03 m ^ m ^ ~ CO > ^ ^ _ CO £ CO £ c/> 2 CO ~ illSNI NVINOSHilWS SBiaVdan libraries SMITHSONIAN INSTITUTION NOlifUllSNI NVINOSHilWS S3iaVH9l ^ ^ ^ Z T CO Z CO M/.. r: v%. . -, /^Tf^ g g 3 # i > ^ ^ 5 • -.^xv 3, - 2 Xioosvi^ >■ ^R I ES^^SMITHSONIAN INSTITUTION N0liniliSNI_NVIN0SHilWs‘^S3 I HVa 8 ll^LI B RAR I Es‘^SMlTHSONlAN ^INSTITUTIOr CO „ ^ ^ ^ z \ ^ ^ ^ z CO ixi /fifi'jCf . clj tj) x£5o>Jo\ cu q: < ^ iliSNi NviNOSHims S3 lava an libraries Smithsonian institution NoiiniiiSNi nvinoshiiws ssiavaai r- z '” ^ [I 2 I- 2 O m 5 . vro 2 O z) W ■'a*\ > (?£:’ > (sE; :;il) 2 2 w« m x», CO m xSiiDcJi/ ^ m xg^oc^/ w ^ ^ in *■. E — to Z: CO £ c/5 _ — \RIES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHimS S3iaVaail LIBRARIES SMITHSONIAN INSTITUTIO .-.to 2: CO z CO z ..■• CO r>- < 2 < 5 X ■ - : J> xgUllS^/ 5 ^ S * 2 w ‘ 2 CO ^ iliSNi NVlNOSHil^^s S3iavaan libraries Smithsonian institution NoiiniiiSNi nvinoshiiws saiavaai — CO ^ CO ^ to — CO \RIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillAIS S3iavaan LIBRARIES SMITHSONIAN INSTITUTIO! Z r- z (-, z r- Z i~ ^0 f , m ^ ID ' z ID _ CO £ CO ' _ CO _ LliSNI NVINOSHillNS S3iaVaan libraries SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHIIIMS S3iavaai 2^ (/) (F\ -»> /rt ^ frt 2 Xgcosi^ > 2 2 ' YT" > CO Z CO ■ Z CO ■■■' Z CO -t. IiRIES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHIIWS S3iavaai1 LIBRARIES SMITHSONIAN INSTITUTIOI CO n CO — '-.CO — fo — CO Q v>c^osv^ — o ■ XjjvAsiA^ Q \L''o5iS!^''!p^ “ " r-1 ''^osv'^ LIISNI "^NVINOSHimS ^S3 1 a Va 8 11 “*LI B R AR I ES SMITHSONIAN INSTITUTION NOIinillSNI ""nVINOSHIIWS ^S3 I ava 8 I r" _ z ^ 2 ^ f“ . z "‘•'x ™ V. o CD 2 . tjj 2 ^ o . 03 2 CD ^ . VCD 2 ol > 33 ISix m "^ 2 XigCTTyiX' CD rn ^ m CO X_ Z in ^ in ± — CO -. _ iRIES SMITHSONIAN INSTITUTION NOIinillSNI NVINOSHimS S3iaVa8l1 LIBRARIES SMITHSONIAN INSTITUTIO .• w z V CO z CO z •.-. CO 2: / V— 5 ^ z A^TfJiKs -H , A./ z H /skfW-CN 3' z; ^ > XgUy^2X 2 > Z CO ' 2 CO Z CO * Z — CO iiiSNi NViNosHiiws S3 1 ava an libraries Smithsonian institution NoiiniiiSNi nvinoshii^^js saiavaa “ cn 2 ^ 2 ^ 2 '.V ^ m z in 'il m CO ' ± (j) ^ CO RARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHimS SHIdVHan ^ Z T CO 2 CO 2 < X. 5: < LIBRARIES SMITHSONIAN INSTITUT 2 , CO O X O CO ■■'■' z CO — z CO -c- CO • ifUliSNI NVINOSHilWS SBIdVaan libraries SMITHSONIAN INSTITUTION f^O'm iSN I _ NWINOSHil WS S3IBVB1 “ •- CO ^ - CO - CO _ V CO ClI XiATItTTX i.i X UJ CO o z RARIES SMITHSONIAN INSTITUTION NOliniliSN! NVINOSHillAIS SHldVaail LIBRARIES SMITHSONIAN^INSTITUT I i / / - 'J rn ^t\ •- /.vW»rAx\ ^ ^ ^ ^ '".'>/ xW XI f- ^ m X ^2 C^ XAon^ - m — (/) \ Z — CO ± CO iniiiSNi NViNOSHims saiavaan libraries smithsonian institution NoiiniiiSNi nvinoshiiws saiava “2P //^ //^ \J} ^ wy - as. (/) -. aS. t/; ^ RARIES SMITHSONIAN INSTITUTION NOlifUliSNI NVINOSHilkMS SBIdVaan LIBRARIES SMITHSONIAN INSTITUT ^ 2 ^ — . C/) ^ ^ ^ — CiJ tn tn ^k-\ CU CO - CO iniliSNI NVlNOSHillNS SBIdVdan LIBRARIES SMITHSONIAN INSTITUTION NOliOiliSNI NVINOSHillAIS S3iaVH f^.Z r- Z ^ Z r^ , Z m 2 xfo^x 2 m 2 rr, .'Jv 2 V' — \j/ __ V// _ V// — IRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIWS S3iavaai1 LIBRARIES SMITHSONIAN INSTITUl fn -»» tn -T t r\ — CD ? XJWS^C?/ ’ z CO •'“ Z CO Z CO c/J iniiiSNi_NviNOSHiiws S3iavaan libraries smithsonian institution NOllnlllS^ nvinoshiiws S3iava ^ \ ^ ^ ^ ^ ^ ^ 12 ^ ^ CO ^ ClI ^ X'oo7;oX CjJ X^Ivcirg^ CO ClJ C z! m Z m O “ O ^ Xit-Asv^^X O _ ^X|aas*1^ — — Z^2_J Z _J IRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIWS SBiavaail LIBRARIES SMITHSONIAN INSTITUT z r-_z r- t- z ^ m ^ x^ocii^sx m — CO ± \ ^ CO — linillSNI NVINOSHillAIS S3!avaail libraries SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHillAIS S3iava Z CO Z ..•■ CO 2; CO Z CO Z CO ^ Z w ^ z CO z IRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHIIIMS S3iavaai1 LIBRARIES SMITHSONIAN INSTITUT . . CO = CO z CO : CO •> t ■ BULLETIN 0/ CARNEGIE MUSEUM OF NATURAL HISTORY EARLY TERTIARY ADAPISORICIDAE AND ERINACEIDAE (MAMMALIA, INSECTIVORA) OF NORTH AMERICA LEONARD KRISHTALKA Section of Vertebrate Fossils NUMBER 1 PITTSBURGH, 1976 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 1, pages 1-40, figures 1-13, tables 1-7, appendix tables A,B,C Issued June 16, 1976 Price: $2.50 a copy CONTENTS Abstract 4 Introduction 4 Systematics 7 Family Adapisoricidae 7 M c Kennatherium 7 “ Diacodon" minutus 10 Scenopagus 10 Ankylodon 16 Macrocranion 18 Talpavus 21 Adapisoricidae: Summary 26 Family Erinaceidae 28 Litolestes 28 Leipsanolestes 30 ?Erinaceidae Entomolestes granger i 32 Erinaceidae: Summary 32 Acknowledgements 33 References Cited 34 Appendix A 36 Appendix B 39 Appendix C 40 ABSTRACT The differentiation and relationships among the four earliest known groups of lipotyphlan insectivores concerns the early Tertiary families Adapisoricidae, Erinaceidae, Nyctitheriidae and Geolabididae. The first two families compose the Erinaceomorpha and are considered here. North American adapisoricids comprise McKennatherium (M. ladae, and an undescribed new species), Scenopagus (S. edenensis, S. priscus, S. curtidens), Ankylodon, Talpavus (T. nitidus, T. duplus, new species, Talpavus sp.), and Macrocranion (M. nitens). The erinaceids include Litolestes (L. ignotus, and tentatively, L. lacimatus and L. notissimus), Leipsanolestes seigfriedti and perhaps Entomolestes grangeri. Leptacodon ladae is here referred to McKennatherium and Entomolestes nitens to Macrocranion. ‘’''Leptacodon'' jepseni and “’Diacodon" minutus do not belong in their respective genera. The former may be an adapisoricid. The latter is probably referrable to Adunator, a European Paleocene condylarth. McKennatherium, Scenopagus, Talpavus, and Ankylodon form a lineage of adapisoricids separate from Macrocranion, which may have originated in Europe, and which adds to the evidence of a Euramerican fauna during the early Eocene. Litolestes ignotus may be central to the radiation of later Tertiary erinaceids, and, along with Ankylodon, may have possessed five lower premolars — possibly a primitive character among erinaceomorphs. Adapisoricids and erin- aceids share a common origin from a Cretaceous non- paleoryctid-leptictid group of insectivores, which may also have been basal to the radiation of primates, dermop- terans, and ungulates. INTRODUCTION This study is an elaboration of part of a doctoral dissertation (Krishtalka, 1975) that concerned the systematics and relationships of the oldest known lipotyphlans: the Adapisoricidae, Erinaceidae, Nycti- theriidae, and Geolabididae. That project began as an effort to identify the insectivores from the Powder Wash local fauna, early Bridgerian of Utah, of which the adapisoricids are a mainstay of the discussions in this paper. The systematics and relationships of the Order Insectivora are among the most perplexing and least understood problems in mammalian evolution. Many workers, among them Gregory (1910), Vandebroek (1961), Van Valen (1967), Thenius (1969), and most recently, Butler (1972) have attempted synthetic re- views of insectivore phylogeny and classification. Insectivores were first regarded as a distinct group by Illiger (1811) and later divided by Haekel (1866) into the suborders Lipotyphla and Menotyphla. The former comprised shrews, moles, and hedgehogs, or those insectivores that lack an intestinal caecum. The Menotyphla combined tree shrews, elephant shrews, and flying lemurs, or those animals in which a caecum did occur. Although the Menotyphla retain many primitive characters that are lost in the Lipotyphla, they are not a natural group. Flying lemurs and elephant shrews have been relegated to the separate orders Dermoptera (Gill, 1872) and Macroscelidea (Butler, 1956), respectively. The suggested relation- ships of the tupaiids have swung between primates and insectivores, and Butler (1972) favored an ordinal ranking (Scandentia) for them as well, at least until their affinities are better understood. Romer (1966) grouped all fossil non-lipotyphlans and some living menotyphlans as the Suborder Pro- teutheria, essentially a wastebasket taxon. Butler (1972) advocated retention of the Proteutheria, minus the living menotyphlans, as a separate order for the poorly understood fossil non-lipotyphlans, thus hmit- ing the concept of the Insectivora to the Lipotyphla. The resultant Order Lipotyphla would include the living families Erinaceidae, Soricidae, Talpidae, Solen- odontidae, Tenrecidae, and Chrysochloridae, and the extinct Adapisoricidae, Nyctitheriidae, Geolabididae, Nesophontidae, Micropternodontidae, Apternodonti- dae, and Plesiosoricidae. Butler’s Order Proteutheria includes the Palaeoryctidae, Lepticitidae, Apatemyi- dae, Pantolestidae, and Plagiomenidae. Living lipotyphlans are united by the absence of an intestinal caecum, by a reduced jugal bone, an ex- panded maxilla in the orbital wall of the skull replacing the palatine, a mobile proboscis served by a series of muscles that affect the form of the skull, a reduced pubic symphysis, absence of the medial internal carotid artery, and an auditory bulla composed mainly of the basispenoid (Butler, 1972). The fossil lipotyphlans are grouped together and with living forms mainly on the basis of dental evidence, although the known cranial remains also support affinities with the Lipotyphla. Butler (1956, 1972) recognized a fundamental dichot- omy among lipotyphlan insectivores — the Erinace- omorpha and Soricomorpha (Saban, 1954) — which he suggested may have originated in the Paleocene. The erinaceomorphs include the Adapisoricidae and the Erinaceidae. In these groups the infraorbital canal is long, the eye is large, and the coronoid process extends posterodorsally. The soricomorphs include the families Geolabididae, Nyctitheriidae, Soricidae, Talpidae, Ple- 4 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 5 siosoricidae, Dimylidae, Micropternodontidae, Soleno- dontidae, Nesophontidae and, tentatively, Apterno- dontidae. These groups possess a short, infraorbital canal, a sni.all eye, and a more nearly vertical coronoid process (Butler, 1956; 1972; personal commun., 1975). The amount of taxonomic shuffling of nyctitheres, geolabidids, adapisoricids, and erinaceids into various familial, subfamilial, and generic groupings in only the last decade is indicative of the poor understanding of the systematics of these early Tertiary insectivores. Biases of collecting and investigation are moot points in this problem. Most of the early work describing early Tertiary insectivore evolution stems from the study of single local mammalian faunas. Rarely did one investigator have at his disposal all of the pertinent fossil insectivore material from known early Tertiary localities. As a result, taxa were duplicated in the litera- ture and their relationships remained muddled. This problem is magnified when considering relationships between North American and European insectivores (e.g., Russell, et al., 1975). Some recent work, how- ever, has concentrated on the systematics of specific groups like nyctitheriids (Robinson, 1968a), geo- labidids (McKenna, 1960b; Lillegraven and McKenna, MS, and European adapisoricids (Russell, et al., 1975). Remains of insectivores from the following localities were used in this study (see Fig. 1): Torrejonian; (1) Gidley Quarry, Lebo Formation, Montana. (2) Rock Bench Quarry, Polecat Bench Forma- tion, Wyoming. Tiffanian: (3) Cedar and Silver Coulee Quarries, Polecat Bench Formation, Wyoming. (4) Mason Pocket, San Jose Formation, Colorado. (5) Bear Creek, Fort Union Series, Montana. Wasatchian : (6) Four Mile localities, Wasatch Formation, Colorado. (7) Powder River localities, Wasatch Forma- tion, Wyoming. (8) Big Horn Basin, Will wood Formation, Wyoming. (9) Lysite and Lost Cabin, Wind River Forma- tion, Wyoming. (10) Huerfano localities, Huerfano Formation, Colorado. (11) Almagre localities, San Jose Formation, New Mexico. Bridgerian : (12) Powder Wash, Green River Formation, Utah. (13) Bridger Basin, Bridger Formation, Wyoming. (14) East Fork, Tepee Trail Formation, Wyoming. Uintan: (15) Myton Pocket, Uinta Formation, Utah. (16) Badwater localities, ?Tepee Trail Forma- tion, Wyoming. All measurements are given in millimeters. For those species represented by abundant material, only the observed range in size is presented in the tables, but measurements of each individual specimen appear in the appendices. Although the tooth nomenclature follows Van Valen (1967), Szalay (1969), and Krishtalka (1973), three terms used here in reference to the morphology of P4 in early Tertiary lipotyphlans need further clarification: a premolariform P4 is characterized by a large, dominant protoconid and an extremely short, usually unicuspid talonid. A small paraconid and metaconid may occur, but they are poorly differentiated from the protoconid. On a semimolariform P4 the talonid is much nar- rower than, but nearly as long as, that on Mi, and usually bears two or three cuspules. The paraconid may be better developed than on a premolariform P4, but is lower than the metaconid and projects from the anterior part of the base of the protoconid. A submolariform P4 is essentially like Mi, except for a narrower talonid and smaller cusps on the trigonid. Characteristically, the paraconid occurs at approximately the same height as the metaconid, and, in this regard, more closely approaches the condition on Ml. A premolariform bears only two cusps, a protocone and paracone. A semimolariform is more like Mi in possessing a metacone as well. The recent reclassification of mammals (McKenna, 1975) and its raison d’etre pertains to this and other studies of mammalian phylogeny, especially those which are based almost exclusively on dental remains of fossil mammals and mammalian dental homologies. According to McKenna (1975), the primitive tokothere post-canine dental formula is dPjPlPjP^dPjMlMf, in which dP^MjM^ have conventionally been referred to as MJM2M3, respectively. Recent suggestions of the occurrence of five premolars in Litolestes-WkQ eri- naceids, Plagiomene-like dermopterans (Schwartz and Krishtalka, 1976), and some plesiadapiform-tarsiiform 6 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I Fig. 1. Geographic location of localities of early Tertiary insectivores described in this paper. 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 7 primates (Schwartz and Krishtalka, 1976; Schwartz, in press) — groups that also have three molars— may require an emendation of the dental formula proposed as primitive for tokotheres. A moot point is the homology of the fourth tooth in the lower jaw — the alleged “premolariform canine” — in Litolestes ignotus and Plagiomene multicuspis. This problem is discussed in detail in Schwartz and Krishtalka (1976) and in the sections in this paper dealing with Litolestes and Ankylodon. The use of the conventional PI4 in this paper is a nomenclatorial compromise. Few of the known early Tertiary erinaceomorphs are represented by adequate dental remains necessary to establish dental homolo- gies. Litolestes and Ankylodon are exceptions, although the deciduous or permanent nature of each of the putative five premolars is uncertain. Also uncertain are the homologies of the premolars in those erinaceo- morphs that have four, and the loci involved in the reduction to four premolars from the suggested primi- tive complement of five. The abbreviations used in this paper are as follows : AMNH, American Museum of Natural History CM, Carnegie Museum of Natural History PU, Princeton University TTU, Texas Tech University UCM, University of Colorado Museum UCMP, University of California Museum of Paleontology YPM, Yale Peabody Museum L, length W, width AW, anterior width (width of trigonid) PW, posterior width (width of talonid) All photographs are stereoscopic scanning electron micrographs and present occlusal views. SYSTEMATICS Family Adapisoricidae (Schlosser, 1887) Schlosser (1887) erected this family of insectivores to include the European Paleocene genera Adapisorex (Lemoine, 1883) and Adapisoriculus (Lemoine, 1885). Simpson (1928) added Entomolestes and Leipsanolestes to this group and relegated Adapisoriculus to the mar- supials, although the latter has since been identified as a tupaiid (Van Valen, 1965a). Later, Simpson (1945) abandoned the concept of the Adapisoricidae and regarded Adapisorex as a leptictid, Entomolestes as an erinaceid, and Leipsanolestes as a subgenus of Lep- tacodon. Van Valen (1967) revived and broadened the Adapisoricidae to include four subfamilies (Adapis- orcinae, Geolabidinae, Nyctitheriinae, Creotarsinae), but acknowledged that many of the generic allocations were tentative. The nyctitheres (Robinson, 1968a) and geolabidines (Butler, 1972) were subsequently removed from the Adapisoricidae and raised to familial rank, an action with which Novacek (in press) and I agree. The Creotarsinae, as Van Valen (1967) acknowledged, is not a natural group. Of the genera originally assigned by Van Valen to the Adapisoricidae, only the following (in addition to Ankylodon) are here recognized as belonging to that family: McKennatherium (yncXndmg Leptacodon ladae), Scenopagus, Talpavus, and Macrocranion (including Entomolestes nitens, Aculeodens, and Messelina). Ses- pedectes and Proterixoides may also belong here or in the Erinaceidae. Both genera are currently under investigation elsewhere (Novacek, MS). Butler (per- sonal commun., 1975) and I believe that the genus Adapisorex, upon which the family is based, is an erinaceid. Diacodon minutus may be a condylarth, and is discussed in detail in this paper. A number of European forms from the Paleocene and Eocene (Amphilemur, Gesneropithex, Amphidozo- therium, Adunator and Paschatherium) that previously had been referred by some authors to the Adapisorici- dae have recently been shown to lack affinity with this family of insectivores (Russell, et al., 1975). McKennatherium Van Valen, 1965b McKennatherium libitum, named by Van Valen (1965b) as a new genus and species of paromomyid primate, was later shown to be a junior synonym of Leptacodon ladae (Szalay, 1968). Van Valen (1967) correctly suggested that L. ladae and the genotype of Leptacodon, L. tener, warranted generic distinction, a point overlooked by Russell, et al. (1975) in their review of North America adapisoricids. Accordingly, McKennatherium is a valid genus, although not a primate. Its systematics have never been formalized, nor its relationships described. M. ladae is here identified as a primitive Torrejonian adapisoricid, very near the ancestry of Scenopagus. 8 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I McKennatherium ladae (Simpson, 1935), new combination (Fig. 2; Table 1) Leptacodon ladae Simpson, 1935 McKennatherium libitum Van Valen, 1965b. type: USNM 9640, RP4-M3, Gidley Quarry, Fort Union Formation, Montana. REFERRED SPECIMENS: AMNH 35954, RP3-M3; and tentatively PU 14774, LP3-M1; PU 14776, RP2, P4, M2.3; PU 14780, LP4-M2; PU 17722, LP3-M3; PU 18500, LP3-M3; PU 19836, RP4-M2. localities: Gidley Quarry, Fort Union Formation, Montana; Rock Bench Quarry Beds, Polecat Bench Formation, Wyoming. KNOWN distribution: Torrejonian of Montana and Wyoming. description; P2 and P3 are double rooted. P2 is essentially a tall, asymmetrical, elongated cusp with a long, concave, posterior slope and a shorter, convex, anterior slope. A minute cuspule occurs at the base of the posterior slope, and a tiny lingual ridge at the base of the anterior one. P3 is dominated by a high, nearly symmetrical protoconid. A tiny paraconid occurs on the anterior part of the base of the protoconid. The talonid is very short, and consists of a single cuspule that is separated from the posterior face of the trigonid by a labiolingual groove. P4 of M. ladae is semimolariform, with a tall protoconid, a slightly lower metaconid on its lingual face, a low, anterobasal paraconid, and a basined talonid. The talonid is approximately one-half of the width of the trigonid, bears three cusps, and has a straight posterior margin. The paraconid varies in size and in the degree of separation from the antero- basal part of the protoconid. The lower molars of M. ladae are very Scenopagus- like, especially resembling S. curtidens. Character- istically, the paraconid is compressed to a strong lophid that extends lingually from the protoconid, and is appressed to the anterior face of the metaconid. The paraconid and the metaconid are closer, and the anteroposterior compression of the trigonid is greater on M2-3 than on Mj. The protoconid and meta- conid are conical, almost bulbous cusps. The talonid is approximately as wide as the trigonid on Mi, but extends less labially on M2-3. The hypoconulid is medial, and on M3 projects posteriorly from the elongated talonid. The hypoconid wears flat on M1-3, the entoconid remains high, and the cristid obliqua meets the trigonid labial to the protoconid-meta- conid notch. M3 is reduced in M. ladae as it is in S. curtidens and S. priscus. remarks: The molars on the Princeton specimens are slightly wider than on AMNH 35954 and are tentatively referred to M. ladae, until a larger sample is available. M. ladae clearly differs from Paleocene Leptacodon sensu stricto (i.e., L. tener, L. packi and L. tnunusculuni) in the lophid-like paraconid and compressed trigonid on the lower molars, the more robust, bulbous cusps, a more external cristid obliqua, and a hypoconid that becomes flat with wear. A new and larger species of McKennatherium, based on material in the Princeton collection from Rock Bench, Polecat Bench Formation, Wyoming, will be described elsewhere. The molars of McKennatherium are virtually identi- cal in cusp morphology to those of Scenopagus. How- ever, P4 in McKennatherium is semimolariform, where- as that of Scenopagus is premolariform. Premolariza- tion of P4 may be a characteristic trend in lineages of erinaceomorphs (Clemens, 1973), and is implied in the Table 1 . Dimensions of lower teeth of McKennatherium ladae P4 Ml M2 M3 L W L AW PW L AW PW L AW PW AMNH 35954 1.5 0.9 1.6 1.2 1.2 1.5 1.2 1.2 1.6 1.1 0.9 PU 14774 1.8 1.1 1.7 1.4 1.3 — — — — — — PU 14776 1.6 1.0 — — — 1.7 1.3 1.4 1.7 1.3 1.0 PU 14780 1.6 1.1 1.7 1.3 1.3 1.6 — 1.4 — — — PU 17722 1.6 1.0 1.7 1.3 1.3 1.7 1.4 1.3 1.6 — 0.9 PU 18500 1.7 1.1 1.7 1.4 1.4 1.7 1.5 1.4 1.8 1.3 1.0 PU 19386 1.6 1.0 1.7 1.2 1.2 1.8 1.4 1.3 — — — Mean 1.63 1.03 1.68 1.30 1.28 1.67 1.36 1.33 1.68 1.23 0.95 Fig. 2. McKennatherium ladae. (A) AMNH 35954, RP3-M3, Fort Union Formation, Montana: approx. X 10; (B) PU 17722, LP3-M3,^ Polecat Bench Formation, Wyoming, approx. X 12. Fig. 3. “Diacodon” minutus, PU 19395, RP3-M3, Polecat Bench Formation, Wyoming, approx. X 12. 10 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I transition from McKetmatherium to Scenopagus. McKennatherium appears to be a primitive Torrejonian representative of an adapisoricid lineage that included the species of the Eocene Scenopagus and Ankylodon. “ Diacodon'' minutus Jepsen, 1930 (Fig. 3) The generic and familial status of this species is in doubt, having been referred to the leptictids Diacodon (Jepsen, 1930) and Palaeictops (Gazin, 1956), and to Leptacodon, ?Metacodontidae (McKenna, 1960a). Some investigators (C. B. Wood; K. Rigby, personal commun., 1974, 1975) have noted that “D.” minutus is similar in some respects to hyopsodontid con- dylarths. The type of minutus was recovered from the Silver Coulee Beds, Polecat Bench Formation, late Tiffanian, of Wyoming. Additional material referrable to this species is known only from the type locality and the older, early Tiffanian, Cedar Quarry beds of the Polecat Bench Formation. Diacodon'" minutus is not a species of the leptictids Diacodon or Palaeictops or of the nyctitheriid Lepta- codon. Its lower dentition is virtually identical to the type of Adunator lehmanni (Wa 368) from the Tiffanian Walbeck fauna of Europe (Russell, 1964), an observa- tion also noted by Russell and C. B. Wood (personal commun., 1975). However, the known parts of the upper dentition of “D.” minutus are not quite as similar to those of Adunator. Relative to Adunator, Mi'2 of “Z>.” minutus are more transverse, with taller, more nearly conical cusps, a deeper labial emargina- tion, and a weaker hypocone. A problem inherent in this comparison is the degree of certainty of associa- tion of the upper and lower dentitions in each of the two taxa. In the original description of “D.” minutus, Jepsen (1930) stated that the remains of the upper and lower dentitions were associated, inasmuch as they were found inches apart. The association of upper and lower teeth of Adunator is less convincing (Russell, 1964). Although Russell (1964; in Russell, et al., 1975) originally identified Adunator as a leptictid, his recent restudy of the genus indicates that its affinities lie more with the condylarths (C. B. Wood, personal commun., 1975) — a conclusion I reached indepen- dently concerning minutus (Krishtalka, 1975). The lower molars of Adunator-" D." minutus are also similar to those of the adapisoricid McKenna- therium ladae, except for a number of moot features: the talonid on M1-2 is slightly elongate, and M2 is larger than Mj. The paraconid on M2 is extremely compressed and is merged with the anterior face of the metaconid. The metaconid, in turn, is greatly expanded anteriorly, in lingual view, so that it is much larger and more bulbous than the protoconid. These exceptions to adapisoricid affinities in the lower molars of Adunator- " D.’’’’ minutus are features associated with some Tiffanian hyopsodontid condylarths like Hapla- letes. However, the molars in Haplaletes have much lower crowns and are more nearly bunodont and robust. P4 in Adunator- " D." minutus is submolariform, with a well-developed paraconid as high as the metaconid. As such, P4 differs from the semimolariform P4 of McKennatherium, and is closer to P4 of Haplaletes. Similarly, the paraconid is strong on P3 in “D.” minutus, weaker in Haplaletes, and absent from P3 of McKennatherium. On the basis of the lower denti- tion alone, Adunator-"" D.’’’’ minutus appear to be adapisoricid-like hyopsodontids, and may be con- generic. Scenopagus McKenna and Simpson, 1959 After McKenna and Simpson (1959) described Scenopagus mcgrewi as a new erinaceid insectivore, McGrew (1959) and McKenna (1960a) noted that a number of the species included by Matthew (1909) in Nyctitherium warranted generic separation and bore specific resemblance to Talpavus and Scenopagus. Following these suggestions, Robinson (1966, 1968a) referred Nyctitherium priscum and N. curtidens to Scenopagus priscus, and arrangement accepted by Russell, et al. (1975). Earlier, Robinson {in McKenna, et al., 1962) had advocated the synonymy of the leptictid Diacodon edenensis (McGrew, 1959) with Scenopagus mcgrewi and combined the two as S. edenensis. Thereafter, Scenopagus material recovered from Bridgerian and Late Wasatchian localities (Robinson, 1966; McGrew and Sullivan, 1970; West, 1973) was referred either to the larger S. edenensis or to the smaller S. priscus. Examination of the sample of Scenopagus from Powder Wash and the Bridger Formation now allows clearer definition of the species of Scenopagus and the ancestry of the late Eocene and Oligocene Ankylodon. Scenopagus edenensis (McGrew, 1959) (Fig. 4; Tables 2, 3) type: AMNH 55685, P4-M3, Tabernacle Butte, Bridger For- mation (Upper), Wyoming. REFERRED SPECIMENS! P"*: TTU-P- 7145, 7146, 7147, 7148, 7149. MU TTU-P- 7150, 7151, 7152, 7153, 7154, 7155, 7156, 7157, 7158, 7159, 7175. M^: TTU-P- 7170, 7171, 7172, 7173, 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE II 7174, 7176, 7177, 7178, 7179, 7167, 7168, 7169. M^: TTU-P- 7190, 7191, 8232, 8233, 8234. DP4: TTU-P- 8235, 8236, 8237, 8238, 8239, 8240, 8241, 8242, 8243. M, : TTU-P- 7161, 7162, 7163, 7164, 7165, 7166. M2: TTU-P- 7180, 7181, 7182, 7183, 7184, 7185, 7186, 7187. M3: TTU-P- 7188, 7189. CM 13739, LP4-M1; CM 13732, RM2-3; CM 13734, RP2M1; CM 13728, RM2.3; CM 6449, RM1.3; CM 13731, RMi; CM 13740, RM2; CM 6444, LP3-M1; CM 6441, RP4, M,; CM 17318, RP3-M1; CM 13737, LP4; CM 13730, RP2, P4-M2; CM 13729, LM1.2; CM 13735, LP4; CM 6485, RP2-M2; CM 13733, LM2; CM 13736, M3; CM 13738, RP4; CM 6443, RM2.3; CM 17317, RM2.3; CM 13741, P4-M2; CM 6433, P4-M2; YPM 13612, LM3; YPM 16910, RM2-3; YPM 14934, LM,_2; YPM 14932, RM2; YPM 16911, RMi;AMNH 11428, RM1.2; TTU-P- 5571, RM1.2; TTU-P- 5572, LM2.3. localities: Powder Wash, Green River Formation, Utah; Dry Creek, Lone Tree, Henry’s Fork, Phi! Mass Ranch, Bridger Formation (Upper), Wyoming; Grizzly Buttes, Black’s Fork, Bridger Formation (Lower), Wyoming; Lost Cabin, Wind River Formation, Wyoming; locality 111, Huerfano Formation, Colorado. KNOWN distribution: Bridgerian, Wyoming; late Wasatchian, Wyoming and Colorado. description: Much of the lower antemolar denti- tion is preserved on CM 13730, CM 6485, CM 17318, and CM 13734, partial mandibles from Powder Wash. Pi and P2 are single-rooted, essentially unicuspid teeth, with an incipient paraconid on the antero-lingual face of the protoconid, and a tiny cingular heel. P3 is double-rooted and lacks a metaconid. The paraconid is not a distinct cusp, but a flange projecting from the anterolingual part of the base of the large rounded protoconid. A paracristid extends from the apex of the protoconid to the paraconid flange. The talonid, which is extremely short, is developed as a weak labiolingual ridge separated by a narrow groove from the posterior wall of the trigonid. A tiny cuspule (?hypoconid) occurs on the talonid ridge and is linked to the trigonid by a faint crest (?cristid obliqua). P4 is premolariform, with a prominent trigonid and a shorter, unicuspid talonid. The base of the talonid is wider than the trigonid, but the dorsal width of the talonid, from the cristid obliqua to the internal edge of the crown, spans only the lingual one-third of P4. The protoconid is the major cusp of the trigonid. A small, conical paraconid arises from the anterolingual part of the base of the protoconid. The two cusps are linked by a paracristid, and share a continuous, curving, labial-anterolabial face, but are demarcated lingually by a shallow, vertical notch. The metaconid, on the lingual face of the protoconid, is approximately one-half the height of the latter and slightly higher than the paraconid. The cristid obliqua is short and extends directly posteriorly from the trigonid to a small hypoconid. The entocristid slopes downward, lingually and forward from the hypoconid to the posterolingual part of the base of the trigonid, and completes enclosure of the shallow-basined talonid. Characteristically, a deep furrow occurs on the ex- ternal wall of the talonid from the cristid obliqua to the base of the crown. The lower molars of S. edenensis lean lingually and are essentially rectangular in occlusal outline, with a slight constriction between the talonid and trigonid. The trigonid, approximately one-third higher than the talonid, is wider than the talonid on M2-3 but is as wide as, or narrower than, the talonid on Mj. M3 is unreduced, has an elongate talonid, and is longer than Ml or M2. The protoconid and metaconid are high, rounded cusps, barely compressed anteropos- teriorly, with the metaconid slightly larger and taller. The paraconid is not a distinct cusp. Rather, it is compressed into a strong, broad paracristid that runs lingually from the anterior part of the base of the protoconid and ends anterior to and merged with the base of the metaconid. Characteristically, in lingual view the notch between the paracristid and metaconid is much higher than the talonid notch at the junction of the entocristid and base of the metaconid. In Macrocranion { = Entonwlestes nitens) the paracristid notch is only slightly higher than, or at the same level as, the talonid notch. The trigonid on Mi of S. edenensis is less compressed anteroposteriorly than on M2-3, and is more nearly triangular. The talonid basin on M1-3 is deep, the hypoflexid shallow. The cristid obliqua is strong and meets the trigonid wall labial to the ventral notch between the protoconid and metaconid. The entoconid is the highest and sharpest of the talonid cusps and is rarely worn, whereas the hypoconid is lower and more nearly rounded, becoming flat with wear. The hypo- conulid, smaller than, and distinct from, the entoconid and hypoconid, occurs only slightly lingual to the midline, and on M3, projects posterodorsally. Two mental foramina occur in S. edenensis: one below P4 and the other below the anterior root of P3. P^ through M^ of S. edenensis have been adequately described elsewhere (McKenna and Simpson, 1959). Isolated M^s identified from Powder Wash are nar- rower and shorter than M^, possess a strong, flaring parastylar salient, and lack a hypocone. As on Mi-2, the protocone on M3 is compressed anteroposteriorly, and the conules are evident, but fairly small. The buccal edge of the crown, oriented anterolabially, is defined by a weak ectocingulum, and, unlike MI-2, is n A 12 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 13 not emarginate. The paracone, much larger and higher than the extremely reduced metacone, is linked to the tip of the parastylar salient by a weak preparacrista. Small, short anterolingual and posterolingual cingula occur basal to the protocone. A number of isolated DP'^s have tentatively been identified from the Powder Wash collection, based on morphological and size association with known parts of the dentition of S. edenensis. DP^s thought to pertain here are molariform teeth, triangular in occlusal view and much longer buccally than lingually, relative to P“^, Mi-2. The paracone and metacone are rounded and conical, with the paraconid slightly taller and larger. The protocone and the conules are sharp, crescentic cusps with equally distinct proto- cristae and conulecristae. As on Ml, the ectoflexus is shallow, and the posterior cingulum is developed as a broad shelf that terminates in a hypocone lingual and basal to the protocone. However, the hypocone is much smaller relative to Mi-2, and the lingual contour of the crown is continuous. On Mi-2 of 5. edenensis, a valley occurs lingually between the bases of the protocone and hypocone. A small parastyle occurs on the parastylar salient of DP'i, which juts anteriorly beyond the remaining anterior margin of the crown. As on P4, Mi-2, the post-metacrista is developed as a high crest running from the metacone to the tip of the expanded metastylar salient. A preparaconulecrista extends to the parastyle and de- fines a narrow paracingulum along the base of the paracone. Similarly, the postmetaconulecrista deline- ates a strong metacingulum along the posterior part of the base of the high postmetacrista. remarks: The distinctions between S. edenensis, S. prisons, and another species, S. curtidens, are dis- cussed in the next two sections. The large sample of S. edenensis from Powder Wash (early Bridgerian) exhibits a range of variation that excludes AMNH 56035, holotype of “S. mcgrewi” (late Bridgerian), and raises the question of the validity of the synonymy of S. mcgrewi and S. edenensis (see Table 3). Along with AMNH 56035, left maxilla with P3-M2, McKenna and Simpson (1959) tentatively referred AMNH 56034, partial left lower jaw with M3, to S. mcgrewi. Both specimens were recovered from the same excavation in the Bridger Formation, near Tabernacle Butte. The authors contended that the lower jaw with M3 was suitable in size and structure for occlusion with the maxilla. In the same paper, Simpson noted that the holotype of Diacodon edenensis (McGrew, 1959), a lower jaw with M1-3, resembled Scenopagus more than the leptictid Diacodon, although its occlusal fit with the holotype maxilla of S. mcgrewi was poorer than that of the referred jaw of S. mcgrewi. Robinson {in McKenna, et al., 1962) concurred with Simpson concerning the affinities of Diacodon edenensis and referred that species to Scenopagus. However, Robinson discounted as individual variation Simpson’s observed differences between the lower jaw of S. mcgrewi and D. edenensis, and synonymized all the material as Scenopagus edenensis. The large sample of S. edenensis from Powder Wash makes it difficult to attribute to individual variation alone the rather large difference in size between the type maxilla of S. mcgrewi and the upper limit of the size range of the upper dentition of S. edenensis, although the tentatively referred lower jaw of S. mcgrewi (AMNH 56034) falls within the upper limit of the range in size of M3 of S. edenensis. Two alternative answers are suggested: The tentatively referred lower jaw of S. mcgrewi is actually S. edenen- sis, and AMNH 56035 remains the type and only known specimen of S. mcgrewi’, or the large range in variation of the lower dentition of S. edenensis is artificial and actually represents two distinct popu- lations. There is some basis for the latter suggestion. The type material of S. mcgrewi and S. edenensis was recovered from deposits in the upper part of the Bridger Formation. The referred sample of S. edenensis came from the early Bridgerian Green River Forma- tion. Examination of the measurements of lower denti- tions referred by Robinson (1966) and West (1973) to S. edenensis shows that specimens recovered from the Upper Bridger deposits fall within the upper part of the size range, whereas those from the Lower Bridger horizons, as well as those from the temporally equiv- alent Powder Wash locality, make up the lower part of the size range. Thus, the Upper Bridger material may represent typical S. edenensis, whereas that from lower Bridgerian and equivalent localities may repre- sent a smaller population of S. edenensis, or a smaller species of Scenopagus. Unfortunately, almost all re- mains of upper dentition referrable to Scenopagus have been recovered from early Bridgerian horizons in Wyoming and Utah. Apart from the type maxilla of S. mcgrewi, only West (1973) has reported upper teeth of Scenopagus from Upper Bridger localities. •^Fig. 4. Scenopagus edenensis. Green River Formation, Utah. (A) CM 17318, RP3-M1, approx. X 12; (B) CM 6449, RM1.3, approx. X 12; (C) CM 6433, LP4-M2, approx. X 14. 14 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 and these correspond well to the size range of upper teeth from Powder Wash that are here identified as S. edenensis. Except for size, all material from early and late Bridgerian horizons referred here and else- where to Scenopagus edenensis and S. mcgrewi is structurally identical. Consequently, all the material is referred to S. edenensis, although we may be dealing with a series of temporal subspecies from the Late Wasatchian through the Late Bridgerian. Scenopagus priscus {yidiTsh, 1872) (Fig. 5; Tables 2, 3) type; YPM 13509, Bridger Formation, Wyoming. REFERRED SPECIMENS: P^: TTU-P- 7051; Ml; TTU-P- 7060, 7078, 7075, 7074, 7077, 7062, 7073, 7080, 7076, 7059, 7057, 7053, 7066, 7068, 7065, 7055, 7069, 7071, 7056, 7072, 7079, 7064, 7067, 7070, 7063, 7061, 7058; CM 30784; M^: TTU-P- 7101, 7102, 7116, 7112, 7106, 7109, 7107, 7115, 7108, 7113, 7110, 7114, 7103, 7104, 7111, 7105, CM 30783; Mi: TTU-P- 7094, 7098, 7099, 7097, 7096, 7100, 9082, 7092, 7093, 7081, 7083, 7085, 7090, 7095, 7087, 7086, 7088, 7120, 7121, 7119; M2: TTU-P- 7124, 7125, 7117, 7118, 7089, 7122, 7129, 7131, 7132, 7130, 7126, 7127; M3: TTU-P- 7128, 7123, 7137, 7140, 7138, 7139, 7135, 7142, 7134, 7141, 7143, 7136. CM 13753, LP4-M1; CM 13749, LM,.2; CM 13750, RP4-M1; CM 13744, P4; CM 17316, M2.3; CM 13746, LMi; CM 13743, LM2.3; CM 13748, M1.2; CM 6486 RP4-M2; CM 17319, LM2.3; CM 13751, LP4-M2; CM 13752, LP4-M1; AMNH 56060, LMjo; AMNH 55156, LP4-M1. localities: Locality III, Huerfano Formation, Colorado; Powder Wash, Green River Formation, Utah; Henry’s Fork and Hyopsodus Hill, Bridger Formation (Upper), Wyoming. KNOWN distribution: Late Bridgerian, Wyoming; early Bridgerian, Utah; late Wasatchian, Colorado. DESCRIPTION AND REMARKS! S. priscus closely re- sembles S. edenensis in known parts of the dentition, but is much smaller and bears a number of diagnostic differences in cusp morphology. On P4 of S. priscus, the paraconid is flange-like and the talonid is usually narrower than the trigonid. M3 of S. priscus is much reduced relative to M2, the ectoflexus on M2 is deep, and the lingual groove between the bases of the hypocone and protocone on M1-2 is weak. In contrast, the paraconid on P4 is more nearly cuspate and dis- tinct in S. edenensis, and the talonid is generally wider than the trigonid. M3 of S. edenensis is unreduced, but elongate relative to M2. The labial emargination of M2 is relatively shallow, and the lingual indentation between the bases of the hypocone and protocone on Mi'2 is pronounced. McGrew and Sullivan (1970) first described putative upper teeth of S. priscus. Although these teeth from Bridger A are very similar to upper molars of S. priscus from Powder Wash, they are significantly larger and are here referred to S. curtidens, a species of Scenopagus described below. Apart from size, the Mis and M2s discussed by McGrew and Sullivan differ from S. priscus in possessing a more pronounced indentation between the bases of the hypocone and protocone, and a shallower emargination of the labial border of the crown. AMNH 11488, previously referred to S. priscus (Robinson, 1966), is here and elsewhere (Russell, et al., 1975) identified as Talpavus nitidus. The latter authors also referred AMNH 55156 from Huerfano to Talpavus sp. In contrast to Talpavus, P4 on AMNH 55156 bears a basined talonid. I follow Robinson (1966) in retaining that specimen in S. priscus. The distinctions between Scenopagus (especially S. priscus) and Talpavus are described below in the section dealing with the latter genus. The type of S. priscus, listed by Robinson (1966: Table 4) as YPM 15309, should read YPM 13509. Scenopagus curtidens (Matthew, 1909), new combination (Fig. 6; Tables 2, 3) Nyctitherium curtidens Matthew, 1909 type: AMNH 12055, partial left mandible with M1.3, from Henry’s Fork locality, Bridger Formation (Upper), Wyoming. referred specimens: YPM 15254, RP4-M3; YPM 16913, LMi_2; YPM 14939, LM1.2; YPM 13610-a, RP4; YPM 13610-b, LP4-M2; YPM 13610-2, LM1.2; YPM 15255, LM2.3; YPM 16912, LP4-M1; YPM 16914, RM^; AMNH 12058, RM2.3; AMNH 12062, LM1.3; AMNH 12064, RP4-M1; AMNH 11491, RMi; AMNH 59643, RMi; AMNH 48183, RP4, M2; TTU-P- 3928, RM1.3; TTU-P- 7054, RMi; UCMP 96155, LM1.3. localities: Henry’s Fork, Twin Buttes, Lone Tree, and Dry Creek, Bridger Formation (Upper), Wyoming; North Leavitt Ranch, Bridger Formation (Lower), Wyoming; unknown locali- ties, Bridger Formation, Wyoming; Cattail Spring, and Point Gulch, Bridger Formation (Upper or Lower), Wyoming; Quarry 88, Almagre facies, San Jose Formation, New Mexico; Alheit Quarry, Hiawatha Member, Wasatch Formation, Colorado; Friars Formation, California. known distribution: Bridgerian, Wyoming; Wasatchian, New Mexico; early Wasatchian, Colorado; early Uintan, California. description: The type and referred specimens ex- hibit features characteristic of Scenopagus, but mor- phologically intermediate between S. edenensis and Fig. 5. Scenopagus priscus. Green River Formation, Utah (A) CM 13752, LP4-M1, approx. X 20; (B) CM 13749, LM1.2, approx. X 15;^- (C) CM 17319, LM2.3, approx. X 14; (D) CM 30784, RMU approx. X 12 (E) CM 30783, LM^, approx. X 12. Fig. 6. Scenopagus curtidens, Bridger Formation, Wyoming. (A) YPM 15254, RP4-M3, approx. X 12; (B) YPM 16914, RM^, approx. X 10. 15 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 16 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. I Table 2. Dimensions of lower teeth of Scenopagus edenensis, S. curtidens and S. priscus P4 Ml M2 M3 L W L PPN PW L AW PW L AW PW S. edenensis 2.0- 1.4- 2.1- 1.7- 1.7- 2.1- 1.8- 1.8- 2.2- 1.7- 1.2- 2.4 1.6 2.4 1.9 2.0 2.7 2.2 2.0 2.7 1.8 1.5 Mean 2.06 1.48 2.23 1.80 1.83 2.25 1.96 1.84 2.38 1.72 1.34 N. 10 9 13 13 12 22 19 19 19 9 9 SD. 0.8 0.12 0.09 0.16 0.22 0.14 0.11 0.35 CV. 3.93 5.78 5.07 9.01 9.79 7.41 5.97 14.79 S. curtidens 1.5- 1.1- 1.7- 1.4- 1.5- 1.7- 1.5- 1.4- 1.5- 1.2- 0.9- 1.6 1.2 2.0 1.6 1.6 1.9 1.7 1.6 1.7 1.3 1.2 Mean 1.52 1.18 1.82 1.49 1.53 1.77 1.58 1.46 1.60 1.26 1.04 N. 6 6 11 11 14 12 11 12 5 7 5 SD. 0.07 0.08 0.07 0.06 0.10 0.12 CV. 3.88 6.00 4.78 3.79 6.32 8.24 S. priscus 1.2- 0.9- 1.3- 0.8- 0.9- 1.4- 1.1- 0.8- 1.3- 0.8- 0.6- 1.4 1.1 1.5 1.3 1.3 1.6 1.3 1.2 1.5 1.1 0.8 Mean 1.30 0.96 1.43 1.04 1.11 1.45 1.20 1.04 1.33 0.95 0.72 N. 6 7 31 31 29 18 18 17 15 15 12 SD. 0.07 0.07 0.06 0.06 0.07 0.03 0.04 0.07 0.05 CV. 5.40 7.23 5.87 4.09 6.39 3.39 3.48 7.92 7.25 S. priscus. As in Scenopagus, the paraconid and meta- conid on the premolariform P4 are lower than (and merged with) the protoconid, and the talonid is essen- tially unicuspid, short and basined. On M1-3 the para- conid is compressed into a low, broad paracristid, and the metaconid and protoconid are large, rounded cusps, with the metaconid slightly taller. The cristid obliqua originates labially on the posterior wall of the trigonid. The hypoconid, lower than the high, conical entoconid, becomes flat with wear and the hypoconulid occurs lingual to the midline. The talonid extends farther labially than the trigonid on Mi, but is nar- rower than the trigonid on M2-3. Two mental foramina occur on the mandible, one below P4 and P3, respec- tively. S. curtidens is intermediate in size between S. priscus and S. edenensis, but is closer to S. priscus in the smaller, less cusp-like nature of the paraconid on P4, a talonid narrower than the trigonid on P4, and a reduced M3 relative to M2. Also, the single M2 re- ferred here to S. curtidens, and the upper molars de- scribed by McGrew and Sullivan (1970) as S. priscus, are virtually identical to upper molars of S. priscus, except that they are larger and possess a deeper valley between the lingual faces of the hypocone and proto- cone, and a shallower labial emargination. YPM 13610-a, 13610-b, and 14939, previously re- ferred to S. priscus (Robinson, 1966) are here placed in S. curtidens. remarks; S. edenensis, S. curtidens, and S. priscus were contemporaneous during the Bridgerian and Late Wasatchian. Only S. curtidens has been recognized from earlier deposits. The reduction of the M3 in S. curtidens and S. priscus may be a derived feature, rela- tive to the retention of an unreduced M3 in S. edenen- sis. With premolarization of P4, the species of Sceno- pagus could have radiated from a Paleocene adapisori- cid like McKennatheriwn ladae. Ankylodon The genus Ankylodon is known from the Late Eocene (Setoguchi, MS thesis) and Oligocene [(Galbreath, 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 17 Table 3. Dimensions of upper teeth of “Scenopagus mcgrewi, ” S. edenensis, S. curtidens and S. prisms p4 Ml m2 m3 L W L W L W L W ‘S. mcgrewi” 2.8 3.3 2.8 3.7 2.5 3.9 est. est. S. edeiiesnsis 2.0 2.6- 2.1- 3.0- 2.0- 3.1- 1.5- 2.6- 2.9 2.4 3.3 2.2 3.5 1.9 2.8 Mean 2.0 2.69 2.25 3.18 2.10 3.31 1.67 2.73 N. 6 7 8 11 6 7 3 3 SD. 0.09 cv. 2.66 S. curtidens 1.5 2.5 S. priscus 1.4 2.0 1.4- 1.9 1.1- 1.9- 1.6 2.2 1.4 2.2 Mean 1.47 2.06 1.32 2.09 N. 1 1 15 17 6 7 SD. 0.07 0.07 CV. 4.81 3.64 1953; Lillegraven and McKenna, (MS); Patterson and McGrew, 1937)] of North America. Excellent material comprising most of the upper and lower dentition of a Chadronian species of Ankylodon is presently being described by Lillegraven and McKenna (MS). The authors have kindly allowed me access to the manu- script and casts of the material. The dental morphology of Ankylodon mirrors that of Scenopagus as follows; Pj and P2 are single-rooted and unicuspid. P3 and P4 are double-rooted, possess a large protoconid, a small paraconid, and a short, shallow-basined talonid with a tiny hypoconid. A metaconid is lacking on P3 and is present, but lower than the protoconid on P4. The talonid is wider than the trigonid on Mj but narrower on M2-3. On M1-3 the paraconid is anteroposteriorly compressed into a strong low paracristid, the meta- conid is taller and larger than the protoconid, the cristid obliqua is labial on the posterior wall of the trigonid, the hypoconid is lower than the high conical entoconid, and becomes flat with wear, and the hypo- conulid occurs slightly lingual to the midline. On M3 the talonid is elongate, and the hypoconulid projects posterodorsally. As in Scenopagus, P^ of Ankylodon lacks conules and bears a tall, central paracone, a lower lingual protocone, a small hypocone on the postprotocrista, a small anterior parastylar salient, a large metastylar expansion, and a high postmetacrista. The crown is essentially T-shaped with an anteroposteriorly ex- panded lingual end. M1-2 are very transverse, with a large valley separating the bases of the hypocone and protocone lingually. The postcingulum is strong and shelf-like and bears a large hypocone. The postmeta- crista is high, the buccal margin of the crown is deeply emarginate medially, and the stylar salients are ex- panded. M3 lacks a hypocone, is smaller than M2, and possesses a large paracone and an extremely reduced metacone. As Russell, et al. (1975) also noted, the features that distinguish Ankylodon from Scenopagus are not unique, but represent the elaboration of certain characters in Ankylodon beyond the condition seen in Scenopagus. Compared to Scenopagus, Pi of Ankylodon is relatively higher, but P2 is reduced. P3 possesses a taller proto- conid, a more distinct paraconid, and a longer, wider, better-developed talonid. Similarly, on P4, the proto- 18 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 conid and metaconid are higher and more conical. The paraconid, although not cuspate, is much larger and projects more anteriorly from the base of the trigonid, and the talonid is more elongate. On M1-2 of Ankylodon, the trigonids are more anteroposteriorly compressed, the protoconid and metaconid are taller, sharper cusps; the talonid is longer, the hypoflexid is deeper, and the cristid obliqua is more arcuate. On known parts of the upper dentition, Ankylodon differs from Scenopagus as follows: P^, much more trans- verse (buccolingually) and elongate (anteroposte- riorly), with longer parastylar and metastylar salients, resembles a “T” instead of a triangle in occlusal out- line, and is more nearly semimolariform. The paracone on P3 is much larger and taller, the protocone is more prominent, and a small metacone is present. On P^, the lingual half of the crown is longer (anteroposte- riorly), the protocone and paracone are taller, more spire-like, and the postmetacrista is higher and bears a tiny metacone. On Mi'2 the hypocone and post- cingulum are considerably enlarged, the lingual fur- row between the hypocone and protocone is deeper and wider, the conules are larger, and the buccal edge of the crown is more emarginate. The metastylar salient is more expanded on Ml, the parastylar area is greater on M2, and on M^ the parastylar salient is greatly reduced relative to Scenopagus. In effect, features that are diagnostic of Scenopagus are equally descriptive of Ankylodon, except for larger and higher, more nearly conical cusps, sharper, broader cingula, larger metastylar and parastylar salients on Ml and M2, respectively, and greater anteroposterior compression of the trigonid on M1-3. These differences between Bridgerian Scenopagus and Uintan- Chadronian Ankylodon may be a consequence of a continuation of similar trends in the same dental characters seen in the evolution of Wasatchian- Bridgerian species of Scenopagus from the Paleocene McKennatherium. Butler (1972) suggested, and I con- cur, that Ankylodon is the latest known adapisoricid. Comparison of Ankylodon with Scenopagus indicates that the ancestry of Ankylodon lies among the Bridgerian species of Scenopagus — a conclusion that Robinson and McKenna have also come to inde- pendently (McKenna, personal commun. 