, 1975:25. Cynomys churcherii Burns and McGillivray, 1989:2637. Holotype. — AMNH 2715, a partial, somewhat distorted skull preserving the palate with some teeth. Horizon and Type Locality. — Late Pleistocene, possibly Sheridanian; locality given only as "Niobrara River, near Grayson, Nebraska" (Hay, 1921). Referred Specimens. — See the appendix. (ieologic and Geographic Range. — Late Irvingtonian (Sheridanian) to Late Rancholabrean of the northern and central Great Plains and adjacent portions of the central Rocky Mountains. Emended Dia ;nosis. — SubgenusL^/zcorra^'^///-- omys, much larger than all other members of subge- nus; like Cynomys leucurus and C. parvidens, con- sistently exhibits well developed bridge between ectolophid and talonid of M3 and extreme reduction or loss of anterior deflection of hypoconid on M3. Description Skull and upper dentition. — Average cranial measurements for two samples of Cynomys niohrarius are given in Table 9 (C. niohrariuslC. spispiza, C. churcherii). The most distinctive fea- ture of this taxon is its large size. In addition, its skull differs from Recent species in shape. Based on the interpretation of PCI presented previously, C. niohrarius appears to have a relatively narrower posterior palate and postorbital region, a pattern evident with inspection of Table 9. Most qualitative features of the skull and upper dentition are shared with other white tails and are described in the account of the subgenus. Cynomys niohrarius resembles C. leucurus and C. parvidens. but differs from C. gunnisoni in that it typically exhibits a reduced jugal triangle and a more flattened dorsal profile of the occipital plate. Lower jaw and dentition. — Average mandibu- lar and lower dental measurements for three samples oi Cynomys niohrarius are provided in Table 1 2 (C. niohrarius. C. spispiza. C. churcherii). Other than size, C niohrarius differs in fev/ respects from C. leucurus and C. parvidens. However, it differs from C. gunnisoni in the more developed bridge between the ectolophid and talonid on M3, and in the greater reduction of the anterior deflection of the hypoconid on M3. The latter feature can be seen in Fig. IOC. Comments All three fossil species synonomized here clearly are members of the subgenus Leucocrossuromys. Hay (1921) did not formally assign Cynomys niohrarius to either subgenus, but he pointed out several similarities between the type and C. leucurus. However, the characters he used were not diagnos- tic. Dalquest (1967) considered this species to be related to the subgenus Cynomys but presented no supporting evidence. Fortunately, the unerupted P^ can be seen through the opening left by the dP\ and SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 29 the conformation of the ridge bounding the anterior cingulum is that of Leucocrossuromys. Likewise. Green (1960) did not fonnally assign C. spispiza to either subgenus but suggested that it might be closer to C. leucwus than to C. ludovicianus. All preserved features of the lower dentition support white tail affinities. This taxon initially was thought to be from the Tertiary (Green. 1960) but subsequently was found to be Late Pleistocene in age (Green. 1963). Cynomys chiircherii was assigned to Leucocrossur- omys in the original description (Burns and McGillivray, 1989). an assignment abundantly sup- ported by cranial and dental evidence. I decided to synonomize these three fossil forms at the species level based on two lines of evidence. First, the types and referred fossil samples resemble each other in general size and morphology and appear to be within the range of variation expected of a single species. Second, all three forms occupied the same general geographic region during the Late Pleistocene, namely, the northern Great Plains. The one example of significant variation within Cynomys niohrarius is in dental shape as reflected in PC2 derived from the analysis of P4-M2 (Table 9). Cynomys churcherii has significantly wider teeth relative to tooth length (large, negative values on PC2 ) than the southern C. niobrarius and C spispiza. Thus, the northern sample appears to represent a distinct geographic form, divergent from southern populations of C. niobrarius. I have followed stan- dard taxonomic practice and recognize the northern and southern morphs as separate subspecies. Cynomys niobrarius niobrarius Hay, 1921 Cynomys niobrarius Hay, 1921:615. Cynomys spispiza Green, 1960:545. Cynomys cf. leucwus: McDonald and Anderson. 1975:25. Holotype. — As for species. Horizon and Type Locality. — As for species. Referred Specimens. — See the appendix. Geologic and Geographic Range. — Known from the Sheridanian through Late Rancholabrean of the central and northern Great Plains and adjacent areas approximately to the present northern bound- ary of the United States. Emended Diagnosis. — Cynomys niohrarius, differing from C. niobrarius churcherii in relatively narrower lower cheek teeth. Cynomys niobrarius churcherii Burns and McGillivray, 1989 Cynomys churcherii Burns and McGillivray, 1989:2637. Holotype.