1975). The lower dentition of Ankylodon, completely pre- served on the AMNH “dark” specimen (Lillegraven and McKenna, MS), consists of three molars and seven antemolar teeth. As traditionally interpreted, the lower dental formula is I2-3C1P1-4M1-3. How- ever, the homologies of the anterior teeth are un- certain, and alternate interpretations of the lower antemolar dental formula are possible. Since the tips of both I2 and I3 are broken off, the identification of these teeth as incisors is tentative. I2 is round and I3 is oval in cross section. Both teeth are enlarged and project anterodorsally. I3, larger than I2, is concave along the posteromedial and posterolabial face of the crown, and bears a posterobasal cuspule. Relative to I2, the morphology of I3 may be interpreted as caniniform. The third tooth in the jaw — the alleged Ci — is struc- turally identical to, and as premolariform as, the alleged Pi, and as such, is the first of five premolariform teeth. These identifications can yield a lower antemolar dental formula of either I3C1P1-5, or, if the two anterior lower teeth are incisors, l2-3Pi-5- Some of the premolars may be retained deciduous teeth, as has been postulated for other preptotheres (McKenna, 1975). A fourth possible interpretation of homologies, which also pertains to the discussion below of the erinaceid Litolestes ignotus, concerns the identification of the third tooth in the jaw — the premolariform “canine.” McDowell (1958: 157, Fig. 19A), has identi- fied the deciduous Ci in the insectivore Nesophontes as a premolariform, single-rooted tooth, which is very similar to the premolariform “canine” of Anky- lodon (and Litolestes ignotus). However, calling this tooth dCi in Ankylodon (in an antemolar dental formula of l2-3dCiPi-4) is premature, since the homol- ogy of this tooth locus in Nesophontes is perhaps also uncertain. The alleged permanent canine of Neso- phontes is not caniniform, but, like its milk predecessor, is premolariform (McDowell, 1958: 157, Fig. 19C), and may represent a permanent premolar with a deciduous premolariform predecessor. The lower ante- molar dental formual of Ankylodon is tentatively identified as: I2, I3 or Ci, dPiP2P3P4P5. Clearly, until more complete material of the anterior dentition of Ankylodon and other adapisoricids is recovered, the homologies of the anterior teeth remain unsettled. Macrocranion Weitzel, 1949 The genus Entomolestes was described by Matthew (1909) from the Middle Eocene Bridger Formation, with E. granger! as the type species. Subsequently, Matthew (1918) named a second species, Entomolestes nitens, from the Early Eocene Willwood Formation. Since then, definite and tentative occurrences of E. nitens have been recorded from the Wasatch Forma- tion, Colorado (McKenna, 1960a), the early Wasatch- ian Powder River local fauna, Wyoming (Delson, 1971), the Lysite and Lost Cabin Members, Wind 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 19 River Formation, Wyoming (Guthrie, 1967, 1971), the Wasatchian Almagre facies of the San Jose Forma- tion, New Mexico (Robinson, 1968b), and Sparnacian- Cuisian localities in France (Russell, et al., 1975). No additional material referrable to E. granger! has been recovered since its original description. Robinson (1968b) and Butler (1972) suggested that E. nitens was generically distinct from E. granger! and should be referred to Scenopagus. Guthrie (1971) disagreed, citing differences between E. mtens and Scenopagus in the structure of the upper molars, and retained the use of E. mtens. From examination of the type, paratype, and a large amount of material referred to E. mtens, it is clear that Robinson (1968b) and Butler (1972) were correct in observing the generic disparity between E. granger! and E. mtens. Perhaps the affinities of the former, as discussed in a later section, are with the Erinaceidae. Although P4 in both species is similarly premolariform, the talonid is much shorter on that of E. granger!. M1-3 of E. granger! have less robust cusps, comparatively shorter talonids with more nearly straight posterior margins, more triangular trigonids, reduced hypoconulids, and shallower talonid basins, compared to E. mtens. As described below, E. mtens is also distinct from Scenopagus, but is virtually identical to Macrocranton tenerum { = MesseUna) hitherto known only from the middle Eocene of Europe. Macrocranton was shown by Tobien (1962) to be a senior synonym of Aculeodens and was referred to the Amphilemuridae of McKenna (1960a), although Van Valen (1967) considered the genus to be a creotarsine adapisoricid. MesseUna tenera (Tobien, 1962) was recently, and I think cor- rectly, referred to Macrocranton (as M. tenerum) by Russell, et al. (1975). These authors have also com- mented on the remarkable similarity between E. mtens and Macrocranton, especially in the lower denti- tion. Except for the occurrence of a posterior cingulid on Mi-2 of tenerum, distinction between that species and E. nttens is difficult. Nonetheless, Russell, et al. (1975), claimed that the presence of a tiny metacone on P'* of E. nttens from Four Mile, as well as the greater transverse nature of Mi-3 compared to Macrocranton warranted generic separation of the two species. But, as described below, P^ of E. nttens from the Almagre does not possess a metacone. Although 1 have not examined the original material of upper molars of Entomolestes cf. nttens from Europe (Russell, et al., 1975), the illustrations imply that part of the hypodigm (notably, Mu-118-L, Av-841-Bn) may be referrable to Scenopagus, upper molars of which are indeed more transverse than those of E. nttens. M> of M. tenerum as figured in Tobien (1962) is similar in crown propor- tion to that of E. nttens. The reconstruction and figures of M2-3 of M. tenerum are apparently inaccurate (Russell, et al., 1975). E. nttens and E. cf. nttens from the early Eocene of North America and Europe are here considered species of Macrocranton. Macrocranion nitens (Matthew, 1918), new combination (Fig. 7; Tables 4, 5) type: AMNH 15697, partial right mandible with P4-M2 from the Willwood Formation, Big Horn Basin, Wyoming. paratype: AMNH 14674, partial right mandible with 1/2 P4- M3 from the Lysite Member, Wind River Formation, Wyoming. REFERRED SPECIMENS: AMNH 48187, RM3; AMNH 48174, RP4-M3; AMNH 48188, frag. P4; AMNH 48179, RP4-M2; AMNH 48189, LP4-M2; AMNH 48181, frag, talonid; AMNH 48176, LP4-M3; AMNH 48182, RM2; AMNH 48172, LM2.3; AMNH 48178, LP4-M3; AMNH 48180, RP4-M2; AMNH 48184, RM1.2; AMNH 48177, LP^-M^; AMNH 48175, RMi-2; TTU-P- 4211, RP4-M2; CM 12398, LMi-J; CM 22019, RM^; CM 22014, LM2-3. localities: Quarry 88, Almagre facies, San Jose Formation, New Mexico; Lysite and Lost Cabin Members, Wind River Formation, Wyoming; Willwood Formation, Big Horn Basin near Otto, Wyoming. KNOWN distribution: Wasatchian, New Mexico, Wyoming. DESCRIPTION : The description presented here is based on the holotype, the paratype, and the single quarry sample from the Almagre. Teeth anterior to P3 and P4 are unknown. P3 is a single-rooted, very small, peg-like tooth, with one laterally compressed, anterior- leaning cusp. P4 is variable in structure, more so than the lower molars, and more than has been noted in previous descriptions. It is a large, two-rooted premolariform tooth with a three-cusped trigonid and lower, shorter, talonid. The protoconid is large and robust, leans posteriorly, and occupies most of the basal area of the trigonid. The paraconid is low, conical, and arises from the anterolingual part of the base of the proto- conid. The two cusps are continuous labially, but are demarcated lingually by a weak vertical groove. A paracristid extends down the anterior face of the protoconid to the paraconid, joining the two cusps along the entire height of the paraconid. The meta- conid, slightly higher than the paraconid, but only one-third the height of the protoconid, occurs low on the posterolingual face of the protoconid, and varies in size from a small bulge (on the holotype AMNH 15697) to a conical cusp (on the Almagre 20 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I specimens). In the latter condition a V-shaped notch separates the metaconid and protoconid. The structure of the talonid of P4 is most variable. On all specimens examined, the talonid is lo\v, as wide as, or wider than, the trigonid basally, and approximately one-half as long. An extremely weak cristid obliqua extends directly posteriorly from the trigonid to a raised point (?hypoconid) on the posterior rim of the talonid, and delimits the labial margin of a shallow talonid basin. The width of the basin from the lingual margin of the talonid to the cristid obliqua represents only the lingual one-third of the total basal width of the talonid. A hypoflexid groove for occlusion of the paracone of P"* slopes very gently labioventrally from the cristid obliqua to the basal margin of the talonid. Since the cristid obliqua is very weak, it is almost obscured by wear on many of the specimens. On some of the material from the Almagre, an addi- tional cuspule occurs on the raised posterior rim of the talonid basin in the position of an entoconid. A precingulum, when present, is weak and short. These variations seem taxonomically insignificant, and in each case, the P4 is associated in a partial mandible with molars of M. nitens. The molars of M. nitens are large, robust teeth that lean lingually, and are almost rectangular in occlusal outline. The talonid is longer and only slightly lower than the trigonid on M1-3, and wider than the trigonid on Mj-2. a slight lateral constriction occurs basally between the trigonid and talonid. The protoconid and metaconid are low, conical, almost bulbous cusps, with the metaconid slightly larger and taller. A strong, broad paracristid runs anterolingually from the ante- rior face of the protoconid, and terminates well anterior to the metaconid, so that the trigonid is open lingually. The paraconid is a barely discernible bulge, anteroposteriorly compressed at the lingual end of the paracristid. The talonid on M1-3 bears a large, rounded ento- conid, almost as tall as the metaconid, and a much lower hypoconid that becomes flat with wear. On Mi-2, the hypoconulid is low, small, and occurs slightly lingual to the midline on the same wall as the entoconid. On M3 the hypoconulid is medial, distinct, as large as the entoconid, and projects posterodorsally. The cristid obliqua is long, broad, and strong, and originates directly below the notch between the proto- conid and metaconid. The hypoflexid notch is shallow, the precingulid is very weak, and a postcingulid does not occur. Characteristically, on Mi the labial part of the base of the trigonid bulges anterolabially and the paracristid is sharply angled at the anterior part of the base of the protoconid. This lends the Mi trigonid a very squared appearance in occlusal view, relative to the more nearly triangular trigonid on M2-3. Simi- larly, the base of the talonid below the hypoconid on Mi"2 is expanded labially and posteriorly. Upper dentitions of M. nitens, although noted from Lost Cabin (Guthrie, 1971) and from the Almagre (Robinson, 1968b), have never been adequately de- scribed. Two specimens, AMNH 48177 and AMNH 48175, comprising P4'M3 and Ml -2 respectively, occur in the single quarry sample of M. nitenshom the Almagre. Only the labial half of P"i is preserved on AMNH 48177. The broken crown bears a very large, tall, and conical paracone, a small anterolabial parastyle, and a short posterolabial metastylar salient. The paracone is approximately twice as high and large as that on Mi-2. There is no evidence of a metacone on P^, although a weak crest runs vertically along the poste- rior face of the paracone to the tip of the metastylar salient. The buccal margin of the crown is shallowly convex labially. Mi-2, essentially rectangular in occlusal outline, are not transverse and are only weakly constricted antero- posteriorly across the conules. The buccal borders of the crowns are shallowly emarginate and the lingual margins are oriented posterolingually. The parastylar and metastylar areas are small and not expanded, although the parastylar salient on Ml juts anteriorly beyond the remaining anterior margin of the crown. The paracone and metacone, low, subequal, conical cusps, are well separated and occur close to the labial border of the crown, leaving virtually no stylar area. The protocone is low, pyramidal, only slightly com- pressed anteroposteriorly, and forms the anterolingual corner of the crown. The hypocone, rounded and conical, is only slightly lower than, but almost as large as, the protocone, and forms the posterolingual corner of the crown. A shallow lingual valley between the protocone and hypocone divides the lingual length of the crown into approximately equal halves. A weak crista extends anterolabially from the hypocone to the anterior wall of the trigon below the preprotocrista. The conules are low, but well developed, conical and subequal, and are linked to the protocone by strong protocristae. The preparaconulecrista and postmeta- conulecrista are weakly developed, but extend to the parastyle and metastyle, respectively, and delineate a narrow paracingulum and metacingulum. The post- metacrista is weak, low, and not crest-like. A short narrow precingulum occurs on Mi‘2. 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 21 M3, triangular in occlusal aspect, has a greatly reduced metastylar area and a relatively tiny metacone, approximately two-thirds the size of that of M2. The paracone and parastylar salient are unreduced relative to M2. M3 lacks a hypocone, and the postcingulum is very short. remarks: Macrocranion nitens and Scenopagus are easily distinguished in known parts of the dentition, although P4 and the upper molars are similar in both forms. In contrast to Scenopagus, the metaconid on P4 of M. nitens is more nearly separate from the protoconid, the cristid obliqua is considerably weaker, the posterolabial part of the base of the talonid is expanded, the posterior rim of the talonid basin is raised, and the slope of the hypoflexid groove is lower. Compared to Seenopagus, M1-3 of M. nitens are lower, their bases are fattened labially (exodaeno- donty), the trigonid is much lower relative to the talonid, the talonid is proportionately wider and con- siderably longer than the trigonid, and the cusps are lower and more bulbous. In M. nitens, the paraconid is better developed, more anterior and not as close to the metaconid, the trigonid on Mj is open lingually and more nearly square in occlusal outline, the hypo- conid is less compressed anteroposteriorly, the cristid obliqua is longer and meets the trigonid more lingually, and the precingulid is weaker. P4 of M. nitens differs from Scenopagus in having a convex labial margin, very small stylar salients, and a much weaker, lower paracone-metastylar crest. The upper molars of M. nitens are less transverse, more nearly rectangular in occlusal outline, and less con- stricted across the conules. The labial margins of the molars are less emarginate, the stylar areas are much smaller and not expanded, and the cusps are lower and more nearly conical. In M. nitens, the hypocone is higher relative to the protocone and occupies more of the lingual area of the crown, the postmetacrista is much weaker and less crest-like, and the metacingulum is narrower. Macrocranion cf. M. nitens (Fig. 8; Tables 4, 5) REFERRED specimens: AMNH 59673, LP4-M3; AMNH 59662, RP4; AMNH 59656, RM2-3; AMNH 59663, RP4; AMNH 80071, RMi; AMNH 59679, RP4-M1; AMNH 59651, RP4; AMNH 59688, LP3-4; AMNH 88846, LMi; AMNH 88834, RMi; AMNH 88839, LM|; AMNH 88830, LM1.2; AMNH 56318, RP3.4; AMNH 59678, RP^-M^; AMNH 88837, LME localities: E. Albeit, Despair and Kent Quarries, Four Mile local fauna, Hiawatha Member, Wasatch Formation, Colorado; Powder River local fauna, Bozeman locality, Wyoming. KNOWN distribution: Early Wasatchian, Colorado and Wyoming; Sparnacian and Cuisian, France. DESCRIPTION AND REMARKS: McKenna (1960a) and Delson (1971) noted the occurrence of a species of Entomolestes cf. E. nitens in the Four Mile and Powder River faunas, respectively, which was 20% smaller than the type and paratype of E. nitens. Although the Almagre sample of M. nitens increases the known variation in size of the species, the original and addi- tional material from Four Mile is here referred to Maerocranion cf. M. nitens. This species is virtually identical to M. nitens in the morphology of the known parts of the dentition. The greatest difference in size occurs in the P4, especially in the talonid. Recovery of a sizeable sample of M. nitens from early Wasatchian horizons should resolve the specific status of the Four Mile and Powder River material. Although the two specimens of upper dentition from Four Mile (AMNH 59678, 88837) agree in size with M. nitens from the Almagre, these are tentatively referred to cf. M. nitens on the basis of the morphology of P'*. P"* on AMNH 59678 bears a well-developed paracone, a paracone- metastylar crest and a wear facet on this crest, where a tiny metacone occurred. This is similar to the condi- tion on UCMP 44085 from Four Mile, which was described and figured (McKenna, 1960a) as Ento- ino/estes cf. E. nitens. P^ of M. nitens from the younger Almagre deposits lacks a metacone. Larger samples of M. nitens and cf. M. nitens may prove the presence of a tiny metacone on P'* to be variable. If not, how- ever, the implied loss of the metacone on P“^ during the Wasatchian in M. nitens would be consonant with the general trend of premoiarization of P4 in early Tertiary erinaceomorphs. The isolated teeth from the Sparnacian and Cuisian of France, which Russell, et al. (1975) described as Entomolestes cf. nitens, exhibit an inordinate range of variation in size. Their structure is virtually identical to Macrocranion nitens and M. cf. M. nitens from North America, including the occurrence of exodaenodonty along the labial margin of Mi-2- It seems probable that the two species of Macrocranion recorded here are also represented by the material from the early Eocene of France. Talpavus Marsh, 1872 This genus has also had a checkered history. T. nitidus was named by Marsh (1872) on the basis of two jaw fragments, one with 1/ 2M1-M2, and the other with a single premolar. The latter proved to belong to Peratherium, a marsupial (Robinson 1968b), whereas 22 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 23 Table 4. Dimensions of lower teeth of Macrocranion nitem and Macrocranion cf. M. nitens P4 Mj M2 M3 L W L AW PW L AW PW L AW PW M. nitens 1.6- 1.1- 1.9- 1.3- 1.5- 1.8- 1.4- 1.5- 1.9- 1.2- 1.3- 2.0 1.3 2.1 1.5 1.6 2.1 1.6 1.7 2.2 1.4 1.4 Mean 1.77 1.20 1.94 1.43 1.54 1.96 1.49 1.60 2.06 1.30 1.35 N. 6 7 9 8 9 11 10 10 5 5 5 SD. 0.09 0.07 0.04 CV. 4.84 5.00 2.94 Macrocranion cf. M. nitens 1.4- 0.9- 1.6- 1.3- 1.4- 1.7- 1.4- 1.5- 1.7 1.2 1.3 1.7 1.1 1.8 1.4 1.5 1.8 1.5 1.6 Mean 1.47 0.97 1.76 1.36 1.47 1.73 1.43 1.53 N. 6 6 7 7 7 3 3 3 Table 5. Dimensions of upper teeth oi Macrocranion nitens and Macrocranion cf. M. nitens p4 Ml m2 m3 L W L W L W L W M. nitens AMNH 48175 — 2.5 1.8 2.5 AMNH 48177 1.8 — 1.8 2.5 1.7 2.7 — — - CM 12398 2.0 2.6 1.7 2.5 1.3 1.8 CM 22019 1.9 2.5 Mean 1.90 2.53 1.78 2.55 Macrocranion cf. M. nitens AMNH 59678 2.2-^ 1.9 2.3 1.8 2.5 AMNH 88837 1.9 2.3 Fig. 7. Macrocranion nitens. (A) Type, AMNH 15697, RP4-M2, Willwood Formation, Wyoming, approx. X 12; (B) AMNH 48178, LP4-M^, San Jose Formation, New Mexico, approx. X 12; (C) AMNH 48177, LP^-M^, San Jose Formation, New Mexico, approx. X 16. Fig. 8. Macrocranion cf. M. nitens, AMNH 59673, LP4-M3, Wasatch Formation, Colorado, approx. X 12. 24 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 the former, YPM 13511, was referred by Matthew (1909) to Nyctitherium. After T. nitidus, based on YPM 13511, was re-established as distinct from Nyctitherium (McKenna, et al., 1962), only four addi- tional specimens were referred to that genus: AMNH 55686, a partial lower jaw with M3 of Diacodon bacchanal is from the late Bridgerian “Hyopsodus” Hill locality; CM 12061, partial lower jaw with P4-M3 from Myton Pocket, Uintan of Utah (McKenna, et al., 1962): AMNH 55662, partial lower jaw with P4-M3 from the Huerfano local fauna, late Wasatchian of Colorado (Robinson, 1966); and YPM 16334, partial lower jaw with P4-M2 from the upper part of the Bridger Formation, Wyoming (Robinson, 1968b). A second species of Talpavus, T. sullivani, was described by Guthrie (1967; 1971) from the Lysite and Lost Cabin Members of the Wind River Forma- tion, Wyoming. A number of the teeth referred to this species are here identified as Scenopagus. The remaining portion of the hypodigm is distinct from Talpavus and needs restudy. Robinson (1968b) also referred the Paleocene Leipsanolestes seigfriedti Simpson, 1928, to Talpavus although that species had previously been allied with Entomolestes (McKenna, 1960a). Leipsano- lestes is distinct from both Talpavus and Entomolestes and is discussed in a later section. The confusion and distinction between the allegedly similar T. nitidus and Entomolestes grangeri were dealt with by Robinson (1968b). Delson (1971) recog- nized Robinson’s criteria, but inexplicably synony- mized the species of Scenopagus with Talpavus. The major problem in defining Talpavus is the fragmentary nature of the type specimen, YPM 13511, and the resulting paucity of diagnostic criteria. How- ever, four features on the type appear to me to be distinctive and indicative of its affinities: the para- conid, preserved only on M2, is compressed anteropos- teriorly into a thin, high crest that flares anterolin- gually at its lingual end; the entoconid is much higher than the hypoconid; the talonid and trigonid are of equal width; the hypoconulid, small and slightly lingual in position, occurs at the posterolabial part of the base of the entoconid and is delineated from that cusp by a sharp notch. The juncture of the hypoconulid and hypoconid is not as well marked, and with wear, the two cusps are joined by a hypocristid. The first two characters align the type of T. nitidus with adapisoricids; the latter ones distinguish it from Scenopagus priscus, the adapisoricid most similar in size and structure to T. nitidus. On M2 of S. priscus the talonid is narrower than the trigonid, the hypoconulid is larger and more clearly separate from the hypoconid, and the paraconid is formed as a broader, flatter lophid. Talpavus nitidus Marsh, 1872 (Fig. 9; Table 6) REFERRED SPECIMENS: YPM 16334, LP4-M2; CM 13717, LP4-M2; AMNH 11488, LM2.3; AMNH 55686, RM3. localities: Dry Creek, Henry’s Fork, and locality 5, Bridger Formation (Upper), Wyoming; Powder Wash, Green River Formation, Utah. KNOWN DISTRIBUTION : Bridgerian, Wyoming, Utah. remarks: I concur with Robinson’s (1968b) de- scription of T. nitidus, which was based on the type and YPM 16334. The latter specimen and the other material referred here resemble the type, and differ from S. priscus in those features noted above: the paraconid on M2 is thinner, more blade-like, the talonid is as wide as the trigonid, and the hypoconulid is smaller and not as clearly separated from the hypo- conid. Apart from the structure of M2, examination of the referred material of T. nitidus discloses addi- tional diagnostic criteria, especially a narrow, pre- molariform P4 with an unbasined talonid. Significantly, the talonid of P4 also lacks cusps — a major distinction between T. nitidus and other material ascribed below to Talpavus. The structure of P4, the paraconid on Mj, and the triangular, uncompressed trigonid on Mi also distinguish T. nitidus from S. priscus. In S. priscus, P4 is wider, the talonid is basined, the metaconid is less well developed, the trigonid on Mi is more nearly square in occlusal outline, and the Mi paraconid is a stronger lophid, more nearly vertical, and closer to the metaconid. Talpavus cf. T. nitidus (Fig. 10; Table 6) referred specimens: CM 12122, LP4-M3; CM 13715, RP3.4; CM 13716, LP4-M1; CM 13718, RP3.4, CM 13754, RMi; CM 26087, LP3-M1; CM 26260, LM1.3; CM 26261, RM1.2; CM 31180, LP3-M3; CM 31181, RM3; CM 26267, RP4-M3; AMNH 55226, LP4-M3. localities: Greybull horizon, Willwood Formation, Big Horn Basin, Wyoming; Powder Wash, Green River Formation, Utah; Huerfano Formation, Colorado. known distribution: Early Wasatchian, Wyoming; early Bridgerian, Utah; late Wasatchian, Colorado. Fig. 9. Talpavus nitidus, CM 13717, RP4-M2, Green River Formation, Utah, approx. X 20. Fig. 10. Talpavus ci. T. nitidus. (A) CM 12122,^- LP4-M3, and (B) CM 26260, LM1.3, Willwood Formation, Wyoming, approx. X 12; (C) CM 26087, LP3-M1, Green River Formation, Utah, approx. X 12. Fig. 11. Talpavus duplus, new species. Type, CM 12061, LP4-M3, Uinta Formation, Utah, approx. X 12. 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 26 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I remarks: Only two features distinguish this material from T. nitidus: slightly smaller size and the occurrence of two cuspules (?entoconid and hypoconulid) on the raised posterior margin of the talonid of P4. P3, pre- served on CM 13715, CM 13718, CM 26087, and CM 31180, is smaller than P4, double-rooted, somewhat laterally compressed, and bears a small paraconid, a protoconid and a short talonid. As in T. nitidus, P4 of cf. T. nitidus is narrow, and its talonid is short, un- basined, and lacks a cristid obliqua. The P4 metaconid is almost as large and tall as the protoconid, and the paraconid is low, rounded, and well developed. The trigonid on Mj is not compressed, but triangular in occlusal aspect, and the paraconid is blade-like and canted anterolingually, so that the trigonid is open lingually. The trigonid on M2 is more compressed than on Ml, and is as wide as the talonid. The hypoconulid on Mi-2 is slightly more separated from the hypoconid in comparison to the condition in T. nitidus, but is smaller than in S. priscus. The occurrence of cuspules on the talonid of P4 may prove variable with recovery of additional Talpavus material. The structure of the known parts of the dentition of Talpavus implies that Talpavus had an origin in common with Scenopagus, especially S. priscus, from a McKennatheriuin-Vike adapisoricid, and differentiated with respect to the loss of the talonid basin on P4. AMNH 55226 from Huerfano was referred by Robinson (1966) to Talpavus cf. T. nitidus — an action recently questioned by Russell, et al. (1975) in their confusing analysis of Talpavus. Although wear on P4 of AMNH 55226 has obliterated the details of the talonid structure, the latter appears to have been unbasined. Similarly, breakage and subsequent repair of the trigonids on M1-2 have produced artificially elongate molars. Nevertheless, as in Talpavus, the M2 talonid is as wide as the trigonid, the hypoconulid on M1-2 is smaller than in S. priscus and is joined by a hypo- cristid to the hypoconid, P4 is much narrower in rela- tion to Ml than is the case in S. priscus, and the paraconid on M2 flares anterolingually. Talpavus duplus, new species (Fig. 1 1 ; Table 6) type: cm 12061, RP4-M3, and only known specimen. TYPE locality: Myton pocket, Uinta Formation, Utah. KNOWN distribution: Uintan of Utah. etymology: duplus, L., double, referring to the double paraconid on P4 and the close resemblance to T. nitidus. diagnosis: Largest known species of Talpavus. Two cuspules form the paraconid on P4. DESCRIPTION AND REMARKS: The features that define Talpavus are well expressed on CM 12061. On P4 the talonid is unbasined, but bears two cuspules in a fashion similar to P4 of Talpavus cf. T. nitidus. The metaconid on P4 is almost as large as the protoconid. The paraconid is formed as a ridge running antero- lingually from the anterior face of the protoconid. Two cuspules occur on the ridge, one proximal to the protoconid and one forming the tip of the paraconid. Apart from their larger size, M1-3 of T. duplus are similar to those of T. nitidus and cf. T. nitidus. The paraconid on Mi, although worn, appears to have been closer to the metaconid than in the latter species. However, features of T. duplus that are typical of Talpavus are the uncompressed trigonid on Mi, M2 tal- onid and trigonid of equal width, and the small hypoconulid joined to the hypoconid by a hypocristid on Mi-2. T. duplus appears to be a Uintan representa- tive of the same lineage that gave rise to the Wasatch- ian-Bridgerian species of Talpavus. Adapisoricidae: Summary The five genera, McKennatheriuni, Scenopagus, Ankylodon, Macrocranion, and Talpavus, seem to com- prise a clade of lipotyphlans distinct from other early Tertiary Erinaceomorpha and Soricomorpha. The major trends in the origin and radiation of Paleocene and Eocene Adapisoricidae appear to be premolariza- tion of the fourth premolar and the development of more nearly rectangular molars with lower, more bunodont cusps — trends shared with early erinaceids and many other early Tertiary mammals. In general, these are adaptations that emphasize crushing rather than puncturing and shearing during mastication. Specifically, P4 in adapisoricids is primitively semi- molariform, but becomes premolariform in Eocene species. M1-3 are characterized by a compressed, lophid-like paraconid that never joins the metaconid. The protoconid and metaconid are low, conical, almost bulbous cusps. The talonid is wide, moderately basined, and bears a high entoconid and a lower hypoconid that flattens with wear. The hypoconulid is median or barely lingual in position, and is not reduced. The hypoflexid notch is shallow; i. e., the cristid obliqua meets the trigonid below and labial to the notch between the protoconid and metaconid. Labial, lingual, or posterior cingulids are not developed. The cusps on the upper molars are relatively low and conical. The stylar shelf is very narrow, without stylar cusps, and the hypocone and conules are well developed. In occlusal view, Mi-3 are essentially rec- 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 27 tangular and somewhat or moderately transverse. as yet unknown in Paleocene species, is premolariform in Eocene adapisoricids. North American Paleocene adapisoricids are rep- resented by McKennatheriwn ladae and a larger, con- temporaneous species from the Rock Bench Quarries, Polecat Bench Formation. The Eocene genus Sceno- pagus, including S. edenensis, S. cwtidens, and S. priscus, seems descended from McKennatherium, and ancestral to the late Eocene and Oligocene Ankylodon, the youngest known genus of the Adapisoricidae. Talpavus, represented by at least two species during the Eocene, has lost the talonid basin on P4. Macro- cranion may have had a European origin, but one species, M. nitens, occurs in early Eocene deposits of western North America. There is ample evidence that extensive faunal interchange occurred between North America and Europe during the Paleocene and early Eocene across a North Atlantic land corridor (Savage, 1971; Szalay and McKenna, 1971; McKenna, 1971, 1972a, 1972b; Dawson, et al., in press). The bulk of recent work concerning early Tertiary lipotyphlans alludes to an ancestral-descendant rela- tionship between adapisoricids and erinaceids, an inference that may have seemed possible if the Adapis- oricidae were defined in a wastebasket sense (Van Valen, 1967, for example). As here understood, how- ever, the two groups, by mid-Paleocene time, had differentiated to an extent that implies a considerable period of independent evolution prior to their earliest known (and contemporaneous) occurrence. Knowledge of their ancestry is clouded and will probably remain so until the record of North American Cretaceous euther- ians improves, and until a number of early Paleocene faunas, which are currently under investigation, are described. Clemens (1973) has examined and tenta- tively refuted the possible relationships between McKennatherium and the late Cretaceous leptictid Gypsonictops. It also seems unlikely that adapisoricids originated from Late Cretaceous palaeoryctids, e. g., Cimolestes, Procerberus, or Batodon, for reasons similar to those stated by Clemens (1974) concerning the ancestry of the paromomyid primate Purgatorius. Palaeorictid lower dentitions are characterized by tall Table 6. Dimensions of lower teeth of Talpavus nitidus, Talpaviis cf. T. nitidus, and T. duplus, new species P4 Ml M2 M3 L W L AW PW L AW PW L AW PW T. nitidus YPM 13511 — — 1.2 1.6 1.1 1.1 CM 13717 1.1 0.8 1.4 1.1 1.1 1.4 1.2 — AMNH 11488 1.4 1.2 1.2 1.3 1.0 0.8 Talpavus cf. T. nitidus CM 12122 1.1 0.7 — — — 1.2 — 1.0 1.1 . — CM 13715 1.2 0.8 CM 13716 1.2 0.7 1.4 — 1.1 CM 13718 1.1 0.8 CM 13754 1.5 1.0 1.1 CM 26087 1.1 0.7 1.4 1.0 1.0 CM 26260 1.3 1.1 1.0 1.1 1.0 1.0 1.2 0.9 0.6 CM 26261 — 1.1 1.2 1.2 1.1 1.1 CM 31180 1.1 0.8 1.3 1.0 1.0 1.1 1.0 1.0 1.2 0.9 0.7 CM 31181 1.2 0.9 0.8 CM 26267 1.0 0.7 1.4 1.1 1.0 1.2 0.9 0.9 1.2 0.9 0.7 AMNH 55226 1.2 0.8 1.4 1.1 1.1 1.4 1.1 1.1 1.4 1.0 0.7 Mean 1.12 0.75 1.38 1.05 1.06 1.20 1.02 1.02 1.24 0.92 0.70 T duplus CM 12061 Type 1.4 0.9 1.6 1.1 1.2 1.6 1.2 1.2 1.6 1.1 1.0 28 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I trigonids and high, sharp cristids and cusps. Their upper molars are very transverse and lack hypocones or posterior cingula — features contrary to the more bunodont dental adaptations of adapisoricids. Some of the material of Pwgatorius from the Puercan Garbani Locality, Montana, is, according to Clemens (1974), of definite insectivoran aspect, and in my view, especially like McKeiviatherium. The late Cretaceous erinaceotan from Alberta (Fox, 1970) seems to be a representative of a non-paleoryctid — leptictid group of insectivores that may have been ancestral to adapis- orcids — a suggestion that Clemens (1974) also makes with reference to the ancestry of Purgatorius. In terms of dental morphology, the same Cretaceous stock that gave rise to adapisoricids (and other Lipotyphla) was probably also basal to the radiation of primates, dermopterans, bats, and tupaiids (the Archonta of Gregory, 1910, minus macroscelidids) as well as un- gulates (McKenna, 1975). Family Erinaceidae Fischer Von Waldheim, 1817 Many of the known early Tertiary insectivores have at one time or another been referred to this family. Simpson’s (1945) Erinaceidae included Enlomolestes, Prolerixoides, Metacodon, and Ankylodon. Subse- quently, Scenopagus, Centetodon, and Talpavus were also placed here (McKenna and Simpson, 1959; McKenna et al., 1962; Robinson, 1966). Van Valen (1967) limited the Erinaceidae to undoubted post- Eocene forms and relegated the above genera to the Adapisoricidae. These are now more strictly defined in this and other studies. The early Tertiary North American Erinaceidae, as here understood, includes Litolestes, Leipsanolestes, possibly Entomolestes grangeri, and an unpublished form from the late Eocene of Wyoming (Setoguchi, M. S. thesis; McKenna, personal commun., 1974). Their shared-derived characters include a premolari- form P4, rectangular molars that progressively de- crease in size from M1-3, a dorsally flattened but cuspate paraconid on M2-3, a greatly reduced hypo- conulid, and a V-shaped talonid basin. 1 agree with Butler (personal commun., 1975) and Russell ( 1964) that the European Paleocene Adapisorex is an erinaceid rather than an adapisoricid. Litolestes Jepsen, 1930 L. ignotus, from the late Tiffanian of Wyoming, was described by Jepsen (1930) from partial remains of the lower dentition, and was referred, with question, to the Insectivora. In subsequent descriptions of two additional species, L. notissimus (Simpson, 1936) from Scarritt Quarry and L. lacmatus (Gazin, 1956) from Bison Basin, Litolestes was identified as a hyopso- dontid condylarth. Van Valen (1967) however, re- turned Litolestes to the Insectivora as a creotarsine adapisoricid. Although Litolestes is here described as an erinaceid, the affinities ofL. notissimus and L. lacunatus are not as clear. As discussed below, they do not resemble typical hyopsodontids. Rather, they may represent late Pale- ocene echinosoricine-like erinaceids. Litolestes ignotus Jepsen, 1930 (Fig. 12; Table 7) type: PU 13362, RP4-M3, Polecat Bench Formation, Wyoming. REFERRED specimens: PU 13354, LP4-M2; PU 13974, RP4-M3; PU 14064, RP3-M2 erupting; PU 19362, LP3-M3; PU 19387, LP2-M2. locality: Silver Coulee Quarries, Polecat Bench Formation, Wyoming. KNOWN DISTRIBUTION : Late Tiffanian, Wyoming. DESCRIPTION AND REMARKS: The lower antemolar dentition of L. ignotus has been dealt with elsewhere (Schwartz and Krishtalka, 1976) and need not be redescribed in detail. I2-3 of L. ignotus are comb-like with digitate, trilobed crowns. The crowns on lower incisors of later Tertiary and Recent erinaceids are not divided, but spatulate. However, incisors with divided crowns are not characters that imply phyletic relation- ships, inasmuch as these have evolved independently in many groups of mammals including Dermoptera, Macroscelidea, Carnivora, Notoungulata, and a variety of living and extinct insectivores (cf. Rose, 1973). The fourth tooth in the jaw, traditionally identified as the canine in L. ignotus, is small, single-rooted, and fully premolariform. On this and other bases, Schwartz and Krishtalka (1976) concluded that this tooth may represent the first of five lower premolars in L. ignotus (and in the dermopteran Plagiomene multicuspis). As noted previously concerning Ankylodon, in addition to identifying this tooth as a premolariform canine or a premolar, a third interpretation — a deciduous canine — is possible, since the alleged dCi in another insectivore, Nesophontes, is also small, single-rooted, and pre- molariform. However, the homology of that locus in Nesophontes is uncertain, since both “dCi” and “Ci” are premolariform and perhaps represent a deciduous premolar and its permanent successor, respectively. Accordingly, the tooth immediately posterior to I3 in L. ignotus (and Plagiomene) is here tentatively identi- fied as dPi, a premolariform deciduous premolar that 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 29 is retained in an antemolar dental complement of Iil2l3dPiP2P3P4P5. However, in the following portion of this paper, the standard post-incisor nomenclature of CiPi-4 is employed for reasons outlined in the introduction and to facilitate comparisons of canine and premolar morphologies. P4 is premolariform and as large as M^, as in typical early erinaceomorphs. The protoconid is dominant, the paraconid is small and anterolingual, and the metaconid is low on the posterolingual slope of the protoconid. The talonid is short, as wide as or wider than the trigonid, and the talonid basin is lingual and very narrow — approximately one-fourth or one-fifth of the width of the talonid. A tiny cuspule occurs at the posterior tip of the cristid obliqua. The labial portion of the talonid slopes ventrolabially away from the cristid obliqua. The molars decrease markedly in size from Mi to M3. The protoconid and metaconid are low and conical. Significantly, the paraconid on Mi is slightly compressed, but is more nearly erect and cuspate than on M2-3. On the latter, the paraconid is more nearly medial and flattened so that it projects ante- riorly. The talonid is as long and wide as the trigonid on Mi-2, but is elongate on M3 because of a posteriorly projecting hypoconulid. The entoconid is higher than the hypoconid, but neither cusp is formed as a distinct cone. Rather, they are elongate anteroposteriorly, with flat internal walls that slope ventromedially so that the talonid basin is not rounded, but a V-shaped valley. The internal walls of both cusps extend to the trigonid and close the talonid basin lingually and labially. The bottom of the talonid valley is formed as a midline, which runs from the trigonid to the posterior margin of the talonid. On M1-2 the hypo- conulid is extremely reduced and occurs on the posterior edge of the entoconid wall, just lingual to the midline. The posterior edge of the talonid is nearly straight. These features of L. ignotus are equally diagnostic of undoubted later Tertiary erinaceids and, most espe- cially, a new mid- or late-Eocene species from the Tepee Trail Formation, Wyoming, currently being described by McKenna (personal commun., 1974). L. ignotus seems slightly more primitive than the latter in lacking posterior cingulids on the lower molars, in possessing higher protoconids on P2-3, and a slightly more developed talonid on P4. In contrast to L. ignotus, the lower canine in L. notissimus and L. lacunatus is large and recurved, P2 and P3 are more elongate and bear a small low paraconid, the talonid on P4 is relatively longer and wider, and the metaconid and protoconid on M1-3 are slightly more bulbous. The hypoconulid is not reduced on M1-3 of L. notissimus and L. lacunatus, the talonid basin is more nearly rounded than V-shaped, and the cristid obliqua is oriented less nearly parallel to the entocristid, compared to that of L. ignotus. Russell, et al. (1975), alluded to the condylarthran nature of Litolestes, although they retained that genus in the Adapisoricidae. One of the problems of their review is the failure to discuss the genotype, L. ignotus. Only L. notissimus was considered. That species, as well as L. lacunatus, superficially resemble hyoposodontids, mainly in the semibulbous nature of the molar proto- conids and metaconids, compared to those of L. ignotus. However, in hyoposodontids known to me the molars are considerably more robust and less angular in occlusal outline. M2 is invariably larger than M i and M3, the molar paraconids are reduced to a simple ridge between the protoconid and metaconid, and P4 is semimolariform or submolariform. As in L. ignotus and the Tepee Trail erinaceid, the lower molars of L. notissimus and L. lacunatus are not robust, they decrease progressively in size from Mi to M3, the paraconid on Mi is slightly compressed but is not reduced, and the paraconids on M2-3 are dorsally flattened but cuspate. L. notissimus and L. lacunatus are stratigraphically older than L. ignotus. As erinaceids, they have more plesiomorphic features than L. ignotus, e.g., the un- reduced hypoconulid and a more nearly rounded talonid basin on the lower molars, and a slightly longer talonid on P4. Of interest is the large lower canine of L. notissimus and L. lacunatus, a feature absent in L. ignotus (Schwartz and Krishtalka, 1976) and later erinaceines (Butler, 1948), but retained in many echinosoricines (Butler, 1948). L. notissimus and L. lacunatus may represent the early echinosoricine condition in which a canine and four premolars are derived from a Kennalestes-\\\c& dental complement of a canine and five premolars. L. ignotus, on the other hand, with loss of the lower canine and retention of five premolars, may mirror the origin of the erinaceine lower dentition. The known parts of the upper dentition of L. ignotus and L. notissimus (AMNH 33942, 33944) are less erinaceid-like than those of the lower dentition. P4 in these species lacks a metacone, and is premolari- form. Mi-3 are typically adapisoricid-like : molars wider than long and shorter lingually than labially, narrow stylar shelves, rectangular occlusal outline 30 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 with a shallow labial emargination and a constriction across the conules, conical or subconical paracone and metacone, well-developed conules with internal wings, a hypocone, and a weak anterior cingulum — characters that are plesiomorphic for many Paleocene eutherians like primates, lipotyphlan insectivores, pantolestids, apatemyids, and condylarths, and not indicative of phylogenetic relationships among taxa within these groups. With regard to adapisoricids, two of the derived characters of upper molars of early erinaceids are the progressive reduction in size from Mi to M3 and the posterolabial position of the hypocone relative to that of the protocone — characters that occur in the Tepee Trail erinaceid. In Litolestes, the hypocone on Mi‘2 occurs more nearly posterior than posterolingual to the protocone. M2 of Litolestes is not larger than Mi to the degree seen in adapisoricids and hyopsodontids. Rather, M2 is shorter than Mi, and is slightly wider only in the area of the stylar shelf. In these respects the upper dentition of Litolestes approaches the typical erinaceid condition but is primitive relative to the upper dentition of the erinaceid from Tepee Trail. Leipsanolestes Simpson, 1928 This genus is known only from the lower dentition of one species. Simpson (1928) originally described L. seigfriedti as an adapisoricid, but later (1929) referred Leipsanolestes to a subgenus of Leptacodon. McKenna (1960a) argued for inclusion of L. seigfriedti in Ento- molestes, whereas Robinson (1968b) placed the species in Talpavus. Russell, et al. (1975), provisionally re- tained Leipsanolestes as a valid genus in the Adapis- oricidae. As described below, the molar morphology of Leipsanolestes is very similar to that of Litolestes ignotus, and differs from M. nitens and Talpavus, especially in those features that distinguish early erinaceids from adapisoricids. Leipsanolestes seigfriedti Simpson, 1928 (Fig. 13; Table 7) type; AMNH 22157, RM2-3, Fort Union Formation, Montana. REFERRED SPECIMENS: CM 11517, LP4-M1 ; CM 11519, RP4- Mi; CM 11532, LMi; CM 11553, RP4-M3; CM 11662, LP4, M2; CM 11664, LP4; CM 11676, RM|; CM 11704, Rl/2Mi- M2; AMNH 56317, RM1.3. localities; Bear Creek, Fort Union Formation, Montana; Reclusa Blowout, Powder River local fauna, Wyoming. KNOWN distribution: Late Tiffanian, Montana; early Wasatchian, Wyoming. DESCRIPTION AND REMARKS: Teeth anterior to P4 are unknown. P4 is semimolariform, with the talonid almost as long as the trigonid. The paraconid is bulbous, anterolingual, and as large as the metaconid. The metaconid occurs slightly higher than the para- conid on the posterolingual face of the protoconid. The talonid bears two tiny cusps, an entoconid and hypoconid, which, respectively, form the posterior tips of the long and parallel entocristid and cristid obliqua. The talonid basin between these two cristids occupies approximately one-half the width of the talonid. The molars of Leipsanolestes seigfriedti are low and rectangular, and decrease in size from Mj to M3, although not as radically as in Litolestes ignotus. The paraconid is fully cuspate and is flattened dorsally to produce a low angular shelf. The protoconid and metaconid are blunt, conical cusps, and the talonid basin is developed as a V-shaped valley. The internal walls of the entoconid and hypoconid are flat, oriented ventromedially toward the talonid basin, and extend anteriorly to meet the trigonid. The hypoconulid is extremely reduced on M1-2, but is discernible on the posterior edge of the entoconid, just lingual to the midline. As a result, the posterior margin of the talonid on Mi-2 is nearly straight, as is the case in L. ignotus, Entomolestes grangeri, and the late Eocene erinaceid from Tepee Trail. Although Leipsanolestes seigfriedti and Litolestes ignotus were contemporaneous erinaceids, the former is more primitive in the retention of a semimolariform P4 and the less-marked reduction in size of the molars from Ml to M3. However, the structure of the molars of L. seigfriedti is unquestionably erinaceid-like. This species may represent an early lineage of erinaceids that, with Litolestes, shared an early Paleocene or older ancestry from a form with a semimolariform P4. Delson (1971) erroneously referred AMNH 56317, partial right mandible with M1-3 from the early Eocene Powder River local fauna, to Entomolestes cf. nitens. The molars are virtually identical to those of L. seigfriedti except for the complete loss of the hypo- conulid on the Eocene form. This specimen is here Fig. 12. Litolestes ignotus, PU 19387, Polecat Bench Formation, Wyoming, approx. X 12. Fig. 13. Leipsanolestes seigfriedti. (A) CM^ 11553, RP4-M3, Fort Union Formation, Montana; (B) AMNH 56317, RM1.3, Wasatch Formation, Wyoming; both approx. X 12. 32 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. I assigned to L. seigfriedti, although recovery of addi- tional material may determine that AMNH 56317 warrants specific distinction from L. seigfriedti. The confusion concerning the affinities of Leipsano- lestes was not resolved in the recent discussion by Russell, et al. (1975). The authors compared this genus only to Entomolestes’’’ nitens, noted a number of diagnostic differences and concluded that the two species shared a special affinity among adapisoricids. In comparison to Macrocranion nitens, P4 of Leipsan- olestes is semimolariform rather than premolariform. The talonid is more elongate and the talonid basin occupies one-half the width of the talonid. M1-3 of Leipsanolestes (and other erinaceids) are distinct from those of M. nitens (and other adapisoricids) in that they are more rectangular in occlusal outline and show a progressive reduction in size. The talonid on Mj-2 of Leipsanolestes is not longer than the trigonid, the hypoconulid is extremely reduced, the talonid cusps occur posteriorly along the nearly straight posterior margin of the talonid, and the paraconid is a low, fully cuspate, dorsally flattened shelf rather than a compressed lophid. Scenopagus and Talpavus differ from Leipsanolestes in the structure of P4, and in possessing higher, sharper, molar cusps, and a deeper hypoflexid notch. As discussed below, among species previously clas- sified by most authors as adapisoricids, Entomolestes grangeri bears the closest resemblance to Leipsano- lestes. ?Erinaceidae Entomolestes Matthew, 1909 Entomolestes grangeri Matthew, 1909 The taxonomic history of this genus was discussed in an earlier section. The genotype, E. grangeri, is known only from the holotype, AMNH 1 1485, partial left mandible with P3-M3, from the upper part of the Bridger Formation, Middle Eocene of Wyoming. Apart from Robinson (1968b), previous considerations of Entomolestes failed to distinguish between E. grangeri and nitens, which I have referred above to Macrocranion. Unfortunately, the molars on AMNH 11485 are damaged, obscuring potentially diagnostic features. However, the lower dentition bears a number of discernible erinaceid characters. Five alveoli are pre- served anterior to P3, and are interpreted by Robinson (1968b) to have been filled with I2-3, Ci, P1-2, all single-rooted. Robinson also noted, however, that the alveolus for P2 is elongate and slightly hourglass- shaped, so that P2 may have had two roots, as in other erinaceids. P3 is double-rooted and has a low, blade-like, procumbent crown. P4 is typically pre- molariform with an extremely short talonid and a single cuspule at the posterior tip of the cristid obliqua. The molars are low, rectangular in occlusal view, and lean anterolingually. M3 is reduced, compared to Mi-2, which are of equal size. The talonids are as wide as the trigonids and not elongate, whereas the opposite is true in M. nitens. The paraconid, damaged on M^ and M3, is low, somewhat compressed, and oriented anterolingually on M2. The metaconid and protoconid are quite worn, but appear to have been well developed and conical. The hypoflexid notch is shallow, as in other erinaceomorphs, since the cristid obliqua origi- nates labially on the posterior wall of the trigonid. The entoconid is high, elongated anteroposteriorly, and has a flat internal wall that slopes ventromedially to the bottom of the talonid basin. A tiny notch occurs lingually at the junction of the high entocristid and the posterolingual face of the metaconid. The hypoconid is worn flat. The hypoconulid on M1-2 is reduced and occurs on the posterior edge of the entoconid, just lingual to the midline, as is also the case in L. ignotus and Leipsanolestes. As a result, the talonid cusps are lined up along the nearly straight posterior margin of the talonid. These characters imply the erinaceid affinities of E. grangeri. Although Mj and M2 are equal in size, the molars of E. grangeri most closely resemble those of Leipsanolestes. Clearly, more complete E. grangeri material is needed before its relationships can be accurately assessed. Erinaceidae: Summary The two or possibly three known genera that com- prise the North American early Tertiary record of erinaceids can be readily distinguished from adapis- oricids and soricomorphs on the basis of dental remains. P2-4 are double-rooted, and P4 is premolari- form. The molars are rectangular in occlusal outline and decrease in size from Mi to M3. The paraconid, more nearly erect and lingual on Mi, is flattened dorsally into a shelf-like cusp on M2-3 and occurs closer to the midline. The protoconid and metaconid are low, conical cusps. The talonid basin is formed as a V-shaped valley by the flat internal walls of the hypoconid and entoconid. These cusps are also elon- gated anteriorly to the trigonid. The hypoconulid is extremely reduced on M1-2 and the hypoflexid notch is shallow. 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 33 In the earliest known erinaceid, Litolestes, some of these characters, like the structure of the talonid and the reduced hypoconulid, are less well expressed and imply the primitive condition. Litolestes and Leipsano- lestes seem to represent two early, divergent lineages of erinaceids. Species of the former seem suitable morpho- logically and temporally to be representatives of early erinaceine and echinosoricine erinaceids. L. ignotus may be close to the ancestry of the late Eocene erinaceid from Tepee Trail, Wyoming. Table 7. Dimensions of lower teeth of Litolestes ignotus and Leipsanolestes seigfriedti P4 Ml M2 M3 L W L AW PW L AW PW L AW PW L. ignotus PU 13354 1.5 1.2 1.6 1.2 1.3 1.4 1.2 PU 13974 1.7 1.2 1.7 1.3 1.3 1.5 1.3 1.3 1.3 1.0 0.9 PU 14064 1.6 1.2 1.7 1.2 1.2 1.5 1.2 1.2 PU 14333 1.7 1.4 1.7 1.5 1.5 PU 19362 1.7 1.4 1.8 1.5 1.5 1.5 1.3 1.3 1.4 1.2 1.0 PU 19387 1.8 1.3 1.8 1.4 1.4 1.6 1.4 1.3 1.4 1.0 0.9 Mean 1.67 1.28 1.72 1.35 1.36 1.48 1.28 1.28 1.37 1.07 0.93 L. seigfriedti CM 11517 1.5 1.0 1.8 1.4 1.5 CM 11519 1.7 1.1 1.8 1.3 — CM 11532 CM 11662 1.6 1.3 1.6 1.4 1.4 1.5 1.3 1.3 CM 11676 CM 11704 1.8 1.3 1.5 1.6 1.4 1.4 AMNH 56317 1.7 1.2 1.3 1.6 1.3 1.4 1.4 — 1.1 Mean 1.60 1.13 1.74 1.32 1.42 1.57 1.33 1.37 ACKNOWLEDGEMENTS During the course of this study I have become deeply indebted to many individuals, most especially to Dr. Craig C. Black, for his warm friendship, con- stant encouragement, and critical counsel. Dr. Black, Dr. Peter Robinson (University of Colo- rado), Dr. Malcolm C. McKenna (American Museum of Natural History), and Dr. Mary R. Dawson (Carnegie Museum of Natural History), originally suggested this study to me as a significant subject for a doctoral dissertation in vertebrate paleontology, and provided generous access to and loan of fossil material in their care. I am also grateful to them and to the following for thoughtful and stimulating discussions concerning early Tertiary insectivore evolution: Dr. Percy M. Butler (Royal Holloway College, London), Dr. Jason A. Lillegraven (San Diego State University), Dr. Jeffrey H. Schwartz (University of Pittsburgh), Michael J. Novacek (University of California, Berke- ley), and Craig B. Wood (Harvard University). Dr. Lillegraven, Dr. McKenna, and Mr. Novacek kindly allowed the use of information from manuscripts in press. The pertinent insectivore material from the Polecat Bench collection was made available to me by Dr. Donald L. Baird (Princeton University), and Craig Wood. I wish to thank Jack Capenos, Crucible Research Division, Colt Industries, Pittsburgh, for the scanning electron micrographs, and Elizabeth Hill for typing the manuscript. This study was supported by teaching and research assistantships from the Department of Biology, the Graduate School and the Institute of Museum Re- search, Texas Tech University, and by a post-doctoral fellowship from Carnegie Museum of Natural History. References Cited Butler, P. M. 1948. On the evolution of the skull and teeth in the Erinacei- dae, with special reference to fossil material in the British Museum. Proc. Zool. Soc. Lond., 118:446-500. 1956. The skull of Ictops and the classification of the Insec- tivora. Proc. Zool. Soc. Lond., 126:453-481. 1972. The problem of insectivore classification, pp. 263-265. I?i Joysey, K. A. and T. S. Kemp, eds.. Studies in verte- brate evolution. Winchester Press, New York. Clemens, W. A. Jr. 1973. Fossil mammals of the type Lance Formation, Wyom- ing. Part III. Eutheria and summary. Univ. Calif. Publ. Geol. Sci., 94:1-102. 1974. Purgatorius, an early paromomyid primate (Mammalia). Science 184: 903-905. Dawson, M. R., R. M. West, W. A. Langston and J. H. Hutchison. [In Paleocene terrestrial vertebrates: northernmost occur- Press] rence, Ellesmere Island, Canada. Science. Delson, E. 1971. Fossil mammals of the early Wasatchian Powder River local fauna. Eocene of northeast Wyoming. Bull. Amer. Mus. Nat. Hist., 146(4) :*305-364. Fischer Von Waldheim, G. 1817. Adversaria zoologica. Mem. Soc. Imp. Nat., Moscow, 5:368-428. Fox, R. C. 1970. Eutherian mammal from the early Campanian (late Cretaceous) of Alberta, Canada. Nature, 227:630-631. Galbreath, E. C. 1953. A contribution to the Tertiary geology and paleontology of northeastern Colorado. Univ. Kansas Paleont. Contribs., Vertebrata, 4:1-120. Gazin, C. L. 1956. Paleocene mammalian faunas of the Bison Basin of south-central Wyoming. Smithson. Misc. Coll., 131(6): 1-57. Gill, T. 1 872. Arrangement of the families of mammals with analytical tables. Smithson. Misc. Coll., ll(l):l-98. Gregory, W. K. 1910. The orders of mammals. Bull. Amer. Mus. Nat. Hist., 27:1-254. Guthrie, D. A. 1967. The mammalian fauna of the Lysite Member, Wind River Formation (early Eocene), of Wyoming. Mem. So. Calif. Acad. Sci., 5:1-53. 1971. The mammalian fauna of the Lost Cabin Member, Wind River Formation (Lower Eocene) of Wyoming. Ann. Carnegie Mus., 43(4):47-l 13. Haekel, E. 1866. Systematische Einleitung in die allgemeine Entwid- klungsgeschichte. Generelle Morphologic der Organis- men, Georg Reimer, Berlin, Bd. 2, pp. XVII-CLX. Illiger, C. 181 1. Prodromus systematis mammalium et avium additis terminis zoographicis utriudque classis. C. Salfeld, Berlin. 301 pp. Jepsen, G. L. 1930. Stratigraphy and paleontology of the Paleocene of northeastern Park County, Wyoming. Proc. Amer. Philos. Soc., 69:463-528. Krishtalka, L. 1973. Late Paleocene mammals from the Cypress Hills, Alberta. Special Publ. Mus. Texas Tech Univ., 2:1-77. 1975. Systematics and relationships of early Tertiary Lipo- typhla (Mammalia, Insectivora) of North America. Doctoral Dissertation, Texas Tech University, Lubbock. University Microfilms, Ann Arbor, Mich. Lemoine, V. 1883. Etude sur le Neoplagiaulax de la faune eocene inferieur des environs de Reims. Bull. Soc. Geol. France, 3® Serie, 11:249-271. 1885. Etude sur quelques mammiferes de petites taille de la faune cernaysienne des environs de Reims. Bull. Soc. Geol. France, 3® Serie, 13:203-217. Lillegraven, J. a. and M. C. McKenna [MS] Phylogenetic history and microevolutionary mosaicism of Centetodon. McDowell, S. B. 1958. The Greater Antillean insectivores. Bull. Amer. Mus. Nat. Hist., 115(3):1 13-214. McGrew, P. O. 1959. The geology and paleontology of the Elk Mountain and Tabernacle Butte area, Wyoming. Bull. Amer. Mus. Nat. Hist., 117:117-176. McGrew, P. O. and R. Sullivan. 1970. The stratigraphy and paleontology of Bridger A. Univ. Wyo. Contrib. Geol., 9(2):66-85. McKenna, M. C. 1960a. Fossil Mammalia from the early Wasatchian Four Mile fauna. Eocene of northwest Colorado. Univ. Calif. Publ. Geol. Sci., 37(1):1-130. 1960b. The Geolabidinae, a new subfamily of early Cenozoic erinaceoid insectivores. Univ. Calif. Publ. Geol. Sci., 37:131-164. 1971. Fossil mammals and the Eocene demise of the De Geer North Atlantic dispersal route. Geol. Soc. Amer. Abstr. with Programs, 3(7):644. 1972a. Was Europe connected directly to North America prior to the middle Eocene? In Dobzhansky, T., M. Hecht and W. C. Steers, eds.. Advances in Evolu- tionary Biology, 6:179-189. 1972b. Eocene final separation of the Eurasian and Green- land-North American landmasses. 24th Internatl. Geol. Congr., Sect. 7:275-281. 1975. Toward a phylogenetic classification of the Mammalia. In Szalay, F. S. and W. P. Luckett, eds., Phylogeny of the Primates; and interdisciplinary approach. Plenum Press: New York. McKenna, M. C. and G. G. Simpson. 1959. A new insectivore from the middle Eocene of Taber- nacle Butte, Wyoming. Amer. Mus. Novitates, 1952: 1-12. 34 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 35 McKenna, M.C., P. Robinson, and D. W. Taylor. 1962. Notes on Eocene Mammalia and Mollusca from Tabernacle Butte. Wyoming. Amer. Mus. Novitates, 2102:1-33. Marsh, O. C. 1872. Preliminary description of new Tertiary mammals. Amer. J. Sci., Ser. 3, 4:122-128, 202-224. Matthew, W. D. 1909. The Carnivora and Insectivora of the Bridger Basin. Mem. Amer. Mus. Nat. Hist., 9:289-567. 1918. A revision of the lower Eocene Wasatch and Wind River faunas. Part V. Insectivora (continued), Glires, Edentata. Bull. Amer. Mus. Nat. Hist., 38:565-657. Novacek, M. J. [In Insectivores of the later Eocene (Uintan) of San Diego Press] County, California. Bull. So. Calif. Acad. Sci. [MS Prolerioxides and Sespedectes (Adapisoricidae, Insec - thesis] tivora) from the later Eocene of southern California. Patterson, B., and P. O. McGraw. 1937. A soricid and two erinaceids from the White River Oligocene. Publ. Field Mus. Nat. Hist., geol. ser., 6:245-272. Robinson, P. 1966. Fossil Mammalia of the Huerfano Formation, Eocene of Colorado. Bull. Peabody Mus. Nat. Hist., 21 :l-85. 1968a. Nyctitheriidae (Mammalia, Insectivora) from the Bridger Formation of Wyoming. Univ. Wyo. Contrib. Geol., 7:129-138. 1968b. Talpavus and Entomolestes (Insectivora, Adapisorici- dae). Amer. Mus. Novitates, 2339:1-7. Romer, a. S. 1966. Vertebrate Paleontology, 3rd ed. Univ. Chicago Press. 468 pp. Rose, K. D. 1973. The mandibular dentition of Plagiomene (Dermoptera, Plagiomenidae). Breviora 411 :1-17. Russell, D. E. 1964. Les mammiferes paleocenes d’Europe. Mem. Mus. Nation. d’Hist. Nat., n.s. ser. C, 13:1-324. Russell, D. E., P. Louis, and D. E. Savage. 1975. Les Adapisoricidae de I’Eocene Inferieur de France. Reevaluation des formes considerees affines. Bull. Mus. Nat. d’Hist. Nat. 327 (45):129-194. Saban, R. 1954. Phylogenie des insectivores. Bull. Mus. Nation. d’Hist. Nat., ser. 2, 26:419-432. Savage, D. E. 1971. The Sparnacian-Wasatchian mammalian faima, early Eocene, of Europe and North America. Abh. Hess. L.-Amt Bodenforsch, 60:154-158. SCHLOSSER, M. 1887. Die Affen, Lemuren, Chiropteran, Insectivora, Mar- supialier, Creodontan und Camivoren des europaishen Tertiars. Teil 1, Beitr. Palaont. Oesterreich-Ungams, 6:1-224. Schwartz, J. H. [In Dental development, homologies and primate phy- Press] logeny. J. Human Evol. Schwartz, J. H., and L. Krishtalka. 1976. The lower antemolar teeth of Litolestes ignotus, a late Paleocene erinaceid (Mammalia, Insectivora). Annals Carnegie Mus., 46: 1-6. Setoguchi, T. [MS The late Eocene marsupials and insectivores from the thesis] Tepee Trail Formation, Badwater, Wyoming. Dept. Geol., Texas Tech Univ., 1973. Simpson, G. G. 1928. A new mammalian fauna from the Fort Union of southern Montana. Amer. Mus. Novitates, 297:1-15. 1929. A collection of Paleocene mammals from Bear Creek, Montana. Ann. Carnegie Mus., 19(2) :1 15-122. 1935. New Paleocene mammals from the Fort Union of Montana. Proc. U. S. Natl. Mus., 83(2981) :221-244. 1936. A new fauna from the Fort Union of Montana. Amer. Mus. Novitates, 873:1-27. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85:1-350. SZALAY, F. S. 1968. The beginnings of Primates. Evolution, 22:19-36. 1969. Mixodectidae, Microsyopidae, and the insectivore pri- mate transition. Bull. Amer. Mus. Nat. Hist., 140: 193-330. SZALAY, F. S., AND M. C. McKeNNA. 1971. Beginning of the age of mammals in Asia: the late Paleocene Gashato fauna, Mongolia. Bull. Amer. Mus. Nat. Hist., 144(4):269-318. Thenius, E. 1969. Stammesgeschichte der saugetiere (einschliesslich der Hominiden). Kukenthal’s Handbuch der Zoologie, 8, Lieferung 47:1-368. Walter de Gruyter, Berlin. Tobien, H. 1962. Insectivoren (Mamm.) aus dem Mitteleozan (Lutetium) von Messel bei Darmstadt. Notizbl. Hessischen Landes. Boden forsch. Wiesbaden, 90:7-47. Vanderbroek, G. 1961. The comparative anatomy of the teeth of lower and non- specialized mammals. International Colloquium on the Evolution of Lower and Non-specialized Mammals. Brussels, Koninklijke Vlaamse Academie voor Weten- schappen, Letteren en Schone Kunsten van Belgie, pp. 215-320. Van Valen, L. 1965a. Treeshrews, primates and fossils. Evolution, 19(2): 137-151. 1965b. A middle Paleocene primate. Nature, 207 (4995) : 435-436. 1967. New Paleocene insectivores and insectivore classifica- tion. Bull. Amer. Mus. Nat. Hist., 135-217-284. Weitzel, K. 1949. Neue Wirbeltiere (Rodentia, Insectivora, Testudinata) aus den Mitteleozan von Messel bei Darmstadt. Abh. Senckenberg. Naturf. Ges., 480:1-24. West, R. M. 1973. Geology and mammalian paleontology of the New Fork-Big Sandy area, Sublette County, Wyoming. Fieldiana : Geology, 29:1-193. 36 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 Appendix A. Dimensions of lower teeth of Scenopagus edenensis, S. curtidens and S. priscus P4 Ml L W L AW S. edenensis YPM 13612 YPM 14932 YPM 14934 YPM 16910 2.3 — AMNH 11428 2.1 1.7 CM 13739 2.0 1.6 1.9 CM 13732 CM 13734 CM 13728 2.0 1.4 — — CM 6444 2.1 1.5 2.3 1.9 CM 6449 2.1 1.7 CM 13731 CM 13740 CM 6441 2.0 1.5 2.3 1.7 CM 17318 2.2 1.5 2.4 1.9 CM 13737 2.0 1.5 CM 13730 CM 13729 2.0 — 2.1 CM 13735 2.0 1.4 CM 6485 CM 13733 CM 13736 2.2 1.4 1.8 CM 13738 CM 6443 CM 17317 TTU-P-5571 TTU-P-5572 TTU-P-7161 2.1 1.5 2.3 1.7 TTU-P-7162 2.3 1.9 TTU-P-7163 2.1 1.8 TTU-P-7164 2.3 1.8 TTU-P-7165 2.1 1.7 TTU-P-7166 TTU-P-7180 TTU-P-7181 TTU-P-7182 TTU-P-7183 TTU-P-7184 TTU-P-7185 TTU-P-7186 TTU-P-7187 TTU-P-7188 TTU-P-7189 2.3 1.9 S. curtidens YPM 16913 1.8 1.6 YPM 14939 YPM 136 10-1 A 1.5 1.1 1.9 1.5 YPM 13610-lB 1.5 1.2 1.8 1.5 YPM 13610-2 1.7 1.4 YPM 15255 YPM 15254 1.5 1.2 1.8 1.5 YPM 16912 AMNH 12058 AMNH 12055 1.5 1.2 1.7 1.6 M2 Ms PW L AW PW L AW PW 2.5 1.8 1.4 2.6 1.9 1.9 1.9 2.5 2.0 1.8 2.7 2.2 2.0 2.7 — — 1.8 2.2 2.0 1.9 2.2 2.0 1.8 2.3 1.7 1.3 2.2 — — 2.3 1.7 1.4 1.7 2.1 1.9 1.8 2.4 1.8 1.5 1.7 2.3 2.0 1.8 2.3 2.0 1.9 2.0 2.1 — 1.9 2.2 1.9 1.8 2.2 2.1 1.8 2.4 1.7 1.4 2.1 1.9 1.8 2.3 1.7 1.3 — — 1.8 — 1.7 1.3 2.0 — — 2.3 1.8 2.0 1.7 1.9 1.7 1.9 2.2 2.0 1.9 2.1 1.8 — 2.3 1.9 1.8 2.3 1.9 1.8 2.2 1.9 1.8 2.3 2.0 1.9 2.3 1.9 1.8 2.2 2.0 1.9 2.4 1.7 1.2 2.2 1.7 1.3 1.6 1.8 1.6 1.4 1.5 1.8 1.6 1.5 1.6 1.8 1.6 1.5 1.5 1.7 1.5 1.4 1.8 — 1.4 1.6 1.2 0.9 1.6 1.8 1.6 1.5 1.6 1.2 1.1 1.5 1.8 1.6 1.5 1.6 1.3 1.2 1.5 1.8 1.7 1.5 — 1.3 — 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 37 Appendix A. Dimensions of lower teeth of Scenopagus edenensis, S. curtidens and S. priscus, continued P4 Mj L W L AW S. curtidens AMNH 12062 AMNH 12064 1.5 1.2 1.8 1.5 AMNH 11491 2.0 1.5 AMNH 48183 AMNH 59643 1.6 1.2 2.0 1.5 UCMP 96155 1.7 1.4 TTU-P-3928 — — TTU-P-7054 1.8 1.4 S. priscus AMNH 56060 1.4 1.1 AMNH 55156 1.3 1.0 1.5 1.3 CM 13753 1.3 1.0 1.5 1.2 CM 13749 1.5 1.2 CM 13750 1.3 0.9 1.5 1.2 CM 13744 CM 17316 CM 13746 CM 13743 1.1 1.5 1.1 CM 13748 1.5 1.1 CM 6486 CM 17319 1.2 0.9 1.5 1.1 CM 13751 1.3 0.9 1.4 1.3 CM 13752 1.4 0.9 1.5 1.1 TTU-P-7094 1.4 1.0 TTU-P-7098 1.3 0.9 TTU-P-7099 1.4 l.I TTU-P-7097 1.4 1.0 TTU-P-7096 1.3 0.8 TTU-P-7100 1.3 0.9 TTU-P-7082 1.4 0.9 TTU-P-7092 1.5 1.0 TTU-P-7093 1.3 0.9 TTU-P-7081 1.4 0.9 TTU-P-7083 1.4 0.9 TTU-P-7084 1.5 1.1 TTU-P-7085 1.5 1.0 TTU-P-7090 1.4 1.0 TTU-P-7095 1.4 1.0 TTU-P-7087 1.5 1.0 TTU-P-7086 1.5 1.0 TTU-P-7088 1.4 1.0 TTU-P-7120 1.4 0.9 TTU-P-7121 1.4 1.0 TTU-P-7119 TTU-P-7124 TTU-P-7125 TTU-P-7117 TTU-P-7118 TTU-P-7089 TTU-P-7122 TTU-P-7129 TTU-P-7131 TTU-P-7132 1.5 1.1 M2 M3 PW L AW PW L AW PW 1.5 1.7 1.6 1.4 1.2 1.5 1.5 1.9 1.5 1.6 1.6 1.5 1.7 1.5 1.4 1.7 1.3 1.0 1.6 1.7 1.6 1.5 1.5 1.3 1.0 1.5 1.2 1.5 1.3 1.2 1.3 0.9 — 1.3 1.2 1.2 1.4 1.3 1.1 1.3 1.4 1.3 1.1 1.1 1.1 1.5 1.3 0.9 0.8 1.1 1.5 1.5 1.1 1.0 1.2 1.4 1.2 1.0 1.5 1.2 1.0 1.3 1.0 0.7 1.3 1.4 1.3 1.1 1.3 1.2 1.0 1.2 1.1 0.9 1.0 1.0 1.1 1.0 1.1 1.1 1.0 1.0 1.1 1.1 1.1 1.0 1.0 1.0 1.5 1.2 1.0 1.4 1.1 0.8 1.6 1.2 0.9 1.5 1.2 1.0 1.4 1.2 1.1 1.4 1.2 1.1 1.4 1.1 1.0 1.5 1.3 1.1 38 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. I Appendix A. Dimensions of lower teeth of Scenopagus edenensis, S. curtidens and S. priscus, continued P4 Ml M2 M3 L W L AW PW L AW PW L AW S. priscus TTU-P-7130 1.4 1.1 1.0 TTU-P-7126 1.4 1.2 1.1 TTU-P-7127 — 1.1 — TTU-P-7128 1.3 0.9 0.7 TTU-P-7123 1.4 0.9 0.7 TTU-P-7140 1.5 1.0 0.8 TTU-P-7137 1.4 0.9 0.7 TTU-P-7138 1.3 1.0 — TTU-P-7139 1.3 0.8 0.6 TTU-P-7 1 35 1.3 0.9 0.7 TTU-P-7142 1.4 1.0 0.7 TTU-P-7134 1.3 — — TTU-P-7 141 1.3 1.0 0.7 TTU-P-7 143 1.3 1.1 0.8 TTU-P-7 136 1.3 0.9 0.7 1976 KRISHTALKA: ADAPISORICIDAE AND ERINACEIDAE 39 Appendix B. Dimensions of upper teeth of Scenopagus edenensis and S. priscus M 1 m2 m3 L W L W L W S. edenensis YPM 16911 2.3 3.3 CM 6433 — 2.6 2.3 3.1 2.2 3.5 CM 13741 2.0 2.6 2.2 3.2 2.0 3.4 TTU-P-7145 2.0 2.8 TTU-P-7146 2.0 2.9 TTU-P-7147 2.0 2.6 TTU-P-7148 2.0 2.7 TTU-P-7149 TTU-P-7150 2.0 2.6 2.2 3.2 TTU-P-7155 2.3 3.3 TTU-P-7156 2.4 3.2 TTU-P-7152 — 3.3 TTU-P-7160 — — TTU-P-7159 2.1 3.2 TTU-P-7153 — 3.1 TTU-P-7175 — 3.1 TTU-P-7151 TTU-P-7167 2.2 3.0 2.1 TTU-P-7169 2.1 3.1 TTU-P-7178 — 3.3 TTU-P-7168 — 3.4 TTU-P-7172 2.1 3.2 TTU-P-7179 TTU-P-8232 TTU-P-8233 TTU-P-8234 2.1 3.3 L 1.9 1.5 1.6 W 2.8 2.6 2.8 S. priscus TTU-P-7051 1.4 2.0 TTU-P-7060 1.4 TTU-P'7075 1.5 TTU-P-7074 — TTU-P-7062 1.5 TTU-P-7077 1.4 TTU-P-7080 1.6 TTU-P-7076 1.4 TTU-P-7059 1.4 TTU-P-7057 1.5 TTU-P-7053 1.5 TTU-P-7068 1.6 TTU-P-7065 1.5 TTU-P-7056 — TTU-P-7071 1.4 TTU-P-7079 — TTU-P-7072 1.5 TTU-P-7070 1.4 TTU-P-7063 1.5 TTU-P-7102 TTU-P-7116 TTU-P-7112 TTU-P-7109 TTU-P-7115 TTU-P-7107 TTU-P-7113 2.0 2.1 2.0 2.0 2.1 2.1 2.0 1.9 2.1 2.1 2.2 2.1 2.1 2.0 2.1 2.1 2.0 1.4 1.9 1.3 2.2 1.1 2.0 1.3 2.2 1.4 2.1 1.4 2.2 40 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 1 Appendix C. Dimensions of lower teeth of Macrocranion nitens and Macrocranion cf. M- nitens P4 Ml M2 M3 L W L AW PW L AW PW L AW PW M. nitens AMNH 48172 1.8 1.6 1.9 1.3 1.3 AMNH 48174 2.0 1.3 2.1 1.5 1.6 2.1 1.6 1.6 2.2 1.4 1.4 AMNH 48176 1.9 1.5 1.5 1.9 1.5 1.6 — — 1.4 AMNH 48178 1.7 1.2 1.9 1.4 1.5 2.0 1.4 1.6 2.0 1.2 1.3 AMNH 48179 1.6 1.1 1.9 — 1.5 2.0 1.4 1.5 AMNH 48180 1.8 1.1 1.9 1.5 1.6 1.8 1.5 1.6 AMNH 48182 2.0 1.5 1.6 AMNH 48184 AMNH 48187 1.9 1.4 1.5 2.0 1.5 1.6 2.0 1.2 1.3 AMNH 48189 1.7 1.2 1.9 1.4 1.5 2.0 1.5 1.6 AMNH 15697 1.8 1.2 1.9 1.3 1.6 2.0 1.4 — AMNH 14674 — 1.3 2.1 1.4 1.6 2.0 1.6 1.7 2.2 1.4 — Macrocranion cf. M. nitens AMNH 59673 1.6 1.3 1.5 1.7 1.4 1.5 1.7 1.2 1.3 AMNH 59662 1.4 0.9 AMNH 59656 AMNH 59663 AMNH 80071 1.4 1.0 1.8 1.4 1.5 1.8 1.4 1.5 1.7 1.3 AMNH 59679 1.4 0.9 1.7 1.3 1.4 AMNH 59651 1.4 0.9 AMNH 59688 AMNH 88846 1.7 1.1 1.8 1.4 1.5 AMNH 88834 1.8 1.4 1.5 AMNH 88839 1.8 1.3 1.4 AMNH 88830 AMNH 56318 1.5 1.0 1.8 1.4 1.5 1.7 1.5 1.6 ■A' m' %,1 .1 I I I BULLETIN 0/ CARNEGIE MUSEUM OF NATURAL HISTORY THE CLARK’S CAVE BONE DEPOSIT AND THE LATE PLEISTOCENE PALEOECOLOGY OF THE CENTRAL APPALACHIAN MOUNTAINS OF VIRGINIA DEDICATED TO THE MEMORY OF DR. J. KENNETH DOUTT CURATOR OP MAMMALS, CARNEGIE MUSEUM OF NATURAL HISTORY, FROM 1938 TO 1972 JOHN E. GUILD AY, Section of Vertebrate Fossils Carnegie Museum of Natural History PAUL W. FARM ALEE, Anthropology Department The University of Tennessee at Knoxville HAROLD W. HAMILTON, Section of Vertebrate Fossils Carnegie Museum of Natural History NUMBER 2 PITTSBURGH, 1977 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 2, pages 1-88, figures 1-21, tables 1-23 Issued January 19, 1977 Price: $12.00 a copy Publications Committee Duane A. Schlitter Harry K. Clench C. J. McCoy, Jr. Managing Editor R. E. Porteous ®1977 by the Trustees of Carnegie Institute, all rights reserved. CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE PITTSBURGH, PENNSYLVANIA 15213 CONTENTS Abstract 5 The Site 5 The Deposit 5 Procedure 6 Abbreviations and Acknowledgments 9 Regional Setting 10 History of Site 11 Clark’s Cave Raptors 15 Quantitative Analysis 16 Faunal List 19 Biotic Discussions Flora 24 Decapoda — Crayfish 24 Insecta — Insects 24 Mollusca — Snails and Clams 24 Pisces — Fish 25 Amphibia — Amphibians 25 Reptilia — Reptiles 27 Aves — Birds 27 Mammalia — Mammals 39 Faunal Comparisons — Methods of Deposition 71 Ecological Interpretation 73 Bird Summary 73 Mammal Summary 74 Age of Deposit 77 The Late Glacial Environment 78 Life Zone Integrity 80 Literature Cited 83 Fig. I. Looking north from Entrance No. 2, Clark’s Cave, Bath County, Virginia, across Cowpasture River. Tower Hill Mountain in back- ground. Fossil deposit in passageway left off photo. Sketch by J. R. Senior from H. Hamilton composite photo. ABSTRACT Remains of 142 species of vertebrates, ca 4,984 indi- viduals, and 35 species of invertebrates, ca 5,547 individuals, were recovered from a late Pleistocene “owl roost” in the entrance talus of Clark’s Cave in the central Appalachian Mountains, lat. 38°05'10" N., Bath County, Virginia, U.S.A. Sixty-four percent of the medium-to- small-sized mammals represented in the cave remains are now found either farther north or conform in size to present boreal population equivalents (Bergmann’s Response) and with the New Paris No. 4, Pa. local fauna. Deposition took place during late glacial times, > 10,000 years B P., and ceased before the boreal-to-temperate, post-glacial floral adjustment was completed at the site. Ptarmigan, Lagopus, and least chipmunk, Eutamias, are added to the late glacial fauna of the Appalachians. A spruce/ pine parkland with nearby bog and meadow- lands in a complicated topographical pattern is suggested by the ecological requirements of the faunal components. A contemporaneous mixture of presently allopatric north- ern, midwestern, and temperate Appalachian species is noted, suggesting a richer ecological picture than was apparent at New Paris No. 4, Pa., 240 km to the north. Quantitative analysis suggests that several species of raptors were involved. Comparisons with other Appa- lachian cave and fissure deposits demonstrate the difficulty of ascertaining methods of deposition by faunal analysis only. INTRODUCTION THE SITE Clark’s Cave, 12 km southwest of Williamsville, Bath County, Virginia (U.S.G.S. Williamsville quadrangle 15' series, latitude 38°05'10" N., longi- tude 79°39'25" W.), is located in north-central Virginia on the south bank of the Cowpasture River in the headwaters of the James River drainage (Fig. 2). It is a large cave with a complicated maze of over 2,400 m of cave passages. The six major entrances to the cave occur along a one km strip of high, pic- turesque, limestone cliff that towers some 30 m above the southern bank of the river. The lower Devonian limestone cliffs, (Helderberg group. New Scotland member (Bick, 1962) form a distinctive feature of the landscape (Figs. I and 6). Extensive talus at their base, still active because of steep profile and crum- bling cliff, forms a high, rocky tumble, forested, cool and shaded by the northward-facing cliffs. The cliff face and its numerous cave entrances make ideal roosts for birds of prey. Although no raptors nest in the cave entrances today, the barn owl (Tyto alba), red-tailed hawk (Buteo jamaicensis), and raven (Corvus corax) still frequent them as a temporary roost. These cliffs attracted raptors during the late Pleistocene. The extensive bone deposit here re- ported from entrance No. 2 (Fig. 3) represents but a small portion of a formerly more extensive accumu- lation of nesting and roosting debris. Facing north across the Cowpasture River from the large amphitheater of entrance No. 2 (Fig. 1), about 20 m above the river, one looks across approximately 50 m to a gradually rising pasture at about the same elevation (ca. 448 m, or 1,495 ft). One and one-half km beyond, the southern end of Tower Hill Mountain ascends to approximately 840 m (2,800 ft) elevation. Warm Springs Mountain, 6.4 km west of the cave, rises to an elevation of over 1,100 m. About a kilo- meter up- and down-stream from the cliffs (Figs. 8 and 9) the Cowpasture River broadens considerably, meandering across a flat valley floor, averaging .6 km wide, abruptly bordered by forested hills. These river flats were undoubtedly the source of most of the vertebrate fossils in this vole-dominated assembly. The rock inhabitants probably came from the talus immediately below the cave. THE DEPOSIT The fossil deposit, the Clark’s Cave local fauna (Field Site No. 3, Figs. 4 and 5), was discovered near the top of a loose, unconsolidated talus feeding down from a passageway that entered the amphi- theater of entrance No. 2 from the west (Figs. 4 and 5). Beginning in darkness, this talus cascaded at a 45° angle some 10 m, where it joined the main am- phitheater talus fronting the cliffs. When first ex- amined in the spring of 1972, bone fragments were noted throughout the talus accumulation. But the rare coincidence of unusually heavy rains that sum- mer, coupled with a heavy rockfall from the cliff face above, caused a rock slide that removed three- quarters of the fossiliferous matrix. Only that por- tion that lay at the top of the talus was left intact. An excavation (Figs. 4, 5 and 7) 2.1 m x 1.4 m x 0.5 m, approximately 3.9 cubic m, was made at the top of the talus in the fossiliferous area. The matrix 5 6 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 2. Site location, Clark’s Cave, Bath County, Virginia. Upper right: exaggerated cross-section of mid-Appalachians showing precipitation pattern, modified from Core, 1974. was 90% cliflf-wastage — ungraded, loose, frost-wedged fragments of Helderberg limestone, ranging in size from tiny spalls to large blocks. The remaining 10% consisted of organic remains and a dark-brown, dry soil-like material, probably worked into the raw talus by wind and gravity accretion. Open cracks between the rocks permitted material to filter down through the limits of the talus prior to and during excavation. The detritus varied from 30 cm to 46 cm in depth and rested on sterile cave clay. Organic inclusions (occurring throughout) were seeds, leaves (see floral listing), woodrat droppings, snails, egg- shell fragments, arthropod chitin, and thousands of small bones and teeth. No Recent owl pellets were encountered during excavation. There was no evi- dence of hair or feathers that occur in Recent owl pellet accumulations. These had apparently decom- posed through the years, leaving only a lag deposit of bones and teeth of prey items. PROCEDURE Approximately 540 kg of matrix was dry-screened through a 6 mm mesh (1 /4-inch hardware cloth) and large items were removed and boxed at the site. One hundred-eighty kg of screen-concentrate was bagged and washed through 1 mm mesh (commercial win- dow-screening) at the New Paris field laboratory. All macroscopic organic inclusions were removed. Further sorting of bones and teeth was carried out at Carnegie Museum under the direction of Helen McGinnis and the senior author, and various cate- gories were submitted to specialists for identification (see Faunal List). All specimens, with the exception of a few passenger pigeon remains, are stored in the Section of Vertebrate Fossils, Carnegie Museum of Natural History. Selected pigeon bones are stored in the avian osteology collection. Department of An- thropology, The University of Tennessee at Knox- 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 7 Fig. 3. Map of entrances to Clark’s Cave, Bath County, Virginia. ville. The loose nature of the matrix and its exposed situation made pollen analysis and carbon-14 dating (see Age of Deposit) unreliable. A carbon date was run on small mammal bones, and the woodrat drop- pings were saved for possible pollen content The three primary study objectives were: 1. Identi- fication to species if possible; 2. Establishment of minimum numbers of individuals for each taxon; 3. Assessment of possible size-differences between fossil and Recent population equivalents. In the case of mammals these objectives were accomplished by studying the dentitions and cranial parts. Mam- malian post-cranial material, with a few exceptions like talpid humeri, received only cursory attention. They were not classified or catalogued, but they are stored with the collection. Bird, reptile, amphibian, and fish identifications were based primarily on postcranial elements. The minimum number of individuals of each spe- cies (MNI) was established by the highest replication of any diagnostic elements — tooth, limb bone, etc. These elements differed from species to species. The voles, for example, were represented primarily by isolated teeth and partial maxillae and dentaries. Approximately 17,000 isolated arvicolid molars were recovered from the screen washings. Individual molars were specifically identified, then tallied according to their serial position, first, second, third molars, whether upper or lower, right or left. These categories were then tallied and the greatest number for any given tooth was considered the minimum number of individuals represented for that par- ticular species. The same procedure utilizing various skeletal parts was employed for the entire vertebrate collection. Elements utilized are listed under mate- rials in the individual species accounts. Preservation of bones and teeth was chemically excellent, but mechanically poor. Bone appeared 8 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 4. Excavation plan, Entrance No. 2, Clark’s Cave, Bath County, Virginia from above. unaltered and frequently could not be distinguished from Recent owl pellet material from the area. Some specimens did have a thin carbonate crust, but none were fused. Bones, shells, and egg shells were badly broken, often ground to fragments on the sharp rock matrix by talus creep and excavating and re- covery procedures. Very small specimens — shrew and bat dentaries, small molluscs — fared best, but larger objects were invariably badly fractured. Teeth, large or small, were usually separated from skull parts. No articulated remains were found. Size sorting was not present and all bone distribution, both horizontal and vertical, in the deposit appeared to be random. Measurements less than one centimeter were taken with the aid of a Spencer Cycloptic stereoscopic mi- croscope, using an ocular grid at 10 X. The partic- ular instrument used (No. 4451931) required a correction factor of 0.97 to convert ocular measure- ments to true millimeters at that magnification. Larger measurements were taken with a dial mi- crometer calibrated to 0.1 mm. Responsibility for various sections are as follows: geology and site mapping, Hamilton; birds, Par- malee; mammals, Guilday. Remaining sections are by Parmalee and Guilday jointly. Specialists involved in identifications are listed in Acknowledgements and in the faunal list. To avoid repetition, previously reported bone sites are not cited in the text. References may be found in Table 23 (site references). Avian nomenclature follows the A.O.U. Checklist of North American Birds (5th edition) and its 32nd supplement (Auk 90:411-419). Mammalian nomen- clature is largely from Jones, Carter, and Genoways, 1975. The bird bones from Back Creek Cave No. 2, Bath Co., Virginia, are included in the inventory of bird remains, but are catalogued and listed separately 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 9 under appropriate species. This site, a shallow “rock shelter” 24 km west of Clark’s Cave, is a late Pleis- tocene owl roost deposit similar to that at Clark’s Cave. Most of the mammals have not yet been stud- ied as of this time, but the faunal make-up of the two sites appears to be similar. ABBREVIATIONS AMNH — American Museum of Natural History; B.P. — before present; ca — approximately; CM & CMNH — Carnegie Museum of Natural History; cm — centimeters; CV — coefficient of variation; g — grams; I — Isotopes, Inc.; kg — kilograms; km — kilometers; m — meters; MNI — minimum number of individuals; N — sample size; OR — observed range; SD — standard deviation; SI — Smithsonian Institution; USNM — United States National Museum; SIU — Southern Illinois University; X — arithmetic mean; > — greater than. Capital and lower-case letters followed by on- line numbers [C/c (canine). P/p (premolar). M/m (molar)] refer to upper and lower teeth, e.g., Ml,ml. ACKNOWLEDGMENTS We are indebted to many people who contributed services and expertise to this project, and take great pleasure in acknowledging their assistance. Mrs. W. G. Clark, the owner of Clark’s Cave, kindly allowed us to excavate and collect on her property. Able assis- tance during excavation and field screening was provided by Lee Ambrose, Alan Bailey, Mr. and Mrs. Nevin C. Davis, Karen Downing, Gwen Foster, Mary Ann Gross, Rita and Allen Hamilton, Paul, Helen, and Mimi Imblum, Mr. and Mrs. William King, and Mr. and Mrs. Robert New. Processing of matrix at the New Paris field laboratory under the co-direction of Allen D. McCrady and Harold W. Hamilton in- volved so many willing volunteers that space permits only a general, but warm thank-you to all. Helen 10 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 2 McGinnis deserves our special thanks for her many months of laboratory sorting and for the study and identification of the anuran remains. We thank our many colleagues for their assistance: Plants: Pel Brunett, Chippewa Nature Center, Mid- land, Mich. Molluscs: Leslie Hubricht, Meridian, Miss.; Dr. William J. Clench, Harvard University. Decapods: Dr. Horton H. Hobbs, Jr., USNM. In- sects: Dr. Everett D. Cashatt, Illinois State Museum. Fish:'!. N. Todd and Dr. Robert R. Miller, University of Michigan. Reptiles and amphibians: Dr. Clarence J. McCoy, Jr., CMNH. Birds: Dr. Mary H. Clench and Dr. Kenneth C. Parkes, CMNH; Dr. Storrs L. Olson, Dr. Richard L. Zusi, and Dr. Alexander Wet- more, USNM; Dr. Amadeo M. Rea, University of Arizona, Tucson. Mammals: Dr. Elaine Anderson, Denver; Caroline A. Heppenstall, Dr. J. K. Doutt, and Dr. Duane A. Schlitter, CMNH; Dr. Karl Koop- man, AMNH; Dr. E. C. Galbreath, SIU; Wayne E. Clark, Assistant Archaeologist, Virginia State Li- brary (who supplied archaeological data). Illustra- tions are by Erica Hanson (Fig. 19), James R. Senior (Figs. 1 - 5, 16) and Nancy Perkins (Figs. 1 1 - 15, 17, 18, 20, 21). Elizabeth Hill, Wendy Pollock, Kathleen Guilday, Albert Kollar, and Ronald Wilson provided laboratory or clerical assistance. We wish to thank Dr. Mary R. Dawson, CMNH, Dr. Charles O. Hand- ley, Jr., USNM, and Dr. Holmes A. Semken, Univer- sity of Iowa, for reviewing the manuscript. Dr. Hand- ley also surveyed the Recent small mammals from the area, and we thank him for his many notes on the Recent vegetation, birds and mammals of the cave area, and for the loan of specimens. The authors are indebted to their wives, Alice, Barbara, and Rita, for things innumerable. Research was con- ducted under NSF Grant No. GB 42258, awarded to the senior author. REGIONAL SETTING The Cowpasture River rises in the central Appala- chian Mountains. It meanders southwestward for some 72 km in an intermontane valley to join the James River, which flows east 400 km to the Atlantic Ocean. For most of its length, the valley of the Cow- pasture River is broad and flat, averaging perhaps 0.8 km in width. In Bath County, Tower Hill Mountain, rising to 970 m, borders the northwestern side of the valley. The mountain crest follows an anticlinal ridge of the resistant Middle Silurian Clinton Formation, a heterogeneous mixture of sandstones and shales. The southeastern rim of the valley is formed by Shen- andoah Mountain, a resistant synclinal core of Up- per Devonian Chemung sands and conglomerates. The river follows the more easily eroded trough be- tween these mountains. Sixteen km northeast of the cave Shenandoah Mountain rises to over 1,060 m but loses altitude to the south. Approximately two km directly south of the cave, a dissected range of hills, “the spurs,” rises to 760 m. About 1.5 km upstream from Clark’s Cave, the river, flowing along the flat valley floor (Figs. 9 and 10), has worn through the structurally older Ridgeley sandstone and flows through an abruptly tightened valley, narrow and precipitous, carved out of the underlying Lower Devonian Helderberg limestone. The Cowpasture valley retains this gorgelike charac- ter past the cave mouth and for another km down- stream, where the valley suddenly broadens again (Fig. 8) as the river emerges once more into shale lands. High mountainous country west of the Cowpasture River valley, rising to over 1,000 m, diverts much of the moisture from the prevailing westerlies, and the valley lies within their rain shadow. This produces a mild xeric effect that has resulted in the formation of a distinct topographically and botanically defined provincial subdivision: the shale barrens. Shale barrens occur along the crests of low shale hills that lie just to the east of the plateau country, from south- ern Pennsylvania to southern Virginia (Keener, 1970). The present climate of the area is temperate and mild. Observations over a forty-year period at Hot Springs, Bath County, 20 km southwest of the cave, show a temperature range from 36.7° to -28.9° C. (98° to -20° F.) with a January average of 0.2° C. (32.4° F.), and a July average of 20.8° C. (69.5° F.). There is no definite wet or dry season, but pre- cipitation at Hot Springs is highest in June and low- est in January, with a yearly average of 105 cm. The weather is controlled by a succession of cyclonic pressure cells from the west, with (especially in the summer months) warm, moist air masses from the Caribbean, resulting in a varied day-to-day climatic pattern. Western Virginia is too far inland for its climate to be directly affected by the Atlantic Ocean except during rare hurricane phenomena. Winters are generally mild, except in the mountain and plateau highlands. Yearly snowfall averages 63.5 cm and usually does not provide continuous winter cover (Hibbard, 1941). Just prior to Colonial settlement, the area sur- rounding the cave was completely forested with an oak-dominant vegetation (Braun, 1950; Hack & Good- 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 11 lett, 1960; Clarkson, 1966; Core, 1974). The slope on which the cave is located was undoubtedly covered, then as now, with a mesic Appalachian cove forest: hemlock (Tsuga), arborvitae (Thuja), tulip poplar (Liriodendroti), and basswood (Tilia). This same for- est is found in gorges on Tower Hill Mountain and Warm Springs Mountain, west of the cave. The na- ture and composition of the forest is in subtle con- cordance with local variations in elevation, exposure, drainage patterns, and soils. As a general rule, val- ley floors and higher mountain summits receive the most moisture, the former from precipitation plus slope drainage, the latter through precipitation alone. Valley floors of the area originally were covered with a white-oak (Quercus u/6a>dominant forest, while the dryer mountain slopes were covered with a chestnut/ chestnut-oak (Castaneaj Quercus /jmnw^j-dominant forest. In higher, moister, moun- tainous areas, and especially in high valley heads, the forest became richer in species of maple (Acer) and birch (Betula). Spruce f/’/ccaj-dominated woodland occurs, or did occur, on many of the higher summits to the west, but not near the Cowpasture River valley. The dis- tinctive shale barrens, open woodlands of low scrub oak (Quercus ilicifolia) and pines (Pinus, spp.j with much bare ground, occupy the low hills bordering the southeastern rim of the valley (Fig. 9), forming a distinct biotype characterized by many low-order plant endemics. The shale barrens appear today as a distinct, open, mildly xeric habitat. They would cer- tainly have been in the cruising range of the Clark’s Cave raptors if they were in existence during the late Pleistocene. The presence of plant endemics argues that the barrens are at least that old, and they may have supported some of the Midwestern forms pres- ent in the deposit, like the sharp-tailed grouse (Pedio- ectes phasianellus), the least chipmunk (Eutarruas minimus), and the thirteen-lined ground squirrel (Spermophilus tridecemlineatus). The floor of the Cowpasture River valley, like that of the broad Shenandoah valley to the east, lies within the Carolinian Life Zone (Dice, 1943). This zone occupies the valley floors of western Virginia up to about 460 m and ascends to 600 m in open country. Its upper limit may dip as low as 370 m, how- ever, in forested ravines (Murray, 1945). All the mountain ridges, with the exception of a few summits above 1,200 m, lie in the Alleghenian or Transition Zone. According to Murray, the theoretical upper limits of the Alleghenian Life Zone should lie be- tween 915 m and 1,070 m in Virginia. But biological zonation is not a simple function of altitude. It varies with slope exposure, air drainage, local pre- cipitation, and soil characteristics. Zones merge gradually, and the upper limits of the Transition Zone in the Virginia and West Virginia mountains range from 1,000 m (Cranberry Glades, W. Va.) to 1,364 m (Elliott’s Knob, Va.). Areas of Canadian Zone, spruce-dominated flora top a few of the higher peaks. “While there is a good deal of Canadian Zone ter- ritory in the high Allegheny Plateau of West Virginia [700,000 acres before logging. Brooks, 1943, p. 25], and a fair area of it on the great peaks of the Smokies along the North Carolina and Tennessee line there is little or no territory in Virginia which can be called pure Canadian. On White Top and Mt. Rogers, the two highest mountains in Virginia which reach 5,519 [1,672 m] and 5,720 feet [1,723 m] respectively, we have some small areas which are practically Cana- dian, and on Middle Mountain in Highland County we have some territory that approaches it.’’ (Murray, 1945 : 20) After logging, the Canadian Zone spruce stands are replaced by Transition Zone mixed hard- wood/coniferous forest. It can be seen that great topographic and biotic diversity, ranging from mountain tops to valley floors, and great variations in subsurface water, ranging from the xeric shale hills of the shale barrens to river flood plains, were present within the cruising range of the Clark’s Cave raptors. Thus the inter- pretation of the fossil fauna from a regional point of view is complicated. HISTORY OF SITE The history of man’s occupation of the Cowpasture River valley may be related to the age of the fossil deposit. Unmolested, the nesting of birds of prey at the site could proceed with little interruption. But the arrival of man in the valley may have brought nesting to an end because the former raptor roost at entrance No. 2 was easily accessible from the river. The fauna recovered from the site suggests that de- position ceased prior to the establishment of the modern regional biota about 10,000 to 11,000 years ago, and current archaeological evidence suggests that paleo-Indian occupation of western Virginia ex- tended at least that far back in time. No direct evidence of such an early occupation has been re- corded from the Cowpasture valley, but an impor- tant paleo-Indian complex, the Thunderbird Site (44 WR 11) (Gardner, 1974), is known from the Shenandoah valley, about 70 km to the east. This site, on an alluvial fan of the South Fork Shenandoah ' lu 12 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 6. View looking southeast, from across Cowpasture River, at Clark’s Cliffs, early Devonian Helderberg limestone. Entrance No. 2 and fossil site off photo to right. Note active, wooded talus. Hamilton photo. 1977 GUILDAY, PARMALEE, AND HAMILTON. CLARK’S CAVE BONE DEPOSIT 13 Fig. 7. Mary Ann Gross (lower left) on excavation site in unconsolidated talus of west passage. En- trance No. 2, Clark’s Cave, Bath County, Virginia. 14 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 8. Cowpasture River, 2 kilometers downstream from Clark’s Cave, Bath County, Virginia. Cave is behind hill to right. Note V-shaped gorge as river traverses Helderberg limestone, and flat valley floor as it emerges into shale lands. River, is centered around an aboriginal jasper quarry from which paleo-Indian hunters periodically re- furbished their tool kits. The easy access to Clark’s Cave roosts may have invited exploitation and dis- ruption at an early date, forcing the birds to seek more secure nesting sites. Four prehistoric archaeological sites in the Cow- pasture River valley itself are recorded in the Virginia State Library archaeological survey site records (site nos. 44 HL 12, 44 HL 14, 44 BA 1, 44 BA 36). One, 44 BA 1, is in a setting quite similar to the Clark’s Cave cliff’s. It is a rockshelter in a small gorge formed where the Cowpasture River traverses a tongue of Devonian limestone that crosses the river valley 13 km down stream from Clark’s Cave. Other small Indian campsites are recorded from the nearby Bull- pasture and Jackson River valleys (MacCord, 1973a, 1973b). The valley was surveyed in the early 1740’s fol- lowing a short period of Indian wars, and per- manent European settlement began. Timber was cleared first in the valley floor and then in the surrounding mountains. The valley soon became a farming and livestock area. The cave became the site of a thriving saltpeter industry established to satisfy the demand for the substance in gunpowder manufacture. Clark’s Cave is reputed to have had over 200 men mining cave earth (Faust, 1964). With the decline of the saltpeter industry at the close of the Civil War in 1865, the site was left relatively un- disturbed. It was surrounded by farming activities and was occasionally visited by the public. It later became a popular resort of weekend cave explorers, but was closed to caving in 1973. In summary, Clark’s Cave and the Clark’s cliff's have been exposed to potential human exploitation for at least the past 10,000 years. Any evidence of an actual Indian occupation at any of the entrances of 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 15 Clark’s Cave would have been destroyed by mining activities, and we are indeed fortunate that site No. 3 was not destroyed as well. Only its location, perched high on a talus slope in a dead-end passage, preserved it. CLARK’S CAVE RAPTORS The fossil bone deposit accumulated in the twilight zone of the cave around the roosts and nests of car- nivorous cliff/ cave-frequenting birds, primarily owls. The bulk of the deposit represents prey collected from the neighboring environs and is the result of the build-up of food debris, primarily owl pellets, through the course of many years. The variety of prey, noc- turnal and diurnal, field and forest, vertebrate and invertebrate, argues for not one, but probably many species of birds of prey. Owls, hawks, eagles, vul- tures, ravens — all cliff-frequenting species — may have contributed at one time or another to the deposit. Raptor bones actually recovered included those of two accipiters, one falcon, and three owls — the great- horned, Bubo virginianus; the long-eared owl, or short-eared owl, Asio, sp.; and saw-whet owl, Aego- lius acadicus, the latter possibly a prey item. Large owls like the snowy owl, Nyctes scandiaca, and the great-horned owl prefer larger game than do smaller species like the short-eared owl; the barred owl, Strix varia; or the barn owl, Tyto alba. When the prey items of the Clark’s Cave deposit are class- ified according to the size of the living animals, the collection resembles that of the prey of medium- sized field-hunting owls, the barn owl, and the short- eared owl (Table 1). Mammals of rabbit size or larger comprised only 0.5% of all prey items analyzed. This accords with the figures for the medium-sized owls. In the larger barred owl, rabbits increase to 6.5%; in the still larger snowy and great-horned owls, from 24% to 46%. They are unrepresented in the diet of smaller species like the screech owl, Otus asio. The candidates most likely to be responsible for the bulk of the deposit are the eared owls (Asio, sp.) and the barn owl. The high percentage of field forms in the deposit, primarily voles of the genus Microtus, suggests a predator that hunts in open country, like the barn owl and the short-eared owl. Basically the short-eared owl (A. flameus) is an open grassland-roosting bird, while the long-eared owl (A. otus) is a pine/ brush thicket-roosting species. Table 1. Diet of various owls from eastern North America. 1 Species Owl total length % Rabbit class % Squirrel class %Mice class %A11 bird N Nyctea scandiaca snowy owl 20" 24.0 12.0 20.0 39.0 251 Bubo virginianus great horned owl 20" 46.0 11.0 33.0 17.0 2,714 Strix varia barred owl 17" 6.5 7.0 78.0 8.0 521 Tyto alba barn owl 14" .1 1.2 95.0 1.4 8,151 Asio otus long-eared owl 13" .5 .3 94.0 2.4 2,112 Asio flammeus short-eared owl 13" .3 5.0 93.0 2.0 263 Aegolius funereus boreal owl 10" — — 93.0 7.0 15 Otus asio screech owl 8" — — 83.0 6.0 73 Aegolius acadicus saw-whet owl 7" — — 96.0 4.0 21 owl, ?species Clark’s Cave, sample Pleistocene .5 2.8 95.0 2.2 4,562 In - Stomach/pellets examined except for Clark’s Cave, where it indicates animals in fossil prey population. Raw data from Latham, R.M., 1950. minimum number of 16 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 9. Mildly xeric shale barren habitat on low shale hills of southern side of Cowpasture River valley. Fig. 10 taken from same spot. Right edge of photo overlaps left edge of Fig. 10. The barn owl is associated primarily with the Caro- linian Life Zone. It occurs no farther north than southern New York State, and avoids the higher por- tions of the Appalachian Mountains. It does occur at the site today. Both feathers and fresh pellets were picked up along the cliff face by Parmalee in July 1974. The predominantly boreal character of the fossil-prey species, many of which do not occur in the Appalachians today, reflecting the effects of the Wisconsinan glaciation, do not appear to be likely barn owl prey, given the present ecological requirements of that species. If the long-eared owl inhabited the cave it could have been responsible for a large portion of the boreal meadow voles in the de- posit. It is possible, however, that the barn owl, despite its modern temperate range, may have inhab- ited the cave during the late Pleistocene. Small egg-shell fragments were distributed ran- domly throughout the deposit. Dr. Mary H. Clench compared them with modern eggs from the CMNH collections, but they could not be identified with certainty. At least some of the fragments appear to be from birds of large size e.g., eagle or vulture. In summary, the Clark’s Cave fossil deposit rep- resents nest debris of some medium-size owl, or owls, with additions from other raptorial species through- out its long depositional history. QUANTITATIVE ANALYSIS All birds and small mammals recovered from the cumulative body weights of the 1 14 species of birds fauna are Recent species, so that the Clark’s Cave and mammals considered to be raptor prey. The rela- fossil fauna can be expressed quantitatively by the tive contribution to the raptor diet of the various 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 17 Fig. 10. Oak ( 0MercM^)-forested hills along northern side of Cowpasture River valley, 2 kilometers upstream from Clark’s Cave, Bath County, Virginia. Note broad flood plain, probable source of many of the fossil voles from the cave deposit. Cave out of sight around far left hill. River flows along base of forested hills from right to left. species or species groups represented can be ascer- tained. Average live weight of individuals of each species considered raptor prey (Faunal List, Table 4) was multiplied by the minimum number of individ- uals of that species recovered. The resulting cumula- tive weights had little meaning as far as expressing actual totals consumed or the true relative caloric value of each species in the raptor diet, because of the many variables involved. But when these figures are compared, they give some indication of relative contributions to the raptor diet. Weights were ob- tained from the literature and represent average approximations. Various authorities did not always agree. Mammal weights were taken from Banfield, 1974, Doutt et ai, 1973, and Youngman, 1975. Bird weights were furnished by Dr. Kenneth Parkes, who stresses that “average” figures, especially for birds with great sexual weight disparity, like the turkey or sharp-shinned hawk, are rough indeed. Despite the unavoidable crudeness of the analysis, trends emerge that probably parallel the true picture. The total estimated biomass represented by the recovered fossils was 277.82 kg, 73.5% (204.31 kg) mammal remains, 26.5% (73.51 kg) birds. Results are summarized in Table 2 by families. Voles (Arvico- lidae) contributed most heavily both in terms of individuals (45.2%) and in terms of body weight (36.6%). Although birds accounted for only 4.6% of the combined numbers of individual birds and mam- mals, grouse (Tetraonidae) accounted for 13.6% of the estimated biomass, ranking second only in impor- tance to the voles. Bats, ranking second in terms of individual animals recovered (34.0%), contributed only 5.5% to the esti- mated biomass. Voles, rabbits and hares (Leporidae), and grouse accounted for 63% of the estimated biomass. Other small rodents (Cricetidae, Sciuridae, Zapodidae) comprised an additional 15%. Although there were more species of birds than mammals, they — with the exception of grouse, pigeons (Colum- 18 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 bidae), and ducks (Anatidae) contributed little to the raptor diet. In terms of biomass, hares ranked third (Table 2) because of their high individual body weight, but they contributed only 0.5% of the number of individ- ual birds and mammals recovered. Of the 15 most common species, in terms of numbers of individuals recovered from the site (Table 3), not a single bird ranked. Eight species of small rodents, 4 bats, and 2 shrews (12.5% of the 120 species of birds and mam- mals from the site) accounted for 86% of the 4,555 individuals recovered. In summary, the deposit was overwhelmingly com- posed of small rodents and bats, and in terms of individuals, all other species found were present in incidental numbers. But in terms of biomass, voles, grouse, and hares were most important in their contribution to the avian diet. Table 2. Biomass and minimum numbers of individuals from Clark’s Cave.^ Estimated Live weight Family kg % 1. Arvicolidae — voles 101.59 36.6 2. Tetraonidae — grouse 37.78 13.6 3. Leporidae — hares 35.47 12.8 4. Cricetidae — deer mice and wood rats 23.10 8.3 5. Sciuridae — squirrels 19.93 7.2 6. Vespertilionidae — bats 15.22 5.5 7. Columbidae — pigeons 9.00 3.2 8. Anatidae — ducks 8.54 3.1 All others (40 famiUes) 27.19 9.7 1 Avian and mammalian families ranked in order of representation, Numbers of individuals Family Individuals % 1. Arvicolidae - voles 2,060 45.2 2. Vespertihonidae — bats 1,554 34.0 3. Cricetidae — deer mice and wood rats 272 5.9 4. Soricidae — shrews 231 5.0 5. Sciuridae - squirrels 117 2.5 6. Tetraonidae — grouse 75 1.6 All others (42 families) 253 5.5 Table 3. Fifteen most abundant vertebrates^ in Clark’s Cave. Species Individuals % 1. Myotis lucifiigns/sodalis/keenii — little brown bats c. 877 19.3 2. Microtus pennsylvanicus — meadow vole c. 658 14.5 3. Microtus xanthognathus - yellow-cheeked vole 511 11.2 4. Eptesicus fuscus — big brown bat 363 8.0 5. Clethrionomys gapperi — red-backed vole 305 6.7 6. Microtus chrotorrhinus — rock vole c. 292 6.4 7. Microtus pinetorum - woodland vole 170 3.7 8. Myotis leibii — Leib’s bat c. 138 3.0 9. Myotis grisescens -gray bat c. 138 3.0 10. Peromyscus nianiculatus — deer mouse c. 117 2.6 11. Peromyscus leucopus - white footed mouse c. 104 2.3 12. Blarina brevicauda — short-tailed shrew 97 2.1 13. Sorex cinereus — masked shrew 67 1.5 14. Synaptomys borealis — northern bog lemming 61 1.3 15. Phenacomys intermedins — heather vole 34 0.1 104 additional species 618 14.3 1 Ranked by numbers of individuals. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 19 Table 4. Plant and animal remains, Entrance 2, Site 3, Clark’s Cave. Scientific name Common name MNI FLORAI (Identified by F. Brunett) 1. Tsuga cf. canadensis hemlock — needles 2. Thuja occidentalis arbor vitae — branchlets, seeds 3. Cary a, ? species hickory — nut fragment 4. Quercus, ?species oak — acorn fragment 5. Celtis occidentalis hackberry c. 150 seeds 6. Phytolacca decandra pokeweed c. 35 seeds 7. Vitis, ? species grape 12 seeds 8. Nyssa sylvatica black gum 9 seeds CRUSTACEA (Identified by H.H. Hobbs, Jr.) 1. Cambants cf. bartonii crayfish 4 2. Cambarus cf. longidus crayfish 2 insectaI (Identified by E.D. Cashatt) 1. Dicaelus, Impedes Carabid (ground) beetle 1 2. Galerita cf. bicolor Carabid (ground) beetle 1 3. Calasorrm, ?species Carabid (ground) beetle 2 Carabidae, ?species Carabid (ground) beetle 9 4. Onthophagus cf. janus scarab beetle 5 Onthophagus, ?species scarab beetle 1 5. Canthon, ?species scarab beetle 1 6. Balbocerosoma, ?species scarab beetle 1 7. Trox, ?species trogid (hide) beetle 1 8. Elateridae, ?species elaterid (click) beetle 1 Coleoptera, ?species beetles, unidentified 6 9. Membracidae, ?species leafhopper 1 10. Calliphoridae or Sarcophagidae fleshflies pupae 11. Vespidae, ?species social wasps 2 GASTROPODAI Snails 4,363 (Identified by L. Hubricht and W.J. Clench) 1 . Hendersonia occulta land snail 2 2. Vallonia costata land snail 5 3. Gastrocopta armifera land snail 1 4. Gastropoda contracta land snail 1 5. Vertigo tridentata land snail 3 6. Vertigo gouldi land snail 3 7. Columella simplex land snail 1 8. Catinella, ?species land snail 1 9. Anguispira alternata land snail 5 10. Discus cat skillensis land snail 7 ^Not considered owl prey. 20 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Entrance 2, Site 3, Clark’s Cave (continued). Common name MNI Table 4. Plant and animal remains, Scientific name GASTROPODA (continued) 1 1 . Helicodiscus parallelus 12. Helicodiscus inermis 13. Helicodiscus jacksoni 14. Stenotrerm hirsutum 15. Stenotrerm fraternum 16. Triodopsis burchi 17. Triodopsis vulgata 1 8 . Triodopsis juxtidens 19. Strobilops labyrinthica 20. Spirodon carinata bivalviaI (Identified by W.J. Clench) 1. Sphaerium striatinum 2. Pisidium dubium OSTEICHTHYS (Identified by J.E. Guilday) 1. Anguilla cf. bostoniensis 2. Esox cf. americanus or niger 3. Semotilus cf. corporalis 4. Nocomis cf. raneyi Cyprinidae, unidentified 5. Catostomus, ?species 6. Moxostorm, ?species Catostomidae, unidentified 7. No turns, ?species Ictaluridae, unidentified AMPHIBIA (Identified by H. McGinnis) 1. Bufo (amerieanus gtoup) 2. Hyla cf. crucifer 3. Rana cf. catesbiana 4. Rana cf. clamitans 5 . Rana cf. pipiens 6. Rana cf. palustris 7. Rana cf. sylvatica Rana, unidentified 8. Ambystoma, ?species 9. Plethodontidae or Salamandridae REPTILIA (Identified by J.E. Guilday) land snail 12 land snail 2 land snail 2 land snail 3 land snail 1 land snail 1 land snail 1 land snail 5 land snail 7 freshwater snail c. 4,300 Clams 140 fingernail clam 1 fingernail clam 139 Bony fishes 57 American eel 1 pickerel 1 creek chub river chub 26 unidentified minnows white sucker \ redhorse sucker ) 18 sucker stonecat small catfish Frogs, toac s, 11 salamanders 328 toad 141 peeper 20 bullfrog 5 green frog 2 leopard frog 22 pickerel frog 22 wood frog 42 frog 38 mole salamander c. 4 salamander c. 32 Reptiles 37 1. cf. Testudinidae turtle 1 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 21 Table 4. Plant and animal remains, Entrance 2, Site 3, Clark’s Cave (continued). Scientific name REPTILIA (continued) 2. Sceloporus cf. undulatus 3. Eumeces cf. laticeps 4. Colubridae, ?species 5. Crotalidae, ?species AVES (Identified by P.W. Parmalee) 1 . Podilymbus podiceps 2. Botaunis lentiginosus 3. Anas cf. platyrhynchos or rubripes 4. cf. Anas crecca 5. cf. Anas discors 6. Lophodytes cucullatus 7. Mergus, ? species Anatidae, ?species 8. Accipiter striatus 9. Buteo cf. plat ypt eras Accipitridae, ?species 10. Falco sparverius 11. cf. Canachites canadensis 1 2. Bonasa umbellus C, canadensis or B. umbellus 1 3. Lagopus cf. mutus 14. cf. Pedioecetes phasianellus Tetraonidae, ?species 1 5 . Colinus virginianus 16. Meieagris gallopavo 17. cf. Rallus limicola 18. Porzana Carolina 19. Gallinula chloropus Rallidae, ?species 20. Pluvialis dominica 21. Philohela minor 22. Capella gallinago 23. cf. Actitis macularia 24. cf. Tringa solitaria 25. cf. Limosa, ?species Scolopacidae, ?species 26. Ectopistes migratorius 27. Coccyzus, ? species 28. Otus asio 29. Bubo virginianus Common name fence lizard broadheaded skink non-poisonous snake rattlesnake or copperhead Birds pied-billed grebe American bittern mallard or black duck green-winged teal blue-winged teal hooded merganser merganser ducks, unidentified sharp-shinned hawk broad-winged hawk hawks, unidentified American kestrel spruce grouse ruffed grouse spruce or ruffed grouse c. rock ptarmigan sharp-tailed grouse grouse, unidentified bobwhite wild turkey Virginia rail sora rail common gallinule rail, unidentified American golden plover American woodcock common snipe spotted sandpiper sohtary sandpiper godwit sandpiper/plover passenger pigeon cuckoo screech owl great horned owl MNI 13 2 . 20 1 Gram live weight Total gram Uve weight 219 2 135 270 1 475 475 2 1,160 2,320 1 350 350 1 400 400 2 610 1,220 1 c. 750 c. 750 5 c. 700 c. 3,500 1 135 135 1 400 400 2 c. 575 c. 1,150 2 115 230 14 470 6,580 16 510 8,160 30-35 c. 490 17,150 1 450 450 3 735 2,205 6 c. 540 c. 3,240 2 165 330 1 5,450 5,450 1 85 85 2 80 160 1 600 600 1 c. 83 c. 83 1 145 145 4 155 620 1 100 100 2 40 80 2 60 120 3 350 1,050 7 c. 75 c. 525 30 300 9,000 1 60 60 1 160 160 1 1,375 1,375 22 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 4. Plant and animal remains. Entrance 2, Site 3, Clark’s Cave (continued). Scientific name Common name MNI Gram live Total gram weight live weight AVES (continued) 30. Asia cf. flammeus or otiis 3 1 . A egolius acadicus 32. Chordeiles minor 33. Chaetura pelagica 34. Megaceryle alcyon 35. Colaptes auratus 36. Dryocopus pileatus 37. cf. Centurus carolinus 38. Sphyrapicus variiis 39. Dendrocopos villosus 40. Dendrocopos pubescens Picidae, ?species 41. Ernpidonax, ?species 42. Eremophila alpestris 43. Petrochelidon pyrrhonota 44. Perisoreus canadensis 45 . cf. Cyanocitta cristata 46. Corviis brachyrhynchos 47. Pams, ?species 48. Pams bicolor 49. Sitta cf. canadensis 50. Certhia familiar is 5 1 . Cistothoms cf . platensis 52. ci. Toxostonw rufum 53. Turdus migratorius 54. Cathams, ?species 55. cf. Sialia sialis 56. cf. Ant hus spinoletta 51. Bombycilla cedronim 58. cf. Dendroica coronata 59. Seiurus, ?species Parulidae, ?species 60. cf. Dolichonyx oryzivoms 61. cf. Sturnella, Ispecies 62. Agelaius phoeniceus 63. cf. Icterus spurius 64. Piranga, ?species 65. cf. Pinicola enucleator 66. Loxia, ?species 67. Pooecetes gramineus 68. Junco hyemalis Fringillidae, ?species short-eared or long-eared owl saw-whet owl common nighthawk chimney swift belted kingfisher common flicker pileated woodpecker red-bellied woodpecker yellow-bellied sapsucker hairy woodpecker downy woodpecker woodpeckers, unidentified flycatcher horned lark cliff swallow gray jay blue jay common crow chickadee tufted titmouse red-breasted nuthatch? brown creeper short-billed marsh wren? brown thrasher robin thrush eastern bluebird water pipit cedar waxwing yellow-rumped warbler water thrush warblers, unidentified bobolink meadowlark red-winged blackbird orchard oriole tanager pine grosbeak crossbill vesper sparrow dark-eyed junco sparrows, unidentified 3 400 1,200 1 85 85 1 80 80 1 22 22 1 140 140 3 135 405 2 280 560 1 78 78 1 50 50 1 50 50 1 26 26 2 c. 70 c. 140 1 c. 12 c. 12 1 40 40 9 23 207 1 75 75 1 85 85 1 412 412 1 10 10 1 21 21 1 11 11 2 8 16 1 9 9 1 70 70 2 80 160 2 c. 32 c. 64 1 31 31 1 23 23 1 31 31 1 13 13 1 20 20 2 c. 12 c. 24 1 34 34 1 97 97 1 60 60 1 21 21 1 29 29 1 57 57 1 30 30 1 25 25 2 20 40 2 c. 18 c. 36 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 23 Table 4. Plant and animal remains, Entrance 2, Site 3, Clark’s Cave (continued). Scientific name Common name MNI Gram live Total gram weight Uve weight MAMMALIA Mammals 4,343 204,315.0 (Identified by J.E. Guilday) 1 . Sorex arcticus arctic shrew 13 8.3 107.9 2. Sorex cinereus masked shrew 67 4.1 274.7 3. Sorex dispar long-tailed shrew 4 5.0 20.0 4. Sorex fiimeus smoky shrew 10 8.0 80.0 5. Sorex palustris water shrew 7 12.3 86.1 Sorex, ?species shrews, unidentified 26 c. 7.5 c. 196.0 6. Micro sorex hoyi pygmy shrew 7 3.2 22.4 1. Blarina brevicauda short-tailed shrew 97 19.3 1,872.1 8. Parascalops breweri hairy-tailed mole 12 52.0 624.0 9. Scalopus aquaticus eastern mole 1 102.0 102.0 10. Condyhira cristata star-nosed mole 13 56.5 734.5 1 1 . Myotis lucifugus or sodalis little brown bat c. 877 x 12. Myotis keenii Keen’s bat c. 7.5 c. 8,647.0 13. Myotis leibii small-footed bat c. 138 1 4 . Myo tis grisescens gray bat c. 138 15. Pipistrellus subflavus eastern pipistrelle 26 4.5 117.0 16. Eptesicus fuscus big brown bat 363 17.5 6,352.5 17. Plecotus cf. townsendii big-eared bat 9 7.5 67.5 18. Lasiurus borealis red bat 3 12.0 36.0 19. cf. Sylvilagus transitionalis New England cottontail 1 969.0 969.0 20. Lepus americamis snowshoe hare 23 1,500.0 34,500.0 21. Tamias striatus eastern chipmunk 24 97.0 2,328.0 22. Eutamias minimus least chipmunk 3 42.9 128.7 23. Marmota monax woodchuck 2 2,850.0 5,700.0 24. Spermophilus tridecemlineatus 13-lined ground squirrel 5 150.0 750.0 25. Sciurus cf. carolinensis gray squirrel 3 520.0 1,560.0 26. Tamiasciuriis hudsonicus red squirrel 25 185.0 4,625.0 27. Glaucomys volans southern flying squirrel 19 60.0 1,140.0 28. Glaucomys sabrinus northern flying squirrel 28 107.0 2,996.0 Tamias or Glaucomys squirrels, unidentified 8 c. 88.0 c. 704.0 29. Peromyscus maniculatus deer mouse c. 117 21.0 c. 2,457.0 30. Peromyscus leucopus white-footed mouse c. 104 22.0 c. 2,288.0 31. Neotoma floridana eastern woodrat 51 360.0 18,360.0 32. Clethrionomys gapperi red-backed vole 305 24.0 7,320.0 33. Phenacomys intermedius heather vole 34 33.0 1,122.0 34. Microtus pennsylvanicus meadow vole c. 658 35.9 23,622.0 35. Microtus chrotorrhinus rock vole c. 292 35.0 10,220.0 36. Microtus xanthognathus yellow-cheeked vole 511 90.0 45,990.0 37. Microtus pinetorum woodland vole 170 25.6 4,352.0 38. Ondatra zibethicus muskrat 6 1,050.0 6,300.0 39. Synaptomys cooperi southern bog lemming 23 28.3 650.9 40. Synaptomys borealis northern bog lemming 61 33.0 2,013.0 24 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 4. Plant and animal remains, Entrance 2, Site 3, Clark’s Cave (continued). Scientific name MAMMALIA (continued) Common name MNI Gram live weight Total gram live weight 4 1 . Zapus hudsonius meadow jumping mouse 22 18.0 396.0 42. Napaeozapus insignis woodland jumping mouse 15 24.0 360.0 43. Erethizon dorsatum porcupine 1 * * 44. Cams cf. dims dire wolf 1 * * 45. Ursus americanus black bear 1 * ♦ 46. Procyon lotor raccoon 1 * * 47. Maries americana marten 1 661.0 661.0 48. Mustela nivalis least weasel 7 41.0 287.0 49. Mustela erminea ermine 4 54.0 216.0 Mustela, (frenata or erminea) weasel, unidentified 1 c. 60.0 c. 60.0 50. Mustela vison mink 2 620.0 1,240.0 5 1 . Mephitis mephitis striped skunk 1 1,660.0 1,660.0 52. Cervus elaphus elk 1 * * 53. cf. Odocoileus virginianus white-tailed deer 1 * * BIOTIC DISCUSSIONS FLORA Plant remains, probably of Recent origin, were incidental in the deposit. Some (hackberry seeds) were probably introduced by cliff-frequenting ro- dents, others (hemlock needles, etc.), by wind. All eight species grow in the area today. (Floral List, Table 4). DECAPODA—Crayfish The few crayhsh represented were probably raptor food remains. Two species, represented by claw fragments, were present. Cambarus cf. bartonii and C. cf. longulus are both present in the Cowpasture River today. The former ranges from Georgia north to New Brunswick, the latter from the Yadkin River of North Carolina to the James River drainage. Dr. Hobbs comments “The punctations on the fingers are somewhat larger than those typical of either species, but if these remains had passed through the alimentary canal of some mammal [or bird, ed.] this may have been responsible for the disproportionately greater solution of these setae-bearing depressions” (letter, 3/7/73). INSECTA — Insects Insect remains were scarce and fragmentary. Twenty-nine of the 32 individuals identified were beetles. Scarab beetles, hide beetles, and flesh flies are attracted to decaying organic matter like raptor roost litter. Wasps probably used the cliffs as nest sites protected from the weather. All remains ap- peared to be Recent. MOLLUSCA — Snails and Clams Approximately 4,363 gastropods were recovered, some 4,300 of which were the small freshwater snail Spirodon carinata, common in the river today. They may have formed a minor food item or been inad- vertently introduced clinging to nesting materials collected by birds or woodrats. Nineteen species of land snails were represented by only 63 shells in this seemingly lime-rich environment. The site may have been too dry or the decaying owl pellets too acidic to attract them. All species are still present in the area and are broadly distributed in the Appala- chians today. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 25 The two species of fingernail clams (Sphaeriidae) were probably also inadvertent inclusions. PISCES— Fish Remains of 57 small fish, seven species, 1% of the total vertebrates, were recovered. They were uni- formly small and minnow-size, although a few of the suckers reached an estimated length of 30 cm. Identifications were based upon diagnostic cranial elements, and in the case of catfish, pectoral fin spines as well. Approximately 4,000 unidentified small fish vertebrae were recovered. All identified species are present in the James River drainage today. The fish are catalogued under CM 29689. Fish formed a minor food item and may have been taken by bankside scavenging or inadvertently in- troduced as stomach contents of other prey items. But at least some owls do actively fish. Dr. Claude W. Hibbard (letter, 10/ 15/62) reported finding, as a boy, screech owls (Otus asio) caught in jump traps set in shallow water that had been baited with shiny tin- foil to attract curious raccoons. The owls apparently mistook the shiny foil for fish and struck. AMPHIBIA — Amphibians Order Anura — Frogs and Toads Family Bufonidae Bufo (americanus group, sensu Blair) — Toad material: cm 29582: 141 left, 133 right ilia. MNI = 141 individuals.. Family Hylidae Hyla ef. crucifer — Spring Peeper material: cm 29581: 20 left, 17 right ilia. MNI = 20 indi- viduals. Family Ranidae Rana cf. catesbeiana — Bullfrog material: cm 29574: 5 left, 2 right ilia. MNI = 5 individuals. Rana cf. clamitans — Green Frog material: cm 29575: 1 left, 2 right ilia. MNI = 2 individuals. CM 29576: (R. catesbeiana or clamitans): 2 left, 2 right ilia. MNI = 2 individuals. Rana cf. pipiens group — Leopard Frog material: cm 29577: 13 left, 22 right ilia. MNI = 22 individuals. Rana cf. palustris — Pickerel Frog material: cm 29578: 18 left, 22 right ilia. MNI = 22 indi- viduals. Rana cf. sylvatica — Wood Frog material: cm 29579: 41 left, 42 right ilia. MNI = 42 indi- viduals. Rana, species indeterminate material: cm 29580: 36 left, 29 right fragmentary ilia. MNI = 36 individuals. remarks: a minimum of 141 toads and 151 frogs, 6% of the total vertebrates, were recovered from the site. Identifications were based primarily on char- acters of the ilium, although frontoparietals and sacra were also studied. The following ilial characters were used; Rana catesbeiana: steep posterior slope of ilial prominence, non-sinuate outline of the prominence viewed dorsally; large size. R. clamitans: steep posterior slope of ilial prominence; sinuate outline of the prominence viewed dorsally (except in immature specimens). R. pipiens: moderately steep angle of ilial prominence; pres- ence of a ridge and pit; ilial prominence blade-like to moderately knob-like (may be some overlap with R. sylvatica). R. palustris: gently sloping ilial prominence and well-defined ridge (Holman, 1967). Some specimens could be R. pipiens or R. sylvatica. Hyla crucifer: The hylid ilia from Clark’s Cave are referable to Hyla on the basis of the ventral acetabular expansion, which is wider than in Acris or Pseudacris. The angle between the shaft of the ilium and the acetabular expansion is obtuse, unlike Acris and Pseudacris. Hyla crucifer is unusual for the genus in that dorsal protuberance is almost always above the anterior half of the acetabular fossa, or completely anterior to it (Lynch, 1966). This is the case in all Clark’s Cave specimens. Considering the boreal nature of the deposit, the possibility of mink frog, R. septentrionalis was considered. Thirteen of the 237 Rana ilia had a keel on the shaft, a character noted on three Recent specimens of the mink frog from the Carnegie Mu- seum collections. These also lacked a pit. However, all of the Clark’s Cave specimens may be variants of R. pipiens, R. palustris, or R. sylvatica. The peepers, H. crucifer, 20 individuals, are arbo- real, small, and probably not available for pre- dation by raptors except perhaps during spring breeding aggregation. Their occurrence at the site is incidental. Of the remaining 131 frogs, 5 species of the genus Rana, only 7%, were “deep-water” frogs (river, lake, pond), bull, and green frogs. The other three species, leopard, pickerel, and wood frog, especially the lat- ter, are more terrestrial, frequenting swampland, wet meadows, and woodland. All congregate in standing water swampland during spring breeding aggrega- tion. All are common in the area today. The pickerel frog appears to be especially common in and around 26 BULLETIN CARNEGIE MUSEUM OF NATURAE HISTORY NO. 2 cave streams in Virginia today (Holsinger, 1964). Birds of prey, like most earnivores, are oppor- tunists, and these frogs were probably prey items. The relatively small number suggests that they were taken sporadically. But frogs would be available only during the warmer portion of the year, so that their relatively low numbers may be a poor reflection of their possible seasonal importance. Almost one-half of the anurans were toads of the genus Bufo. Two species occur in western Virginia at the present time, Bufo ainericanus, the American toad, and Bufo woodhousei fowleri, Fowler’s toad. The two species could not be separated on the basis of the recovered fossils and both may be represented. Although remains of the eastern spadefoot toad, Scaphiopus holbrooki, were recovered from the Natural Chimneys, Va. deposit, no evidence of this species was noted at Clark’s Cave. Nocturnal, terrestrial and slow-moving, toads would appear to be susceptible to owl predation. Frogs and “amphib- ians” have been reported from owl food remains (Latham, 1950), but toads are rarely specified. Howell, 1932, and Munro, 1929, refer to toads in the stomachs of burrowing owls, Speotyto cunicularia, so that the skin secretions, so distasteful to mammals, apparently do not deter at least the burrowing owl. We would like to thank Harold D. Mahan, Director, Cleveland Museum of Natural History, for the bur- Table 5. Measurements (in mm)l of Bufo (ainericanus group) Sacra. Locality X OR N Width of centrum at condyles 28.89 22.5-36.0 33 11.5-32.5 — 34.5 1 Length of centrum Clark’s Cave, Va., late Pleistocene 25.34 New Paris No. 4, Pa., late Pleistocene Southwestern Pa., Recent CM 37153, Bufo americanus 33.0 Width of centrum, anterior end 22.0 16.0-33.0 31 10.0-23.5 ~ 28.0 1 Height of centrum, anterior end Clark’s Cave, Va., late Pleistocene 14.1 9.0—17.5 33 New Paris No. 4, Pa., late Pleistocene 6.0—15.0 — Southwestern Pennsylvania, Recent CM 1)1 \53, Bufo americanus 15.0 1 1 Measurements by H. McGinnis. Clark’s Cave, Va., late Pleistocene New Paris No. 4, Pa., Southwestern Pa., Recent CM 31 \ 53, Bufo americanus 20.0-34.5 32 11.5-31.0 Clark’s Cave, Va., late Pleistocene New Paris No. 4, Pa., late Pleistocene Southwestern Pa., Recent CM 37 1 5 3, americanus 1977 GUILDAY, PARMALEE, AND HAMIETON: CEARK’S CAVE BONE DEPOSIT 27 rowing owl observations. Measurements of Bufo sacra from Clark’s Cave appear to fall within the range of Recent Bufo amer- icanus, and are larger than those of the smaller of the two size groups from New Paris No. 4, Pa., identified as B. a. copei. Order Urodela — Salamanders A minimum of 36 salamanders, estimated by verte- bra and limb bone counts, are catalogued under CM 29692 (see Faunal List). They were undoubtedly incidental food items. Although salamanders are common in the area, the immediate excavation site appears too well-drained and dry to attract them today. REPTILIA — Reptiles Order Chelonia — Turtle A single unidentified carapace fragment (CM 29695) is the only evidence of turtle in the deposit. Because of the selection bias of the birds of prey it is not clear whether the absence of turtles from the site is a reflection of cooler environmental conditions (as at New Paris No. 4, Pa.) or immunity from predation. Order Squamata — Lizards and Snakes A minimum number of 13 fence lizards (CM 29584) and two broad-headed skinks (CM 29585) were based on characters of the dentary and compared directly with Recent specimens from the Carnegie Museum collections. Both species occur at the site today. The fence lizard, also reported from the Natural Chim- neys, Va. deposit, reaches its northern limits today in central Pennsylvania, slightly north of the range of the broad-headed skink. Lizard remains probably post-dated the primary boreal deposit. Snakes, ca 2,800 vertebrae (CM 29693, 29694), an estimated 20+ individuals, accounted for only 0.4% of the total vertebrate MNI. One partial vertebra was from a medium-sized crotalid. All others were from colubrids of small size. Snakes were a minor item of owl diet. Large snakes appear to have been deliber- ately avoided. Colubrids, especially natricines, were common in the boreal New Paris No. 4 deposit, so those at Clark’s Cave may have been contempo- raneous with the bulk of the late Pleistocene fauna at the site. AVES— Birds Bird remains from the Clark’s Cave deposit totaled 3,600 pieces. Approximately 212 individuals, belong- ing to 68 species in 34 families, were represented. In addition, slightly over 700, or 72% of the bird-bone sample identified to order, consisted of indeter- minate passerines. Avian remains from Back Creek Cave No. 2 consisted of 415 pieces, at least 17 species belonging to nine families, a minimum of 55 individ- uals. The bulk of the remains resulted from the feeding activities of predatory birds. Such bone concentra- tions provide an index to the relative abundance and variety of a faunal assemblage, but the condition of the individual bones is often adversely affected by feeding breakage, the effects of digestive fluids, post- depositional breakage and rodent damage. A large percentage of the bones are fragmented and the diagnostic characters of the articulating ends of many are missing. In the case of the avian material from these two cave deposits, it was apparent that the larger the species (e.g. waterfowl, grouse, turkey), the greater the extent of bone fragmentation. This, combined with the difficulty of distinguishing be- tween elements of closely related species, made many identifications only tentative. Factors that prevented species identification, or osteological characters that made certain critical determinations possible, will be discussed under the various families or species of birds below. Order: Podicipediformes Family: Podicipedidae — Grebes Podilymbus podiceps (Linnaeus) — Pied-billed Grebe material: cm 29613. Paired quadrates: incomplete sternum, coracoids, tarsometatarsus, ulna. MNI = 2 individuals. REMARKS: The pied-billed grebe is a common migrant and fairly common winter resident in the state, occuring on both lakes and rivers. Although all elements (except the quadrates) were incomplete, they are in all probability referable to this species. Order: Ciconiiformes Family: Ardeidae — Herons and Bitterns Botaurus lentiginosus (Rackett) — American Bittern material: cm 29631. Incomplete left humerus. MNI = 1 individual. Remarks: Both ends of this humerus were miss- ing, but the overall dimensions, location of the deltoid crest, and form/ position of the impression of the brachialis anticus muscle compare favorably with that of the American bittern. Probably an uncommon migrant and summer resident in western Virginia. 28 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Order: Anseriformes Family: Anatidae — Swans, Geese, and Ducks Anas platyrhynchos Linnaeus — Mallard, and Anas rubripes Brewster — Black Duck — or both material: cm 29614. 2 incomplete right humeri. MNl = 2 individuals. Anas Linnaeus — Duck material: cm 29615. Distal ends of right tibiotarsus and coracoid. MNI = 1 individual. cf. Anas crecca Linnaeus — Green-winged Teal material: cm 29616. Proximal end of left humerus. MNI = 1 individual. cf. Anas discors Linnaeus — Blue-winged Teal material: cm 29617. Incomplete left humerus. MNI = 1 individual. Lophodytes cucuUatus (Linnaeus) — Hooded Merganser material: cm 29618. 1 proximal end, 1 distal end and shaft of right humeri. MNI = 2 individuals. Mergus Linnaeus — Merganser material: cm 29678. Incomplete proximal end of left hu- merus. MNI = 1 individual. Duck/ Merganser spp. material: cm 29619. 23 non-diagnostic or fragmentary ele- ments including coracoids, ulnae, femora, tibiotarsi, humeri. remarks: All waterfowl elements were incom- plete, making specific determinations especially tenuous. One indeterminate duck tibiotarsus and sternum and two humeri fragments (CM 29619) were those of small species, possibly teal, while several other pieces probably represent mergansers, some of the larger puddle ducks (Anas) or both. All water- fowl species represented at Clark’s Cave could be found today along the Cowpasture River, primarily during migration. Because of the large size of mal- lards, the black duck, and mergansers (Mergus), it would have taken a raptor the size of a large hawk (Buteo) or great horned owl (Bubo) to capture and transport these birds to the cave entrance. Order: Falconiformes Family: Accipitridae — Hawks and Harriers cf. Accipiter striatus Vieillot — Sharp-shinned Hawk material: cm 29620. Incomplete right coracoid. MNI = 1 individual. Buteo platypterus (Vieillot) — Broad-winged Hawk material: cm 29621. Fragmented distal end of right tibio- tarsus. MNI = I individual. Hawk sp. material: cm 29622. Claw. remarks: Hawks were poorly represented at Clark’s Cave, and the paucity of their remains possibly suggests that the cave entrance was not generally used as a roosting/ feeding locale. The tibiotarsus fragment, little more than the tendinal bridge and lacking the condyles, is suggested as being that of broad-winged hawk, based on existing pro- portions. The claw was from a large hawk the size of a red-tailed hawk. Murray (1952:40) considers the sharp-shinned hawk a fairly common resident in the Lexington area, but an uncommon resident in south- western Virginia. This small “bird hawk’’ may have been a major contributor of the numerous passerines in the deposit. Family: Falconidae Caracaras and Falcons Falco sparverius Linnaeus — American Kestrel material: cm 29623. 1 complete and 1 proximal half of left carpometacarpal; 1 proximal and I distal section of right tar- sometatarsal. MNI = 2 individuals. material (Back Creek Cave No. 2): CM 29700. Incomplete left carpometacarpus. MNI = 1 individual. REMARKS: This small falcon, a fairly common resident or migrant in Virginia, preys upon insects, small birds and mammals. If this raptor frequented the cave entrances, it could have also been a sig- nificant contributor of small animals to the deposits. Order: Galliformes Family: Tetraonidae — Grouse and Ptarmigan cf. Canachites canadensis (Linnaeus) — Spruce Grouse material: cm 29624. 14 carpometacarpals, 23 tarsometatarsi, 2 humeri, 2 ulnae, 1 coracoid, 4 tibiotarsi. MNI = 14 individuals. material (Back Creek Cave No. 2): CM 29701. 10 tarsometa- tarsi. MNI = 6 individuals. Bonasa umbellus (Linnaeus) — Ruffed Grouse material: cm 29625. Complete (or sections of) items includ- ing 1 jaw, I femur, 4 coracoids, 2 tibiotarsi; 24 tarsometatarsi, 14 humeri, 3 ulnae, 26 carpometacarpals. MNI = 16 individuals. material (Back Creek Cave No. 2): CM 29702. 8 carpo- metacarpals, 10 tarsometatarsi. MNI = 6 individuals. C canadensis and/or B. umbellus — Spruce or Ruffed Grouse, or both material: cm 29626. A total of approximately 272 incom- plete, fragmented, or non-diagnostic elements. MNI — 30 to 35 individuals. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 29 MATERIAL (Back Creek Cave No. 2): CM 29703. A total of approximately 80 incomplete, fragmented, or non-diagnostic elements. MNl = 9 individuals. remarks: Osteological differences between spruce and ruffed grouse are extremely subtle and, taking into account overlap in size between males and females within and between the species, the majority of elements of these two grouse are specifi- cally inseparable. In general, however, certain bones of the spruce grouse appear slightly more delicate (comparing birds of the same sex) than those of the ruffed grouse. For example, the tarsometatarsus of the spruce grouse is shorter and the shaft narrower or “pinched” just posterior to the distal foramen. The pneumatic foramen in the spruce grouse humerus appears proportionally larger and more rounded (oval in Bonasa). The carpometacarpus of the ruffed grouse appears heavier or stouter compared with one of equal length from the spruce grouse, and the intermetacarpal process in ruffed grouse appeared better developed in the comparative specimens ex- amined. While species identification based on other bones are more tentative it is felt that these deter- minations are valid and that the spruce grouse as well as the ruffed grouse is represented at both caves. The ruffed grouse is still common in the more heavily forested regions of western Virginia. The spruce grouse, on the other hand, “a bird of the northern wilderness, of thick and tangled swamps, and of spruce forests, where the ground is deep in moss and where the delicate vines of the snowberry and twinflower clamber over moss-covered stubs and fallen, long-decayed tree trunks” (Bent, 1932: 121), was unknown in the state in historic times. Wet- more (1959) identified remains of C. canadensis from Natural Chimneys, Va. With the added records of this bird from the two Bath Co. caves, it is apparent that at one time this boreal grouse was definitely a part of the avifauna of western Virginia. It has also been reported from the late Pleistocene Ladds Quarry local fauna, Bartow Co., Ga. (Wetmore, 1967). Lagopus cf. mutus (Montin) — Rock Ptarmigan material: cm 29627. Complete left humerus and tarsometa- tarsus. MNl = 1 individual. Fig. II. Modern range of spruce grouse, Canachites canadensis (Linnaeus), adapted from Godfrey, 1966. Present in Clark’s Cave local fauna. 30 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 MATERIAL (Back Creek Cave No. 2): CM 29704. 1 complete right and 2 left tarsometatarsi. MNl = 2 individuals. REMARKS: The recovery of ptarmigan elements at both Bath Co. cave sites was especially significant in light of the birds’ present distribution and habitat requirements. Northern Quebec and Newfoundland represent the most southern reaches of the bird’s range in eastern North America, approximately 1,800 km north-northeast of Bath Co. Ptarmigan are birds of the open tundra, and their occurrence as far south as west-central Virginia is indicative of the former presence of open or semi-open expanses of tundra that must have covered the northern Appala- chian Mountains during full-glacial to late Pleis- tocene times. The recovery of caribou (Rangifer tarandus L.) elements from Sullivan Co. caves in eastern Tennessee (Guilday, Hamilton, and Parma- lee, 1975) also serves to substantiate the former southern extension of a tundra-like habitat and re- lated climatic conditions to at least the central Appala- chians. In a summary of their findings based on pollen data. Maxwell and Davis (1972:506) state that “When the Wisconsin ice sheet stood at its maxi- mum position, tundra vegetation bordered the ice sheet. In the eastern United States, tundra extended at least 300 kilometers due south of the ice border at 2,700 feet (800 meters) elevation on the Allegheny Plateau. Spruce and jack (and/or red) pine forest grew at lower elevations in Virginia.” Anatomical differences among the species of Lagopus also are extremely subtle, and many of the fragmentary or abraded elements are difficult to distinguish from those of Bonasa or Canachites. The large, open circular pneumatic foramen of the humerus appears to be one of the more valid char- acters separating ptarmigan from the other two genera of grouse. The humerus and tarsometatarsus from Clark’s Cave (Fig. 13) were verified as L. mutus by Alexander Wetmore (pers. comm. 1/29/73). The latter element is especially difficult to identify to species, however, and it is possible that the willow ptarmigan, L. lagopus (Linnaeus), is represented. Both species occupy a similar range in northeastern Canada, but the willow ptarmigan prefers areas of Fig. 12. Modern ranges of rock ptarmigan, Lagopus mutus (Montin), and willow ptarmigan, Lagopus lagopus (Linnaeus), adapted from Godfrey, 1966. Ptarmigan fossils present in Clark’s Cave local fauna. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 31 cm 1 2 3 4 5 e li Fig. 13. Lagopus cf. mutus (Montin), Clark’s Cave local fauna, Bath County, Virginia, CM 29627. Upper: left tarsometatarsus; lower: left humerus. Stunted and scattered trees — a habitat perhaps char- acteristic of the Allegheny and Shenandoah moun- tains in the late glacial times. Occasional winter wanderers have been recorded as far south as Maine (Kenduskeag) and northern New York (Lewis County), 680 km north of Clark’s Cave (5th A.O.U. “Check-list,” p.l3 1). However, because of certain subtle but possibly significant osteological diflferences noted by Storrs Olson (letter Olson/ Parmalee, 10/24/75) between the cave elements and those of the rock and willow ptarmigan skeletons in the collections of the USNM, specific determinations should be considered ten- tative until these diflferences can be resolved. cf. Pediocetes phasianellus (Linnaeus) — Sharp-tailed Grouse material: cm 29628. 2 right and 2 left carpometacarpals; sections of 3 coracoids, 1 ulna, 1 tarsometatarsus, 1 tibiotarsus, 1 humerus, 1 femur. MNI = 3 individuals. MATERIAL (Back Creek Cave No. 2): CM 29705. 1 nearly complete right and 1 left carpometacarpus, 4 distal ends of tibio- tarsi. MNI = 4 individuals. REMARKS: Remains of at least four sharp-tailed grouse were identified from Natural Chimneys, Va. The former presence of this grouse in west-central Virginia is also of special interest, since today its closest range proximity to the Bath County sites is northern Wisconsin and Michigan and southern Ontario (Fig. 14). Osteologically, it and the prairie chicken, Tympanuchus cupido (Linnaeus), are ex- tremely similar. The synsacrum is considered by some avian osteologists to be the only specifically diag- nostic element (e.g. Wetmore, 1959), so the Clark’s Cave and Back Creek Cave No. 2 material lends it- self only to tentative determination. Presence of this species, based upon the synsacrum, was definitely 32 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 established at New Paris No. 4, Pa. during the late Pleistocene. The overall greater length or more robust limb elements tend to differentiate bones of either Pedio- ecetes or Tympanuchus from those of Bonasa and Canachites. Nevertheless, another 4 tarsometatarsi and 2 carpometacarpals (CM 29629) from Clark’s Cave compare closely with those of sharp-tailed grouse (Pedioeceles tarsometatarsi are propor- tionately shorter and less robust than Tympanuchus), yet could conceivably be from large male ruffed grouse. The sharp-tailed grouse is unknown from the state in historic times, and the prairie chicken (heath hen, T. c. cupido) “possibly” occurred in eastern Virginia along the Atlantic seaboard (5th A.O.U. “Check-list”: 136). However, the sharp-tailed grouse is certainly the “boreal” species of the two, and con- sidering its habitation of forested regions and espe- cially areas of low thickets and open glades or savannas, it would have been well adapted to the late Pleistocene habitat of the Appalachians. Family:Phasianidae — Quails and Pheasants Colinus virginianus (Linnaeus) — Bobwhite material: cm 29630. Complete (or sections of) items in- cluding 3 tarsometatarsi, 3 coracoids, 2 humeri, 2 carpometa- carpals, sternum, radius. MNI = 2 individuals. MATERIAL (Back Creek Cave No. 2): CM 29706. Radius, ulna; 2 incomplete carpometacarpals, I tarsometatarsus, phalanx 1. MNI = 2 individuals. remarks: a common bird in the open and semi- open brushy areas throughout the state, but rare in heavily forested regions. Family: Meleagrididae — Turkeys Meleagris gallopavo Linnaeus — Turkey material: cm 29632. Phalanx I; sections of scapula, humerus, tarsometatarsus, radius. MNI = 1 individual. remarks: Except for the complete phalanx I, the turkey elements are fragmented and, in the case of the humerus shaft, rodent gnawed. It was reported originally to have been abundant throughout the state (Murray 1952:44). However, even the largest of raptors would have difficulty in capturing adult turkeys, and therefore the paucity of turkey remains at owl roost sites like Clark’s Cave is not unexpected. Fig. 14. Modern range of sharp-tailed grouse, Pedioecetes phasianellus (Linnaeus), adapted from Godfrey, 1966. Present in Clark’s Cave local fauna. 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 33 Order; Gruiformes Family: Rallidae — Rails, Gallinules, and Coots cf. Rallus limicola Vieillot — Virginia Rail material; cm 29633. Distal third of right tibiotarsus. MNI = 1 individual. Porzana Carolina (Linnaeus) — Sora Rail material; cm 29634. Incomplete sections of 2 tarsometa- tarsi, 1 tibiotarsus, I coracoid, 2 humeri, 2 carpometacarpals. MNI = 2 individuals. remarks: Apparently a fairly common migrant and nesting bird locally in western Virginia (Blacks- burg), the Virginia rail would not be unexpected in the marshy grasslands bordering the Cowpasture River. The sora, although less numerous in the western part of the state than it is along the coastal tidewater marshes, is reportedly not uncommon locally in inland marshes and river flood plains. A fragmented right carpometacarpus (CM 29635) of what appears to be an additional species of small rail (yellow or black?) was recovered at Clark’s Cave. cf. Gallinula chloropus (Linnaeus) — Common Gallinule material; cm 29636. Distal end of lower mandible. MNI = 1 individual. remarks: This section of bill is too fragmentary to permit an unquestionable determination but, on the basis of comparative specimens examined, it appears to be this species rather than coot {Fulica americana) or purple gallinule ( Porphyrula martinica ). There are numerous migration and nesting records of the com- mon gallinule in Virginia (Murray, 1952:46,47). Wet lowlands bordering the Cowpasture River could have afforded this gallinule, as well as other marsh- dwelling species like the rails and woodcock, suitable nesting habitat. Order: Charadriiformes Family: Charadriidae — Plovers and Turnstones Pluvialis dominie a (Muller) — American Golden Plover material; cm 29637. Complete left tarsometatarsus. MNI = 1 individual. remarks: Murray (1952:48) states that the Ameri- can golden plover is rare inland, but he lists fall occurrence of this bird at Roanoke and Blacksburg. Family: Scolopacidae — Woodcock, Snipe, and Sandpipers Philohela minor (Gmelin) — American Woodcock material; cm 29638. Complete (or sections of) items in- cluding 6 tarsometatarsi, 2 femora, 2 scapulae, 3 coracoids, 2 carpometacarpals, 2 humeri, 2 quadrates, radius, ulna. MNI = 4 individuals. remarks: The American woodcock is reported to be a rare to uncommon summer resident throughout most of Virginia, but Murray (1952:49) does list young observed in Shenandoah Park, Bath, and Alle- ghany counties. Marshes 1.6 km south and 3 km northeast of Clark’s Cave could have provided suit- able habitat. Capella gallinago (Linnaeus) — Common Snipe material; cm 29639. Complete left coracoid. MNI = 1 in- dividual. cf. Ac tins macularia (Linnaeus) — Spotted Sandpiper material; cm 29640. 2 tarsometatarsi, 2 femora, 1 carpometa- carpal, 3 coracoids; incomplete sections of 2 bumeri, I tibio- tarsus. MNI = 2 individuals. ITringa solitaria Wilson — Solitary Sandpiper material: cm 29641. 2 right distal end sections of tarsometa- tarsi. MNI = 2 individuals. Indeterminate spp. — sandpipers/ plovers material: cm 29642. Sections of sternum, 2 humeri, 2 ulnas, radius, 3 tibiotarsi, 3 tarsometatarsals, carpometacarpus, 10 coracoids. MNI = 6 or 7 individuals. REMARKS: Elements of the numerous small species of plovers and sandpipers are often difficult to identify, especially when incomplete or abraded. Several complete, or nearly complete small sand- piper bones from Clark’s Cave compared most favorably with those of the spotted and least sand- pipers. Five of the indeterminate coracoids fall within the size range of the spotted/ solitary/ least group. In Virginia the spotted sandpiper is a common local summer resident. One was observed by C. O. Handley, Jr. (pers. comm. 10/7/74) on the bank of the Cowpasture River at Clark’s Cave on August 13, 1974. All three species occur in varying numbers as transients or migrants within the state. cf. Limosa Brisson — Godwit material: cm 29643. Sections of 3 right distal tibiotarsals, right coracoid, proximal left humerus. MNI = 3 individuals. REMARKS: In addition to these five fragmentary elements, three other large “shorebird” bone pieces (coracoid, 2 ulnae) from Clark’s Cave fall within the godwit-willet size range. These remains, because of their fragmented or eroded condition, could not be specifically identified with certainty. Although both the marbled godwit, Limosa fedoa (Linnaeus), now on the increase, and the Hudsonian godwit, L. hae- 34 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 mastica (Linnaeus), were formerly common along the coast during migration, there are apparently no inland records. The willet, Catoptrophorus semi- palmatus (Gmelin), a species similar osteologically and in size to the godwits, was reported from Natural Chimneys, Va. by Wetmore. However, after careful comparison with Limosa and Catoptrophorus, the Clark’s Cave material, especially the tibiotarsi, appears to be one of the godwits. Order; Columbiformes Family: Columbidae — Pigeons and Doves Ectopistes migratorius (Linnaeus) — Passenger Pigeon material; cm 29644. A total of approximately 275 elements, complete and incomplete; wing and leg bones predominate. MNI = 30 individuals. MATERIAL (Back Creek Cave No. 2): CM 29707. One bill and 24 wing and leg elements. MNI = 8 individuals. remarks: Considering the former abundance of the passenger pigeon in eastern North America, it is not surprising to find quantities of their remains in early hawk or owl roost deposits like Clark’s Cave and Back Creek Cave No. 2. Murray (1952:61) refers to several accounts relating the pigeons’ former abundance in Virginia, including Bath County, and mentions a roost (in 1872) in Buckingham County that reportedly covered an area of four square miles. Remains of E. migratorius at Clark’s Cave accounted for nearly 28% of the total number of identified bird bones (excluding indeterminate passerines). Until the last two decades of the 19th century, this pigeon appears to have been one of the major food items of raptors occupying these cave sites. Order: Cuculiformes Family: Cuculidae — Cuckoos and Roadrunners Coccyzus Vieillot — Cuckoo material: cm 29645. Incomplete right carpometacarpus. MNI = 1 individual. REMARKS: The yellow-billed cuckoo, Coccyzus americanus (Linnaeus), is reported to be consider- ably more common in Virginia than the black-billed cuckoo, C. erythropthalmus (Wilson), although both species could be expected in the vicinity of Clark’s Cave. Even if the carpometacarpus had been com- plete, it is doubtful that it could have been identified to species. Order: Strigiformes Family: Strigidae — Typical Owls cf. Otus asio (Linnaeus) — Screech Owl material: cm 29646. Section of a premaxilla and nasal process. MNI = I individual. Bubo virginianus (Gmelin) — Great Horned Owl material; cm 29647. Incomplete section of left articular; sclerotic bone, possibly referable to this species. MNI = 1 indi- vidual. Asio otus (Linnaeus) — Long-eared Owl, Asio flammeus (Pontoppidan) — Short-eared Owl, or both material: cm 29648. Sections of 3 tarsometatarsi, humerus, coracoid, 2 femora, 4 ulnae. MNI = 3 individuals. Aegolius acadicus (Gmelin) — Saw-whet Owl material: cm 29649. Complete tarsometatarsus; sections of radius, ulna, femur, coracoid. MNI = 1 individual. Owl sp. material: cm 29650. Claw, 15 metapodials. material (Back Creek Cave No. 2): CM 29708. Section of sternum (screech owl? : CM 29708), distal end of tibiotarsus (Aegolius sp.? : CM 29720), claw (CM 29721), and premaxilla fragment, possibly A. acadicus (CM 29719). REMARKS: Both the screech owl and the great horned owl are fairly common locally throughout the state. Murray (1952:63-64) reports the long- eared owl as a rare resident, the short-eared owl as rather common in the coastal areas, also infre- quent in suitable habitat throughout the mountains as a migrant and winter visitor, and the saw-whet owl as being a rare winter visitor in Virginia (C. O. Handley, Jr. saw one near Clark’s Cave, Panther Gap, Sept. 7, 1975). Except for metapodials, the complete tarsometatarsus of a saw-whet owl, and the nearly complete tarsometatarsus of Asio, all other elements of these birds were fragmentary or eroded. The latter bone is probably referable to Asio otus, a species of owl associated with dense stands or thickets of coniferous trees. Clark’s Cave is presently (1974) being used as a feeding/ roost site by barn owls, but no remains of this bird were encountered in the deposit. Although the long-eared owl is capable of taking prey the size of ruffed grouse and cottontails, most of its food consists of small rodents and, to a lesser extent, small birds. Screech owls utilize both small birds and rodents, while the saw-whet owl feeds primarily on small rodents. The great horned owl is known to take mammals like skunks and por- cupine, and birds as large as the Canada goose, turkey, and the red-tailed hawk have been recorded as prey of this owl. Data on the food habits of the great horned owl provided by Bent (1938:306- 312) are indicative of the variety of prey species 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 35 remains that might be expected at a roost site. It is apparent from the variation in size of the animals represented in the Clark’s Cave deposit that both large and small raptorial birds con- tributed. Order: Caprimulgiformes Family: Caprimulgidae — Goatsuckers Chordeiles minor (Forster) — Common Nighthawk material: cm 29651. Incomplete right humerus and right coracoid. MNl = I individual. remarks: a common summer resident and mi- grant over much of the state. Order: Apodiformes Family: Apodidae — Swifts Chaetura pelagica (Linnaeus) — Chimney Swift material: cm 29652. Complete left carpometacarpus. MNI = 1 individual. remarks: a common summer resident throughout the state. Order: Coraciiformes Family: Alcedinidae — Kingfisher Megaceryle alcyon (Linnaeus) — Belted Kingfisher material: cm 29653. Incomplete right humerus, left car- pometacarpus, and left coracoid. MNl = 1 individual. remarks: The belted kingfisher occurs as a per- manent resident in Virginia, varying in abundance locally depending upon season and general habi- tat. Birds were observed by C. O. Handley, Jr. (letter Handley/ Parmalee, 10/7/74) along the Cowpasture River in September, 1974. Wetmore reported it from Natural Chimneys, Va. Order; Piciformes Family; Picidae — Woodpeckers Colaptes auratus (Linnaeus) — Common Flicker material: cm 29654. 1 complete left carpometacarpus, in- complete elements including I ulna, 4 carpometacarpals, 3 tar- sometatarsi, femur, coracoid. MNl = 3 individuals. material (Back Creek Cave No. 2): CM 29709. 3 incomplete carpometacarpals, femur, tarsometatarsus and 2 humeri (juv- enile ?; also a tibiotarsus and tarsometatarsus section of a juvenile woodpecker, possibly flicker). MNI = 2 individuals. Dryocopus pileatus (Linnaeus) — Pileated Woodpecker material: cm 29655. 2 incomplete right humeri, right scap- ula. MNI = 2 individuals. cf. Centurus carolinus (Linnaeus) — Red-bellied Woodpecker material: cm 29656. 1 partial right and 1 partial left humerus. MNI = 1 individual. ? Melanerpes erythrocephalus (Linnaeus) — Red-headed Woodpecker material (Back Creek Cave No. 2): CM 29710. Distal end of left humerus. MNl = I individual. Sphyrapicus varius (Linnaeus) — Yellow-bellied Sapsucker material: cm 29657. Proximal end of left humerus. MNl = I individual. Dendrocopos villosus (Linnaeus) — Hairy Woodpecker material: cm 29658. Proximal end of right humerus, nearly complete right tibiotarsus. MNI = 1 individual. MATERIAL (Back Creek Cave No. 2): CM 29711. Com- plete left carpometacarpus and tarsometatarsus, distal end of right tarsometatarsus. MNI = 1 individual. Dendrocopos pubescens (Linnaeus) — Downy Woodpecker MATERIAL: CM 29659. I incomplete left carpometacarpus, distal end of left tarsometatarsus. MNI = 1 individual. Woodpecker ssp. MATERIAL: CM 29660. Sections of femur, radius, carpometa- carpus, and 2 tarsometatarsi. MNI = probably 2 individuals. MATERIAL (Back Creek Cave No. 2): CM 29712. Sections of 4 tarsometatarsi (1 juvenile), tibiotarsus, carpometacarpus and ulna. MNl = 3 individuals. REMARKS: All species of woodpeckers represented in these two cave deposits are considered permanent residents in Virginia, and all may be found today in Bath County. The indeterminate woodpecker ele- ments all appear to be within the size range of the red-bellied/ red-headed/ hairy woodpecker group. Woodpeckers appear to be particularly susceptible to predation by raptorial birds, and their remains often occur in cave and rock-shelter deposits. To illustrate, approximately 400 bird bones were re- covered at the Raven Rocks Site, Belmont County, Ohio (Shane and Parmalee, in press). At least 15 species were represented and nearly 12% of the bones were those of woodpeckers. Of these, at least one-third were nestlings or juveniles. Order: Passeriformes Family: Tyrannidae — Flycatchers Empidonax Cabanis — Flycatcher MATERIAL: CM 29661. Complete right humerus. MNl = I in- dividual. REMARKS: This element compares closely with the yellow-bellied flycatcher, E. flaviventris (Baird and 36 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 2 Baird), and the Acadian flycatcher, E. virescens (Vieillot). The former species is an uncommon mi- grant in Virginia, the latter a common summer resident. C. O. Handley, Jr. heard several Acadian flycatchers singing along the Cowpasture River at Clark’s Cave in June and August, 1974 (letter Handley/ Parmalee, 10/7/74). Family: Alaudidae — Larks cf. Eremophila alpestris (Linnaeus) — Horned Lark material: cm 29662. Complete left coracoid. MNI = 1 indi- vidual. remarks: Although this coracoid compares well with the horned lark, a bird of the open grasslands, identification must remain tentative. The coracoid has proved to be specifically non-diagnostic in a large number of passerine birds, especially in those groups containing numerous closely related species. Family: Hirundinidae — Swallows Petrochelidon pyrrhonota (Vieillot) — ClilT Swallow material: cm 29663. Complete (or sections of) items includ- ing 13 humeri, 4 ulnae, carpometacarpus, coracoid, and tar- sometatarsus. In addition, there are 12 fragmented or abraded swallow elements (CM 29664) that are probably referable to this species. MNI = 9 individuals. remarks: The cliff swallow, a local summer resident in Virginia, may still be found nesting in natural cliff sites, although nests of colonies are most often observed in barn eaves and under bridge floors and railings. Apparently one or more of the hawks or owls utilizing Clark’s Cave preyed upon local nesting colonies of these swallows. It is of interest to note, however, that Bent (1942:480) states that “Preda- ceous birds cannot be considered as serious enemies of the cliff swallow.’’ At least eight cliff swallows were represented in the avifauna of Natural Chim- neys, Va. C. O. Handley, Jr. saw them flying near Clark’s Cave in 1974. Family: Corvidae — Jays, Magpies, and Crows Perisoreiis canadensis (Tinnaeus) — Gray Jay material: cm 29665. Complete left ulna, proximal end of left humerus. MNI = I individual. REMARKS: Although the ulna recorded here as P. canadensis may be subject to question because this element in general lacks good diagnostic characters, both it and the section of humerus (a good diagnostic bone) compare favorably with gray jay. The former occurrence of this bird as far south as the Shen- andoah Mountains area was established when re- mains of it were encountered in the Natural Chim- neys, Va., deposit. The gray jay is associated with boreal coniferous forests, and its former presence in Bath and Augusta counties (as an established resi- dent and not as a casual visitor) is suggestive of a once-cooler climate and a possibly different spruce- fir forest association. Cyanocitta cristata (Linnaeus) — Blue Jay material: cm 29666. Incomplete left carpometacarpus. MNI = I individual. material (Back Creek Cave No. 2): CM 29713. Proximal half of right humerus, incomplete right carpometacarpus and femur. MNI = 1 individual. Corvus brachyrhynchos Brehm — Common Crow material: cm 29667. Fragmentary distal end of left humerus, proximal end of right tibiotarsus, right scapula. MNI = 1 indi- vidual. remarks: Both the blue jay and common crow occur throughout Virginia as permanent residents and may be seen occasionally today in the vicinity of Clark’s Cave. Family: Paridae — Titmice Pams Linnaeus — Chickadee material: cm 29668. Proximal half of right humerus, frag- ment of proximal end of left humerus. MNI = 1 individual. Pams bicolor Linnaeus — Tufted Titmouse material: cm 29669. Proximal end of left humerus. MNI = 1 individual. REMARKS: Both the Carolina chickadee, Pams carolinensis Audubon, and the black-capped chick- adee, Pams atricapillus Linnaeus, can be expected in the vicinity of Clark’s Cave. Chickadees and the tufted titmouse are fairly-common to common statewide. Family: Sittidae — Nuthatches cf. Sitta canadensis Linnaeus — Red-breasted Nuthatch material: cm 29670. Complete right humerus. MNI = 1 individual. REMARKS: This nuthatch and the white-breasted nuthatch, Sitta carolinensis Latham, both occur in western Virginia, varying in abundance locally from rare to common. Family: Certhiidae — Creepers Certhia familiaris Linnaeus — Brown Creeper material: cm 29671. 2 complete left tarsometatarsi. MNI = 2 individuals. REMARKS: Murray (1952:76) records the brown 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 37 GRAY JAY Fig. 15. Modern range of gray jay, Perisoreus canadensis (Linnaeus), adapted from Godfrey, 1966. Present in Clark’s Cave local fauna. creeper as a generally common winter resident and migrant in Virginia. It is also a local summer resi- dent in the mountains. C. O. Handley, Jr. observed several nests at Mountain Lake, Giles County. Family; Troglodytidae — Wrens ? Cistothorus platensis (Latham) — Short-billed Marsh Wren material: cm 29672. Right humerus, lacking a portion of the proximal end. MNl = 1 individual. remarks: Although this humerus section com- pares favorably with C. platensis, identification can only be tentative because of its incompleteness. During spring and fall migrations this bird is a rare transient, as is the longbilled marsh wren, Telma- todytes palustris (Wilson), in the western and other inland regions of the state. Family: Mimidae — Mockingbirds and Thrashers cf. Toxostoma rufum (Linnaeus) — Brown Thrasher material: cm 29673. Proximal half of left humerus; incom- plete right coracoid, possibly referable to T. rufum. MNI = 1 individual. remarks: a common to abundant bird throughout the state. Family: Turdidae — Thrushes, Solitaires, and Bluebirds Turdus migratorius Linnaeus — Robin material: cm 29674. Complete right and left carpometa- carpus, right ulna, radius and tarsometatarsus; incomplete ster- num, proximal left humerus, right carpometacarpus. In addition, a complete left femur and the distal halves of a right tibiotarsus, tarsometatarsus, and humerus are also probably robin. MNl = 2 individuals. Hylocichla Baird — Thrush material: cm 29675. Nearly complete right and left humeri. MNI = 2 individuals. cf. Sialia sialis (Linnaeus) — Eastern Bluebird material: cm 29676. Incomplete proximal half of left hu- merus. MNI = 1 individual. remarks: The robin is an abundant summer resident and migrant, and locally common winter resident in the state. The two humeri determined as Hylocichla fall within the size range of the veery/ 38 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 hermit/ gray-cheeked /Swainson’s thrush group, all of which may be found in the vicinity of Clark’s Cave at one season or the other. The eastern blue- bird is a permanent resident in Virginia, primarily inhabiting the semi-open brush and grassland areas. Family: Montacillidae — Wagtails and Pipits ? Anthus spinoletta (Linnaeus) — Water Pipit material: cm 29717. Incomplete right coracoid. MNI = 1 individual. remarks: The distal or head portion of this ele- ment compares favorably with the water pipit. The overall head shape and extremely deep groove be- tween the glenoid facet and the bicipital attachment (external view) appear diagnostic. However, deter- mination is tentative because of the fragmentary condition of the coracoid. Water pipits are considered to be irregular migrants and winter visitors in most of Virginia (Murray, 1952:84). Family: Bombycillidae — Waxwings Bombycilla cedrorum Vieillot — Cedar Waxwing material: cm 29677. Complete right humerus. MNI = 1 individual. remarks: a permanent resident, but erratic in occurrence in western Virginia. Family: Parulidae — Wood Warblers cf. Dendroica coronata (Linnaeus) — Yellow-rumped Warbler material: cm 29679. Nearly complete right humerus. MNI = I individual. ? Seiurus Swainson — Waterthrush material: cm 29680. Nearly complete right humerus. MNI = 1 individual. Warbler ssp. material: cm 29680. Complete (or sections oO items in- cluding 2 humeri, 2 carpometacarpals, 2 tibiotarsi, 2 ulnae, cor- acoid. MNI = 2 individuals. MATERIAL (Back Creek Cave No. 2): CM 29722. Nearly com- plete left humerus possibly referable to golden-winged warbler, Vermivora chrysoptera (Linnaeus). MNI = 1 individual. remarks: Murray (1952:92) records the myrtle warbler (yellow-rumped warbler) as an “abundant transient everywhere; uncommon to abundant winter visitor as far west as Lexington; occasional in winter at Blacksburg.” In addition to these few elements identified as warbler, many of the fragmentary or abraded bones recorded as Indeterminate Passerines are probably from species within this Family. Family: Icteridae — Meadowlarks, Blackbirds, and Orioles cf. Dolichonyx oryzivorus (Linnaeus) — Bobolink MATERIAL: CM 29681. Proximal end of right humerus. MNI = 1 individual. cf. Sturnella Vieillot — Meadowlark MATERIAL: CM 29682. Distal end of left humerus. MNI = 1 individual. Agelaius phoeniceus (Linnaeus) — Redwinged Blackbird material: cm 29683. Complete left humerus, right ulna and carpometacarpus. MNI = 1 individual. cf. Icterus spurt us (Linnaeus) — Orchard Oriole material: cm 29684. Complete right humerus. MNI = 1 individual. remarks: The four icterids identified from Clark’s Cave occur as permanent residents, migrants, or nesting birds in the western part of the state. Identifications based on incomplete elements or those lacking good diagnostic characters (e.g., the ulna) are recorded as tentative. However, after com- parison of elements with these species and those of closely related blackbirds (e.g., Baltimore oriole and orchard oriole), it is felt that the determinations are valid. Family: Thraupidae — Tanagers Piranga Vieillot — Tanager material: cm 29685. Premaxilla. MNI = 1 individual. remarks: The scarlet tanager, Piranga olivacea (Gmelin), is widespread and an abundant summer resident. The summer tanager, Piranga rubra (Lin- naeus), is an uncommon and local resident in western Virginia. On the basis of this incomplete mandible section from Clark’s Cave, it was not possible to identify the species. Family: Fringillidae — Grosbeaks, Finches, Sparrows, and Buntings cf. Pinicola enucleator (Linnaeus) — Pine Grosbeak material: cm 29686. Anterior section of lower mandible. MNI = 1 individual. material (Back Creek Cave No. 2): CM 29714. Anterior section of lower mandible. MNI = 1 individual. remarks: Although similar in size and general structure to the mandibles of cardinal, rose-breasted grosbeak, and evening grosbeak, these two cave specimens compare most closely with P. enucleator, based on the wider, more U-shaped inner angle of the symphysis and the distance of the splenial de- 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 39 pression from the tip. Murray (1952:105) records one specimen collected in Shenandoah Park and one observed at Richmond. If the determination of these bill sections is correct, the former occurrence of this grosbeak in western Virginia represents local winter stragglers or perhaps reflects the cooler climate and spruce forests of the late Pleistocene — a habitat type now occupied by this bird. Loxia Linnaeus — Crossbill material: cm 29718. Section of premaxilla. MNl = 1 in- dividual. MATERIAL (Back Creek Cave No. 2); CM 29723. Complete right humerus; distal end of a right humerus, probably referable to Loxia. MNl = 2 individuals. remarks: The red crossbill, Loxia ciirvirostra Linnaeus, is considered an erratic visitor at all sea- sons in the Virginia mountains, and the white-winged crossbill, Loxia leucoptera Gmelin, an infrequent winter visitor (Murray, 1952:106). Indeterminate sparrow ssp. material: cm 29687. 2 incomplete carpometacarpals, 2 humeri, coracoid, remarks: An additional right humerus, tenta- tively determined as vesper sparrow, Pooecetes gramineus (Gmelin), was the only complete diag- nostic element recovered of this group of fringillids, which are especially difficult to identify osteologi- cally. The vesper sparrow is listed as a summer resi- dent from Richmond west (Murray, 1952:109). Junco cf. hyemalis (Linnaeus) — Dark-eyed Junco material: cm 29688. 3 complete humeri, premaxilla. MNl -2 individuals. material (Back Creek Cave No. 2): CM 29715. Complete left humerus. MNl = 1 individual. remarks: Juncos are common winter residents throughout the state, and permanent residents in the mountains above 758 m. MAMMALIA — Mammals Order: Insectivora — Insectivores Family: Soricidae — Shrews Genus Sorex Linnaeus — Long-tailed shrews Sorex arcticus Kerr — Arctic shrew material: cm 24540: 1 partial skull. CM 24581: 13 left, 12 right mandibles. MNl = 13 individuals. Sorex cinereus Kerr — Masked shrew material: cm 24580. 67 left, 61 right mandibles. MNl = 67 individuals. Sorex dispar Batchelder — Big-tailed or rock shrew material: cm 24583. 2 left, 4 right mandibles. MNl = 4 individuals. Sorex fumeus Miller — Smoky shrew material: cm 24582. 9 left, 10 right mandibles. MNl = 10 individuals. Sorex palustris Richardson — Water shrew material: cm 24587. 3 left, 7 right mandibles. MNl = 7 in- dividuals. Sorex, .^species material: cm 24584: 17 left, 12 right fragmentary large mandibles (S'. fumeusjS. arcticusjS. palustris)', 9 left, 10 right fragmentary small mandibles (S. cinereus jS. dispar). CM 24585: unidentified skull fragments. MNl = 26 individuals. REMARKS; All five species of Sorex from the de- posit inhabit at least some portion of the Appala- chian Mountain chain. Sorex arcticus, the second commonest soricid from the Clark’s Cave deposit, occurs only as far south as New Brunswick, 1,440 km northeast of Clark’s Cave. It ranges across the con- tinent, from eastern Quebec to Alaska, in boreal habitats ranging from coniferous forest to tundra. Banfield, 1974, suggests that it is typical of sub- climax or transitional vegetation stages, and that while it is frequently taken in bogs and marshes, it favors a habitat slightly drier than that of other species of shrews. The water shrew, S. palustris, still occurs in the Appalachian Mountains south, at increasingly greater elevations, into the Great Smok- ies, approximately 240 km southwest of Clark’s Cave (Linzey & Linzey, 1971). It has been taken in Ran- dolph County, W. Va., and in all probability does occur in suitable mountain streamside habitats above 760 m in western Virginia (Handley & Patton, 1947). It is no longer present in the Clark’s Cave area, but has recently been trapped within 15 km of the cave at a higher altitude (Pagels and Tate, 1975). Despite the riverside location of the site, S. palustris remains were not common (Table 6). Only S. dispar remains were fewer in number, reflecting the specialized habitats of these two shrews. The peculiar habitat of the big-tailed shrew, S. dispar, confined to and generally deep into rocky talus, contributes to its apparent rarity. Handley (1956b) trapped it near the crest of Big Mountain at 1,200 m elevation, Giles County, Va., southwest of Clark’s Cave. It has subsequently been found to be fairly common in four areas near Mountain Lake, Giles County, and also has been taken on Whitetop 40 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 and Clinch Mountains. It may be expected in cool, moist, boulder and talus fields, above 600 m in the mountains (Handley & Patton, 1947). It occurs in the Clark’s Cave talus today and was trapped there in 1974 by C. O. Handley, Jr. and associates. 5. dispar occurs throughout the Appalachian chain, from Maine to Tennessee and North Carolina (Hall & Kelson, 1959). A smaller species, S. gaspensis, occurs in the Gaspe Peninsula of eastern Quebec, 240 km north of the range of S. dispar. If S. gaspensis is a subspecies of S. dispar, as some have suggested (Jackson, 1928; Banfield, 1974) this suggests that, in the northern limits of its range, S. dispar is not as closely confined to a talus habitat, but finds the mic- rohabitat it requires in surface forest litter rather than in the labyrinthine cold air sinks of rock talus refuges of the South. Goodwin states (Jackson, 1928:91) that the habits of S. gaspensis appear to be similar to those of S. palustris. Although not directly compared with S. gaspensis, the Clark’s Cave speci- mens agree in size with those of S. dispar (Table 7). S. dispar may therefore have been more broadly distributed during past boreal phases. This is sug- gested by its presence at New Paris No. 4, Pa. (4 individuals), in a region of well-drained rolling hills not physiographically suited for S. dispar today. The modern range of S. dispar suggests that, at least during Wisconsinan times, it was confined to the Appalachian region and was unable to cross or cir- cumvent the St. Lawrence/ Great Lakes water barrier into east-central Canada following glacial retreat. The masked shrew, S. cinereus, the most widely distributed North American insectivore today, found in most of northern North America from Alaska to Labrador south, and in the mountains to New Mexico and North Carolina, was by far the commonest long- tailed shrew recovered from the deposit. It is possible that some of the remains may be those of the southern Bachman’s shrew, N. longirostris. Identification was based solely on lower jaws, because of the fragmen- tary condition of the remains. Except for a smaller average size, lower jaws of these two species cannot be differentiated. S. longirostris is basically a shrew of the southern lowlands, although it has been taken in western Virginia (Handley & Patton, 1947). The Clark’s Cave collection was referred to S. cinereus because of its larger size — somewhat larger than modern S. cinereus from the central Appalachians (Table 7) — and the boreal aspect of the accompanying fauna. Sorex cinereus was also the commonest long- tailed shrew at New Paris No. 4, Pa., a site of similar age, where the possible presence of N. longirostris did not arise. The smoky shrew, S. fumeus, the commonest Sorex in the mountains of western Virginia today, and the commonest long-tailed shrew in the central and southern Appalachians (Smith et al, 1974; Linzey & Linzey, 1971; Handley & Patton 1947), was not common in the deposit. Seven S. cinereus were recovered for every S. fumeus, and its remains were somewhat less common than those of the arctic shrew (Table 6). S. fumeus is common in the deciduous and deciduous/coniferous forests of southeastern Canada and New England. It occurs south in the Appalachian Mountains and flanking plateau areas to northern Georgia (Hall & Kelson, 1959). 5 fumeus appears to be more dependent on mature forest cover than the two commonest long-tailed shrews from the de- posit, S. cinereus and S. arcticus. Its relative scarcity in mid-Appalachian late Pleistocene deposits sug- gests more open woodlands in comparison to the dense forest cover that now prevails. Four speci- mens of S. fumeus were trapped by C. O. Handley, Jr. and associates in the Clark’s Cave talus in 1974. Microsorex hoyi (Baird) — Pygmy shrew material: cm 24579. 7 left, 4 right mandibles. MNl = 7 individuals. remarks: The rarest of the Appalachian long- tailed shrews today, the pygmy shrew appears to have been not as rare in late Pleistocene periglacial sites from eastern North America. It accounted for 6.5% of the long-tailed shrews from Clark’s Cave (Table 6), 9.7% from Natural Chimneys, Va., 1 1 .8% from Robin- son Cave, Tenn., and 19% from New Paris No. 4, Pa. The latest review (Long, 1972a) lists only 31 modern specimens, from eastern North America, of the races thompsoni and winnemana. By way of contrast, at least 46 individuals have been recorded in that area from late Pleistocene cave sites: Bootlegger Sink, Pa. (2), New Paris No. 4, Pa. (11), Natural Chimneys, Va. (7), Clark’s Cave, Va. (7), Meyer Cave, 111. (1), Welsh Cave, Ky. (1), and Robinson Cave, Tenn. (17). Microsorex hoyi exhibits a clear Bergmann’s Re- sponse. Size increases with latitude. The largest representatives are from northern Canada and Alaska. The smallest races, montanus, in the Rocky Mountains, and winnemana in the central and south- ern Appalachian region, mark the southern limits of its Recent range. Surprisingly, measurements of fossil specimens from boreal Wisconsinan faunas in the East agree with those of the Recent M. h. thomp- soni still in the area. Unlike other insectivores from these deposits which have larger northern races 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 41 today, such as Blarina brevicauda, measurements show no trend toward larger size in Microsorex. The eastern races of this shrew, M. h. thompsoni and M. h. winnemana, were originally described as separate species (Baird, 1857; Preble, 1910). Jack- son, 1928, recognized only M. hoyi, with thompsoni and winnemana as subspecies. Long, 1972a, regarded thompsoni and winnemana as closely related to each other and distinct from M. hoyi. He grouped these two eastern forms under the species M. thomp- soni, with winnemana as a subspecies. All of the characters differentiating M. thompsoni from M. hoyi are variable throughout the range of M. hoyi. Long, 1972a, and 1974, states that the elevation of thompsoni (includes winnemana) to species level is tentative and that, “this view needs further study in critical areas.” The apparent small size of these late Wisconsinan fossils suggests that M. thompsoni may indeed be a valid species, and that the larger northwestern M. hoyi may have spread into eastern Canada from the west during postglacial times. Except for the Gaspe Peninsula, where it is replaced by M. hoyi, the north- ern limit of M. thompsoni appears to be defined by the Great Lakes-St. Lawrence waterway, which may have acted as a barrier to its northward spread in postglacial times. Unfortunately, fossil specimens are invariably lower jaws or skull fragments. Taxo- nomic reviews of living species are based primarily on external characters and complete crania, so that comparison is difficult. The pygmy shrew, despite its rarity, occurs throughout the Canadian/ Hudsonian boreal forest from Alaska to Labrador, and south along the Appa- lachians in deciduous hardwood forests to the Great Smokies (Long, 1974). Although always found in a forested context and usually fairly close to boreal mesic habitats it may be trapped in such a variety of situations from swamp to dry open country that its presence at the site furnishes no specific ecological clues (Long, 1972b). There are only two published Recent records from the state of Virginia, both in the Piedmont, well east of the Appalachians proper: one in Fairfax Co., Va., and one at Altavista, Campbell County, 110 km southeast of Clark’s Cave, both in deciduous forest (Handley & Patton, 1947). As an example of their modern rarity. Long, 1972a, was able to document only seven specimens south of Maryland. C. O. Handley, Jr., USNM, (pers. comm.) has subsequently acquired three more Virginia speci- mens. Blarina brevicauda (Say) — Short-tailed Shrew material: cm 24541; 19 partial skulls, 3 maxillae; 26 left, 22 right mandibles. CM 24578: 67 left, 75 right mandibles; skull fragments. MNl = 97 individuals. remarks: The short-tailed shrew, the commonest one in Virginia today, was also the commonest in- sectivore from the Clark’s Cave deposit. It accounted for 42% of the seven species of shrews recovered from the cave, (42% from Natural Chimneys, Va., and 38.9% from New Paris No. 4, Pa.). From a paleoenvironmental point of view, osteo- logical remains of this shrew can furnish important clues. Short-tailed shrews vary markedly in size throughout their modern geographic range. Average size increases with latitude. The largest races, B. b. brevicauda and B. b. manitobensis, are from Min- nesota and Manitoba. A smaller race, B. b. kirt- landi, occurs in the mid-Appalachians today. A much smaller species, Blarina carolinensis, occurs in the southern lowlands of eastern North America. A subspecies, B. b. churchi, larger than the mid-Appala- chian B. b. kirtlandi, is found on the mountain sum- mits of the southern Appalachians — apparently adapted to a cooler environment, a reflection of the same physiological forces that shaped the over-all increase in size with increasing latitude. The short-tailed shrews from the Clark’s Cave deposit were markedly larger than present day mid- Table 6. Frequencies (%) of species of Sorex Microsorex from three late Pleistocene Appalachian cave deposits. 1 Species Clark’s Cave, Va. Natural Chimneys, Va. New Paris No. 4, Pa. Sorex arcticus 12.0 (13) 8.3 (6) 10.3 (6) Sorex cinereus 62.0 (67) 62.5 (45) 60.3 (35) Sorex dispar 3.7 (4) — 6.8(4) Sorex fumeus 9.3 (10) 13.9 (10) 1.7(1) Sorex palustris 6.5 (7) 5.5 (4) 1.7(1) Microsorex hoyi 1( ) = MNI. 6.5 (7) 9.7 (7) 19.0(11) 42 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 7. Measurements (in mm) of various species of Sorex. Locality and age X OR SD CV N Total mandible length, condyle to anterior point of dentary Sorex arcticus Clark’s Cave, Va., late Pleistocene 9.3 9.1— 9.5 9 New Paris No. 4, Pa.,1 late Pleistocene 8.9 8.7— 9.3 5 S. cinereus Pennsylvania, modern! 7 28 7.1— 7.7 .10 1.37 22 Clark’s Cave, Va., late Pleistocene 7.57 7. 3-8.0 .21 2.77 18 New Paris No. 4, Pa.,! late Pleistocene 7.6 7. 2-7. 8 .27 3.55 17 S. dispar Clark’s Cave, Va., late Pleistocene 8.4 8.3— 8.4 3 New Paris No. 4, Pa.,! late Pleistocene 8.2 2 S. fumeus Clark’s Cave, Va., late Pleistocene 9.36 8. 9-9. 8 6 S. palustris Clark’s Cave, Va., late Pleistocene 10.08 1 New Paris No. 4, Pa.,! late Pleistocene 9.5 1 p4 — m3, antero-posterior crown length^ S. arcticus Clark’s Cave, Va., late Pleistocene 4.64 4.56—4.76 7 New Paris No. 4, Pa.,! late Pleistocene 4.4 4.2— 4.5 6 S. cinereus Pennsylvania, modern! 3 59 3.6— 3.9 .07 1.89 20 Clark’s Cave, Va., late Pleistocene 3.98 3. 6-4. 4 .13 3.40 35 New Paris No. 4, Pa.,! late Pleistocene 3.9 3.7— 4.3 .16 4.09 29 S. dispar Clark’s Cave, Va., late Pleistocene 4.36 1 New Paris No. 4, Pa.,! late Pleistocene 4.3 1 S. fumeus Clark’s Cave, Va., late Pleistocene 4.85 1 New Paris No. 4, Pa.,! late Pleistocene 4.7 1 S. palustris Clark’s Cave, Va., late Pleistocene 5.04 4. 9-5. 2 3 New Paris No. 4, Pa.,! late Pleistocene 5.2 1 !Uata from Guilday et al., 1964. ^Defined as Cj - M3 in Guilday et al., 1964 (See Repenning, 1967 for correct terminology). 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 43 Table 8. Measurements (in mm)l of Microsorex . Taxon and locality X OR N Palatal length Recent; M. thompsoni^ 5.60 5.30-5.84 5 M. cf. thompsoni, 5.24 1 New Paris No. 3, Pa. M. t. winneinana 1 5.25 1 M. h. hoyi^ 5.67 5.21-6.50 24 M. h. alnonim^ 5.60 5.54-5.66 6 M. h. eximius^ 5.86 5.70-5.98 3 M. h. washingtoni^ 5.56 1 M. h. montamis^ 5.4 5.26-5.54 3 Late Pleistocene; New Paris No. 4, Pa. 5.20 5. 1-5. 3 4 Welsh Cave, Ky. 5.20 1 Robinson Cave, Tenn. 5.10 Maxillary breadth 4.80-5.30 3 Recent: M. h. montanus^ 3.9 3.9-4.0 10 M. cf. thompsoni, New Paris No. 3, Pa. 4.17 1 Late Pleistocene: New Paris No. 4, Pa. 4.15 4.0-4. 3 4 Bootlegger Sink, Pa. 4.1 1 Welsh Cave, Ky. 4.1 P4 - M3 1 Recent: M cf. thompsoni. New Paris No. 3, Pa. 3.39 1 Late Pleistocene; New Paris No. 4, Pa. 3.45 3.4-3.5 4 Robinson Cave, Tenn. 3.5 3. 2-3. 8 3 Total length of mandible with il Recent: M. cf. thompsoni. New Paris No. 3, Pa. 8.15 1 Late Pleistocene: New Paris No. 4, Pa. 7.97 7.6-S.4 9 Clark’s Cave, Va. 8.12 Total length of dentary 8.05-8.24 3 Recent: M. cf. thompsoni. New Paris No. 3, Pa. 6.6 1 Late Pleistocene: New Paris No. 4, Pa. 6.5 6.4-6.6 10 Bootlegger Sink, Pa. 6.3 1 Clark’s Cave, Va. 6.5 6. 3-6.8 7 Robinson Cave, Tenn. 6.5 5. 9-7.0 7 1 Measurements from Long, 1972a. 44 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 8. Measurements (in mm)^ of Microsorex (continued). Taxon and locality Recent: M. cf. thompsoni. X Height, ascending ramus OR N New Paris No. 3, Pa. 3.0 1 Late Pleistocene: New Paris No. 4, Pa. 2.94 2.9-3.0 9 Bootlegger Sink, Pa. 3.0 1 Clark’s Cave, Va. 2.99 2.91-3.10 7 Welsh Cave, Ky. 2.9 1 Robinson Cave, Tenn. Recent: M cf. thompsoni. 3.2 p4 — m3 2.9-3.S 20 New Paris No. 3, Pa. 3.39 1 Late Pleistocene: Bootlegger Sink, Pa. 3.49 1 Clark’s Cave, Va. 3.51 3.39-3.69 5 Welsh Cave, Ky. Recent: M. cf. thompsoni. 3.6 ml - m3 1 New Paris No. 3, Pa. 2.91 1 Late Pleistocene: New Paris No. 4, Pa. 2.86 2.8-3. 1 8 Bootlegger Sink, Pa. 2.91 1 Clark’s Cave, Va. 2.97 2.8-3.2 8 Welsh Cave, Ky. 2.8 1 Robinson Cave, Tenn. 2.9 2.8-3. 1 12 ^Measurements from Long, 1972a. Appalachian specimens, in keeping with the boreal aspect of the deposit. Four of the five cranial mea- surements taken average larger than those of B. b. brevicauda from Minnesota (Table 9). Maxillary breadth of 15 Clark’s Cave skulls average only 0.14 mm smaller than the type and allotype of B. b. manitobensis (Anderson, 1947:23), the largest modern subspecies. Three specimens (of 15) ex- ceeded this. The total length of the lower jaw averaged 8.6% larger, and the length of the lower toothrow (p4-m3) 7.6% larger than modern B. b. kirtlandi. The Clark’s Cave fossil Blarina are com- parable in size to the largest living subspecies and to the larger of the two size groups from New Paris No. 4, Pa. The Clark’s Cave sample, in addition to being composed of specimens much larger than those now in the area, appears to be homogeneous, i.e., not racially mixed. Total variation of the length of the lower jaw amounted to only 11% (% increase of largest to smallest), with a low coefficient of varia- tion of 3.00, comparable to that of a Recent Pennsylvania sample of B. b. kirtlandi (Guilday et al., 1964:151). In contrast, in the New Paris No. 4, Pa., Blarina sample, composed of a mix- ture of large late Pleistocene and smaller Re- cent specimens, total variation of that mea- surement rose to 20%, with a coefficient of varia- tion of 5.74. At Meyer Cave, 111., where the Blarina were markedly heterogenous, a mixture of large B. b. brevicauda and smaller B. b. cf. kirtlandi, total variation was 25%, over twice that shown by the Clark’s Cave sample. This is one of the few species in the deposit yield- 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 45 Table 9. Measurements (in mm) of Blarina brevicauda, various localities. Locality X OR SD CV Recent: Minnesota 1 8.4 Maxillary breadth 8.0-9.2 0.35 4.13 Pennsylvania 1 7.8 7.0-8. 3 0.22 2.80 Cowpasture River valley, Bath Co., Va. Elevation 440-485 m USNM 489606-489610, 489704-489714 7.6 7.2-7.8 0.18 2.37 Warm Springs Mountain, Bath Co., Va. Elevation 644-1106 m USNM 489880-489884 7.8 1 00 b White Sulphur Springs, Greenbriar Co., W. Va.2 7.66 7.2-8.02 — — Pisgah Forest, Transylvania Co., N. C.2 7.53 7.2-8.0 __ Late Pleistocene New Paris No. 4, Bedford Co.,Pa.l 8.50 00 1 00 — — Natural Chimneys, Augusta Co., Va.l 8.31 8.0-8.7 Clark’s Cave, Bath Co., Va. 8.66 8.2-9. 1 0.26 3.06 Recent: Minnesota! Total length, mandible (including incisor) 16.2 14.9-17.6 0.67 4.13 Pennsylvania! 