— PMA R85.9.12, a virtually com- plete skeleton with skull, lower jaws, and complete dentition. Horizon and Type Locality. — From burrows, dated at 22,000 to 33,000 yr B.P, intruding the Miocene Hand Hills Fomiation; "Winter site gravel pit, in the Hand Hills, 28 km NE of Drumheller, Municipal District of Starland, Alberta" (Bums and McGillivray, 1989). Referred Specimens. — See the appendix. Geographic and Geologic Range. — Known only from the mid-Wisconsinan of the Hand and Wintering Hills, southern Alberta. Emended Diagnosis. — Cynomys niobrarius, differing from C. niobrarius niobrarius in the rela- tively wider lower cheek teeth. Comments The position of the boundary between Cynomys niobrarius churcherii and C. niobrarius niolvarius is not clearly defined. Cynomys niobrarius has been recovered from deposits of Sheridanian, Early Rancholabrean (Sangamonian), and Late Rancho- labrean age in the Medicine Hat sequence of south- eastern Alberta (Stalker and Churcher, 1970). Un- fortunately, only two of these specimens are suffi- ciently complete for use in the PCAofP4-M'). These resemble the southern form more than C niobrarius churcherii in tooth shape, but the sample is too small to allow for statistical comparisons. Cynomys sp. I was unable to study reported prairie dog fossils from three important localities. These fossils either could not be found or were under study by other investigators. Stalker and Churcher (1970) listed Cynomys cf. meadensis as part of the Wellsch Valley Local Fauna 30 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 from southern Saskatchewan. Specific identifica- tion almost certainly was based on the age of these fossils — C. meadensis once was thought to be very old(Hibbard, 1956) — andtheWellsch Valley speci- mens probably have nothing to do with C. meadensis (= C. liidovicianus, as synonomized in this paper). However, if these fossils represent the genus Cynomys (which is not certain), they would be of great interest because of the age (probably Sappan, possibly younger; Churcher, pers. comm.) and geographic position of the locality. Semken (1966) reported Cynomys cf. gunnisoni from the Kentuck Local Fauna, McPherson County, Kansas, a fauna generally thought to be Sappan in age. He reported the presence of a bridge connecting the ectolophid and talonid on M3, thus these fossils probably represent an early member of the subgenus Leucocrosswomys. The record is of considerable interest because of the age of the fauna. Barnosky and Rasmussen (1988) listed two un- named species of Cynomys from Porcupine Cave, Park County, Colorado. The site is significant be- cause of its age (near 400 ky B.P. ) and location (high elevation — 2900 m — in the central Rocky Moun- tains). These fossils may shed light on regional prairie dog biogeography and evolution. In addition to these faunas, three specimens that I have examined deserve comment. A right lower jaw with P4-M3 (FHSU VP-6931 ) recovered from the Williams Farm locality. Rice County, Kansas, is thought to be "Illinoian" in age (Holman, 1984). The teeth are somewhat worn, but the shape of P4 and presence of a bridge connecting the ectolophid and talonid on M3 appear to support white tail affinities. The fossil, however, exhibits greater anterior deflec- tion of the hypoconid on M3 and smaller overall size than typical oi Cynomys niohrarius. The conforma- tion of the M3 hypoconid resembles C. gunnisoni, but the fossil is larger than typical of that species. This fossil may simply reflect normal variability in Illinoian C. niohrarius. Alternatively, it may repre- sent a somewhat earlier stage in the evolution of white tails on the Great Plains, in which case the fauna probably predates the Illinoian. A left lower jaw with P4-M3 (FHSU VP-7065) was recovered from Harper 2 1 C, McPherson County, Kansas, a locality of uncertain but possibly Early Rancholabrean age. In contrast to the Williams Farm specimen, this fossil exhibits black tail charac- teristics of the P4 and the hypoconid on M3, suggest- ing affinities with the subgenus Cynomys. However, the M3 bridge is extremely well developed, more so than in any other black tail that I have examined. Thus, the taxonomic placement of this specimen remains uncertain. A left lower jaw with P4-M3 (UWYG 3392) was recovered from Chimney Rock Animal Trap, a mixed Late Pleistocene/Holocene locality in Larimer County, northcentral Colorado (Hager, 1972). The specimen clearly represents a white tail, but is smaller than any specimen of Cynomys niohrarius I have examined (MDALV, 13.6; compare with Table 12). However, UWYG 3392 resembles all three Recent white tails. Most likely, this fossil represents C. leucurus, present locally today. It is almost certainly Holocene in age. PHYLOGENETIC RELATIONSHIPS Hypotheses of relationships among Recent and fossil species of Cynomys are given in Fig. 12. Analysis of 19 characters (Table 13) using PAUP resulted in three equally parsimonious trees, shown in Figs. 1 2 A-C. Each of these trees requires 27 steps and has a consistency index of 0.963. These three phylogenetic hypotheses are very similar, differing only in the positions oilCynomys vetus and C. hihhardi relative to each other. There are many missing characters for these two species (Table 13), making it difficult to resolve their rela- tionships. Two trichotomies are evident in all trees (C. spenceri-ludovicianus-me.xicanus and C. niohrarius-leucurus- parvidens), reflecting identi- cal sets of character states among species in each triad (Table 13). An alternative phylogenetic hypothesis, which I currently favor, combines character and stratigraphic data (Fig. 1 2D ). Cynomys spenceri and C. niohrarius are considered ancestral in their respective lineages based on stratigraphic occurrence and appropriate ancestral morphology. Cynomys sappaensis is con- sidered ancestral to later black tails, not the sister group of all advanced black tails and white tails as suggested by Figs. 12 A-C. This relationship is suggested by the probable presence of a white tail in deposits of similar age at the Kentuck locality (Semken, 1966), suggesting that the split between subgenera had already occurred. Not suprisingly, ?C. vetus and C. hihhardi, which are least derived morphologically, are also the oldest species of prai- rie dogs currently known. Available evidence sug- gests that ?C. vetus is somewhat more primitive than C. hihhardi. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 31 The relationships among Recent species sug- gested in Fig. 12 are consistent with other hnes of evidence. Close relationships have been suggested between Cynomys leiicwus and C. parvideiis based on karyotype and other features (Pizzimenti and Nadler, 1972), and between C. mexicamis and C. ludovicianus based on genie data (McCullough and Chesser, 1987). The monophyletic relationship of C. gimnisoni to other white tails is consistent with immunologic evidence (McCullough et al., 1987). B TO v> JD 3 J3 qj '-£ ;i c 0) ctl §: TO U c a <0 C TO U S 3 <0 C TO E ^ < S 05 c Ul C o 3 O U) vf T3 £ c 3 3 U 3 s s niobrarius Fig. 12. (A-C) Hypotheses of relationships among fossil and Recent species of Cynomys obtained by an analysis of cranial, mandibular, and dental characters (see Table 1 3 and text). Derived characters are given below for each tree: if no state is indicated, the transition is from state 0 to 1. (A) a: 2(1), 5, 7, 11, 13, 14(1), 15, 17(1): b: 12(0): c: 3, 6, 9: d: 8( 1), 17(2): e: 8(2): f: 2(2), 10: g: 4(1), 14(2), 16, 18(1), 19(1): h: 1: i: 2(0), 4(2), 18(2), 19(2). (B)a-b: same as tree A, nodes a-b:c: 3,6, 8(1), 9, 17(2): d-h: same as tree A, corresponding nodes e-i. (C)a: same as tree A, node a: b: 8(1), 17(2): c: 12(0); d: 3, 6, 9: e-i: same as tree A. nodes e-i. (D) Hypothesis which incorporates stratigraphic position of taxa. SUMMARY Eight species of prairie dogs, genus Cynomys. are recognized in the fossil record. ICynomys vetus and C. hihhardi are early (Late Blancan and/or Early Irvingtonian), primitive forms of uncertain subge- nus. Cynomys sappaensis, described herein from the Early Irvingtonian of Nebraska, probably is a primi- tive member of the subgenus Cynomys. Another new species, C. spenceri, is known from the Late Irvingtonian and Early Rancholabrean of the Great Plains. It is a small but advanced member of the subgenus Cynomys probably ancestral to C. ludovicianus. known from the Late Rancholabrean and Holocene. A small, mid-Wisconsinan prairie dog, known only from southeastern New Mexico, is tentatively referred to the extant C. mexicanus. Cynomys gunnisoni, known from the Middle Irvingtonian to Holocene, and C niobrarius. known from the Late Irvingtonian and Rancholabrean, are fossil representatives of the subgenus Lf'ncocrai'^wr- omys. The latter underwent geographic differentia- tion into northern (C. n. churcherii) and southern (C. //. niobrarius) subspecies. Two extant species, C parvidens and C. leucurus. are derived from C. niobrarius and have no pre-Holocene fossil record. 32 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Table 13. Data matrix showing distribution of 19 cranial and dental characters, and brief descriptions of states for each character. Missing characters are given as "9." Characters are as follows: 1 = On skull roof, posterior margins of premaxillae usually project (0) to the level of, ( 1 ) well posterior to the posterior margins of the nasals; 2 = Jugal angle (0) flattened, (1) weakly triangular, (2) strongly triangular; 3 = Viewed anterodorsally, anterior margin of squamosal root (0) concave, (1) flattened; 4 = Dorsal margin of occipital plate usually (0) domed, (1) moderately flattened, (2) strongly flattened; 5 = Ventral wall of infraorbital foramen usually (0) thin and horizontal, (1) robust and inclined lateroventrad from its medial end; 6 = Zygomatic plate, anterior view, (0) weakly, (1) strongly concave; 7 = Tooth rows (0) subparallel, ( 1 ) strongly convergent posteriorly; 8 = Hypsodonty of cheek teeth (0) intermediate. (1) high, (2) very high; 9 = Occlusal outline of P-^ (0) circular, (1) anteriorly flattened; 10 = Ridge bounding anterior cingulum on P^ (0) low and incomplete, ( 1 ) well developed and complete; 1 1 = Labial portion of protoloph on M ^ -M-"* (0) lacks, ( I ) usually exhibits strong posterior expansion; 1 2 = M-^ (0) weakly, ( 1 ) strongly elongate relative to M-^; 1 3 = Portion of lower jaw beneath the diastema (0) slender and long, ( 1 ) relatively short and moderately or very deep; 14 = Lower cheek teeth (0) narrow. ( 1 ) moderately wide, (2) extremely wide; 15 = p^ protolophid (0) incomplete, ( 1 ) well developed and complete; 16 = P4 hypoconid (0) large and expanded labially, ( 1 ) reduced and not strongly expanded labially; 1 7 = M3 metalophid (0) incomplete, ( 1 ) complete but low, (2) complete and high; 18 = Anterior deflection of M3 hypoconid (0) well developed, (1 ) reduced but usually present, (2) absent or very weak; 19 = M3 bridge between ectolophid and talonid platform (0) absent, ( 1 ) sometimes present and variably developed, (2) consistently present and well developed. Characters Taxa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Hypo, ancestor 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 9 ICynomys vetus 0 1 0 9 1 0 9 0 0 9 0 9 9 9 9 9 9 9 C. hibhardi 9 9 9 9 9 9 0 9 9 9 0 1 0 0 C. sappaensis 9 9 9 9 9 1 9 9 0 2 0 0 C. spenceri 0 2 0 2 1 0 2 0 0 C. ludovicianus 0 2 0 2 1 0 2 0 0 C. mexicanus 0 2 0 0 1 0 2 0 0 C. gunnisoni 1 1 1 T 0 2 1 2 1 1 C. niohrariiis 0 0 2 2 0 2 1 2 2 2 C. leucurus 0 0 0 9 0 0 1 2 2 2 C. parvidens 0 0 2 2 0 2 1 2 2 2 LITERATURE CITED Alvarez, T. 1 983. Notas sobre algunos roedores fosiles del Pleistocenoen Mexico. Esc. Nac. Cienc. Biol., Mexico 27:149-163. Anderson, E. 1968. Fauna of the Little Box Elder Cave, Converse County, Wyoming. Univ. Colorado Stud., Earth Sci. Sen 6:1-59. Barnosky, a. D. and D. L. Rasmussen. 