15.15 14.6-16.0 0.32 2.09 Cowpasture River valley, Bath Co., Va. (data above) 15.3 14.8-15.9 0.16 1.04 Warm Springs Mountain, Bath Co., Va. (data above) 15.2 14.6-15.6 Late Pleistocene: New Paris No. 4, Bedford Co., Pa.! 15.3 13.8-17.2 0.88 5.74 Clark’s Cave, Bath Co., Va. 16.6 15.5-18.4 0.50 3.00 Recent: Cowpasture River valley, Bath Co., Va. (data above) Length of dentary (from condyle) 12.4 11.9-13.1 0.47 3.79 Warm Springs Mountain, Bath Co., Va. (data above) 12.2 11.9-12.4 Late Pleistocene: Clark’s Cave, Bath Co., Va. 13.3 12.3-15.2 0.54 4.05 Ipata from Guilday et al., 1964. ^Data from Ray, 1967, Table 2. N 21 77 16 5 10 10 2 4 15 19 17 16 5 29 52 15 5 60 46 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 9. Measurements (in mm) of Blarina brevicauda, various localities (continued). Locality X OR SD CV N p4 — m3, Antero-posterior Crown Length^ Recent: Minnesota 1 6.29 5. 8-6.6 0.19 3.02 20 Pennsylvania! 6.06 5.7-6. 3 0.19 3.13 24 Cowpasture River valley, Bath Co., Va. (data above) 6.1 5.9-6. 3 0.22 3.60 16 Warm Springs Mountain Bath Co., Va. (data above) 6.2 6.0-6. 3 5 White Sulphur Springs, Greenbrier Co., W. Va.2 5.85 5.48-6.19 — — 10 Pisgah Forest, Transylvania Co., N. C.2 5.78 5.47-5.98 — 10 Late Pleistocene: New Paris No. 4, Bedford Co., Pa.! 6.14 5.5-6.8 0.39 6.48 36 Natural Chimneys, Augusta Co., Va.! 6.43 6. 2-6.6 0.13 2.02 9 Clark’s Cave, Bath Co., Va. 6.56 6. 2-6.9 0.14 2.13 57 Recent: Minnesota! Width, mandibular condyle 4.16 3.9-4.6 0.19 4.56 22 Pennsylvania! 3.85 3.7-4. 1 0.13 3.37 24 Cowpasture River valley, Bath Co., Va. (data above) 3.89 3.7-4. 2 0.18 4.62 16 Warm Springs Mountain, Bath Co., Va. (data above) 3.89 3. 8-4.0 5 Late Pleistocene: New Paris No. 4, Bedford Co., Pa.! 3.91 3. 3-4.5 0.30 7.67 37 Clark’s Cave, Bath Co., Va. 4.01 3.49-4.65 0.20 5.03 13 ^Data from Guilday et al., 1964. 20ata from Ray, 1967, Table 2. ^Defined as C1-M3 in Guilday et al., 1964. (See Repenning, 1967 for terminology). ing clues to the minimum date of deposition. Blarina occurred continuously from boreal times to the present in the area. During this time span it dimin- ished in size. If deposition had continued undimin- ished, remains of this shrew would have been expeeted to reflect this. They did not. From an in- spection of the measurements in Table 9 and the low coefficient of variation from the Clark’s Cave sample, it is apparent that the deposit was composed pri- marily of large individuals — the minimum measure- ments recorded exceed the average of Recent B. b. kirtlandi. This size difference is clearly brought out by comparing the measurements of the Clark’s Cave Blarina sample with those from Recent examples 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 47 from the Cowpasture River valley taken by Charles O. Handley, Jr. and his students (Table 9). It would appear that all short-tailed shrews, and by extension the bulk of the deposit, accumulated prior to any diminution in size, and that deposition threrefore ceased at that spot by the close of late Pleistocene times, or if deposition continued to occur, it was not on a sustained basis. The large size of the short-tailed shrew remains is indicative of more boreal conditions, but its wide choice of habitats, occupying practically all terres- trial habitats within the forested East, are too broad to provide any specific ecological clues. Short- tailed shrews range north to the shores of James Bay. Remains of this shrew were conspicuously absent from the lower levels of the stratified New Paris No. 4, Pa., deposit. These lower levels produced only boreal mammals, and it appears that at that site, the short-tailed shrew was absent during full glacial times, becoming common only with the return of more temperate conditions. But it may have been present at that time at Clark’s Cave, two degrees latitude to the south. Further collections may clarify the situation. Family: Taipidae — Moles Parascalops breweri (Bachman) — Hairy-tailed Mole material: cm 24537. 2 partial skulls; 8 left, 7 right mandibles; II left, 12 right humeri. MNl = 12 individuals. Scalopus aquaticus (Linnaeus) — Eastern Mole material: cm 24539. 1 partial right mandible, no dentition; 1 left humerus. MNl = I individual. Condylura cristata (Linnaeus) — Star-nosed Mole material: cm 24538. 12 left, 4 right partial mandibles; 11 left, 13 right humeri; I rostrum. MNl = 13 individuals. remarks; Moles (0.9% of all terrestrial mammals) were rare in the deposit. Their secretive habits pro- tect them from aerial predation. The eastern mole, Scalopus, was represented by a single individual. The hairy-tailed mole, Parascalops (46%), and the star-nosed mole, Condylura (50%), were present in equal numbers. This is in striking contrast to the situation at Sheep Rock Shelter, Pa., 300 km north of Clark’s Cave, where Scalopus accounted for 89% of all moles, Condylura 9%, and Parascalops but 2.8%. The modern ecological situation is much the same at both sites, a narrow intermontane flood plain be- tween high, parallel, oak-forested ridges in the Atlantic drainage. The high numbers of Scalopus at Sheep Rock, compared with the situation at Clark’s Cave is, we believe, due to the relative age of the two sites. Sheep Rock Shelter is a Recent owl roost uninfluenced by late Pleistocene climatic changes. Both Parascalops and Condylura occur in western Virginia today (Handley & Patton, 1947). They probably do not occur as low as Clark’s Cave or, if so, rarely. Scalopus, however, should occur in the extensive flat flood plain of the Cowpasture River above and below the site within cruising range of the Clark’s Cave owl population. Its relative rarity in the fossil deposit is an indication of the predominantly boreal conditions reflected by the fossil assemblage. The fact that Scalopus was taken in such large numbers relative to Parascalops and Condylura at Sheep Rock indicates that the eastern mole was as susceptible to owl predation as the other two species of moles. Its scarcity in the Clark’s Cave deposit must reflect the true picture at the time of deposi- tion. Moles were relatively more common at Sheep Rock ( 1 2% of mammals against 0.9% at Clark’s). This reflects differences in excavation technique. The total recovery attempted at Clark’s Cave resulted in larger numbers of other small mammals from the sample. Parascalops and Condylura have similar modern ranges in the Appalachian Mountains and adjacent plateau country of eastern North America north to the southern portion of the Canadian Shield. Para- scalops is found no farther north than Maine. Condylura ranges north to James Bay and southern Labrador. Both had more extensive ranges during the boreal phase of the Wisconsinan. Parascalops has been reported from Robinson Cave in central Ten- nessee, and Parascalops and Condylura from Crank- shaft Pit in the Ozark Highlands of southeastern Missouri. The range of Scalopus, on the other hand, is austral, including all the southern and central United States north to Minnesota, Michigan, and Massachusetts. It enters the Appalachians only in areas where it can follow the low altitude flood plains and rivers that traverse the mountain ridges. The one Scalopus humerus measures 14.8 mm in length and 12.5 mm in width, well within the range of S. a. aquaticus, the race living in the state today (Guilday, 1961a:117, Table 1, for comparative measurements). Adult males of both the star-nosed and the hairy- tailed moles are larger than females. The Clark’s Cave samples are too small to adequately express the true extent of variation, but they do suggest sexual size differences as in the living animals. This is reflected by the slightly larger coefficient of varia- tion of p4-m3 of the Clark’s Cave Condylura, com- 48 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 10. Measurements (in mm) of Condylura cristata and Parascalops breweri. Locality and age X OR SD CV N Recent: Pennsylvania, males (CM 23864-71, 6.3 Condylura cristata Alveolar length p4 — m3 6.1-6.6 0.15 2,37 14 23874-5, 23877-80) Late Pleistocene: Clark’s Cave, Va. 6.8 6.S-7.2 0.23 3.37 16 Recent: Pennsylvania, males Maximum length, humerus 13.0 12.8-13.2 6 Late Pleistocene: Clark’s Cave, Va., females? 13.1 12.8-13.4 11 Clark’s Cave, Va., males? 13.9 13.8-14.3 — — 8 Clark’s Cave, Va., sexes combined 13.4 12.8-14.3 0.46 3.46 19 Late Pleistocene: Clark’s Cave, Va., males? Parascalops breweri Maximum length, humerus 14.6 14.0-15.2 10 Clark’s Cave, Va., females? 13.4 13.2-13.7 — — 5 Clark’s Cave, Va., sexes combined 14.2 13.2-15.2 0.65 4.58 15 pared with the same measurement in a sample of Recent males, and in the bimodal distribution of the total length of Condylura humeri. The Clark’s Cave Condylura humeri fell into two size classes, with no overlap in their observed ranges (Table 10): females? 12.8-13.4, males? 13.8-14.3. An alternate explana- tion, that the Clark’s Cave sample is composed of two intermingled populations, a larger late Pleistocene and a smaller Recent form, was rejected because a similar situation could not be demonstrated in those species, Blarina brevicauda, for example, where a significant sexual size difference does not occur. Humerus length of six Recent male Condylura (Table 10), the larger sex, fell within the range of the Clark’s Cave presumed females and were signifi- cantly smaller than those of presumed males from the deposit. The Recent males averaged some 6.5% smaller than the Clark’s Cave males?, suggesting that the late Pleistocene population, as at New Paris No. 4, Pa., was indeed larger in body size than its modern regional counterpart. In the case of Parascalops, little, if any, size differ- ence can be shown. Sexual size differences are sug- gested by the relatively broad observed range of humeri (Table 10), but the sample is too small to be other than suggestive. Order: Chiroptera — Bats Family: Vespertilionidae — Evening Bats At least 1,554 individual bats were represented, and their remains formed 36% of the total mammalian assemblage at the site. Bats are not usually rep- resented in raptor prey. Their high numbers at Clark’s Cave reflected the presence of a large bat colony in the cave itself. Gillette and Kimbrough (1970:265) cite instances of bat predation by raptors, “groups of hawks or falcons working together in a small area with devastating effectiveness on emerg- ing or returning colonies.” The nocturnal feeding habits of bats and owls bring them into contact with greater frequency, and owl predation on bats is well-documented (Hall & Blewett, 1964). The roost site, ideal for birds of prey, is too shallow and exposed to have served as a hibernaculum, and most of the excavated bat bones are presumed prey items derived from the large bat population deep in the cave itself. Genus Myotis Kaup — Little Brown Bats MATERIAL: CM 24599-24619. 2,309 whole or partial mandibles. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 49 1,000 of the most complete are grouped in lots according to alveo- lar length of lower toothrow, c-m3, in 0.1 mm increments. Addi- tional material, unmeasured mandibles, skull fragments etc., under CM 24626-24628. A minimum of 877 M. lucifugus, M. sodalis, or M. keenii based upon 76% of all Myolis. A minimum of 138 M. cf. leibii based upon 12% of all Myotis. A minimum of 138 M. cf. grisescens based upon 12% of all Myotis. remarks: All identifications are based upon lower jaws. Alveolar length of lower toothrow c-m3 of 1,000 specimens is shown in the graph in Fig. 16. The Clark’s Cave Myotis sample is composed of three subgroups based upon size. A trimodal curve results with modes at 5.4, 5.8, and 6.6 mm. These could be related to modern eastern North American species. The smaller group, mode 5.4 mm, about 12% of the Myotis in the sample, is Myotis cf. leibii, Leib’s bat. Although never a common species, Leib’s Myotis, all species, Clark's Cave, Bath County, Virginia, alveolar length cl - m3 N = 1,000 (489 left; 51 1 right) lower jaws. Ordinate number of individuals Fig. 16. Histogram. Measurements in millimeters of mandibles of various species of bats, Myotis, from Clark’s Cave, Bath County, Virginia. 50 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 bat is broadly distributed in the East from Maine and southern Quebec south to western Virginia (Hall & Kelson, 1959). Not listed as a member of the fauna of Virginia by Handley & Patton, 1947, it has sub- sequently been reported from Millboro Cave, Bath County, and Hupman’s Saltpeter Cave, 750 m eleva- tion, in neighboring Highland County (Johnson, 1950), and present but rare, in Starr Chapel and Porter’s Caves, Bath County (Holsinger, 1964). Four specimens were netted at entrance No. 3 of Clark’s Cave in August, 1974, by C. O. Handley, Jr. The larger subgroup, mode 6.6 mm, estimated as 12% of all Myoiis, is M. cf. grisescens, the gray bat. This bat has been reported only in extreme south- western Virginia, Grigsby Cave, in Scott County (Holsinger, 1964). It has been recorded as a fossil north and east of its present range from Cumberland Cave, Md., and Organ-Hedricks, Windy Mouth, and Patton Caves, W. Va. (Handley, 1956a:251). These records, coupled with the evidence of a sizeable former population in Clark’s Cave, suggest a range shift to the southwest since late Pleistocene times. By far the bulk of the Myotis, about 75%, fell within an intermediate range. Inspection of this curve, mode 5.8 mm, indicates that unlike the other two, which appear symmetrical, it is strongly skewed to the left. There are at least two and possibly three species of Myotis represented here: M. lucifugus, the little brown bat; M. sodalis, the social bat; and M. keenii, Keen’s bat. At the present time, M. lucifugus is the commonest northeastern cave bat, M. keenii is mod- erately common, and M. sodalis is uncommon. During the years 1946-1951, six M. lucifugus were collected for each M. keenii in Pennsylvania (Roberts & Early, 1952). There is some evidence that M. keenii may have been relatively more common during the late Pleistocene in Pennsylvania, at least. New Paris sinkhole No. 4 yielded the remains of 302 M. keenii, but only 225 M. cf. lucifugus or sodalis (Guil- day et ai, 1964:151). The fact that the frequency curve is skewed towards the small end of the scale suggests that the commonest component species was probably M. lucifugus (or M. sodalis) and that the larger M. keenii was the minority species. No attempt was made to assign a specific identification to particular specimens because of the extensive overlap in measurements in modern samples and the poor state of preservation of the fossils. Holsinger, 1964, records M. keenii, M. lucifugus, and M. sodalis from Bath County caves. Pipistrellus subflavus (F. Cuvier) — Pipistrelle Bat material: cm 24620. 26 left, 19 right mandibles. MNl = 26 individuals. remarks: Pipistrelles comprised only 1.2% of the total number of bats recovered. In contrast, they formed 53% (45 out of 84 bats) of a mist-net sampling run for three hours at entranee no. 3 of Clark’s Cave in August, 1974. Alveolar length c-m3, mm followed by (N): 4.5(1), 4.6(8), 4.7(5), 4.8(16), 4.9(3), 5.0(5), 5.1(1). Eptesicus fuscus (Palisot de Beauvois) — Big Brown Bat material: cm 24624, 24625. 341 left, 363 right mandibles, partial skulls and fragments. MNl = 363 individuals. remarks: The big brown bat is a common species at the cave today. It comprised 15% (13 out of 84) of the mist-netted bats, August, 1974. Lasiurus cf. borealis (Muller) — Red Bat material: cm 24621. 3 left, 2 right mandibles. MNl = 3 individuals. REMARKS: Although this tree bat is one of the commonest of eastern North American bats, it is not normally a cave-frequenting species and would not be captured with any frequency at the cave entrance by birds of prey. Alveolar length c-m3, mm followed by (N): 5.1(1), 5.3(1), 5,8(2), 5.9(1). Plecotus E. Geoffroy Saint Hilaire — Big-eared Bats material: cm 24622. 9 left, 7 right mandibles. MNl = 9 individuals. REMARKS: Two Recent species of big-eared bats have been reported from Virginia, P. rafinesquii Lesson and P. townsendii Cooper. P. rafinesquii is a bat of the southeastern United States and the Mississippi Valley as far north as Ohio. The sole Virginia record is from Dismal Swamp on the coastal plain (Handley, 1959:161). It has been collected in the Appalachians just west of the state, however, in eastern Tennessee, Kentucky, and West Virginia (Collison Cave, Nicholas Co., Handley, 1959:165). The only Recent record of a big-eared bat from the mountainous western portion of the state is of P. townsendii, the western big-eared bat. There is a thriving population of this bat in the central Appala- chians, separated by some 960 km from the Ozark Mountain segment of its range. According to Handley they have been reported from at least 19 caves from the West Virginia highlands within an area 48 km wide and 64 km long (/6z<7:202) north of Clark’s Cave, with an isolated station 240 km southwest in Taze- well County, Va. Holsinger, 1964, reports P. town- 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 51 sendii from caves of over 600 m elevation from High- land and Bath counties, Virginia. The Clark’s Cave specimens, fragmentary mandibles, could have been either species. Mandibles of a third species, the extinct P. alleganiensis de- scribed from the mid-Pleistocene Cumberland Cave deposit, are not distinguishable from modern species of Plecotus. But Cumberland Cave clearly antedates the Clark’s Cave deposit (Gidley & Gazin, 1938). Bats of the genus Plecotus have also been reported from Frankstown Cave, a Pleistocene fissure deposit in the Ridge and Valley section of central Pennsylvania (Guilday, 1961b). Handley, 1959, regards the disjunct Appalachian distribution of P. townsendii as a relict of a more extensive past distribution. Its isolation is reflected in its recognition as a distinct subspecies, P. t. vir- ginianus. The big-eared bat population of Clark’s Cave was never a large one. The few remains from the fossil deposit, representing only 9 individuals, or 0.6% of the fossil bat fauna, may have been from itinerant individuals. Order: Lagomorpha — Rabbits, Hares, Pikas Family: Leporidae — Rabbits and Hares Sylvilagus cf. transitionalis (Bangs) — New England Cottontail material: cm 24591. Partial left innominate, distal end of left femur. MNl = 1 individual. remarks: The minimum number of all leporids from Clark’s Cave is 27 (1% of all terrestrial mam- mals). All but one, based on the two fragments listed above as S. transitionalis, are referred to snowshoe hare. The S. transitionalis bones are small (Table 1 1), too small for snowshoe hare or common cottontail S. floridanus. The diagnostic supraorbital processes were not preserved, so the identification, based solely on size, is provisional. Handley & Patton, 1947:189, state that S. tran- sitionalis is found at higher elevations throughout the mountainous portions of the state, probably not below 750 m elevation. It is sympatric with Lepus americanus in Virginia today. Bailey, 1946:282, specifically lists Bath County in the modern range. Fossil or subfossil remains have also been reported from New Paris No. 4, Pa., and Ladds Quarry, Ga. Lepus americanus Erxleben — Snowshoe Hare material: cm 24590. 18 left, 14 right premaxillae (14 re- taining first incisors); 32 isolated first upper incisors; 2 left, 2 right frontals; 10 left, 8 right maxillae; 1 occipital; 17 left, 13 right mandibles; 9 atlas vertebrae; 21 left, 15 right scapulae; 12 left, 10 right humeri; 14 ulnae; 4 radius; 21 right, 6 left innomi- nates; femora — 16 proximal, 13 distal ends, 1 complete; 15 left, 8 right tibia; 15 left, 17 right calcanea. MNI = 23 individuals. REMARKS: The snowshoe hare does not occur in the Cowpasture River valley or the surrounding hills today. Handley & Patton (1947:186) state that the snowshoe hare probably occurred at one time throughout the mountainous areas of the state at higher elevations. “Now known to occur only in Highland County, where it is uncommon ... re- stricted to areas of spruce and fir, found in open woods and thickets.” Characters differentiating Lepus from Sylvilagus included shape of the supraorbital processes, the occipital, the posterior margin of incisive foramen, the rugosity of the zygoma, and the relative slender- ness of shafts of humerus, femur, and tibia. Although most of the lagomorph material was so fragmentary that specific, even generic, identifi- cation was not possible, all, with the exception of two elements identified as New England cottontail, CM 24591, are referred to Lepus americanus for the following reasons: (1.) All preserved diagnostic parts were Lepus; (2.) Measurements average slightly larger than late Pleistocene specimens of L. ameri- canus from New Paris No. 4, Pa. The complicating factor in identification is that gross size cannot be used in separating L. ameri- canus from Sylvilagus floridanus. Modern L. a. virginianus from the central Appalachians is easily distinguished from Sylvilagus by virtue of its larger size, but late Pleistocene populations of L. ameri- canus, like those at New Paris No. 4, Pa., are as small as modern Sylvilagus from the same area. They are indistinguishable on the basis of size alone. Order: Rodentia — Rodents Family; Sciuridae — Squirrels Tamias striatus (Linnaeus) — Eastern Chipmunk material: cm 24533. 22 left, 24 right whole or partial man- dibles; 5 left, 4 right maxillae. MNI = 24 individuals. remarks: The eastern chipmunk is common throughout the state at all altitudes and in all types of brushy or forested situations today (see Handley & Patton, 1947 for exceptions). It is also common in the Clark’s Cave deposit, comprising 22% of the eight species of sciurids from the deposit. The Clark’s Cave sample is composed of chip- munks larger in size than present mid-Appalachian forms (Table 12). They agree in size with both the late Pleistocene sample from New Paris No. 4, Pa., 52 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 and the largest living subspecies, T. s. pipilans (Table 12). Tamias striatus from the Appalachians exhibits a negative Bergmann’s Response. It increases in size with decreasing latitude. Specimens from Quebec at the northern extremity of its northeastern range average 5% smaller in total length of skull than specimens from western North Carolina (Table 12). Individuals of T. s. pipilans from Louisiana, the largest subspecies, may be as much as 12.5% larger Table 1 1. Measurements (in mm) of Lepus americanus, Sylvilagus floridamis and Sylvilagiis transitionalis. Locality X OR N Distal width of humerus Lepus americanus. Recent, Pa. Lepus americanus. New Paris No. 4, Pa. 10.1 9.5-10.5 6 late Pleistocene Lepus americanus, Clark’s Cave, Va. 8.3 7.5-9. 1 28 late Pleistocene 8.8 8. 3-9.5 18 Sylvilagus floridanus , Recent, Pa. 8.1 7.7-8.6 10 Sylvilagus transitionalis. Recent, Mass. 7.5 7.3-7.7 2 Anterior-posterior diameter of acetabulum Lepus americanus , New Paris No. 4, Pa., late Pleistocene 7.8 Lepus americanus , Clark’s Cave, Va. late Pleistocene 7.95 Sylvilagus floridamis. Recent, Pa. 8.2 Sylvilagus transitionalis. Recent, Mass. 7.7 Sylvilagus cf. transitionalis, Clark’s Cave, Va., late Pleistocene 6.5 Distal width of femur Lepus americanus. Recent, Pa. Lepus americanus. New Paris No. 4, Pa. 15.4 14.8-16.4 7 late Pleistocene 13.2 12.4-14.3 20 Lepus americanus, Clark’s Cave, Va. late Pleistocene 13.4 12.3-14.5 11 Sylvilagus floridanus. Recent, Pa. 13.4 12.8-13.8 10 Sylvilagus transitionalis , Recent, Mass. Sylvilagus transitionalis, Clark’s Cave, 12.0 11.7-12.3 2 Va., late Pleistocene 11.5 1 Distal width of tibia Lepus americanus. Recent, Pa. Lepus americanus. New Paris No. 4, Pa., 13.5 12.4-14.5 7 late Pleistocene 11.2 11.0-11.5 9 Lepus americanus, Clark’s Cave, Va., late Pleistocene 11.7 11.0-12.8 18 Sylvilagus floridanus , Recent, Pa. 11.2 11.0-11.6 5 Sylvilagus transitionalis. Recent, Mass. 10.3 1 7.0-8.4 20 7.4-8.9 16 7. 5-9.1 9 00 1 q 2 1 Measurements by W. Pollock and A. Guilday. Specimens examined; Lepus americanus. Recent, Pa.; CM 2212, 2211, 2213, 2214, 2215, 2217, 2216, 33286, 33284, 33285, G-644; New Paris No. 4, Pa., late Pleistocene; CM collections; Clark’s Cave, Va., late Pleistocene; CM 24590. Sylvilagus floridamis. Recent; CM 27007, 27008, 27010, 27012, 7951, 7952, 8363, 10673, 19365, 35790, 10137, 7651, G-8. Sylvilagus transitionalis'. no number, Guilday collection. Mass.; Clark’s Cave, Va., late Pleistocene; CM 24591. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 53 in total length of skull (USNM 28459, Ray, 1965: 1020) than average Quebec specimens. Unfortunately, comparable measurements could not be taken from the fragmentary Pleistocene speci- mens, but those from Clark’s Cave average 8% larger in alveolar length of lower toothrow than modern Pennsylvania specimens of T. s. fisheri. Although the eastern chipmunk does become smaller toward the north in the eastern portion of its range, it becomes larger at all latitudes toward the west. It is possible that the superior size of the late Pleisto- cene T. strialus from the mid-Appalachians may be a reflection of the same environmental conditions that accounted for the influx of so many midwestern forms into the Appalachians at that time. At Clark’s Cave and Natural Chimneys, Va., re- mains of nocturnal, arboreal flying squirrels exeeeded those of terrestrial, diurnal ehipmunks 2/1 and 3/1 respectively. At New Paris No. 4, Pa., the reverse was true. Chipmunk remains outnumbered those of flying squirrels by 2/1. Marked ehanges in the relative percentages of species from two or more sites sam- pling the same temporal fauna in the same general geographic area, in this case the mid-Appalachians, reflect either differing habitats, predator preferences, or methods of deposition. The Clark’s Cave and Natural Chimneys deposits were formed largely by owls, which accounted for a large number of noctur- nal flying squirrels in comparison with the number of diurnal chipmunks. At New Paris No. 4, Pa., animals were trapped in an open tumble-in sinkhole, re- sulting in higher numbers of terrestrial, as opposed to arboreal, forms. Predator bias was not a factor at New Paris No. 4, so nocturnal forms were not selected, as in the case of the owl roost sites. Eutamias minimus (Bachman) — Least Chipmunk material: cm 24532. 2 left mandibles with full dentition; 1 left, 1 right mandible with p4-ml. MNI = 3 individuals. remarks: The least chipmunk has not previously been reported from the Appalachians, either fossil or Recent. It occurs in western and central North America, usually in open to brushy, boreal, coni- ferous forest situations, and reaches its greatest Table 12. Measurements (in mm) of Tamias striatus. Locality and taxa X OR SD cv N Recent: T. striatus pipilans^ Alveolar length, p4 — m3 6.76 6. 3-7.4 0.30 4.40 19 New Paris No. 2, Pa. (1,875 yrs. B.P.) 6.28 5. 8-6.8 0.18 2.86 114 Late Pleistocene: New Paris No. 4, Pa. (11,300 yrs. B.P.) 6.74 6. 3-7.1 0.20 6.67 30 Hartman’s Cave, Pa.l 6.90 6. 2-7.7 0.42 6.11 9 Clark’s Cave, Va. 6.81 6.5-7. 3 0.21 3.08 28 Robinson Cave, Tenn. 7.5 7. 3-7. 8 — — 2 Recent: T. s. quebecensis, Quebec^ 480 30’N. lat. 39.1 Total length of skull 37.4-40.0 4 T. s. lysteri, Ontario^ 430 30’N. lat. 39.4 38.0-40.1 7 T. s. fisheri, N.Y.2 410 N. lat. 40.1 38.8-41.0 10 T. s. striatus, N.C.2 340 N. lat. 41.1 39.8-42.6 — — 11 Pleistocene: T. aristus, Ga.l 340 N. lat. 52.7 1 Measurements from Ray, 19651 ; Howell, 1929^; Cameron, 195Q3. 54 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 abundance in open, sandy, pine and spruce park- lands. It is not found in the United States east of Wisconsin. But north of the Great Lakes it occurs across the southern half of Ontario east to extreme western Quebec, 800 km north of Clark’s Cave (Ban- field, 1974), as shown in Fig. 17. Identification, based upon dentitions, both lower and (Back Creek Cave No. 2, Va.) upper, is firm to genus. Referral to E. minimus is based on small size, geographic proximity, and ecological proba- bility. The eastern chipmunk (Tamias), an inhabitant of deciduous forest and brushland, and the least chip- munk (Eutamias) coexist today in a broad area of east-central Canada extending from southern Mani- toba and northern Wisconsin and Minnesota toward the east and north of the Great Lakes for 650 km across the southern half of Ontario (Fig. 18). Where the two occur in the same general area, Tamias prefers deciduous, Eutamias coniferous, forest situations. Their remains also have been found together at the 7,000 to 7,500-year B.P. levels at Fig. 17. Modern range of least chipmunk, Eutamias minimus (Bachman), adapted from Hall & Kelson, 1959. Present in Clark’s Cave local fauna. Table 1 3. Measurements (in mm) of Eutamias cf. minimus, late Pleistocene. Locality and measurement X OR N Clark’s Cave, Va. p4 - m3, alveolar length 5.38 5.33-5.40 4 p4 - m3, occlusal length 4.85 4.75-4.95 4 Back Creek Cave No. 2, Va. P3 - M3, occlusal length c. 5.33 1 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 55 the Itasca Bison Kill Site in northcentral Min- nesota (Shay, 1971). Eutamias minimus is now known from one other Appalachian cave site, a raptor-roost deposit of similar age. Back Creek Cave No. 2, Va., 24 km southwest of Clark’s Cave (CM 29727, maxilla with P3-M2). This strictly diurnal hibernating rodent would not be expected to bulk as large in the deposit as noc- turnal small mammals active the year round. It is possible that this small, mouse-size ground squirrel may have been more common than its remains would indicate. Marmota monax Linnaeus — Woodchuck material: cm 24527. I right 1 left mandible; skull fragments. MNI = 2 individuals. remarks: a common species in late Pleistocene Appalachian cave deposits, this large terrestrial ground squirrel ranges throughout the deciduous and coniferous woodlands of east-central and northern North America north to Alaska and Labrador. Oc- clusal length p4-m3 is 18.9 mm. Spermophilus tridecemlineatus (Mitchill) — 13-lined Ground Squirrel material: cm 24531. 5 left, 5 right whole or partial mandi- bles; 2 right maxillae. MNI = 5 individuals. REMARKS: Remains of the 13-lined ground squirrel are generally distributed in eastern North American periglacial cave sites where they form from 4% to 10% (over 90% at Welsh Cave, Ky.) of the sciurids recovered. Remains are known from at least six other late Wisconsinan sites east of their present range: New Paris No. 4, Pa.; Bootlegger Sink, Pa.; Eagle Cave, W. Va.; Natural Chimneys, Va.; Welsh Cave, Ky.; Robinson Cave, Tenn. They have also been reported from Cumberland Cave, Md., late Kansan in age. The 13-lined ground squirrel was apparently common and widespread throughout the East during late glacial times, a distinct indicator of semi-prairie or parkland conditions. Remains of this nominally midwestern ground squirrel occur with those of the eastern chipmunk and the least chipmunk at Clark’s Cave. These three species of ground squirrels coexist today only in a relatively small area, stretching from northern Wisconsin north and west of the Great Lakes to southern Manitoba, from 1 ,200 km to 2,400 km north- west of Clark’s Cave (Fig. 18). This area defines the present day boundary between eastern coniferous/ deciduous forest and grassland (Cushing, 1965), the Illinoian and Canadian Biotic Zone contact of Dice, 1943. In this area today the chipmunk, Tamias striatus (deciduous woodland preference), the least chipmunk, Eutamias minimus (coniferous woodland preference), and the 13-lined ground squirrel, Sper- mophilus tridecemlineatus prairie preference), find an intermesh of habitats enabling them to occur sympatrically, if not on identical ground, at least in such proximity that they can be harvested by raptors operating from a single roost, as at Clark’s Cave. Alveolar length of P3-M3 from one individual was 7.46 mm. Alveolar length of p4-m3 from eight examples was X 7.76 mm (OR was 7.37-8.2). Sciurus cf. carolinensis Gmelin — Gray Squirrel material: cm 29612. I left M2 (unerupted); I right M2 or M3 (extreme tooth wear); I M3 (slight tooth wear); I left, 1 right p4; I left, 1 right ml; 1 right ml or m2; I left 1 right m3. MNI = 3 individuals. remarks: The gray squirrel found primarily in deciduous mast-producing forest was probably not a member of the boreal fauna at Clark’s Cave. It is common throughout the Appalachians north to New Brunswick. Farther north only the red squirrel, Tamiasciurus hudsonicus, occurs. The persistence of the gray squirrel in western Virginia during full glacial times would have been governed by the type of forest cover. Pollen analysis at Hack and Quarles Pond in the Shenandoah Valley, at approximately the same altitude as both Natural Chimneys and Clark’s Cave, seems to indicate that the western valley floor of the state supported a coniferous boreal forest within which one would not expect the gray squirrel to occur (Craig, 1969). At the New Paris Pa., sinkholes, gray squirrel was the common tree squirrel recovered from the Recent Sinkhole No. 2 fauna. Only red squirrel remains were present in the boreal late Pleistocene Sinkhole No. 4 fauna. One maxilla of gray squirrel was recovered from the late Pleistocene Natural Chimneys, Va., local fauna (CM 7536), in association with five red squirrels. The red squirrel is also the commonest squirrel in the Clark’s Cave fauna (MNI 25). The ten molars recovered were from at least three individuals, judging from the various states of tooth- wear. They agree in size with modern Pennsylvania comparative material and are not large enough to be those of fox squirrel, Sciurus niger. Tamiasciurus hudsonicus (Erxleben) — Red Squirrel material: cm 24534. 13 left, 25 right whole or partial mandi- bles; 8 left, 4 right maxillae; I premaxilla. MNI = 25 individuals. 56 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 18. Modern range overlaps of three species of terrestrial sciurids from Clark’s Cave local fauna. A = overlap least chip- munk, Eutamias minimus (Bachman), and eastern chipmunk Tamias striatus (Linnaeus), B = overlap E. minimus, T. striatus, and 13-lined ground squirrel, Spermophilis tridecemlineatus (Mitchell). remarks: The common large tree squirrel of Appalachian late Pleistocene sites, the red squirrel, accounted for 26% of the sciurids from the deposit. The red squirrel, the most widely distributed North American sciurid, ranges throughout the northern boreal forest from Alaska to Labrador, south in the Rocky Mountains to New Mexico in the West, in the Appalachian Mountains to North Carolina in the East. Commonest in coniferous forests, it is a very adaptable squirrel, and may occur in deciduous hardwood south to Iowa, Illinois, and North Carolina (Hall & Kelson, 1959). It occurs at the site today. The Clark’s Cave material is so fragmentary that the only measurements that could be taken were alveolar length of P4-M3 and p4-m3. P4-M3, one observation, was larger than that of thirty modern examples of T. h. loquax from Pennsylvania (Table 14). It was comparable in size to specimens from northwestern North America. Measurements of p4- m3 also indicate a large form, although this measure- ment is difficult to take with consistency, and is biased toward larger size by the sloping p4 and erosion of the anterior alveolar wall. Alveolar length of p4-m3 may exceed occlusal length by as much as 0.5 mm, but even when this correction is made it is evident that the Clark’s Cave Tamiasciurus material is of larger size than Recent eastern specimens, and is comparable to Recent Northwestern and late Pleistocene New Paris No. 4, Pa. material. Genus Glaucomys Thomas — Flying Squirrels Glaucomys sabrinus (Shaw) — Northern Flying Squirrel material: cm 24529. 28 left, 22 right whole or partial mandi- bles; 8 left, 13 right maxillae. MNI = 28 individuals. Glaucomys volans (Linnaeus) — Southern Flying Squirrel material: cm 24530. 15 left, 19 right whole or partial man- dibles; 1 left, 2 right maxillae. MNI = 19 individuals. remarks: Flying squirrels comprised about 42% of tbe eight species of sciurids from the deposit, and 63% from Natural Chimneys, Va., high figures that clearly reflect the activity of nocturnal birds of 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 57 Table 14. Measurements (in mm) of Tamiasciurus hudsonicus. Age and locality X P4 OR — M3, occlusal length SD CV N Recent: Pennsylvania! 7.37 6.7-8. 1 — — 30 Natishquan R., Que., Hamilton R., Lab.! 7.58 7. 2-8.4 — — 17 Hudson Bay, Que.! 1.15 7. 5-8.1 — — 16 Moorhead, Minn.! 7.85 7. 2-8.4 — — 10 Aklavik, NWT, Seward, Alaska! 8.09 7.7-S.4 — — 9 Late Pleistocene: Clark’s Cave, Va. 8.4 p4 — m3, occlusal length 1 Recent: Pennsylvania! 7.39 6. 8-7.8 — — 33 Natishquan R., Que., Hamilton R., Lab. 7.66 7. 3-8.0 — — 20 Hudson Bay, Que.! 7.84 7.6-8. 2 — — 15 Moorhead, Minn.J^ 7.92 7. 2-8.5 — — 11 Aklavik, NWT, Seward, Alaska! 8.20 8.0-8.4 — — 9 Late Pleistocene: New Paris No. 4, Pa.! 8.20 p4 8.0-8. 3 — m3, alveolar length 4 Late Pleistocene: Clark’s Cave, Va. 8.68 8.15-9.2 .3 3.46 20 ^Measurements from Guilday et al., 1964. prey. Percentages for the diurnal chipmunk {Tamias striatus), 23% and 14% respectively, were much lower. In contrast with these hgures, derived from raptorial bird roost-litter, remains of nocturnal flying squirrels from the pit-trap sinkhole at New Paris No. 4, Pa., comprised only 19% of all squirrels, and were greatly exceeded in numbers by diurnal terrestrial chip- munks (31%). As at other late Pleistocene cave deposits from the central Appalachians, New Paris No. 4, Pa., and Natural Chimneys, Va., there were two species of Glaucomys present. The larger is identified as G. sabrinus, the smaller species as G. volans. They were sharply diflferentiated in size, with no overlap in observed ranges, and had a mean size-differential of 16% in alveolar length of p4-m3 (Table 15). Late Pleistocene representatives of both species are larger than those now present in the central Appalachians. Samples of G. volans from New Paris No. 4, Pa., Natural Chimneys, Va., and Clark’s Cave, Va., average 8% larger than modern Pennsylvania specimens, although there is an overlap in the ob- served ranges (Table 15). Samples of G. sabrinus from the same cave deposits, however, average 14% larger than Recent G. sabrinus from the Appala- chians in alveolar length of p4-m3 with no overlap in their respective observed ranges (Table 15). They do agree in size with specimens from the north- ern and western portions of the Recent range. Alveo- lar length of p4-m3 is a difficult measurement to take with accuracy because the anterior root of p4 slopes anteriorally away from the crown. In fossil specimens the anterior wall of the alveolus of p4 is often broken or eroded in varying degree. This would have the effect of lengthening the measurement. Despite this source of error, the size differential appears to be a real one. Both G. volans and G. sabrinus were larger during late Wisconsinan times in the Appala- chians than are their modern counterparts, a reflec- tion of cooler conditions, expressed as “Bergmann’s Response.” Glaucomys volans occurs today throughout the state at all elevations and in all forest types (Handley & Patton, 1947). Glaucomys sabrinus, on the other 58 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 hand, has been taken in Virginia only at the highest elevations (Handley, pers. comm.). It occurs in the mountains directly west of Bath County, in West Virginia, at elevations exceeding 900 m. It has been taken as far south, in the Appalachians, as Tennessee and North Carolina at altitudes ranging from 1,200 m to 1,500 m (Handley, 1953). Glaucomys volans ranges north to approximately the United States/ Canadian border in eastern North America, and south to the Gulf coast. Glaucomys sabrinus, on the other hand, ranges throughout forested Canada and south along both the Rocky Mountain and Appalachian Mountain chains at increasingly higher altitudes. G. sabrinus is primarily a squirrel of northern coniferous/ hard- wood forests and taiga, while G. volans is more characteristic of deciduous temperate forests. Both overlap in distribution in the northern Great Lakes area and the central Appalachians, and may be found in the same woodlots in some mountain areas. G. volans occurs commonly in the Clark’s Cave area today. The presence of G. sabrinus in such large numbers at the site is certainly indicative of former boreal conditions. The presence of G. volans, however, does not necessarily indicate temperate conditions, as the fossil population is separable by size from its modern equivalent. This presumably has physiological implications. Table 15. Measurements (in mm) of alveolar length lower toothrow (p4— m3) of Glaucomys. Age and locality X OR SD CV N Glaucomys sabrinus (Shaw)l >2 Late Pleistocene: New Paris No. 4, Pa. 7.7 7.6-8. 1 — — 3 Natural Chimneys, Va. 7.8 7. 3-8.4 — — 14 Clark’s Cave, Va. 8.0 7.6-8.6 .01 1.19 30 Robinson Cave, Tenn. 7.8 00 I o — — 8 Recent: Eastern United States; G. s. coloratus 6.9 6.7-7. 1 — 7 G. s. fiiscus 6.7 6.S-6.9 — — 5 G. s. macrotus 6.6 6.0-7.0 — 23 Canada: G. s. sabrinus 7.2 6. 8-7.6 — 4 G. s. makkovikensis 7.1 6.9-7.4 2 Alaska: G. s. zaphaeus 7.6 1.3-1.9 — — 5 Idaho: G. s. bullatus (=bangsi, Mayer, 1941) 8.7 00 b^ 1 00 — — 7 Glaucomys volans (Linnaeus)^ Late Pleistocene: New Paris No. 4 6.4 — 2 Natural Chimneys, Va. 6.5 6.3-6.9 — 17 Clark’s Cave, Va. 6.7 6.3-7. 3 .18 2.67 17 Robinson Cave, Tenn., Armadillo Pit 6.3 6. 2-6.6 — — 5 Robinson Cave, Tenn., Sloth Pit 6.6 6.4-6.8 — 4 Recent: Pennsylvania^ 6.0 5.6-6.4 — 38 1 Measurements from Howell, 1918; Handley, 1953; Guilday et a/., 1964. ^Alveolar length of upper toothrow, when found in the hterature, was converted to alveolar length of lower toothrow by multiplying by 0.94, a constant found to work out with modern Pennsylvania specimens. ^Data from Guilday, 1962; Guilday et al., 1964, 1969. ^Unpublished measurements, J.K. Doutt, CMNH Recent mammal collection. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 59 Family: Cricetidae — Deer Mice and Woodrats Genus Peromyscus Gloger material: cm 24682-28685, 29571. 221 left, 163 right man- dibles; skull fragments and isolated teeth. MNI = 221 individuals. Peromyscus maniculatus (Wagner) — Deer Mouse material: cm 29569. 41 ml’s. An estimated minimum of 117 individuals based upon percentage of identified ml’s of P. man- iculatus and P. leucopus. Peromyscus leucopus (Rafinesque) — White-footed Mouse material: cm 29570. 37 ml’s. An estimated minimum of 104 individuals based upon percentage of identified ml’s of P. maniculatus and P. leucopus. remarks: Remains of both P. maniculatus and P. leucopus were present in the collection in about equal numbers. Identifications were based upon dental characters. The first lower molar of P. leu- copus is slightly larger (Table 16) than that of P. maniculatus with a greater incidence of accessory styles and lophs, a more massive bilaterally sym- metrical anteroconid, and a deeper anterior antero- conid reentrant. Out of a series of 91 Peromyscus ml’s selected at random, 14% could not be assigned to species either because of heavy toothwear or intermediate morphology. Of the remaining teeth. 41 (53%) were identified as P. maniculatus, 37 (47%) as P. leucopus. Three nights of trapping by Dr. Charles O. Hand- ley, Jr. and students of the University of Virginia in August, 1974, sampling talus, hilltop deciduous woods, woods-meadow ecotone, and hayfields surrounding the cave, produced 52 P. leucopus in 2000 trap nights. Peromyscus maniculatus apparently no longer occurs at or near the site. Both species are common woodland forms in the state today. P. leucopus occurs everywhere except in the higher mountain summits. P. maniculatus is confined to cooler mountain forests above an eleva- tion of 750 m (Handley & Patton, 1947). It is possible that the percentage of P. leucopus to P. maniculatus increased in the cave area as the en- vironment changed during the depositional history of the site, and that the 50/50 ratio of the identified material represents an average figure bridging a transition from predominantly P. maniculatus in late glacial times to 100% P. leucopus during the Holo- cene. Stratigraphic evidence from New Paris No. 4 suggests that this did occur in central Pennsylvania. In that fissure deposit, P. maniculatus formed 96% of the total Peromyscus recovered at the 8-10-m depth, but at the 0.25-m level was completely replaced by Table 16. Measurements (in mm) Peromyscus, Clark’s Cave. Late Pleistocene. X OR SD cv N Length ml — m3 Peromyscus, ?species CM 24682 3.4 3. 1-3.7 .14 4.20 48 Length ml P. cf. maniculatus CM 29569 1.53 1.43-1.68 .05 3.26 41 P. cf. leucopus CM 29570 1.59 1.47-1.72 .06 3.86 37 above combined 1.56 1.43-1.72 .06 4.40 78 Width ml P. cf. maniculatus CM 29569 .81 cc OS 1* — 41 P. cf. leucopus CM 29570 .85 .67-.96 — 35 Width ml X 100 Length ml P. cf. maniculatus CM 29569 52.5% 45.8%-61.2% — - 40 P. cf. leucopus CM 29570 53.5% 44.3%-62.5% — 36 60 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 P. leucopus, the only species presently living at the site. No evidence of golden mouse, Ochrotomys nut- talli (Harlan), or harvest mouse, Reithrodontomys humilis (Audubon and Bachman), small rodents with a dental morphology somewhat similar to Peromys- cus, was noted. Both occur sparingly in western Virginia today on the northern edges of their ranges (Handley & Patton, 1947; Wilder & Fisher, 1972). Neotoma floridana (Ord) — Woodrat material: cm 24536. 10 left, 12 right mandibles with at least ml in place; 25 left, 39 right isolated ml’s; 22 left, 34 right mandibles with no dentition; I skull; 3 palates; 21 left, 18 right maxillae, MNl = 51 individuals. remarks: The woodrat is present at the site today. Five were trapped by C. O. Handley, Jr. and party in August, 1974. Woodrat droppings (15,000 estimated by weight) were the commonest organic items in the deposit, other than bones and teeth. The woodrat is common in the southern and central Appalachians at all elevations, wherever suitable cliff/ cave/ talus habitat occurs. It does not range farther north than southern New York and western Connecticut (Hall & Kelson, 1959). Its failure to extend its range farther north in the Appalachians may be due to lack of suitable habitat rather than to climatic conditions per se. Other species of Neo- toma occur in the mountains of western North America north to southern Alaska. Despite its mod- ern temperate distribution, Neotoma was probably a member of the boreal late Pleistocene fauna at the site. Neotoma remains in association with boreal small mammals are known from New Paris sink- hole No. 4, Pa. N. f. magister, the northeastern race, probably occupied most of its present range, south of the glacial terminal moraine, throughout the Wisconsinan glaciation at least. Measurements of Clark’s Cave woodrat dentitions do not differ significantly from those of modern Pennsylvania comparative material (Table 17). In Neotoma, occlusal area increases with tooth wear. Therefore comparisons can only be made between specimens of comparable age. The denti- tions were ranked in three relative age classes: (1) light — wear pattern established on occlusal surfaces, but not yet fully developed; (2) medium — wear pattern fully developed, but reentrants not isolated by cingular wear; (3) heavy — at least some reentrants isolated by cingular wear. Occlusal length of molars in the first wear class is significantly shorter. These were calculated separately. In a sample of 68 ml’s from Clark’s Cave, 51% showed light wear, 44% showed medium wear, and 5% heavy wear. Family: Arvicolidae — Voles (see Kretzoi, 1962) Clethrionomys gapperi (Vigors) — Red-backed Vole material: cm 24519: 41 left, 27 right mandibles with ml; 3 left, 2 right ml’s; 8 left, 7 right mandibles, no dentition; isolated upper molars. CM 24559: right ml (Fig. 19K). CM 24560: right M3 (Fig. 19C). CM 24570: 84 left, 107 right mandibles with ml; 141 left, 168 right ml’s; 154 left, 162 right Ml’s; 786 other molars. MNl = 305 individuals. Table 17. Dental measurements (in mm) Neotoma floridana (Ord). Locality Toothwear X OR N Occlusal length ml — m3 Clark’s Cave, Va. Medium & heavy 9.8 9.3-10.2 5 Pennsylvania, Recent Medium & heavy 9.7 9.5-10.2 15 Occlusal length ml Clark’s Cave, Va. light 3.4 3.1-3.9 34 Clark’s Cave, Va. Medium & heavy 3.9 3.6-4.3 32 Pennsylvania, Recent Medium & heavy 3.8 3.6-4. 1 17 Occlusal length m2 Clark’s Cave, Va. Medium & heavy 3.2 3.0-3.4 14 Pennsylvania, Recent Medium & heavy 3.3 3. 1-3.6 17 Occlusal length m3 Clark’s Cave, Va. Medium & heavy 2.3 2. 1-2.4 11 Pennsylvania, Recent Medium & heavy 2.3 2.1-2.5 15 Specimens examined; Recent, Pa.; CM Mammal No’s. 29170, 73, 74, 76, 77, 78, 79, 88; 32602, 04; 33545 ; 35078, 79, 82, 86, 89, 90; 38638, 41; 39337. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 61 Fig. 19. Occlusal outlines of vole (Arvicolidae) molars, Clark’s Cave local fauna, Bath County, Virginia. Cementum is indicated only for Synaptomys. upper row = right M3’s: A. Synaptomys cooperi, CM 24566. B. Synaptomys borealis, CM 24564. C. Clelhrionomys gapperi, CM 24560. D. Phenacomys intermedius, CM 24562. E. Microlus pinetorum, CM 24558. F. Microtus xanthog- naihus, CM 24553. G. Microlus pennsylvanicus, CM 24554. H. Microtus chrotorrhinus, CM 24555. LOWER ROW = right ml’s. I. Synaptomys cooperi, CM 24565. J. Synaptomys borealis, CM 24563. K. Clelhrionomys gapperi, CM 24559. L. Phenacomys intermedius, CM 24561. M. Microtus pinetorum, CM 24557. N. Microtus xanthognathus, CM 24552. O. Microtus pennsylvanicus or chrotorrhinus, CM 24556. remarks: The red-backed vole is one of the commonest small mammals in the mountains of western Virginia. It is a forest vole of boreal affinity characteristic of northern coniferous forests of the Hudsonian and Canadian Life Zones and ranges south along the mountain summits of the Appala- chians to northern Georgia. It does not occur at the site today despite the inviting appearance of the cool, shaded, cliff-based habitat. Extensive trapping in 1974 in and around the site failed to produce it. While it is common at higher elevations in the moun- tains, it is no longer present at the lower altitude of the Cowpasture River valley. Its presence in the deposit reflects former cooler woodland conditions near the site. Red-backed voles accounted for 15% of the eight species of voles represented at Clark’s Cave, 17% at Natural Chimneys, Va., and 21% at New Paris No. 4, Pa. No direct assessment of forest cover can be ascertained from such figures. At Sinkhole No. 4, Pa., within a vertical span of 4 m of continuous fis- sure fill, numbers of Clelhrionomys increased from 4% at the lower level to 50% at the higher level. Correlated with this was a relative decline of field forms reflecting a period of floral change from open parkland to a closing coniferous forest. The vole assemblage, with the exception of Dicrostonyx at New Paris No. 4, is the same at all three sites — all dominated by M. xanthognathus. Unfortunately, the time span at Clark’s Cave is not known, nor is any stratification apparent in the deposit itself, so that any internal changes in relative numbers of individual species could not be determined. Clelhrionomys remains were almost twice as common in the deposit as those of the woodland vole, Microtus pinetorum (MNI 305:170), the forest- inhabiting vole now present at the altitude of the site. This suggests that the boreal phase of the history of the site was the period of greatest depositional activity or duration. The sub-surface burrowing habits of M. pinetorum do not necessarily protect it differentially from predation. At a similar late Pleistocene owl roost. Natural Chimneys, Va., M. pinetorum was the more common of the two species 62 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 (MNI 47:65). M. pineiorum replaced the red-backed vole at lower altitudes in the central Appalachians as temperate conditions returned following the Wisconsinan glaciation. At New Paris, Pa., sink- holes No. 2 and No. 4 are located only a few meters apart. The Clethrionomys! M. piuetorum ratio changed from MNI 260:12 in the late Pleistocene boreal Sinkhole No. 4 fauna to MNI 0:51 in the Re- cent Sinkhole No. 2 assemblage. In summary, Clethrionomys was a common species in the deposit, exceeded in MNI only by M. xanthognathus, M. pennsylvanicus, and M. chrotorr- hinus, among the voles, and was almost twice as common as remains of M. piuetorum, the woodland vole now present at the site. This suggests that the boreal phase was the period of greatest depositional activity. Phenacomys intermedius Merriam — Heather Vole materi.\l: cm 24515: 4 left, 5 right mandibles with ml; 1 left ml; 3 left, 1 right mandible, no dentition; 7 additional molars. CM 24561: right ml (Fig. I9L). CM 24562: right M3 (Fig. 19D). CM 24571: 3 left mandibles with ml; 3 right mandibles with ml; 3 right mandibles, no dentition; 26 left, 12 right ml; 33 Ml; 13 M2; II M3; 57 additional molars. MNI = 34. remarks: This boreal rodent is no longer found in the eastern United States. It inhabits northern coniferous forest from the southern Yukon east to the Labrador coast. Unlike the northern bog lem- ming, Synaptomys borealis, which has a similar modern geographic distribution in eastern North America, Phenacomys does not occur south of the St. Lawrence estuary in the mountains of New Brunswick and northern New England. Like the red-backed vole, Clethrionomys, the heather vole is not a grazer. Both genera, in con- trast with all other American voles, have rooted molars, ill-suited for a diet of grasses. Heather vole remains at the site are, therefore, indicative of boreal forest of some type. Banfield (1974:193) summarizes its present habitat preferences, “Most have been taken in open, dry, coniferous forests of pine or spruce with an understory of heaths . . . usually near water . . . shrubby vegetation on the borders of forests and in moist, mossy meadows.” It feeds on bark, buds, seeds and foliage of shrubs and forest understory. Heather vole remains have been recovered from two other late Pleistocene cave sites in the state. Natural Chimneys and Back Creek No. 2. They were formerly widely distributed in late Pleistocene boreal faunas from periglacial eastern North America (Guilday & Parmalee, 1972). Rarely taken by modern collectors, the heather vole is also rare in fossil collections: 3.6% of all voles at Natural Chimneys; 1.6% at Clark’s Cave, and 3.8% at New Paris No. 4, Pa. It was the rarest vole in the Clark’s Cave deposit. Both boreal woodland forms, the relative abun- dance of Phenacomys and Clethrionomys remains in late Pleistocene sites from the central Appala- chians are positively correlated: Clark’s Cave, Va., 1.6% and 15%; Natural Chimneys, Va., 3.6% and 17%; New Paris No. 4, Pa., 3.8% and 21% of all voles respectively. Microtus pennsylvanicus (Ord) — Meadow Vole m.\terial: cm 24521: 21 partial skulls plus maxillae and isolated molars for a total of 27 left Ml, 31 right Ml; 30 left M2; 26 right M2; II left M3, 7 right M3. CM 24554: right M3 (Fig. I9G). CM 24573: 138 left Ml, 130 right Ml; 262 left M2, 247 right M2; 294 left M3, 308 right M3. MNI = 316 individuals. Adjusted MNI = 658 individuals (69.3% of 950 Microtus, sp. ml’s. See last paragraph of M. chrotonhimis discussion for explanation of adjusted MNI). REMARKS: The meadow vole is one of Virginia’s commonest small mammals, occurring at all eleva- tions in all meadow and grassy bog situations. The commonest species in the deposit, 24% of all terres- trial mammals, it is also the most widely distributed small vole in North America today. A closely related form, M. agrestris, the field vole, occurs widely in Europe and northern Asia as well. The two popula- tions are now widely separated geographically but are obviously derived from a common stock (Klim- kiewicz, 1970). Although the meadow vole occurred as far south as Florida during late Wisconsinan times (Devil’s Den, Martin, 1974b) it does not occur in the southeastern lowlands today. It ranges from South Carolina and the highlands of Georgia to northern Labrador, including all of Canada south of the tundra, west to the coast of Alaska. Modern barn owl pellets from the Clark’s Cave cliflfs yielded meadow vole remains, and they were trapped in the Cowpasture River valley by C. O. Handley, Jr., and party in 1974. The abundance of M. pennsylvanicus remains in the deposit reflects not only the presence of suitable meadowlands in the river valley and the field- hunting predilections of the owls, but the fact that this was the only rodent in the deposit that was probably taken in undiminished numbers throughout its depositional history. It would have been the species least affected by the climatic change from the boreal late Pleistocene period to modern temperate in the central Appalachians. As long as moist grassy 1977 GUILDAY, PARMALEE, AND HAMIETON; CLARK’S CAVE BONE DEPOSIT 63 areas were present, so was this vole. Microtus penmylvaniciis was also the commonest species of terrestrial mammal recovered from both Natural Chimneys, Va., and New Paris No. 4, Pa. At only one site analyzed to date was it relatively uncommon: the Recent New Paris No. 2, Pa. (3 M. pennsylvanicus, 51 M. pinetorum). This is because animals trapped in New Paris No. 2 were only those whose immediate home ranges included the sinkhole entrance into which they tumbled. Had the deposit been an old owl roost accumulation like that at Clark’s Cave, Microtus pennsylvanicus would have been present in large numbers concentrated by meadow-foraging owls. The dry woodland surround- ing the mouth of New Paris No. 2 does not support M. pennsylvanicus even though it is regionally common. Dentitions of this species from central Appalachian fossil deposits, as well as those of M. xanthognathus and M. chrotorrhinus, agree in having a lower first molar exhibiting a pattern of five closed triangles between the posterior crescent and the anterior tre- foil (ml of M. pinetorum has three closed triangles, the fourth and fifth being broadly confluent). This advanced condition and their abundance in such sites makes them convenient horizon-markers. Microtus molars with five triangles are not present in earlier Pleistocene deposits, e.g., those at Cumberland Cave, Md. (late Kansan), and Trout Cave, W. Va. (Kansan/ Yarmouthian). Instead, ml’s of a simpler evolu- tionary grade, possessing three to four triangles, are present (van der Meulen, unpublished). Microtus chrotorrhinus (Miller) — Rock Vole material: cm 24520: left maxilla with MI-M3; partial skull with left M1-M2, right M2-M3; 3 M3. CM 24555: right M3 (Fig. 19H). CM 24574: 3 left, 3 right Ml; 3 left, 6 right M2; 102 left, 135 right M3. MNI = 140 individuals. Adjusted MNI = 292 individuals (30.7% of 950 Microtus, sp. ml, see last paragraph, this discussion. remarks: The rock vole has never been trapped in Virginia but may possibly occur on some of the higher mountain summits. This animal has one of the most aberrant habitats of any North American Microtus. During the boreal episode of late Wis- consinan times the rock vole enjoyed a much wider geographical distribution in eastern North America. But following the retreat of the Wisconsinan ice sheet, its range in the central and southern Appala- chians has been reduced to small disjunct populations in rocky areas of cool mountain forests above 900 m. Its continued presence in the southern and central Appalachians is due solely to the persistence of such ecological enclaves, but it is absent in most of them now. Post-Wisconsinan range adjustment of M. chrotorrhinus consisted of local extinction and retreat to higher altitudes in the southern portions of its range, coupled with colonization of suitable northern areas as coniferous forest advanced into recently ice-freed areas. There were apparently few or no adaptive changes on the part of the animal itself. R. L. Martin (1973), in a detailed study of 464 Recent specimens from over 55 localities from Labrador to North Carolina, found no indication of clinal variation in the dentition. He did find a random pattern of variation from site to site caused by re- stricted gene flow between widely scattered popula- tions marooned, as it were, on boreal “islands” in the eastern deciduous forests. The late Pleistocene sample of cranial material and dentitions from New Paris No. 4, Pa., is indistinguishable from M. chro- torrhinus in every character studied by Martin, but it averages somewhat larger in length of toothrow than modern central Appalachian material. Unfor- tunately, the M. chrotorrhinus sample from Clark’s Cave was too fragmentary for measurement of other than isolated teeth. The ratio of M. chrotorrhinus to M . pennsylvanicus varies from 4.8% at Natural Chimneys, Va., and 21.5% at New Paris No. 4, Pa., to a high of 30.7% at Clark’s Cave. These varying percentages probably reflect the regional extent of rocky habitat favored by this species. The talus fronting the high Clark’s Cave cliffs appears most favorable for this vole. Extensive trapping by C. O. Handley, Jr. and party in 1974, however, proves that M. chrotorrhinus no longer occurs at the site, despite the presence of cool, damp, forested, rocky tumbles. During a cooler episode, however, it would have made ideal habitat, providing a ready supply of rock voles for predators. The identification and assessment of minimum numbers of individuals of all voles from the deposit, with the exception of M. chrotorrhinus and M. pennsylvanicus, was based upon counts of ml’s. The largest and most readily identifiable tooth in vole dentitions, the ml has a greater percentage of re- covery in the field, because of its larger size, than do the other relatively smaller molars. Unfortunately the ml’s of M. chrotorrhinus and M. pennsylvanicus look alike, and minimum numbers of individuals of these two species had to be based upon the smaller but distinctive M3’s. The combined total of these two species, based upon the more easily lost M3’s was 456. However, a census of all Microtus ml’s that could be referred to either species came to a mini- 64 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 mum of 950 individual animals, indicating a con- siderable loss in recovery of the smaller M3’s. If this were the case, then minimum numbers of M. chrotorrhinus and M. pennsylvanicus could not be legitimately compared to minimum numbers of the other species of voles from the site. To remedy this, an adjusted minimum number of individuals for these two species was obtained by dividing the 950 MNl, based on ml’s, by the relative percentages of M. chrotorrhinus to M. pennsylvanicus as in- dicated by M3 counts. This was also done at New Paris No. 4, Pa. (M. pennsylvanicus - 361, or IS. 5% of Microtus sp.; M. chrotorrhinus = 99, or 21.5% of Microtus sp.), and Natural Chimneys, Va. (M. pennsylvanicus - 121, or 95.2% of Microtus sp.; M chrotorrhinus - 6, or 4.8% of Microtus sp.). Some isolated M3’s are of intermediate morphology and perhaps 5% were misidentified. But the chances of misidentification are reciprocal and this source of error probably does not influence the relative ratios of M. chrotorrhinus to M. pennsylvanicus based upon M3’s from the site. Microtus xanthognathus (Leach) — Yellow-cheeked Vole material: cm 24522: 156 left, 194 right mandibles with ml; 21 left ml, 14 right ml; 5 left m3, 16 right m3; 79 partial skulls, 7 maxillae with 42 left Ml, 42 right Ml, 18 left M2, 25 right M2, 8 left M3, 10 right M3. CM 24552: 1 right ml (Fig. 19N). CM 24553: 1 right M3 (Fig. 19F). CM 24572: 33 left, 42 right mandibles with ml; 288 left ml, 260 right ml, 110 left Ml, 97 right Ml, 31 left M2, 42 right M2, 182 left M3, 205 right M3; 4 partial skulls. MNl = 511 individuals. remarks: The largest species of North American Microtus with an average weight three times that of M. pennsylvanicus, the yellow-cheeked vole was the single most important food item in the deposit. It was exceeded in numbers of individuals only by the meadow vole. Both species outnumbered all other terrestrial small mammals. Despite its abundance, no other mammal in the deposit has adjusted its range so dramatically from that day to this. Remains of this vole have been recovered south of the late Wisconsinan ice moraines from late Pleistocene sites in Missouri, Iowa, Illinois, Kentucky, Pennsylvania, West Virginia, Virginia, and Tennessee. Its range has now shifted north almost 20° in latitude, some 1,900 km, to the taiga of western Canada and Alaska, where it is now rare, local, and seldom collected. The yellow-cheeked vole was also a dominant small mammal at New Paris No. 4, Pa. (344 MNl). As at Clark’s Cave, it was slightly exceeded in num- bers by M. pennsylvanicus (361 est. MNl). When large numbers of M. xanthognathus were encoun- tered at New Paris No. 4, Pa., there was some ques- tion about whether this species, so rare and local today, was present in such relatively large numbers regionally, or whether ecological conditions imme- diately surrounding the sinkhole mouth produced a local condition favorable to M. xanthognathus, creating a “hot spot’’ which did not reflect the re- gional picture. Its abundance at Clark’s Cave, how- ever, clarifies its regional significance. Instead of tumbling into a sink trap, the Clark’s Cave M. xan- thognathus were concentrated from a much larger area. They must have been abundant in the cruising range of the Clark’s Cave raptors, sustaining their numbers over a long period of time, to bulk so large in the deposit. Despite the different accumulation mechanisms at New Paris No. 4, Pa., and Clark’s Cave, Va., and the fact that they are 240 km apart, both sites are located just east of the Appalachian Plateau in the now relatively dry intermontane valleys of the Ridge and Valley section of the Appalachians. Dry shale barrens are characteristic of both areas today, and the general character of the two sites, allowing for the difference in latitude, may have been similar during late Pleistocene times. The only other Virginia localities for fossil M. xanthognathus. Natural Chimneys and Back Creek No. 2, are also in the mountainous western portion of the state. This is probably a reflection of differential distribu- tion of suitable fossil traps. Remains from Boot- legger Sink, Pa. (M. xanthognathus), and Little Kettle Creek, Ga. ( Clethrionomys, Synaptomys, Voorhies, 1974), indicate that boreal small mammals were widely distributed in the eastern Piedmont south and east of their present ranges. The former widespread occurrence of M. xanthognathus in the American Midlands (Hallberg, Semken, and Davis, 1974) and its relative abundance in such sites strongly suggests that this now-rare vole was both widespread and common in eastern North America south of the continental Wisconsinan glaciation — a convenient index fossil for deposits of this age. It appears, how- ever, not to have been present in western North America during this time. It has not been reported from numerous sites of comparable age in the central Rockies (Jaguar Cave, Ida.; Wilson Butte, Ida.; Little Box Elder Cave, Wyo.). Even today the animal is not found in mountainous terrain. It has not been reported from sites older than Wisconsinan. It was not present in the Appalachians in Kansan/Yar- 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 65 mouthian times (Cumberland Cave, Md.; Trout Cave, W. Va.) and the pre-Wisconsinan history of this species is unknown. Unfortunately, this vole is so seldom collected and occurs so sporadically throughout its range, that habitat information of the living animal is sketchy and general. It occurs, randomly and unpredictably, in a variety of boreal woodland situations. Its range is included in the Hudsonian Biotic Province of Dice (Dice, 1943), the taiga, avoiding both barren-ground habitat to the north and the closed coniferous forests of the Canadian Biotic Province of southern Canada. In the taiga, however, it has been found in both moist valley sites and upland well-drained woodlands. It burrows extensively, appears to prefer friable soils, although it may occur in wet sphagnum bogs. It does not occur in grasslands, like M. pemi- sylvanicus, or in rocky wooded talus as does M. chrotorrhinus. All accounts of present habitat (Pre- ble, 1908; Lensink, 1954; Banfield, 1974; Youngman, 1975) mention some form of woodland cover — a thin boreal forest dominated by spruce and jackpine with a ground covering of heaths and sphagnum would accommodate this vole. Although M. xanthognathus and M. chrotorrhinus are both boreal forest forms, with sympatric ranges in the mid-Appalachians during at least Wiscon- sinan times, their ranges are widely separated today. M. xanthognathus occurs in the western sub-Arctic, its known range separated from that of M. chro- torrhinus by approximately 800 km. M. chrotorr- hinus is found only in the eastern sub-Arctic and as a relict form south in the higher Appalachians. Why have these two species gone their separate ways (Fig. 20)? The following sketch agrees with the paleon- tological and geological facts, as far as they are known, and may explain the present distribution of these two species. Both species have been recovered from seven late Pleistocene sites (New Paris No. 4, Pa.; Bootlegger Sink, Pa.; Eagle Rock, W. Va.; Back Creek No. 2, Va.; Natural Chimneys, Va.; Clark’s Fig. 20. Modern ranges of (A) Yellow-cheeked vole, Microtus xanthognathus (Leach); (B) Rock vole, Microtus chrotorrhinus (Miller); superimposed upon Canadian Life Zone; from Hall & Kelson, 1959. Appalachian portion of Life Zone highly general- ized from true fragmented situation. 66 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Fig. 21. Successive stages of continental Wisconsinan glacial recession in central and eastern North America (from Bryson el ai, 1969), and periglacial ranges of yellow-cheeked vole, Microtus xanihogmthus (Leach), (from Hallberg et al., 1974) and rock vole, Microtus chrotorrhinus (Miller). A = ice cover, 7,000 years B.P. B = ice cover, 9,000 years B.P. C = ice cover, 12,000 - 13,000 years B.P. D = Wisconsinan periglacial range of Microtus xanthognathus (Leach). E = Wisconsinan periglacial range overlap, M. xanthognathus, M. chrotorrhinus. Cave, Va.; and Baker Bluff, Tenn.) but all these sites are in or near the Appalachian Mountains. Sites farther west (Welsh Cave, Ky.; Meyer Cave, 111.; Waubonsie local fauna, Iowa; Bat Cave, Mo.; Peccary Cave, Ark.) have produced remains of M. xanthognathus only (Fig. 21). As post-glacial warming set in and the open boreal forest of periglacial central and eastern North America gave way to denser tree cover and more temperate conditions, M. chrotorrhinus was able to survive in the Appalachians by retreating to higher elevations. Its range fragmented in the central and southern Appalachians as suitable rocky boreal habitat became increasingly restricted. Density of tree cover apparently had little effect as long as the animal could resort to its preferred habitat. M. xanthognathus, however, became extinct in the Appalachians, as post-glacial reforestation produced a closed-canopy forest. Continental meltback of the Wisconsinan ice front proceeded much more rapidly in the American Midlands than it did in the East (Fig. 21). Continental ice masses persisted in both Ungava and eastern Keewatin for thousands of years after an ice-free corridor in the Central Plains opened east of the Rocky Mountains north to Alaska. As this corridor opened to the Northwest, it was colonized by M. xanthognathus from the south. The animal’s range shifted to the north in toto, advancing on its northern front, becoming extinct in its southern fringes as reforestation eclipsed its preferred habitat in that area. Unable to spread northeast into the eastern Arctic at this time because of the persistent ice block, or to continue in relict habitats in the Appala- chians, as did M. chrotorrhinus, it could only follow the fortunes of the taiga as it retreated (or ad- vanced?) to higher latitudes in the Northwest. When the eastern ice block melted, sometime after 7,000 years B.P. (Bryson et al, 1969), eastern Keewatin, south of the barren ground, was duly colonized by M. xanthognathus. Access farther east to Ungava was blocked by Hudson Bay and by closed boreal 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 67 forests to the south. By the time Ungava was open to biotic invasion from the south, M. xanthognathus was extinct in the Appalachians. Hudson Bay has also acted as a barrier to the eastern post-glacial spread of the lemming Dicrostonyx torquatus, allow- ing the isolated D. hudsonius to survive in the tundra of Ungava (Guilday, 1963). Microtus pinetorum (LeConte) — Woodland Vole material; cm 24524: 3 left, 4 right mandibles with ml; 7 left, 5 right ml. CM 24557: 1 right ml (Fig. 19M). CM 24558: 1 right M3 (Fig. 19E). CM 24576: 24 left, 18 right mandibles with ml; 136 left, 1 14 right ml; 31 left, 33 right M3. MNl = 170 in- dividuals. remarks: The woodland vole, 8% of all voles in the deposit, was the least common of the four identi- fied species of Microtus (Table 4). Although only M. pennsylvanicus was trapped in the area by C. O. Handley, Jr. and party in 1974, and present in a few barn owl pellets from the cliffs, M. pinetorum is undoubtedly present in the Cowpasture River valley wherever conditions are suitable for this semi- burrower. It is present throughout western Virginia at medium-to-low altitudes, in loose, friable soils of open woodlands, orchards, and field borders. Identification to species is based upon geographic probability. It is conceivable that M. ochrogaster may be represented as well. The woodland vole does not occur in the northern coniferous forests, reaching its northern limits at approximately the U.S./ Canadian border. It is present in all late Pleistocene sites examined to date from the mid-Appalachians, and probably was present during the boreal episode of deposition. It is replaced at higher elevations in the Appala- chians today by the red-backed vole, Clethrionomys gapperi, another woodland form. At New Paris No. 4, Pa. (11,300 years B.P.), 260 Clethrionomys, but only 12 M. pinetorum (4.4% of their combined number) were recovered. At Clark’s Cave, ratio of M. pine- torum to C. gapperi is 36%. This large relative in- crease at Clark’s Cave may be due either to the presence of M. pinetorum in larger numbers during the boreal episode, or to deposition at Clark’s Cave continuing on into more temperate times. We suspect the former. The occurrence of macrofossils of decid- uous trees and shrubs (Quercus, Corylus) in a pre- dominantly boreal spruce (Picea) flora and the apparent association of Phenacomys, Clethrionomys, M. pennsylvanicus, and M. pinetorum in alluvial silts of the Brayton local fauna of southwestern Iowa (Dulian, MS), dated at 12,420+ 180 years B.P., a time near the final retreat of the Wisconsinan ice from Iowa, suggests that M. pinetorum may also have been at Clark’s Cave during the boreal episode of deposition. Ondatra zibethicus (Linnaeus) — Muskrat material: cm 24526. Partial palate with left ml-m2; 4 left, 3 right mandibles with ml; 2 right mandibles, no dentition. MNI = 5 individuals. remarks: This is an abundant semi-aquatic species present at the site today and throughout most of temperate and boreal North America south of the tundra. The low number of muskrats from this riverbank site can be accounted for by their relatively large size and aquatic habits, protecting them from owl predation. Synaptomys cooperi Baird — Southern Bog Lemming material: cm 24517: 1 right mandible with ml; I right mandible, no dentition; I left ml; I right m3. CM 24565: I right ml (Fig. 191). CM 24566: I right M3 (Fig. 19A). CM 24567: 2 right mandibles with ml; 22 left ml’s, 18 right ml’s. MNl = 23 individuals. REMARKS; The southern bog lemming is relatively widespread in meadows, dry fields, and occasional bogs in western Virginia, at all altitudes (Handley & Patton, 1947:169). It is the rarest vole in the area. In the Appalachian Mountain region its range extends from the Great Smoky Mountains in the south, north through New Brunswick to southern Labrador. In the northern portions of its range S. cooperi is nar- rowly sympatric with the southern fringe of the range of the northern bog lemming, V. borealis. Although S. cooperi is more temperate in distribution, the ecological diflferences between the two species in their area of sympatry are not clear. Both species were recovered from New Paris. No. 4, Pa., where a stratigraphic shift in relative numbers made it apparent that S. cooperi gradually replaced 5. borealis in the central Appalachians somewhere near 11,000 years B.P. as the climate ameliorated. S. cooperi accounted for 20% of all Synaptomys recovered at New Paris No. 4, Pa. At Clark’s Cave the proportion of S', cooperi was somewhat greater, 27% of all Synaptomys (N = 84). This may reflect either a difference in the relative ages of the sites, an ecological difference caused by differing physical parameters, or a reflection of a longer period of deposition at Clark’s Cave, extending into temperate times. The S. cooperi remains from Clark’s Cave agree in size with those from other late Pleistocene deposits in the central Appalachians (Table 18), smaller than the population now in the area. This probably 68 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 2 reflects a cold-stressed environment and it mirrors the present size dine of modern S. cooperi, i.e., an inverse size/ latitude correlation. This, together with a relatively low coefficient of variation, suggesting a common gene pool, indicates that the S. cooperi remains at Clark’s Cave were largely laid down prior to the advent of full-temperate conditions at the site. Synaptomys borealis (Richardson) — Northern Bog Lemming material: cm 24516; 7 left, 4 right mandibles with ml; 3 left, 5 right ml; 1 right m2; 1 left, 1 right m3. CM 24563: 1 right ml (Fig. I9J). CM 24564: I right M3 (Fig. 19B). CM 24568: 4 left, 8 right mandibles with ml; 40 left, 43 right ml; 1 left, I right Ml. MNl = 61 individuals. REMARKS: This forest lemming is no longer found in the central or southern Appalachians. Its present range includes the boreal forest and taiga from Alaska to Labrador, south to Minnesota and the White Mountains of New Hampshire, 1,000 km northeast of Clark’s Cave. Its habitat is variously described (Soper, 1942, 1948; Banfield, 1974) as ranging from moist to dry situations — grassy, second- growth spruce and poplar; thick gloomy spruce woods carpeted with sphagnum; grass-clumped willow swamp; small meadows; moist spruce woods; spruce bogs; alpine meadows. Uncommon to rare throughout its range today, the northern bog lemming is commonly found in late Pleistocene deposits from the central Appalachians (New Paris No. 4, Pa.; Bootlegger Sink, Pa.; Eagle Cave, W. Va.; Bowden Cave, W. Va.; Natural Chim- neys, Va.; Back Creek Cave No. 2, Va.; Guy Wilson Cave, Tenn.; Baker Bluff Cave, Tenn.; Robinson Cave Tenn.). Its remains are never present in large numbers, however. At New Paris No. 4 (MNI 71) it comprised only 5.8% of all voles recovered, at Clark’s Cave (MNI 61) only 3% of all voles. Occlusal length of 33 ml from Clark’s Cave, Va., averaged 2.88 mm, OR 2. 5-3. 2 mm, SD .04, CV 1.39. Eighty ml from New Paris No. 4, Pa., averaged 2.91 mm, OR 2.3-3. 2, SD .17, CV 5.84. Family: Zapodidae — Jumping Mice Zapus hudsonius (Zimmerman) — Meadow Jumping Mouse material: cm 29572: 29 mandibles, with at least ml; 12 additional ml; 12 m2; 28 maxillae. MNI = 22 individuals. remarks: The meadow jumping mouse is broadly distributed throughout northern North America, south of permafrost, from southern Alaska east to Labrador, south to Oklahoma, Alabama, and Georgia (Hall & Kelson, 1959:772). It prefers meadow and old-field situations of dense low vegetation, stream banks, and clearings in either open or forested country. Zapus is present at the site today; one was trapped in the hay fields above the cave in August, 1974, by C. O. Handley, Jr. It is present throughout the state wherever there is suitable habitat (Handley & Patton, 1947:182). Measurements of 36 lower first molars are: Total length, M = 1.5 mm; OR = 1.38 mm — 1.70 mm. Measurements of 22 M2’s are: Total length, M = 1.39 mm; OR = 1.26 — 1.57 mm. Napaeozapus insignis (Miller) — Woodland Jumping Mouse material: cm 29573: 15 mandibles with ml; 13 additional ml; 17 m2; 13 maxillae. MNI = 15 individuals. remarks: The woodland jumping mouse does not occur at Clark’s Cave today. It is closely confined to stream banks and woodland edge situations in the spruce/fir and hemlock/northern hardwood forests of eastern North America, from the Lake Superior region east to southern Labrador, thence south along the Appalachian Mountains at increasing altitudes to northern Georgia (Wrigley, 1972). In Virginia it is confined to mountain summit forests of cool, moist, rocky aspect in the western part of the state at higher elevations (Handley & Patton, 1947). Napaeozapus insignis exhibits a pronounced “Berg- mann’s Response.” Individuals from the extreme southern portions of its range average some 12% smaller than those of eastern Canada (Wrigley, 1972). Specimens from the Clark’s Cave deposit have greater dental dimensions than modern mid- Table 18. Measurements (in mm) Synaptomys cooperi Baird, occlusal length ml. Locality X Pennsylvania, Recent 1 2.48 New Paris No. 4, Pa., late Pleistocenel 2.41 Clark’s Cave, Va., late Pleistocene 2.40 Natural Chimneys, Va., late Pleistocene^ 2.39 lOata from Guilday, et al., 1964: 162. OR SD CV N 2. 1-2.7 0.19 7.66 25 2. 3-2.5 0.09 3.73 20 2.2-2. 7 0.11 4.75 33 2.2-2.S 0.05 2.92 14 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 69 Table 19. Dental measurements (in mm) Napaeozapus insignis (Miller). Age and locality X Total length ml OR N Recent: Pennsylvania 1 1.6 1.5-1.8 20 Late Pleistocene: New Paris No. 4, Pa.l 1.8 1.7-2. 1 11 Natural Chimneys, Va.l 1.7 1.6-1.7 6 Clark’s Cave, Va. 1.7 Total length m2 1.6-1.9 27 Late Pleistocene: Clark’s Cave, Va. 1.6 1.5-1.8 21 1 Measurements from Guilday et al., 1964: 168. Appalachian material by some 7%. They compare most closely with late Pleistocene specimens from New Paris No. 4, Pa., and Natural Chimneys, Va. All three sites agree in the boreal nature of their respective faunas. The Napaeozapus from all three sites represent a late Pleistocene population in the mid-Appalachians, larger than the population now present in the area. Family: Erethizontidae — Porcupines Erethizon dorsatum (Linnaeus) — Porcupine material: cm 24528: 1 left mandible with full dentition; 2 upper molars. MNI = 1 individual. remarks: There are no modern records for the porcupine in Virginia. It reaches its present southern limits as a breeding population in northern Penn- sylvania (Doutt et al, 1973). Porcupine remains have been recovered from Natural Chimneys, Va. It was apparently more widespread in karst areas immediately west of the Appalachians where it has been reported from Late Prehistoric archaeological sites as far south as Tennessee and Alabama (Par- malee & Guilday, 1966; Barkalow, 1961; Weigel, 1974). Primarily a denizen of coniferous, or northern hardwood, forests, the porcupine is also strongly attracted to rocky terrain and caves. Family: Canidae Cams, cf. C. dims Leidy — Dire Wolf material: cm 29611: 1 right unciform. MNI = 1 individual. remarks: Large mammal remains are rare in the deposit. The single carpal is 20% larger than a com- parable unciform of a large male timber wolf, C. lupus from Alaska (CM mammal no. DC 1247). It is referred with confidence to the dire wolf because of its size and the late Pleistocene age of the bulk of the collection. The dire wolf is the only species of wolf present in Appalachian deposits of Wisconsinan age. It was replaced by the timber wolf following its extinction at the close of the Pleistocene. Direct comparison was made with the Clark’s Cave unciform and the adjacent carpal, an os magnum, of the Powder Mill Creek Cave, Mo., C. dims skele- ton (Catalog No. P-249, Galbreath, 1964). They were of comparable size. Measurements of CM 29611 are: greatest depth, 22.2 mm; length metatarsal IV facet, 17.4 mm, approx.; width metatarsal IV facet, 17. 1+ mm. (C. lupus DC 1247 = 17.5 mm, 13.1 mm, 15.0 mm, respectively). Family; Ursidae Vrsus americanus Pallas — Black Bear material: cm 24592: 1 m2. CM 29696: 1 phalanx. MNI = I individual. remarks: Definitely not a raptor prey item, the isolated tooth and phalanx were probably deposited by hoarding woodrats. Measurements of the m2 are: length of crown, 15.5 mm; width of crown, 10.7 mm. This agrees well with measurements of 17 sixteenth-through-seventeenth-century archaeo- logical specimens from two late prehistoric sites in West Virginia (46 Pu 31 and 46 Fa 7); length 15.3 mm (13.1-17.7 mm), width 1 1.8 mm (10.7-13.3 mm). The black bear is still found in the mid-Appalachians. Family: Procyonidae Procyon lotor (Linnaeus) — Raccoon material: cm 24593: I right upper deciduous molar. MNI = 1 individual. 70 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 REMARKS: Adults are too large for the majority of raptors. The juvenile represented by this single molar may have been owl prey. Family: Mustelidae Martes americana (Turton) — Pine Marten material: cm 24954: distal half right humerus; distal half right tibia. MNl = I individual. remarks: There is no evidence that the pine marten was ever found in Virginia during historic times. It did occur as far south as Pennsylvania in northern coniferous/ hardwood forests (Rhoads, 1903). Paradiso, 1969, believes it to have been ex- terminated in the mountains of Maryland, “as far back as 85 years ago,” and it once may have occurred in the ridgetop spruce forests of West Virginia. It is a common late Pleistocene cave fossil from the mid-Appalachians and has been reported from Eagle Cave, W. Va., Benedict’s Cave, W. Va. (CM 24698), New Paris No. 4, Pa., Natural Chimneys, Va., and as far south as Robinson Cave in central Tennessee. Genus Mustela Linnaeus Mustela erminea Linnaeus — Ermine material: cm 24597: I left, 2 right maxillae with P4; I left, 2 right P4; 3 left, 2 right mandibles with ml; I right mandible, no dentition; left humerus, left femur, left tibia. MNl = 4 individuals. Mustela nivalis Linnaeus — Least Weasel material: cm 24596: 6 left, 2 right P4; partial skull with P4; 3 left, 2 right mandibles with ml; 1 left, I right ml; I humerus. MNl = 7 individuals. Mustela ?species, cf. M. frenata (female) or M. erminea (male) material: 29699: anterior half of skull with left P4-MI, right P2, right P4-M1. MNl = I individual. Mustela vison Peale & Palisot de Beauvois — Mink material: cm 24595: partial skull; left mandible fragment; right humerus; 2 right femora; right tibia. MNl = 2 individuals. remarks: The relative scarcity of carnivores in the deposit reflects the collecting bias of the birds of prey. Large species were either not represented at all, or by isolated teeth of fortuitous occurrence. A few small carnivores did find their way into the deposit as prey items, their numbers in inverse pro- portion to their size. The two weasels definitely present in the deposit, M. erminea and M. nivalis, are the two smallest North American carnivores. The least weasel is rare in the state today but is probably present throughout the mountain counties of the state (Handley & Patton, 1947). It is a circum- boreal species whose range extends along the Appa- lachian Mountains south to the Great Smokies. M. erminea has not been recorded, either living or fossil, as far south as Virginia. It, too, is a circum- boreal species, but reaches its present southern limits in northern Pennsylvania, where it is un- common. There is one anomalous record from the Piedmont of Maryland (Vazquez, 1956). Late Pleistocene remains of least weasel have been recovered from New Paris No. 4, Pa., Natural Chim- neys, Va., and Back Creek No. 2, Va., in the central Appalachians, and west of the Appalachians, from Meyer Cave, 111., Welsh Cave, Ky., and Robinson Cave, Tenn. M. erminea, however, has not previously been reported from the Pleistocene of eastern peri- glacial North America. It was present in the Conard Lissure, Ark., far south of its modern range (Brown, 1908). M. frenata, the long-tailed weasel, the commonest species in Virginia today, is not definitely repre- sented in the deposit. The possible partial skull listed above is not complete enough for identifica- tion. At least three M. frenata specimens were recovered from Natural Chimneys, Va., and the species was undoubtedly present in the Clark’s Cave area at some period during the accumulation of the deposit. Although the presence of M. erminea points to more boreal conditions in the area at one time, all weasels and mink occur in such a variety of ecological conditions throughout their range that they provide no definite paleoecological clues. cf. Mephitis mephitis (Schreber) — Striped Skunk material: cm 24598: 1 ulna; 1 left, I right lower canine. I basicranial fragment. remarks: The elements are too large for spotted skunk, Spilogale, and the identification is tentative because of the fragmentary nature of the specimens. The presence of the hooded skunk, M. macroura, or the hognosed skunk, Conepatus (Ladds Quarry, Ga.), in such a boreal deposit seems remote. Lamily: Cervidae — Deer cf. Cervus elaphus Erxleben — Elk material: cm 29697: posterior zygopophyses of an anterior lumbar vertebra. cf. Odocoilus virginianus Zimmerman — White-tailed Deer material: cm 29698: left naviculocuboid. 2 thoracic spines. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 71 1 partial third or fourth cervical vertebra. remarks: Identifications are tentative, based upon size comparisons with Recent material. These fragments may have been introduced by Recent woodrat activities, although signs of extensive gnawing were not noted. White-tailed deer are common in the area today. Both elk and deer remains have been reported from prehistoric Indian sites in Bath County approximately 1 1 km west of the cave (MacCord, 1973a, 1973b). Caribou (Rangifer tarandus) are known to have inhabited the Ridge and Valley section of the Appa- lachians as far south as Tennessee during the late Pleistocene (Guilday et ai, 1975) and may have been sympatric with both deer and elk during certain phases of glacial recession. No caribou remains were identified from the Clark’s Cave deposit. Table 20. Measurements (in mm) Mustela erminea and Mustela nivalis, Clark’s Cave. Species X OR N Species length, C - Ml M. erminea 12.1 12.1 2 length, P4 M. erminea 4.4 r-* 1 q 5 M. nivalis 3.36 2. 8-3.7 8 width, P4 M. erminea 2.12 2.0-2. 3 5 M. nivalis 1.75 q 1 8 width, Ml M. erminea 3.65 3.6-3.7 2 M. nivalis 2.65 2.6-2.7 2 length, ml M. erminea 4.2 3.9-4.6 4 M. nivalis 3.48 3.2-3.8 7 width, ml , talonid M. erminea 1.22 1.2-1. 3 4 M. nivalis 1.0 p 1 7 Measurements by Dr. Elaine Anderson. FAUNAL COMPARISONS— METHODS OF DEPOSITION The fossil faunas of two mid-Appalachian late Pleistocene small mammal deposits. New Paris No. 4, Pa., and Clark’s Cave, Va., represent a sufficient number of individual animals to make statistical comparison possible. Both deposits sampled the same regional fauna in the same physiographic area, the Ridge and Valley section of the mid-Appalachians, during approximately the same time interval. The method of deposition was not the same, however. At New Paris No. 4, Pa., mammals were trapped by falling down a narrow, vertical fissure 10 m in depth, which filled with accumulating surface debris. At Clark’s Cave, most mammal remains represent digestive remains, from raptorial birds, deposited on the talus floor in a sheltered cave entrance. Despite the different methods of deposition, both faunas were composed almost exclusively of small mammals up to the size of a hare (98% at New Paris No. 4, 99.5% at Clark’s Cave). In the former instance, size of entrapped mammals was governed by the small 1.5 m2 entrance, in the latter by the selective bias of the Clark’s Cave birds of prey. A casual size analysis of entrapped animals from the two faunas would give no indication of their differing modes of 72 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 deposition. The proportion of small-to-large mammals in a raptor deposit is due to the selection bias of the birds of prey, and is independent of the time involved in forming such a deposit. The proportion of small-to- large mammals in a tumble-in trap, like a sinkhole, will increase with time. The slower the rate of infill, the greater the proportion of small-to-large mam- mals, because of their overwhelming majority in the surrounding fauna. This majority increases their probability of entrapment. Bat remains were common at both deposits (43% MNI, New Paris No. 4; 36% MNI, Clark’s Cave, Va.) but for different reasons. The New Paris No. 4 re- mains represent natural mortality of a resident popu- lation accumulating as infill progressed. The Clark’s Cave bats were raptor prey. Bats are usually rare in owl pellet debris, and their presence in such numbers at the Clark’s Cave roost reflects the special case of an owl roost situated at the mouth of a large cave containing a flourishing bat population. The numbers of bats at New Paris No. 4 and Clark’s Cave do not reflect the difference in the method of deposition at the two sites. There are two clues, however: (1) The sinkhole bat material was exceptionally well-pre- served, yielding complete skulls and semi-articulated remains, contrasting with the fragmented collection from Clark’s Cave; (2) The New Paris No. 4 sample from a relatively small roost-shaft was dominated by only two species of Myotis, with a few Pipistrellus and Eptesicus. There were at least seven species of bats and five genera in the Clark’s Cave deposit, a reflection of the much larger bat roosting area afforded by Clark’s Cave and cliffs. There are minor differences in the composition of the mammalian faunas from the two sites. Those small mammals most suitable for owl prey (shrews, moles, small rodents) were somewhat more common at Clark’s Cave (see Table 21), but larger prey items (hares, rabbits) were relatively scarce. Lagomorphs were twice as common in the sinkhole deposit. These differences reflect the selection bias of the raptors. On the basis of these data alone the method of deposition is not apparent. If one looks at the diversity of the two faunas, more obvious differences appear. Although all major mammal taxa are represented in both faunas, 54 species of mammals were identified from Clark’s Cave, but only 40 from New Paris No. 4. The sink- hole sample was derived only from those animals blundering into the small surface opening, while the raptor-accumulated Clark’s Cave deposit represents a sampling from many habitats within the cruising range of the birds of prey. Three other differences in relative numbers are noteworthy — many birds, few snakes, and presence of fish in the raptor roost deposit, compared with the sinkhole trap. Birds are rarely trapped in vertical fissures, but are common in raptor debris. Two hundred and nineteen birds, of 68 species, 4.8% MNI of com- bined birds and mammals, were identified from Clark’s Cave. At Natural Chimneys, Va., a probable roost site, similar to Clark’s Cave, (79 birds, 40 species, or 9.33% of combined birds and mammals), birds were also a common element. In contrast, only nine birds representing seven species, or 0.3%, were present in the New Paris No. 4 sinkhole deposit, a ratio of one bird for every ten in the Clark’s Cave deposit. There was also a discrepancy in relative numbers of snake remains. Minimum numbers of individuals were not accurately ascertained but the number of snake vertebrae compared to the minimum number of mammals from the two sites was 39.1% at Clark’s Cave, but 70.9% at New Paris No. 4. There was also a difference in the size of the individual snakes involved. Those at Clark’s Cave were small and crotalids were not represented (with the possible ex- ception of one fragmentary vertebra). At New Paris No. 4, however, snakes varied in size from what must have been new-born to large individuals, both colubrids and crotalids. Snakes were a minor food item at Clark’s Cave. The raptors avoided large snakes. At New Paris No. 4, however, there was no such selection mechanism and, at least in the upper portions of the fissure deposit, snakes may have sought out the fissure for hibernation. Turtle and lizard remains in both faunas were either not represented or were negligible. In the case of the turtles at New Paris No. 4 this is probably a reflection of late Pleistocene boreal conditions at the time of infill, for box turtles ( Terrapene Carolina) were common in neighboring sinkholes on the same hillside holding Recent faunal remains. The absence of turtles in an owl roost deposit may be due to raptor selection. Frogs and toads were common to both sites (6.7% MNI Clark’s Cave, 4.9% MNI New Paris No. 4, relative to MNI mammals). To summarize: On the basis of the bone collec- tions alone, and with no knowledge of the geo- logical situation it is doubtful that one could state that New Paris No. 4 was a sinkhole trap deposit 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 73 while Clark’s Cave was a raptor roost. There are a few differences, but they are relative and mostly take on meaning through hindsight. As more of these faunas are studied, however, the differences noted above may take on greater significance. Table 21. Mammal composition of two mid-Appalachian cave deposits. MAMMALS Taxon Talpidae — moles Soricidae — shrews Chiroptera - bats Sciuridae — squirrels Cricetidae - deermouse, woodrat Arvicolidae — voles Zapodidae — jumping mice Erethizontidae - porcupine Leporidae - rabbits Carnivora — carnivores Artiodactyla - hoofed animals Totals — Mammal BIRDS % Total birds to total mammals New Paris No. 4, Pa. Sinkhole Trap Number Number Per cent of of of Species Individuals Individuals 2 3 .10 7 107 3.74 4 1235 43.22 6 83 2.90 3 238 8.33 10 1213 42.57 2 17 .59 1 3 .10 1 52 1.82 3 5 .17 1 1 .03 40 2857 7 9 Clark’s Cave Raptor Roost Number Number Per cent of of of Species Individuals Individuals 3 26 .50 7 231 5.30 8 1554 35.78 8 117 2.69 3 272 6.26 9 2060 47.43 2 37 .85 1 1 .02 2 24 .55 8 19 .44 2 2 .04 53 4343 68 210 4.83 ECOLOGICAL INTERPRETATION The interpretation of the fauna of the Clark’s Cave fossil deposit is clouded by three factors: (1) selective bias of the carnivorous birds responsible for the deposit (influenced by variations in hunting tech- niques, species of raptors, time of the year roost was occupied, size of prey, hunting radius); (2) great topographic diversity providing a variety of en- vironmental niches; (3) lack of definite knowledge of the time interval represented by the deposit within the rapidly changing framework of late glacial times. BIRD SUMMARY Eighty-five species of birds have been reported from three late Pleistocene cave deposits in Virginia: Clark’s Cave; Back Creek Cave No. 2, Bath County; and Natural Chimneys, Augusta County. Nine species were common to all three deposits: spruce grouse, Canachites canadensis', ruffed grouse, Bon- asa umbellus; sharp-tailed grouse, Pedioecetes phasianallus; bobwhite, Colinus virginianus; wood- cock, Philohela minor; passenger pigeon, Ecto- pistes migratorius; flicker, Colaptes auratus; blue jay, Cyanocitta cristata; and junco, Junco hyemalis. Seventeen additional species were present in at least two of the deposits: blue-winged teal. Anas discors; sharp-shinned hawk, Accipiter striatus; broad- wingedf?) hawk, Buteo platypterus; American kes- trel, Falco sparverius; rock(?) ptarmigan, Lagopus cf. mutus; turkey, Meleagris gallopavo; kingfisher, Megaceryle alcyon; red-headed woodpecker, Mel- anerpes erythrocephalus; hairy woodpecker, Den- drocopos villosus; downy woodpecker, Dendrocopos pubescens; cliff swallow, Petrochelidon pyrrhonota; gray jay, Perisoreus canadensis; red-breasted nut- hatch, Sitta canadensis; brown thrasher, Toxostoma rufum; robin, Turdus migratorius; red-winged black- bird, Agelaius phoeniceus; pine grosbeak, Pinicola enucleator. 