1988. Middle Pleistocene arvicoline rodents and environmental change at 2900-meters elevation. Porcupine Cave, South Park, Colorado. Ann. Carnegie Mus. 57:267- 292. Bryant, M. D. 1945. Phylogeny of Nearctic Sciuridae. Amer. Midi. Nat. 33:257-390. Burns, J. A. andW. B. McGillivray. 1989. A new prairie dog, Cynomys churcherii, from the Late Pleistocene of southern Alberta. Can. J. Zool. 67:2633-2639. Ceballos-G., G. and D. E. Wilson. 1985. Cynomys mexicanus. Mamm. Species 248:1-3. Dalquest, W. W. 1967. Mammals of the Pleistocene Slaton local fauna of Texas. Southwestern Nat. 12; 1- 30. Dalquest, W. W. 1988. Fossil prairie dogs (Cynomys) from Texas. Texas J. Sci. 40:3-10. EsHELMAN, R. E. 1975. Geology and paleontology of the early Pleistocene (late Blancan) White Rock fauna from north-central Kansas. Pp. 1-60. in Studies on Cenozoic Paleontology and Stratigraphy. Claude W. HihhardMem. Vol. 4. Univ. Michigan Pap. Paleontol. 13. EsHELMAN, R. E. and M. W. Hager. 1984. Two Irvingtonian (Medial Pleistocene) vertebrate faunas from north- central Kansas. Pp. 384-404. in H. H. Genoways and SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 33 M. R. Dawson (eds.), Contributions in Quaternary Vertebrate Paleontology: A Volume in Memorial to John E. Guilday. Carnegie Mus. Nat. Hist. Spec. Publ. 8. EsHELMAN, R. E. and C. W. Hibbard. 1981. Nash local fauna (Pleistocene: Aftonian) of Meade County, Kan- sas. Univ. Michigan Mus. Paleontol.Contrib. 25:317- 326. Goodwin, H. T. 1990a. Ph.D. dissertation: Systematics, biogeography, and evolution of fossil prairie dogs (genus Cynomys). Univ. Kansas, Lawrence. 291 pp. Goodwin, H. T. 1990b. A large white-tailed prairie dog, Cynomys (Leucocrossuromys), from the Pleistocene of Kansas. Curr. Res. Pleist. 7:107-109. Goodwin, H. T. 1993. Subgeneric identification and bio- stratigraphic utility of late Pleistocene prairie dogs (Cynomys. Sciuridae) from the Great Plains. South- western Nat. 38: 105-1 10. Green, M. 1 960. ATertiary Cynomys from South Dakota. J. Paleontol. 34:545-547. Green, M. 1963. Some late Pleistocene rodents from South Dakota. J. Paleontol. 37:688-690. Hager, M. W. 1972. A late Wisconsin-Recent vertebrate fauna from Chimney Rock Animal Trap, Larimer County, Colorado. Univ. Wyoming Contrib. Geol. 11:63-71. Hager, M. W. 1 975. Late Pliocene and Pleistocene history of the Donnelly Ranch vertebrate site, southeastern Colorado. Univ. Wyoming Contrib. Geol., Spec. Pap. 2:1-62. Hall, E. R. 1981. The Mammals of North America, 2nd ed. John Wiley and Sons, New York. 1 : 1-600 + 90. Harris, A. H. 1985. Late Pleistocene Vertebrate Paleo- ecology of the West. Univ. Texas Press, Austin. 293 pp. Harris, A. H. 1987. Reconstruction of mid-Wisconsin environments in southern New Mexico. National Geogr. Res. 3:142-151. Harris, A.H.I 989. The New Mexican Late Wisconsin — east versus west. National Geogr. Res. 5:205-217. Hay, O. p. 1921. Description of Pleistocene Vertebrata, types or specimens of which are preserved in the United States National Museum. Proc. U. S. Natl. Mus. 59:617-638. Hay, O. p. and H. J. Cook. 1930. Fossil vertebrates collected near, or in association with, human artifacts at localities near Colorado, Texas; Frederick, Okla- homa; and Folsom, New Mexico. Proc. Colorado Mus. Nat. Hist. 9:4^0. Hibbard, C. W. 1942. Pleistocene mammals from Kansas. Bull. Kansas Geol. Surv. 41:261-269. Hibbard, C. W. 1 944. Stratigraphy and vertebrate paleon- tology of Pleistocene deposits of southwestern Kan- sas. Bull. Geol. Soc. Amer. 55:707-754. Hibbard, C. W. 1956. Vertebrate fossils from the Meade Formation of southwestern Kansas. Pap. Michigan Acad. Sci.,Arts, Lett. 41:145-203. Hibbard, C. W. and D. W. Taylor. 1960. Two late Pleis- tocene faunas from southwestern Kansas. Univ. Michi- gan Mus. Paleontol. Contrib. 16:1-223. Hibbard, C. W., R. J. Zakrzewski, R. E. Eshelman, G. Edmond, C. D. Griggs, and C. Griggs. 1978. Mam- mals of the Kanopolis local fauna, Pleistocene (Yarmouth)ofEllsworthCounty.Kansas. Univ. Michi- gan Mus. Paleontol. Contrib. 25:11^4. Hoffmann, R. S. and J. K. Jones, Jr. 1970. Influence of late-glacial and post-glacial events on the distribution of Recent mammals on the northern Great Plains. Pp. 355-394. //; W. Dort, Jr., and J. K. Jones, Jr. (eds.). Pleistocene and Recent Environments of the Centred Great Plains. Univ. Kansas Press, Lawrence. Hollister, N. 1916. a systematic account of the prairie dogs. N. Amer. Fauna 40:1-36. HoLMAN, J. A. 1 984. Herpetofaunas of the Duck Creek and Williams local faunas (Pleistocene: Illinoian) of Kan- sas. Pp. 20-38. /■// H. H. Genoways and M. R. Dawson (eds.). Contributions in Quaternary Vertebrate Pale- ontology: A Volume in Memorial to John E. Guilday. Carnegie Mus. Nat. Hist. Spec. Publ. 8. HooGLAND, J. L. 1981. The evolution of coloniality in white-tailed and black-tailed prairie dogs (Sciuridae: Cynomys leucurus and C. ludovicianus). Ecology 62:252-272. Indeck, J. 1 987. Ph.D. dissertation: Sediment analysis and mammalian fauna from Little Box Elder Cave, Wyo- ming. Univ. Colorado, Boulder. 2 1 0 pp. (abstract seen). Jelinek, a. J. 1960. A late Pleistocene vertebrate fauna from Texas. J. Paleontol. 34:933-939. Johnson, E. 1 974. Zooarchaeology and the Lubbock Lake site. Pp. 107-122. //; C. C. Black (ed.). History and Prehistory of the Lubbock Lake Site. The Museum J. 15:1-160. Lundelius, E. L., Jr. 1 967. Late Pleistocene and Holocene faunal history of central Texas. Pp. 287-319. in P. S. Martin and H. E. Wright, Jr. (eds. ), Pleistocene E.xtinc- tions: A Search for a Cause. Yale Univ. Press, New Haven. Lundelius, E. L., Jr. 1 972. Fossil vertebrates from the late Pleistocene Ingleside fauna, San Patricio County, Texas. Univ. Texas Bur. Econ. Geol., Rept. Invest. 77:1-74. Lundelius, E. L., Jr., C. S. Churcher, T. Downs, C. R. Harington, E. H. Lindsay, G. E. Schultz, H. A. Semken, S. D. Webb, and R. J. Zakrzewski. 1987. The North American Quaternary sequence. Pp. 211-235. in M. O. Woodbume (ed.), Cenozoic Mammals of North America: Geochronology and Biostratigraphy. Univ. California Press, Berkeley. Manly, B. F. J. 1986. Multivariate Statistical Methods: A Primer Chapman and Hall, New York. 159 pp. Martin, L. D. 1969. M.S. thesis: A medial Pleistocene fauna from near Angus, Nuckolls County, Nebraska. Univ. Nebraska, Lincoln. 75 pp. 34 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Martin, L. D. and C. B. Schultz. 1985. Small mammals of the Seneca and Sappa local faunas (post-Ogailala of Nebraska). TER-QUA Symposium Ser. 1:163-179. McCuLLOUGH, D. A. AND R. K. Chesser. 1987. Genetic variation among populations of the Mexican prairie dog. J. Mamm. 68:555-560. McCuLLOL'GH, D. A., R. K. Chesser, and R. D. Owen. 1987. Immunological systematics of prairie dogs. J. Mamm. 68:561-568. McDonald, H. G. and E. Anderson. 1975. A Late Pleis- tocene vertebrate fauna from southeastern Idaho. Tebiwa 18:19-37. Miller, B. B. 1978. A radiocarbon-dated (Wisconsinan) moUuscan fauna from southeastern Kansas. J. Paleontol. 52:993-1002. -Nadler, C. F.,R. S.Hoffmann, and J. J. Pizzimenti. 1971. Chromosomes and serum proteins of prairie dogs and a model of Cynomys evolution. J. Mamm. 52:545- 555. PiNSOF, J. D. 1992. Doctoral dissertation: The late Pleis- tocene vertebrate fauna from the American Falls area, southeastern Idaho. Idaho State Univ., Pocatello. 421 pp. Pizzimenti, J. J. 1975. Evolution of the prairie dog genus Cynomys. Univ. Kansas Mus. Nat. Hist. Occas. Pap. 39:1-73. Pizzimenti, J. J. and C. F. Nadler. 1972. Chromosomes and serum proteins of the Utah prairie dog, Cynomys panideiis ( Sciuridae). Southwestern Nat. 1 7:279-286. Rasmussen, D. L. 1974. New Quaternary mammal locali- ties in the upper Clark Fork River valley, western Montana. Northwest Geol. 3:59-61. Rhodes, R. S., II. 1984. Paleoecology and regional paleoclimatic implications of the Farmdalian Craigmile and Woodfordian Waubonsie mammalian local fau- nas, southwestern Iowa. Illinois State Mus. Rept. In- vest. 40:1-51. Rogers, K. L., C. A. Repenning, R. M. Forester, E. E. Larson, S. A. Hall, G. R. Smith, E. Anderson, and T. J. Brown. 1 985. Middle Pleistocene (Late Irvingtonian: Nebraskan) climatic changes in south-central Colo- rado. National Geogr. Res. 1: 535-565. RoHLF, F. J. 1985. BIOM: a package of statistical pro- grams to accompany the text Biometry. Copyright by F. J. Rohlf. SAS Institute Inc. 1985. SAS User's Guide: Statistics, Version 5 ed. SAS Institute Inc., Gary, N.C. 946 pp. Schultz, C. B. and L. D. Martin. 1970. Quaternary mammalian sequence in the central Great Plains. Pp. 341-353. in W. Dort, Jr., and J. K. Jones, Jr. (eds.). Pleistocene and Recent Environments of the Central Great Plains. Univ. Kansas Press, Lawrence. Schultz, C. B., L. D. Martin, L. G. Tanner, and R. G. Corner. 1978. Provincial land mammal ages for the North American Quaternary. Bull. Univ. Nebraska State Mus. 9:183-195. Schultz, C. B. and L. G. Tanner. 1957. Medial Pleis- tocene fossil vertebrate localities in Nebraska. Bull. Univ. Nebraska State Mus. 4:59-81. Schultz, G. E. 1969. Geology and paleontology of a late Pleistocene basin in southwest Kansas. Geol. Soc. Amer., Spec. Pap. 105:1-85. Semken, H. a.. Jr. 1 966. Stratigraphy and paleontology of the McPherson Eqiius beds (Sandahl Local Fauna), McPherson County, Kansas. Univ. Michigan Mus. Paleontol. Contrib. 20:121-178. Skwarawoolf, T. 1980. Mammals of the Riddell local fauna (Floral Formation, Pleistocene, Late Rancho- labrean). Saskatoon, Canada. Nat. Hist. Contrib. 2:1- 129. Slaughter, B. H., W. W. Crook, Jr., R. K. Harris, D. C. Allen, and M. Seifert. 1962. The Hill-Shuler local faunas of the upper Trinity River in Dallas and Denton counties, Texas. Univ. Texas Bur. Econ. Geol., Rept. Invest. 48:1-75. Slaughter. B. H. and R. Ritchie. 1963. Pleistocene mam- mals of the Clear Creek Local Fauna, Denton County, Texas. J. Graduate Res. Cent. 31:117-131. Stalker, A. MacS. and C. S. Churcher. 1970. Deposits near Medicine Hat, Alberta, Canada. Geol. Surv. Canada, Ottawa (wall chart). Swofford, D. L. 1985. PAUP: Phylogenetic analysis using parsimony. User's manual. Illinois Nat. Hist. Surv., Champaign, Illinois. TiLESTON, J. V. and R. R. Lechleitner. 1966. Some com- parisons of the black-tailed and white-tailed prairie dogs in north-central Colorado. Amer. Midi. Nat. 75:292-316. Van Devender, T. R. and R. D. Worthington. 1977. The herpetofauna of Howell's Ridge Cave and the paleo- ecology of the northwestern Chihuahuan Desert. Pp. 85-106. ;/; R. H. Wauer and D. H. Riskind (eds.). Transactions of the Symposium on the Biological Resources of the Chihuahuan Desert Region. United States and Mexico. Natl. Park Serv. Trans. Proc. Ser. 13. VooRHiES, M. R. andR. G. Corner. 1985. Small mammals with boreal affinities in Late Pleistocene (Rancho- labrean) deposits of eastern and central Nebraska. TER-QUA Symposium Ser. 1:125-142. Wood, A. E. and R. W. Wilson. 1936. A suggested nomenclature for the cusps of the cheek teeth of rodents. J. Paleontol. 10:388-391. Wood, H. E., II, R. W. Chaney, J. Clark, E. H. Colbert, G. L. Jepsen, J. B. Reeside, JR., and C. Stock. 1941. Nomenclature and correlation of the North American continental Tertiary. Bull. Geol. Soc. Amer. 52:1^8. Zeimens, G. and D. N. Walker. 1974. Bell Cave, Wyo- ming: preliminary archaeological and paleontological investigations. Wyoming Geol. Surv., Rept. Invest. 10:88-90. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 35 APPENDIX Referred fossil specimens, organized by taxon, age, location (alphabetically by state, county, local- ity), and element. Catalogue numbers followed by (?) are referred with question. Fossils referred to species known from small samples {ICynomys vetus, C. hihhardi, C. sappaensis) are listed in the species accounts. Institutional abbreviations are explained in Materials and Methods. Because of the large number of fossils involved, I do not give a full description of each. Fossils are identified and grouped by basic element (e.g., L or R lower jaw). Localities marked with (*) are dated based on the prairie dogs (Goodwin, 1993), and specimens from these locali- ties were excluded from samples of temporally defined OTUs (see Materials and Methods). Cynomys spenceri Sheridanian Kanopolis. Ellsworth Co., KS (Hibbard et al., 1978)— LM3: UMMP 60414 Sandahl, McPherson Co., KS (Semken, 1966); L lower jaws: UMMP V40497, 50463. Angus, Nuckolls Co., NE (Schultz and Tanner, 1957; Martin, 1969)— palates: UNSM 33725, 33726; L lower jaws: UNSM 33706, 33715, 33898, 2068-67; R lower jaws: UNSM 33679, 33680, 33681, 33683, 33705, 33706, 33707, 33708, 33709, 33714, 33725, 33793, 33794, 2505-55. Angus-higher, Nuckolls Co., NE — L, R lower jaws, partial skeleton: UNSM 47723 Prairie dog locality, Nuckolls Co., NE — R lower jaws: UNSM 47709, 47710. Slaton, Lubbock County, TX (Dalquest, 1967, 1988)— L maxilla: MWU 6789; R maxillae: MWU 4632. 6633, 6786; 4 L lower jaws: MWU 6652, 6786, 6787 TMM 882^; 2 R lower jaws: MWU 6790, 6791. Early Rancholabrean Mesa de Maya, Las Animas Co., CO (Hager, 1975) — L lower jaw: UWYG 6032. Unnamed, Clark Co., KS— L lower jaw: KUVP 13436. Butler Springs, Meade Co., KS — L lower jaw: UMMP 45974. Alma, Harlan Co, NE — L lower jaw: KUVP unnum- bered (G). Unnamed locality in Sangamon Soil(?), Harlan Co., NE— L lower jaw: UNSM 50778. BeaverCrossing, Seward Co. ,NE — Llowerjaw: UNSM 2701; R lower jaw: UNSM 2702; RM3: UNSM 2702. Roadside Snail site, Seward Co., NE — associated L maxilla, RP4, L lower jaw: UNSM 2713. Sheridanian/Early Rancholabrean Williams Farm, Rice Co., KS (Holman, 1984) — associ- ated L, R maxillae, L. R lowerjaws: UMMP V60230. Cynomys ludovicianus Late Rancholabrean Craigmile. Mills Co., lA (Rhodes, 1984)— RM3: IOWA 46412. *Unnamed, Cheyenne Co., KS — partial skeleton, L, R lowerjaws: KUVP 517. *Keiger Creek. Clark Co., KS — skull, L, R lowerjaws, partial skeleton: FHSU VP-3526. Pyle Ranch, Clark Co., KS (Hibbard, 1944)— L maxilla, L, R lowerjaws: KUVP 6710. Stephenson Ranch, Clark Co., KS (Hibbard, 1944)— L lowerjaw: KUVP 5896. Duck Creek, Ellis Co., KS (Holman, 1984; date based on J. D. Stewart, pers. comm.) — partial skeleton, skull fragment, R lowerjaw: KUVP unnumbered. *South Pit. Finney Co., KS— R lowerjaw: KUVP 6909. *Unnamed, Finney Co., KS— L maxilla: KUVP 6818; 5L, 2R lowerjaws: KUVP 6814. *Deer Park, Meade Co., KS (Hibbard, 1956)— L lower jaw: UMMP V60532; RM1-M2: UMMP 3 1963 (type of Cynomys meadensis; Deer Park fauna clearly Blancan in age but prairie dogs are intrusive; see text). *KU-REP-002, Republic Co., KS— R maxilla, R lower jaw fragment: KUVP 73 17; Llowerjaw: KUVP 73 16. *KU-SHD-08, Sheridan Co., KS— L lowerjaw: KUVP 6643. KU-WIL-02, Wilson Co., KS (Miller, 1978)— R lower jaw: KUVP unnumbered. KU-WIL-03, Wilson Co., KS— R lowerjaw: KUVP755 1 . Unnamed, Dundy Co., NE— R lowerjaw: UNSM 50773. Citellus zone, Lincoln Co., NE — L, R lowerjaw: UNSM 30036. Litchfield. Sherman Co., NE (Voorhies and Comer, 1985)— L lower jaw: UNSM 88195; RP4: UNSM 88196; RM3: UNSM 88201. Dark Canyon Cave, Eddy Co., NM (Harris, 1985)— 6 L lowerjaws: TMM 4 1 228- 1 022. - 1 642, - 1 65 1 , - 1 658, - 1671, -2132; 7 R lowerjaws: TMM 41228-1645, - 1674, -1679, -1689, -1696, -2127, UTEP 75-19. 36 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 HowelPs Ridge Cave, Grant Co.. NM (Harris, 1985; Van Devender and Worthington, 1977) — L lower jaw: UTEP 32-539(7). U-Bar Cave, Hidalgo Co., NM (Harris, 1987, 1989)— LM,_2: UTEP 5689-153-272. Folsom, Union Co., NM (Hay and Cook, 1930)— L lower jaw: DMNH 1248. UCM L82009-L8201 O.Blaine Co.. 0K—2R lower jaws: UCM 59066. 59067. BumhamSite,WoodsCo.,OK — L,R lower jaws (Bumham private collection). *Jackson Farm. Clay Co.. TX— 3 R lower jaws: TMM 30973- 1 . Clear Creek, Denton Co., TX (Slaughter and Ritchie, 1963 ) — isolated teeth including diagnostic LP4: SMU 60628. Hickory Creek, Denton Co.,TX (Slaughter et al., 1 962)— R lower jaw: SMU 60296. Lewisville, Denton Co., TX (Slaughter et al., 1962)— palate, 2 L, 3 R lower jaws: SMU 60668. Lubbock Lake, Lubbock Co., TX (Johnson, 1974)— L lower jaw: TTU A5423; R lower jaw: TTU A5062. Ingleside, San Patricio Co., TX (Lundelius, 1972) — 2 skulls: TMM 30967-850, -1014; 2 L, R maxillae: TMM 30967-834, -835; 5 L, 9 R lower jaws: TMM 30967-490, -834, -851,-933, -993. Foley Sands, Wichita Co., TX (Jelinek, I960)— 2 skulls: UMMP 3 1 754, 4225 1 ; 10 L, 4 R lower jaws: UMMP 31759,42252. Northwest Materials, Wichita Co., TX— 3 skulls: UMMP 32360, 32361, 32362; 11 L, 8 R lower jaws: UMMP 32357. Laubach Cave, Williamson Co., TX ( Lundelius, 1 967) — fragmentary skull: TMM 40673-101; L lower jaw: TMM 41343-13. Bell Cave, Albany Co., WY (Zeimens and Walker, 1 974)— 2 L lower jaws: UWYA 4383B, 4384B; 3 R lower jaws: UWYA4385B, 4386B, 4387B. Cynomys cf. Cynomys mexicanus Late Rancholabrean Dry Cave: Lost Valley, Eddy Co., NM (Harris, 1985, 1987)— 2 L lower jaws: UTEP 1-4, -1401 ; 3 R lower jaws: UTEP 1-1030, -1067, -1402. Cynomys {Cynomys) sp. Cudahyan Hall Ash. Jewell Co., KS (Eshelman and Hager, 1 984)— LP4: USNM 304236. Burnett Ranch, Knox Co., TX (Dalquest, 1988)— M3: MWU 12225 (not seen). Late Pleistocene (either Cynomys spenceri or C. ludovicianus) Barnesville, Weld Co., CO— L lower jaw: UCM 59068. Unnamed, Barton Co., KS— skull: KUVP 13434. XI Ranch, Meade Co., KS— L lower jaw: KUVP 6476. Unnamed. Russell Co., KS— 3 R lower jaws: KUVP 6286. Medicine Creek Dam Core Trench, Frontier Co., NE — palate: UNSM 50774. Gosper Co., NE— L lower jaw: UNSM 31250. UNSM-KX-102, Knox Co., NE— palate. L lower jaw: UNSM 31238. League Ranch, Knox Co., TX — R lower jaw: TMM 40475-5. Green Estates, San Patricio Co., TX — 2 L lower jaws: TMM 40605-10. -16. Cynomys gunnisoni Cudahyan Hansen Bluff, Alamosa Co., CO (Rogers et al., 1 985 )— R maxilla, isolated teeth, associated LP4-M3: ADAM unnumbered. Late Rancholabrean IsletaCaves,BernalilloCo.,NM(Harris, 1985)— skulls: UTEP 41-313,-314, UTEP 46-29; 23 L lower jaws: UTEP 41-31 6, -A2066. -A2794, UTEP 46-37, -264, -265, -267, -270, -274, -279, -290, -291, -294, -295, -300, -304, -589, -A2082, -A2083. -A300 1 , - A3004, -A3008, -A3014; 4 R lower jaws: UTEP 41-315, - 317, -A2065, -A2346 (much of this material prob- ably is Holocene in age). Burnet Cave, Eddy Co., NM (Harris, 1985)— L lower jaw: UNSM 21931; R lower jaw: UNSM 22563. Dark Canyon Cave. Eddy Co., NM (Harris, 1985)— 2 L lower jaws: TMM 41228-1012, -1685; 6 R lower jaws: TMM 41228-1, -1011, -1655, -1691, -1694, - 2132. Dry Cave; Animal Fair, Eddy Co., NM (Harris, 1985, 1989)— L.R maxillae: UTEP 22-1674; R lower jaw: UTEP 22-1556: 2LM3: UTEP 22-2053, -2396. Hermit's Cave, Eddy Co!. NM (Harris, 1985)— palate: UNSM 1 8972(7). Howell's Ridge Cave, Grant County, NM (Harris, 1 985; Van Devender and Worthington, 1977) — R lower jaw: UTEP 32-548; R, L M3: UTEP 32-531. U-Bar Cave, Hidalgo Co., NM (Harris, 1987, 1989)— 2 L lower jaws: UTEP 5689-99-21, 5689-120-6; LM,_2: UTEP 5689-109-104; 2 LM3:UTEP 5689-78-20, 5689-78-21. SYSTEMATIC REVISION OF FOSSIL PRAIRIE DOGS 37 Late Pleistocene Unnamed, Brewster Co., TX (Harris, 1985) — partial skull: UTEP 13-1 (?). Cynomys niobrarius niobrarius Sheridanian Mitchell Bluff, Medicine Hat, ALT (Stalker and Churcher, 1970)— L lower jaw: ROM MB-27; L lower jaw and isolated teeth: ROM MB- 135. Sandahl, McPherson Co., KS (Semken, 1966) — L max- illa, L lower jaw: UMMP V61146; 2 R lower jaws: UMMP 45355, 50467. Prairie dog locality, Nuckolls Co., NE — L, R maxillae: UNSM 2059-67; 5 L lower jaws: UNSM 2059- 67(b), 2059-67(c),UNSM JAH-102,JAH-104, JAH- 105. Hay Springs/Rushville/Gordon, Sheridan Co., NE (Schultz and Tanner, 1957)— R maxilla: UNSM 50784; 4L lower jaws: UNSM 3 1 377, 3 1 380, 3 1 388, 31389: 8Rlowerjaws:UMMP41252, UNSM 21304, 31378, 31381, 31383, 31385, 31387, 50783. Early Rancholabrean Lindoe Bluff/Mitchell Bluff, Medicine Hat, ALT (Stalker and Churcher, 1970)— L lower jaw: ROM LB-49; RMi_2, associated LP4-M3: ROM MB-68-476. Mesa de Maya, Las Animas Co., CO (Hager, 1975) — LP4: UWYG 6098. American Falls, Power Co., ID (Pinsof, 1992) — L max- illa: IMNH 65001/16466; R maxilla: IMNH 65001/ 36228; 3 L lower jaws: IMNH 612/40175, 65001/ 36229, 78025/34276; 2 R lower jaws: IMNH 782/ 40183,65001/31194. Cragin Quarry, Meade Co., KS (Hibbard and Taylor, 1960; G. E. Schultz, 1969)— L maxilla: UMMP 35580; L lower jaw: UMMP 46045. Alma, Harlan Co., NE — skull: KUVP unnumbered; 5 L lower jaws: KUVP unnumbered (C,D,E,F,H); 3 R lower jaws: KUVP unnumbered (A,B,I). Sangamon SoiI(?), Harlan Co., NE — L lower jaws: UNSM 50777, 50780. Goins pocket, Lincoln Co., NE — 2 L lower jaws: UNSM 50782. Boxelder Canyon, Lincoln Co., NE — skull fragments, L,Rlowerjaws, post cranial elements: UNSM 30238. Beaver Crossing, Seward Co., NE— RP4: UNSM un- numbered. Riddell, Saskatoon, SKT (SkwaraWoolf, 1980)— R P4: USG VM-54. Sheridanian/Early Rancholabrean Unnamed, Jewell Co., KS— L lower jaw: KUVP 421. Quinn Canyon, Dawson Co., NE — R lower jaw: UNSM 30117. Unnamed, Harlan Co., NE — skull, R lower jaw: UNSM 31249. Unnamed, Lincoln Co., NE — L, R lower jaws: UNSM 30194. Late Rancholabrean Gah Island Bluff, Medicine Hat, ALTA (Stalker and Churcher, 1970)— R lower jaw, LM, ^^ 2- ^^M GIB-MG-19. *Bejewelled Oriental Whitetail, Arapahoe Co., CO — 2 L lower jaws: UCM 58261, 59318. Rainbow Beach, Power Co., ID (McDonald and Ander- son, 1975) — skull, L, R lower jaws, partial skeleton: IMNH 269/29107; Llowerjaw: IMNH 72006/23663; R lower jaw: 72003/24645. *South Pit, Finney Co., KS (Goodwin, 1990b)— skull, L, R lower jaws: KUVP 6908. *Unnamed, Logan Co., KS— palate: FHSU VP-4630; 3 L lower jaws; FHSU VP-3662, VP-4631, VP-4632; R lower jaw: FHSU VP-6636. *KU-MEA-08, Meade Co., KS— R lower jaw: KUVP 4614. *Unnamed, Norton Co., KS— R lower jaw: FHSU VP- 3183. *KU-REP-002, Republic Co., KS— R maxilla, LP3, R, L lower jaws: KUVP 7313; L lower jaw: KUVP 7315; R lower jaw: KUVP 7314. KU-ROO-003, Rooks Co., KS— 2 skulls: KUVP 25 104, 25105; Llowerjaw: KUVP 55595; 2 R lower jaws: KUVP 55596, 55597. KU-ROO-007, Rooks Co., KS— L lower jaw: KUVP 55593; 3 R lower jaws: KUVP 6061 1 , 63 1 10, 63 1 1 1 . Dutton Ranch No. 8, Powell Co., MT (Rasmussen, 1974)— L lower jaw: UMTG 2293. MV 6546, Powell or Granite cos., MT — L lower jaw: UMTG 2318. *Elm Creek, Buffalo Co., NE— 6 L lower jaws: FHSU VP-3143, VP-3144, VP-3145, VP-3146, VP-3147, VP-4I08; 6 R lower jaws: FHSU VP-3148, VP- 3149, VP-3150, VP-315I, VP-3152, VP-4110; iso- lated teeth: FHSU VP-3154, VP-4111. Smith Falls, Cherry Co., NE (Voorhies and Corner, 1985)— R maxilla: UNSM 82020; 3 L lower jaws: UNSM 820 1 6, 82058, 82 169; 2 R lower jaws: UNSM 82017, 82018; LP4: UNSM 82063; LM3: UNSM 82061. Unnamed Peorian loess locality, Dawson Co., NE — R lower jaw: UNSM 30093. 38 UNIV. KANSAS NAT. HIST. MUS. MISC. PUBL. No. 86 Citellus zone, Dawson Co., NE — R lower jaw: UNSM 30281. Citellus zone, Lincoln Co., NE — palate, R lower jaw, partial skeleton: UNSM 30102; R lowerjaw: UNSM 30061. Litchfield, Sherman Co., NE (Voorhies and Corner, 1985)— 3 LM3: UNSM 88197, 88199. 88200. Roosevelt Lake, Tripp Co.. SD (Green, 1960, 1963)— LP3-M3: SDSM 57100 (type of C. spispiza); LP4: SDSM 5939; LM3: SDSM 5940. Natural Trap Cave, Bighorn Co., WY — L, R maxillae, L lowerjaw: KUVP unnumbered. Little Box Elder Cave, Converse Co., WY (Anderson, 1968; Indeck, 1987)— 17 L, 13 R lower jaws: UCM . 23608 to 23611. 23613 to 23621, 23627, 23631 to 23633, 23637, 23638, 23640, 23642, 23643, 23646, 23648, 23650. Late Pleistocene Nussbaum, El Paso Co., CO — palate, R lower jaw: UCM 34665. KU-DEC-00 1 , Decatur Co., KS— brain cast, L maxilla, R lowerjaw, isolated teeth: KUVP 3968. KU-PHI-18, Phillips Co., KS— skull: KUVP unnum- bered. KU-SHD-01, Sheridan Co., KS— L lowerjaw: KUVP 12439. Cynomys niohrarius churcherii Late Rancholabrean Courtney, Hand Hills, ALT (Burns and McGillivray, 1989)— 3 skulls: PMA P86.ll.10, P86.ll.17, P89.22.2; 7 L lower jaws: PMA P86. 11.10, P86. 1 1 .32, P88.20.9, P88.20. 1 0. P88.20.30, P88.20.32, P89.22.4; R lowerjaw: PMA P86.1 1.17. Sinclair, Hand Hills,ALT(ibid.)—skull: PMA P75. 10.1; R lowerjaw: PMA P75.7.1. Winter, Hand Hills, ALT (ibid.)— 13 skulls: PMA P85.9.14, P85.9.33, P85.9.79, P85.9.97, P85.9.242, P86.3.8, P86.3.21, P86.3.25, P86.3.241, P86.3.381, P86.3.391,P86.3.516,P86.9.1;9Llowerjaws:PMA P85.9.14,P86.3.28,P86.3.108,P86.3.143,P86.3.166, P86.3.357, P86.3.376, P86.3.450, P86.3.464; 2 R lower jaws: PMA P86.3.391, P86.9.1. Schowalter, Wintering Hills, ALT (ibid.) — L lowerjaw: PMA P87.8.2. Cynomys {Leucocrossuwmys) sp. Sappan Kentuck, McPherson Co. , KS ( Semken, 1 966 ) — L lower jaw: UMMP 50494; RM3: UMMP 50495 (not seen) Sheridanian/Early Rancholabrean Williams Farm, Rice Co., KS (Holman, 1984)— R lower jaw: FHSU VP-6931(?). QE882.R6 G66 1995 S\sli.iM,ila n\iM'>ii "I •"'■'•il ?<■"' Harvard MCZ Library .. .'*L^.,'i',!'.' 3 2044 062 450 093 REC ^ 72. Leptodactylid frogs of Colombia. By John D. 73. Type and figured spec Museum of Natural H and R. W. Coldiron. P 74. Relationships of pocki By Lawrence R. Hear 75. The taxonomy and pj Duellman and Marini 76. Variation in clutch an 24 May 1985. Paper! 77. Type and figured spe Museum of Natural I Chorn and A. M. Nei 78. Type and figured spj Museum of Natural I 1-14. 5 February 19 DATE DUE 79. Type and figured sp », r-xT 1^ DEMCO, INC. 38-2931 Museum or Natural Ihaloi^. i unx i. x uu.— Ecuador and adjacent ind. — University of Kansas Jtewart, A. M. Neuner northern Great Plains. le 1983. Paper bound. efania. By William E. ■. Paper bound. 1. Pp. 1-76, 15 figures. I University of Kansas •P. Schultze, L. Hunt, J. e University of Kansas ion Anne Jenkinson. Pp. le University of Kansas der, Assefa Mebrate and Robert W. Wilson. Pp. 1-83. 21 November 1986. Paper bound. 80. Phylogenetic studies of north american minnows, with emphasis on the genus CyprineUa (Teleostei: Cypriniformes). By Richard L. Mayden. Pp. 1-189, 85 figures, 4 color plates. 1 June 1989. Paper bound. ISBN: 0-89338-029-6. 81. A Phylogenetic Analysis and Taxonomy of Iguanian Lizards (Reptilia: Squamata). By Darrel R. Frost and Richard Etheridge. Pp. 1-65, 24 figures, 3 appendices. 28 September 1989. Paper bound. ISBN: 0-89338-033-4. 82. Bats of Portugal: Zoogeography and Systematics. By Jorge M. Palmeirim. Pp. 1-53, 39 figures, 24 tables, 1 appendix. 15 March 1990. Paper bound. ISBN: 0-89338-034-2. 83. Altitudinal Ecology oiAgama tuberculata Gray in the Western Himalayas. By Robert C. Waltner. Pp. 1-74, 38 figures, 24 tables. 20 February 1991. Paper bound. ISBN: 0-89338-036-9. 84. Systematics of the West Indian Lizard Genus Leiocephalus (Squamata: Iguania: Tropiduridae). By Gregory K. Pregill. Pp. 1-69, 21 figures, 3 tables, 4 appendices. 30 April 1992. Paper bound. ISBN: 0-89338-041-5. 85. The Amphibamidae (Amphibia: Temnospondyli), with a Description of a New Genus from the UpperPennsylvanianofKansas. By Eleanor Daly. Pp. 1-59, 40 figures, 4 tables. 17 February 1994. ISBN: 0-89338-046-6.