74 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Undue emphasis should not be placed on the presence or absence of a given species of bird, because of predator bias and identification problems inherent in the relatively poor preservation of fragile bird bones. The 26 species listed above were common enough in western Virginia during late Pleistocene times to have appeared in more than one site. All three sites occur in the intermontane valleys of western Virginia and are roughly comparable in age. The avian fauna complements the paleoecologi- cal picture suggested by the mammals. The pres- ence of such relatively sedentary birds as spruce grouse, sharp-tailed grouse, and ptarmigan suggests cooler conditions and an open coniferous parkland — a picture complicated by the topographic diversity of the region, probably presenting a mosaic of open and closed situations. Only the three species of grouse and the passenger pigeon were present in substantial numbers, a situa- tion that may reflect the collecting bias of the raptors responsible for the deposit rather than their true relative abundance. Ptarmigan and passenger pigeon [although the latter formerly occurred seasonally as far north as James Bay, ca. 52° N. latitude (Todd, 1963] were not normally found in any numbers in the same area within historic times. Their co-pres- ence at Clark’s Cave and Back Creek No. 2, assuming they were contemporaneous in the area, which is most likely, was due to the ecological diversity im- posed by the mountainous terrain and the ability of birds, both predator and prey, to transgress the bounds of their “normal” habitats, especially if contiguous habitats like open, tundra-like ridge crests and more heavily wooded intermontane valley parklands or bog-forests were to parallel each other throughout the central Appalachians during glacial maxima. The sharp-tailed grouse and the magpie {Pica pica), the latter found only at Natural Chimneys, represent former eastern extensions of their present western ranges (see Fig. 14). Neither species is found in a closed-forest situation today. Their pres- ence suggests semi-prairie or parkland conditions and reinforces a similar conclusion from western elements in the mammal fauna: — 13-lined ground squirrel and least chipmunk. There have been un- dated, but presumably late Pleistocene, finds of badger, Taxidea taxus, (Bootlegger Sink, Pa., CM collection, unpublished) and grizzly bear, Ursus arctos, (Organ Cave, W. Va., CM 12999) that fit this picture as well. MAMMAL SUMMARY Within Recent times 68 species of mammals are known to have inhabited the mid-Appalachians (Table 22, columns 2 and 3). Fifty-four species were identified from the cave deposit (Table 22, columns 1 and 2). Those Recent mammals missing from the deposit were either too large or formidable (Bison bison, Felis concolor. Lynx rufus, Martes pennanti, Lutra canadensis, Canis lupus, Vulpes vulpes, Urocyon cinereoargenteus. Castor canadensis) to be taken by most raptors, or were of “southern” dis- tribution {ISorex longirostris, Cryptotis parva, Syl- vilagus floridanus, Sciurus niger, Reithrodontomys humulis, Ochrotomys nuttalli, Mustela frenata, Spilogale putorius). The absence of the hoary and silver-haired bats (Lasiurus cinereus, Lasionycteris notivagans) is not surprising considering their mi- gratory habits and rarity in and around caves. Three Recent rodents, black rat (Rattus rattus), Norway rat (Rattus norvegicus), house mouse (Mus mus- culus). Old World forms introduced during the Colonial period, were also absent. Considering its present abundance and its paleogeographic his- tory, the absence of opossum (Didelphis virginianus) remains from the site is probably significant. Al- though reported from the late Pleistocene of Devil’s Den, Fla., and Ladds Quarry, Ga., opossum re- mains have been conspicuously absent from all eastern North America deposits of that period from East Tennessee to Pennsylvania. Even as late as the 16th century, based upon analyses of aborig- inal cultural debris (Guilday, 1958), D. virginianus was rare or absent in the Northeast, and its spread north to southern Canada and New England was a relatively recent phenomenon. Sylvilagus floridanus, Sciurus niger, Reithrodontomys and Ochrotomys, based upon their present distributions, seem out of place in a late-glacial boreal setting, and were probably not in the area at the time the deposit was accumulating. Nine species of mammals found in the deposit do not inhabit the mid-Appalachians today. The dire wolf (Canis dims) is extinct. The remaining eight are all of boreal or mid-Western affinities (Sorex arcticus, Eutamias minimus, Spermophilus tridecemlineatus. 1977 GUILDAY, PARMALEE, AND HAMILTON; CLARK’S CAVE BONE DEPOSIT 75 Synaptomys borealis, Phenacomys intermedins, Microtus xanthognathus, Mustela erminea, Maries americana). If we omit mammals not reasonably considered raptor prey — those species larger than a woodchuck (3 kg±) — thus reducing the selective bias introduced by predator preference, 48 species of medium-to- small-sized mammals were identified from the fossil deposit. Eight of these (16.6%, see above) no loivger occur in the mid-Appalachians. An additional 1 1 (25%) survive at that latitude, at higher elevations in the mountains, and have not been found in the Cow- pasture valley (Condylura cristata, Parascalops breweri, Sorex cinereus, Sorex palustris, Peromyscus maniculatus, Clethrionomys gapperi, Microtus chrotorrhinus, Napaeozapus insignis, Sylvilagus transitionalis, Lepus americanus, and Glaucomys sabrinus). An additional five (10.4%), although still living near the site today, differ in size characteristics from their modern counterparts, agreeing with modern boreal samples of the same species or with those from the late Pleistocene New Paris No. 4 fossil collection (Blarina brevicauda, Tamias stri- atus, Tamiasciurus hudsonicus, Glaucomys volans, and Synaptomys cooperi). The mammalian fauna of the Clark’s Cave deposit is almost identical with that of New Paris No. 4, Pa. Differences can be ascribed to its more southerly location (absence of Dicrostonyx, higher numbers of Sorex fumeus, Microtus pinetorum, and Per- omyscus leucopus), possible accidents of deposition and sampling (presence of Eutamias minimus at Clark’s Cave, but not at New Paris No. 4), or predator bias (absence of Mylohyus and larger variety of bats at Clark’s Cave). The activity of raptors was also apparent in the greater number of birds from Clark’s Cave. Those boreal forms represented at Clark’s Cave but not at New Paris No. 4, — ptarmigan, spruce grouse, gray jay — may simply not have been trapped in the tumble-in sinkhole at New Paris No. 4. The bulk of the recovered fauna, as at New Paris No. 4, suggests that the predominate vegetative cover was an open boreal woodland dominated by conifers. Topographical diversity undoubtedly produced, as it does today, a variety of ecological niches. The only significant difference between the two faunas was the presence of the collared lemming, Dicrostonyx hudsonius, at New Paris No. 4, Pa., but not at Clark’s Cave. As the Clark’s Cave raptors concentrated on voles as a major prey item, the absence of the col- lared lemming at Clark’s strongly suggests the ab- sence of paucity of tundra in the Cowpasture valley or its neighboring mountain walls. Tundra conditions somewhere in the area are suggested by the presence of bones of the more vagile ptarmigan both at Clark’s Cave and at Back Creek Cave No. 2, 24 km to the west. The evidence to date suggests that while tundra may have extended down the Appalachians to at least the latitude of New Paris No. 4, Pa. (Di- crostonyx) and Buckle’s Bog, Md. (pollen profile. Maxwell & Davis, 1972), its presence as far south as west-central Virginia may have been discontinuous. In common with the mammals recovered from Clark’s Cave, no species of bird found solely to the south of Virginia today was recovered. All species occur in the state today or farther to the north and northwest (spruce grouse, sharp-tailed grouse, gray jay, pine grosbeak, and rock (?) ptarmigan). In summary, the recovered mammal fauna agrees in its ecological implications with the bird fauna in the absence of “southern” species and the influx of northern and western forms. A shortcoming of a fossil raptor roost like the Clark’s Cave deposit, is the virtual absence of large mammals because of the selection bias of the birds. There are large late Pleistocene mammals known from western Virginia in local faunas, unaccom- panied by smaller vertebrates, so that the fuller picture must be pieced together from several sites that may or may not have been exactly contempo- raneous. Ground sloth, Megalonyx; mammoth, Mammuthus primigenius; mastodon, Mammut amer- icanum; horse, Equus; caribou, Rangifer tarandus; extinct moose, Cervalces; bison. Bison; extinct muskoxen, Symbos and Bootherium; have been recovered from Saltville, Smythe County, Va., southwest of Clark’s Cave, associated with a date of 13,460 radiocarbon years. The pollen associated with the muskox remains (Ray et ai, 1967) indicated a flora dominated by spruce and pine, and was inter- preted as spruce parkland interspersed with ponds, marshes, and prairies, an environment similar to that suggested by the Clark’s Cave local fauna. These large herbivores were attracted to and mired in saline springs. Molars of the giant beaver {Castoroides ohioensis) have been recovered from Natural Chim- neys, Va. Long-nosed peccary Mylohyus nasutus is known from Natural Chimneys and Back Creek No. 1, Va. There is archaeological evidence from Russell Cave, Ala., that Mylohyus nasutus may have sur- vived as late as 7,000 years ago. No direct evidence links any of these large animals with the Clark’s Cave local fauna, but they probably coexisted with it. 76 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 22. Species of mammals, Clark’s Cave, and Recent mid-Appalachians. ( Present in Clark’s Cave deposit only Sorex arcticus (13) Eutamias minimus (3) Spennophilits tridecemlineatus (5) Synaptomys borealis (6 1 ) Phenacomys intermedins (34) Microtus xanthognathus (511) Present in Clark’s Cave deposit and Recent mid-Appalachians Family Didelphiidae Family Talpidae Condylura cristate (13) Parascalops breweri (12) Scalopus aquaticus ( 1 ) Family Soricidae Sorex cinereus (67) Sorex dispar (4) Sorex fumeiis (10) Sorex paliistris (7) Micro sorex hoyi (7) Blarina brevicauda (97) Family Vespertilionidae Myotis liicifiigiis/sodalis Myotis keenii Myotis leibii (c. 138) Myotis grisescens (c. 138) Eptesicus fuscus (363) Pipistrellus siibflavus (26) Plecotus townsendii (9) Lasiuriis borealis (3) Family Leporidae Lepiis americamts (23) Sylvilagus transitionalis (1) Family Sciuridae Tamias striatus (24) Marmot a monax (2) Tamiasciurus hudsonicus (25) Sciurus carolinensis (3) Glaiicomys volans (19) Glaitcomys sabrinus (28) Family Castoridae Family Cricetidae Peromyscus maniculatus (c. 117) Peromyscus leucopus (c. 104) Neotoma floridana (51) Family Arvicolidae Synaptomys cooperi (23) Clethrionomys gapperi (305) Microtus pennsylvanicus (c. 658) Microtus chrotorrhinus (c. 292) Microtus pinetorum (170) Ondatra zibethicus (6) ) = MNI, fossil deposit. Present in Recent mid-Appalachians only Didelphis virginianus Sorex longirostris Cryptotis parva Lasionycteris noctivagans Lasiurus cinereus Sylvilagus floridanus Sciurus niger Castor canadensis Reithrodontomys humulis Ochrotomys nuttalli 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 77 Table 22. Species of mammals, Clark’s Cave, and Recent mid-Appalachians. ( ) = MNI, fossil deposit (continued). Present in Clark’s Cave Present in Clark’s Cave Present in Recent deposit only deposit and Recent mid-Appalachians mid-Appalachians only Family Muridae Rattus rattus Rattus norvegicus Mus musculus Family Zapodidae Zapus hudsonius (22) Napaeozapus insignis (15) Family Erethizontidae Erethizon dorsatum ( 1 ) Family Canidae Canis cf. dims (1) Family Ursidae Ursus americanus ( 1 ) Family Procyonidae Procyon lotor ( 1 ) Family MusteUdae Manes americana (1) Mustek nivalis (7) Mustek erminea (4) Mustek vison (2) Mephitis mephitis ( 1 ) Canis lupus Vulpes vulpes Urocyon cinereoargenteus Martes pennanti Mustek frenata Spilogale putorius Lutra canadensis Family Felidae Felis concolor Lynx rufus Family Cervidae Cervus ekphus ( 1 ) Odocoileus virginianus ( 1 ) Family Bovidae Bison bison AGE OF DEPOSIT It is clear, from the presence of so many boreal birds and mammals, that the deposit accumulated during a cooler climatic episode. The mammalian fauna, with minor exceptions, is identical with that of New Paris No. 4, Pa. (11,000 years B.P.), even to infraspecific size (Bergmann’s Response). This, plus the relatively superficial nature of the deposit scattered throughout the upper 45 cm of an uncon- solidated cliff talus, indicates a relatively late date, but one predating the Recent fauna and flora. Deposition could not have extended very far, if at all, into Recent times on a sustained basis. In addition to the negative evidence of the absence of introduced or domestic species of birds and mammals, the 78 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 size characteristics of the fossil population samples of species like Sorex cinereus, Blarina brevicauda, Condylura cristata, Tamias striatus, Tamiasciurus hudsonicus, Glaucomys volans, Glaucornys sabrinus, Lepus americanus, Synaptomys cooperi, and Napaeo- zapus insignis are distinct from their Recent mid- Appalachian representatives, paralleling those of New Paris No. 4, Pa. If deposition at the site had continued into Recent times there should have been a size-continuum between the late Pleistocene and Recent representatives within each species, ex- pressed as an increase in variation, larger observed ranges, and higher coelbcients of variation. Such was not the case. Deposition must have halted, for all practical purposes, before these mammals showed any degree of measurable physiological adjustment to the changing environment. Judging from the rate of accumulation of modern “owl-roosts,” the Clark’s Cave deposit was built up in a relatively short time, probably between 20,000 and 11,000 years ago — a concentrated accumulation of an active population of both adult and fledgling raptors. Deposition may have been terminated by rockfalls, but this is not apparent at the site today, and the deposit was not buried under rock debris. The abandonment of the roost as a nesting site may have been hastened by Indian molestation in this easily accessible river-bank nesting site. Wayne C. Clark, Assistant Archaeologist, Archaeological Society of Virginia (letter, Oct. 1, 1975) reports 32 Prehistoric Indian sites from Bath County alone, dating primarily from 8,000 B.C. to 1,600 A.D., with the majority of the large sites dating to the Woodland period, ca. 1,000 B.C. to 1,600 A.D., so that the Clark’s Cave raptors could have been subject to human predation for at least the last 10,000 years. This is speculation, however. Any evidence of a prehistoric occupation in any of the larger entrances of Clark’s Cave would have been destroyed by the Colonial and Civil War saltpeter miners. If we assume that deposition ceased at the site prior to Recent times, there are a few seeming anom- alies— the presence of the woodland vole (Micro tus pinetorum) and the white-footed mouse (Pero- myscus leucopus) in greater numbers than one would expect in a boreal environment suggests either that they were able to adjust to the peri- glacial conditions at this latitude or, perhaps, were pioneer colonizers in the ecological ferment of post- glacial warming. Both species are inhabitants of dry, deciduous, hardwood forests, but occur locally in northern New England today in cool hemlock/ maple woods of Canadian aspect. The presence of a single individual of the eastern mole (Scalopus aquaticus) appears anomalous, based on its present geographic distribution, as well as of birds like the bobwhite, Colinus virginianus (two individuals), and one individual each of orchard oriole Icterus spurius, tanager Piranga ?species, and a question- able red-bellied woodpecker Centurus carolinus. Owls still roost on the Clark’s cliff's, however, and a scattering of Recent remains may have been intro- duced into the deposit. ACM date of 2,260+ 85 years B.P., based on bone carbonate, was obtained from a sample of 215 gms of unsorted bone fragments (1-7224). This date is obviously too recent. The superficial nature of the deposit increased the chances of contamination and the date is unacceptable as far as dating the Clark’s Cave deposit is concerned. THE LATE GLACIAL ENVIRONMENT Late glacial climatic changes and their biological consequences are still known only in broadest out- line. Data have accumulated from geological and biological sources, both marine and terrestrial — frost and ice phenomena, plant fossils including pollen, invertebrate and vertebrate fossils — that point to effects of glacial cooling extending to lower latitudes hundreds of kilometers south of the terminal Wisconsinan moraine. The extent and direction of environmental change in the Appalachians is governed by the regional topography — long, parallel mountain ridges inter- spersed with intermontane valleys extending south- wards in essentially unbroken array from the former glacial front to the Carolinas and Georgia, culminat- ing in the Great Smoky Mountains. The effect of such topography is seen today in the distribution of forest types. “. . . The long southward projection of the northern hardwood forests on the Appalachian summits today probably had a full glacial predecessor in the form of tundra and boreal forest. Such a pro- jection could have been a main avenue for the invasion of boreal elements into the boreal forests 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 79 of the time.” (Flint, 1971:509) Evidence of glacial cooling reflected in the dis- tribution of boreal species of plants during full glacial times (Watts, 1970) has been recorded from as far south as Georgia and northern Florida. We will concentrate on climatic changes in the central Appalachian West Virginia/ Virginia area and their relationship to Clark’s Cave (see V. A. Carbone’s excellent survey of the late-glacial environment in the neighboring Shenandoah Valley, Va., in Gardner, 1974:84-99). There is evidence from several sources for the former occurrence of Wisconsinan-age tundra, i.e., open, treeless grass and sedge expanses and perma- frost in the Appalachian highlands south of the terminal moraine. The most dramatic claim, former alpine glaciation at Boone Fork, Grandfather Moun- tain, N.C. at 36° O' N. (Berkland & Raymond, 1973), has been rejected. It was based upon parallel grooves, worn into rock outcrops by metal logging cables, that were misinterpreted as glacial striations (McKeon, 1974; Hack & Newell, 1974; Berkland & Raymond, 1974). But the presence of block fields and of patterned-ground relicts on ridge crests from Pennsylvania to southern Virginia at altitudes ranging from 707 m in central Pennsylvania (Rian- sares) to 1,615 m (Whitetop Mountain) on the Ten- nessee/Virginia/North Carolina border (Clark, 1968, in Flint, 1971:283) suggests former permafrost activity. These sites are undated and may conceivably have predated the Wisconsinan glaciation. Direct evidence of Wisconsinan age tundra in the central Appalachians rest upon the evidence of plant and animal fossils. A pollen profile extending back into full glacial times is known from Buckle’s Bog, Md., at the headwaters of the Casselman River, at an elevation of 814 m, 170 km north of Clark’s Cave on the eastern rim of the Appalachian Plateau. Maxwell and Davis (1972), studied a 258 cm core from Buckle’s Bog and interpreted the lowermost 66 cm (zone BB-1), dated from 19,000 to 12,700 radiocarbon years, as true tundra. Nonarboreal pollen, over 50% sedges and grasses, predominated. Spruce (10-22%) and pine (5-17%) were the dominant arboreal pollen. The authors construed the relatively high percentage of spruce to pine as indicative of the nearby spruce, perhaps in sheltered valleys within 25 km of the site. Pine pollen, shed in greater abundance and more widely wind-distributed, should have been present in greater relative amounts if spruce were not locally present. The faunal evidence for the former presence of tundra, although not as extensive as that obtained from pollen analysis, is highly suggestive. Direct historical continuity between the Wisconsinan low- latitude periglacial tundra and the Recent eastern Canadian tundra is indicated by the skeletal remains of the Hudson’s Bay collared lemming, Dicrostonyx hudsonius, at 11,300 radiocarbon years, from New Paris No. 4, Pa., at the relatively low altitude of 465 m, just 16 km east of the rim of the Appalachian Plateau. The ptarmigan, Lagopus sp., reported in this paper from Clark’s Cave and nearby Back Creek No. 2 Cave, also suggests the near presence of tundra or tundra-like conditions. Because birds are more mobile than mammals and the possibility exists that the ptarmigan may be the willow rather than the rock ptarmigan, the evidence is not as firm as at New Paris No. 4. But at least, nearby open ground is indicated. Other boreal species, recorded from cave deposits as far south as 36° latitude in Tennessee (the caribou, Rangifer tarandus), 34° latitude Geor- gia (the spruce grouse, Canachites canadensis), and porcupine from the Coleman HA fauna, Fla., 29° latitude, indicate just how far south mammalian adjustments occurred. Changes in forest composition following the Wis- consinan glacial recession were first noted at about 13,600 radiocarbon years as a shift from coniferous to deciduous species in Georgia. But, on the basis of radiocarbon dating, approximately 1,000 more years were required for floral changes to be noted in the higher and more northern Appalachian plateau and ridge provinces as a change from tundra vegetation to boreal woodland (Maxwell & Davis, 1972). The situation at Clark’s Cave at this time was somewhat different. The broad expanse of the Appalachian Plateau flanking the northwestern approaches to the Appalachian ridges served as a wind and precipita- tion shield between them and the continental glaciers (Fig. 2 insert). Consequently the intermontane valleys to the east were protected from the full climatic rigors of the period. Pollen analyses at Hack and Queries ponds, on the floor of the Shenandoah valley, near Staunton, Va. (Craig, 1969), indicate that at the time tundra was noted at Buckle’s Bog and, by extension, from the higher ridges of the Appalachian Plateau and Mountains, the protected valley floors supported an open spruce and pine woodland. And as the tundra gave way to open spruce woodland in the higher Alleghenies between 12,700 and 10,000 radiocarbon years, the spruce woodlands of the protected intermontane valleys were changing to mixed closed-canopied forests 80 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 of spruce, pine and hardwoods. The late Pleistocene vegetation of the Clark’s Cave area during the deposition of the fossil deposit is hard to categorize because of the rugged terrain. Local topographical relief is some 500 m within a few kilo- meters of the site. Topography varies from flat flood plain to cliff and mountainous terrain. The picture derived from the fauna itself can only be a montage, but full glacial conditions in the Cow- pasture River valley can be visualized as a spruce parkland. The flood plain was probably dotted with copses of spruce intermingled with prairie, marshes, and small ponds. The shale hills along the eastern margin of the river valley that now support a xeric shale-barren flora were, in all probability, relatively dry then, perhaps supporting jackpme (Piuus hank- siana) parklands. The presence of marshes or small standing-water ponds is suggested by grebes, bit- terns, ducks, rails, plovers, and sandpipers, (26% of the identified species of birds from the deposit) and the number and variety of frogs. Coniferous to deciduous change in forest com- position proceeded with great rapidity. This was reflected in the changing make-up of the verte- brate fauna. Most of the mammals present during boreal times either migrated to higher latitudes, underwent physiological adjustment, or espe- cially in the case of the larger species, became extinct. Three carbon dates associated with mammalian faunas in Pennsylvania are relevant. At 11,300 radiocarbon years, the fauna of New Paris No. 4, Pa., was dominated by boreal small rodents and insectivores. By 8,570 radiocarbon years, the fauna from New Paris sinkhole No. 3 at New Paris, situated within a few meters of sinkhole No. 4, was a Recent one with no trace of the boreal species that so dominated the latter site. The mammalian fauna associated with a date of 9,240 radiocarbon years at Hosterman’s Pit, Pa., 136 km northeast of New Paris, was like- wise Recent in species composition. This suggests that a radiogenically undated deposit like that at Clark’s Cave can be assigned a rather definite age limit (minimum if boreal in make-up, maximum if composed of Recent species). Following the period of initial coniferous/decid- uous floral turnover, the region has been character- ized by a closed-canopy, oak-dominated deciduous forest. There have been minor climatic changes reflected in relative forest composition, successive submaxima of hemlock (Tsuga), chestnut (Castanea), and hickories (Carya) that probably required little, if any, adjustment on the part of at least the larger mammals. Changes of this magnitude may not be monitored by deposits like the Clark’s Cave fossil assemblage that are not as sensitive as pollen sites to minor climatic oscillations. The Clark’s Cave local fauna with its mixture of forest and meadow forms and its boreal, continental aspect suggests several distinct habitats within the hunting radius of the cave-based raptors. The cli- matic model of Saltzman and Vernekar (1975) suggests that the Full Glacial averaged macro- climate of the Holarctic ice border was cold and dry with reduced precipitation and evaporation. Chan- nel-scouring and downcutting in late glacial fluvial deposits of the South Fork Potomac River, northeast of Clark’s Cave, suggests increased stream velocity and run-oflf attributed to decreased evaporation (Gardner, 1974). This suggests that the floor of the Cowpasture River valley may have been relatively more mesic than the surrounding high-relief uplands because of a higher regional water table and the pooling of slope run-off caused by lower evaporation rates. The extensive flood plain (Fig. 10) probably developed boreal bogs and grass and sedge wetlands, while the relatively xeric uplands supported coniferous scattered woodlands, their density depending upon local hydrologic conditions. Such an environment would satisfy the mixed requirements of the varied fossil assemblage, allowing the raptors to draw from contiguous but contrasting habitats. The most disturbing aspect of the deposit is the lack of an absolute date within the Wisconsinan late glacial period. Only one fully comparable deposit has been studied in the area. New Paris No. 4, Pa. Perhaps this will remedy itself when more sites are developed and temporal faunal changes with refugial survivals are documented in the mid-Appalachians during the late Pleistocene. LIFE ZONE INTEGRITY The view that periglacial biotic adjustment to zone systems is no longer tenable. The initial picture continental glacial oscillations was essentially just presented by pollen analysis could evoke several latitudinal and altitudinal shifting of present life- interpretations because it was possible to study only 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 81 Table 23. Site References. Site Age Type of Site Reference Alabama Russell Cave c. 7 - 9,000 yrs. B.P. Cave Weigel et al., 1974 Arkansas Conard Fissure Irvingtonian Cave, fissure Graham, 1972;Brown, 1908 Peccary Cave, Newton Co. 2,230+ 120 yrs. B.P. Cave Davis, 1969; Quinn, 1972; 16,700+ 250 yrs. B.P. Semken, 1969 Florida Coleman IIA, Sumter Co. “Middle Pleistocene” Sinkhole Martin, 1974 Devil’s Den, Levy Co. Georgia Late Pleistocene/early Holocene Sinkhole Martin & Webb, 1974 Ladds Quarry, Bartow Co. Rancholabrean Fissure Ray, 1965, 1967;Lipps& Ray, 1967; Wetmore, 1967 Idaho Jaguar Cave, Lemhi Co. 10,370+ 350 yrs. B.P. Cave Guilday & Adams, 1967; 11,580+ 250 yrs. B.P. Kurten & Anderson, 1972 Wilson Butte, Jerome Co. 4,890+ 300 yrs. B.P. (M-1087) 15,000+ 800 yrs. B.P. (M-1410) Cave Gruhn, 1961, 1965 Illinois Meyer Cave, Monroe Co. Early Holocene Fissure Parmalee, 1967 Iowa Brayton local fauna. Audubon Co. 12,420+ 180 yrs. B.P. Fluvial Dulian, 1975 Waubonsie local fauna. Mills Co. “late Wisconsinan” Fluvial Hallberg et al., 1974 Kentucky Welsh Cave, Woodford Co. 12,950+ 550 yrs. B.P. (1-2982) Cave, sinkhole Guilday et al., 1971 Maryland Cumberland Cave, Allegany Co. Irvingtonian Cave, fissure Gidley & Gazin, 1938 Minnesota Itasca Bison Site, Clearwater Co. 7 - 8,000 yrs. B.P. Bison kill. Shay, 1971 Missouri Fluvial Bat Cave, Pulaski Co. Crankshaft Cave, Rancholabrean Cave Hawksley, er a/., 1973 Jefferson Co. Late Pleistocene/Recent Sinkhole Parmalee, et al., 1969 Pennsylvania Bootlegger Sink, York Co. Rancholabrean/early Holocene Cave, sinkhole Guilday, et al., 1 966 Hosterman’s Pit, Centre Co. 9,240+ 1,000 yrs. B.P. (M-1291) Cave Guilday, 1967 New Paris No. 2, Bedford Co. 1,875+ 100 yrs. B.P. (1-743) Sinkhole Guilday & Bender, 1958 New Paris No. 3, Bedford Co. 8,570+ 145 yrs. B.P. (1-5313) Sinkhole CMNH unpubl. New Paris No. 4, Bedford Co. 11,300+ 1,000 yrs. B.P. (Y-727) Sinkhole Guilday et al., 1964 Sheep Rock Shelter, Huntingdon Co. 490+ 100 yrs. B.P. (M-1904) Rockshelter Guilday & Parmalee 1965; 8,920+ 320 yrs. B.P. (M-1909) Michels & Smith, 1967 Tennessee Baker Bluff Cave, Sullivan Co. 10,560+ 220 yrs. B.P.) 11,640+ 250 yrs. B.P.j Cave, fissure Guilday et al., 1975 19,100+ 850 yrs. B.P.^X 3495) Guy Wilson Cave, Sullivan Co. 19,700+ 600 yrs. B.P. (1-4163) Cave Guilday et al., 1975 Robinson Cave, Overton Co. Rancholabrean Cave Guilday et al., 1969 Virginia Back Creek Cave No. 1 (Cook Cave), Bath Co. Rancholabrean Rockshelter, CMNH coll., unpubL owl roost 82 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 2 Table 23. Site References (continued). Site Virginia (continued) Back Creek Cave No. 2 (Sheets Cave), Bath Co. Age Rancholabrean Natural Chimneys, Augusta Co. Rancholabrean Saltville 1 3,460+ 420 yrs. B.P. (SI-46 1 ) Type of Site Rockshelter, owl roost Cave, owl roost Paludal Reference CMNH coll., unpubl. Guilday, 1962 Ray et al., 1967 a portion of the fossil flora (those plants possessing identifiable wind-distributed pollen). Only a gen- eralized conception of the sampled paleoenvironment was possible. Anomalous deciduous pollen present in an otherwise boreal pollen profile was always susceptible to the interpretation that it was not indigenous, but wind-transported from a distance. Remains of large Pleistocene mammals, mostly those of extinct herbivores (proboscideans, ovi- bovines, sloths, cervids) and carnivores could supply little to the detailed ecological picture. Pollen analysis supplied the outlines of biotic adjustment in lower latitudes, and convincingly demonstrated that the magnitude of change was much greater than that suggested by earlier workers who lived prior to the advent of palynology as a research tool (Braun, 1950). But the question of the detailed biotic composition of these adjustments, whether they represented simple biome fluctuations or resulted in new and unique combinations, re- mained open until paleontological sites of late Pleis- tocene age, containing remains of ecologically sensitive small vertebrates of known Recent habitat, were discovered and analyzed. Initiated in modern form by the late C. W. Hibbard (Hibbard, 1949), techniques for recovering micro- vertebrates soon demonstrated that periglacial faunas were composed of an amalgam of temperate and boreal species, and that the fossil faunas, even though composed of Recent forms, could not be duplicated today at any latitude. Initial analysis of such sites consisted of constructing overlap maps (like Fig. 18, this paper) in which the Recent ranges of the various recovered species were superimposed on one another and the area of maximum overlap was considered an approximation, or, at the least, an indication of former environments. It is becoming obvious that modern boreal terres- trial environments cannot be considered analogous to late Pleistocene situations in lower latitudes. Even if the climate could be duplicated precisely, which is not possible, differences in topography, geology, and the historical development of their respective biotas could not. Modern eastern North American boreal biotas are composed largely of post-Wisconsinan immigrants occupying deglaciated terrain that had been stripped of living things by glacial ice. Hence a competative edge was given to those species preadapted for a harsh boreal environ- ment and capable of rapid spreading. Southern periglacial faunas, on the other hand, were composed of both boreal forms and those temperate species that could survive either in local refugia or in various degrees of integration in an unglaciated situation. Periglacial habitats were not created and destroyed with the rapidity of those once covered by ice, but adjusted slowly and less dramatically with climatic change, creating a fluid mosaic of environmental niches supporting a richer biota than would be the case if that biota had to start anew (as in once- glaciated areas) with no historical continuity. Work by Dr. Holmes Semken, University of Iowa, and his students on late Pleistocene faunas from the Midwest demonstrated that this is true even for areas of little topographic diversity. Although differ- ent species are involved, the picture becomes even more complicated in the Appalachian Mountains where highlands and lowlands parallel each other alternately for hundreds of kilometers in a north- south direction. The richness of the Clark’s Cave local fauna, compared to the more boreal New Paris No. 4 fauna, 240 km to the north, appears to express this biotic gradient. In addition to this boreal-temper- ate intermesh, a cross-mesh of eastern forest and now mid-western forms occurred, producing a unique fauna that cannot be analyzed in the light of the present-day habitat requirements of each species taken individually. Each component species was responding not only to environmental pressures, but to pressures resulting from new and unique combina- tions of species, so that the definitive environmental picture of that time and place may not be the sum of its component parts as suggested by the Recent ecological requirements of each of the species in- volved. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 83 Literature Cited American Ornithologists’ Union 1957. Check-list of North American birds, 5th ed., Lord Bal- timore Press, Baltimore, Md. 691 pp. Anderson, Rudolph 1947. Catalogue of Canadian Recent mammals. Natl. Mus. Canada Bull. 102. 238 pp. Bailey, John Wendell 1946. The mammals of Virginia. Williams Printing Co., Rich- mond, Va. i-xvi + 416 pp. Baird, S. F. 1857. Reports of explorations and surveys. Part I, Mammals. V. 8, 757 pp. Banfield, a. W. F. 1974. The mammals of Canada. Univ. of Toronto Press, pp. v-xxv; 1-438. Barkalow, Frederick S. Jr. 1961. The porcupine and fisher in Alabama archaeological sites. Jour. Mammal., 42(4):544-545. Bent, Arthur C. 1932. Life histories of North American gallinaceous birds. U. S. Natl. Mus. Bull. 162. 490 pp. 1938. Life histories of North American birds of prey (Part 2). U. S. Natl. Mus. Bull. 170. 482 pp. 1942. Life histories of North American flycatchers, larks, swallows, and their allies. U. S. Natl. Mus. Bull. 179. 555 pp. Berkland, James O., and Loren C. Raymond 1973. Pleistocene glaciation in the Blue Ridge Province, South- ern Appalachian Mountains, North Carolina. Science, I8L651-653. 1974. Letter to the Editor. Science, 184:89-91. Bick, Kenneth F. 1962. Geology of the Williamsville Quadrangle Virginia report of investigations 2, Va. Div. Miner. Resour., Charlottesville. 40 pp. Bole, B. Patterson, Jr., and Philip N. Moulthrop 1942. The Ohio Recent mammal collection in the Cleveland Museum of Natural History. Sci. Publ. Cleveland Mus. Nat. Hist., V(6):83-I8I. Braun, E. Lucy 1950. Deciduous forests of eastern North America. The Blak- iston Co., Philadelphia and Toronto. 596 pp. Brooks, Maurice 1943. Birds of the Cheat Mountains. The Cardinal, 6{2);25-45. Brown, Barnlim 1908. The Conard Fissure, a Pleistocene bone deposit in north- ern Arkansas; with descriptions of two new genera and twenty new species of mammals. Mem. Amer. Mus. Nat. Hist., 9(4): 155-204. Bryson, Reid A., Wayne M. Wendland, Jack D. Ives, and John T. Andrews 1969. Radiocarbon isochrones on the disintegration of the Laurentide Ice Sheet. Arctic and Alpine Res., l(l):l-13. Cameron, Austin W. 1950. A new chipmunk (Tamias) from Ontario and Quebec. Jour. Mammal. 31(3):347-348. Carbone, Victor A. 1974. The Paleo-environment of the Shenandoah Valley. In Gardner, William M., ed.. The Flint Run Paleo-Indian Complex: A preliminary report 1971-1973 Seasons. Occas. Publ. I, Dept, of Anthrop., The Catholic Univ. of Amer. pp. 84-99. Clark, G. M. 1968. Sorted patterned ground: new Appalachian localities south of the glacial border. Science. 161:355-356. Clarkson, Roy B. 1966. The vascular flora of the Monongahela National Forest, West Virginia. Castanea. 3I(1):I-119. Conaway, Clinton H., and Donald W. Pfitzer 1952. Sorex palustris and Sorex dispar from the Great Smoky Mountains National Park. Jour. Mammal. 33( I): 106-108. Core, Earl L. 1974. Vegetation of West Virginia. McClain Printing Co., Parsons, W. Va. 217 pp. Craig, Alan J. 1969. Vegetational history of the Shenandoah Valley, Virginia. In U. S. Contrib. Quat. Res., Geol. Soc. of Amer. Spec. Papers 123, Boulder, Colo. pp. 283-296. Cushing, Edward J. 1965. Problems in the Quaternary Phytogeography of the Great Lakes Region. In Wright, H. E. Jr., and David G. Frey, eds.. The Quaternary of the United States. Prince- ton Univ. Press, pp. 403-416. Davis, L. C. 1969. The biostratigraphy of Peccary Cave, Newton County, Arkansas. Proc. Arkansas Acad. Sci., 23:192-196. Dice, Lee R. 1943. The biotic provinces of North America. Univ. Michigan Press, Ann Arbor. 78 pp. Doutt, j Kenneth, Caroline A. Heppenstall, and John E. Guilday 1973. Mammals of Pennsylvania. 3rd ed. Pa. Game Comm., Harrisburg, Pa. 288 pp. Dulian, James Joseph [MS Paleoecology of the Brayton Local Biota, Late Wis- thesis] consinan of southwestern Iowa. Dept, of Geol., Univ. Iowa. Faust, Burton S. 1964. The Saltpetre Caves of Virginia. In Douglass, Henry H., ed.. Caves of Virginia, Virginia Region of the Natl. Speleo. Soc., Virginia Cave Survey, Falls Church, pp. 31-56. Flint, Richard Foster 1971. Glacial and Quaternary Geology. John Wiley & Sons, Inc. 892 pp. Galbreath, Edwin C. 1964. A dire wolf skeleton and Powder Mill Creek Cave, Mis- souri. Trans. 111. State Acad. Sci. 57(4):224-242. Gardner, William M., ed. 1974. The Flint Run Paleo-Indian Complex; a preliminary report 1971-73 seasons. Occas. Publ. No. 1, Dept, of Anthrop., The Catholic Univ. of Amer. 146 pp. Gidley, John W., and C. Lewis Gazin 1938. The Pleistocene vertebrate fauna from Cumberland Cave, Maryland. U. S. Natl. Mus. Bull. 171:1-99. Gillette, David D., and John D. Kimbrough 1970. Chiropteran mortality. In Slaughter & Walton, eds.. About Bats. South. Methodist Univ. Press, pp. 262-283. 84 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 2 Godfrey, W, E. 1966. The birds of Canada. Natl. Mus. Canada Bull. 203. 428 pp. Gr.aham, Russell William [MS Biostratigraphy and paleoecological significance of thesis] the Conard Fissure local fauna with emphasis on the genus Blarina. Dept, of GeoL. Univ. of Iowa; unpubl. Grlihn, Ruth 1961. The archaeology of Wilson Butte Cave, southcentral Idaho. Occas. Papers Idaho State Coll. Mus. No. 6. 202 pp. 1965. T wo early radiocarbon dates from the lower levels of Wilson Butte Cave, southcentral Idaho. Tebiwa, Jour. Idaho State Mus. 8(2):57. Guii.dav, John E. 1958. The prehistoric distribution of the opossum. Jour. Mam- mal. 39:39-43. 1961a. Prehistoric record of Scalopus from Western Penn- sylvania. Jour. Mammal. 42(1):1 17-1 18. 1961b. Plecoms from the Pleistocene of Pennsylvania. Jour. Mammal. 42:402-403. 1962. The Pleistocene local fauna of the Natural Chimneys, Augusta County, Virginia. Ann. Carnegie Mus., 36(9): 87-122. 1963. Pleistocene zoogeography of the collared lemming (Dicrostonyx). Evolution, 1 7(2): 194-197. 1967. J he climatic significance of the Hosterman’s Pit local fauna. Centre County, Pennsylvania. Amer. Antiquity. 32(2):23 1-232. Gihlday, John E., and Eleanor Adam 1967. Small mammal remains from Jaguar Cave, Lemhi County, Idaho. Tebiwa, Jour. Idaho State Mus. 10(1); 23-36. Guilday, John E., and Martin S. Bender 1958. A Recent fissure deposit in Bedford County, Penn- sylvania. Ann. Carnegie Mus. 35:127-138. Guilday, John E., Harold Hamilton, and Allen D. McCrady 1966. The bone breccia of Bootlegger Sink, York County, Pennsylvania. Ann. Carnegie Mus. 38(8): 145-163. Guilday, John E., H. W. Hamilton, and A. D. McCrady 1969. Jhe Pleistocene vertebrate fauna of Robinson Cave, Overton County, Tennessee. Palaeovertebrata. 2:25-75. Guilday, John E., Harold W. Hamilton, and Allen D. McCrady 1971. I'he Welsh Cave peccaries (Plaiygomts) and associated fauna, Kentucky Pleistocene. Ann. Carnegie Mus. 43: 249-320. Guilday, John E., Harold W. Hamilton, and Paul W. Parmalee 1975. Caribou (Rangifer larandiis L.) from the Pleistocene of Tennessee. Jour. Tenn. Acad. Sci. 50(3):I09-1 12. Guilda'i, John E.. Paul S. Martin, and Allen D McCrad'i 1964. New Paris No. 4: A Pleistocene cave deposit in Bedford County, Pennsylvania. Bull. Natl. Speleo. Soc. 26(4): 121-194. GiTLDAY. John E., and Paul W. Parmalee 1965. Animal remains from the Sheep Rock Shelter (36 Hu I ), Huntingdon County. Pennsylvania. Pa. Archaeologist. 35(l):.34-49. Guilday, John E., and Paul W. Parmalee 1972. Quaternary periglacial records of voles of the genus Phenacomys Merriam (Cricetidae:Rodentia). Quat. Res. 2:170-175. H ack, John J ., and John C. Goodi ett I960. Geomorphology and forest ecology of a mountain region in the central Appalachians. U. S. Geol. Surv. Prof. Paper No. 347. pp. 1-64. Hack, John 1 ., and Wayne L. Newell 1974. Letter to the editor. Science. 184:89. Hallberg, George R., Holmes A. Semken, and L. Carson Davis 1974. Quaternary records of Microtus xanihognathus (Leach), the yellow-cheeked vole, from northwestern Arkansas and southwestern Iowa. Jour. Mammal. 55(3):640-645. Hall, E. Raymond, and Keith R. Kelson 1959. The mammals of North America. The Ronald Press Co., New York. 2 vols. 1 162 pp. Hall, John S., and Charles H. Blewett 1964. Bat remains in owl pellets from Missouri. Jour. Mammal. 45(2):303-304. Handley, Charles O., Jr. 1953. A new flying squirrel from the southern Appalachian Mountains. Proc. Biol. Soc. Washington. 66:191-194. 1956a. Bones of mammals from West Virginia caves. Amer. Midland Nat. 56( 1 ):250-256. 1956b. The shrew Sorex dispar in Virginia. Jour. Mammal. 37(3):435. 1959. A revision of American bats of the genera Euderma and Plecotus. Proc. U. S. Natl. Mus. 1 10(3417):95-246. , AND C. P. Patton 1947. Wild mammals of Virginia. Commonwealth of Virginia Comm. Game and Inland Fisheries. 220 pp. Hawksley, Oscar, Jack F. Reynolds, and Robert L. Foley 1973. Pleistocene vertebrate fauna of Bat Cave, Pulaski County, Missouri. Bull. Natl. Speleo. Soc. 35(3):61-87. Hay, Oliver P. 1920. Descriptions of some Pleistocene vertebrates found in the United States. Proc. LI. S. Natl. Mus. 58(2328): 83-146. 1923. The Pleistocene of North America and its vertebrated animals from the states east of the Mississippi River and from the Canadian provinces east of longitude 95°. Carnegie Inst. Wash. 499 pp. Hibbard, Claude W. 1949. Techniques of collecting microvertebrate fossils. Con- trib. Mus. Pal. Univ. Mich 8:7-19. Hibbard, Foy N. 1941 . Climate of Virginia. U. S. Dept. Agric. Yearbook, Climate and Man. pp. 1159-1169. Holman, J. Alan 1967. A Pleistocene herpetofauna from Ladds, Georgia. Bull. Georgia Acad. Sci. XXV(3): 1 54-166. Holsinger, John R. 1964. The biology of Virginia caves. In Douglass, Henry H., ed.. Caves of Virginia, Va. Region Natl. Speleo. Soc., Virginia Cave Survey, Falls Church, pp. 57-74. Howell, Arthur H. 1918. Revision of the American flying squirrels. N. Amer. Fauna. 44: 1-64. 1929. Revision of the American chipmunks. U. S. Dept. Agric. N. Amer. Fauna. 52:157. 1932. Florida bird life. Fla. Dept. Game and Freshwater Fish, xxiv + 579 pp. Jackson, Hartley H. T. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 85 1928. A taxonomic review of the American long- tailed shrews. N. Amer. Fauna. 51:238 pp. Johnson, David H. 1950. Myotis suhulalus Leibii in Virginia. Jour. Mammal. 31(2);I97. Jones, J. Knox, Jr., Dilford C. Carter, and Hugh H. Genoways 1975. Revised checklist of North American mammals north of Mexico. Occas. Papers, Mus. Texas Tech Univ. No. 28, 7 March. 14 pp. Keener, Carl S. 1970. The natural history of the mid-Appalachian Shale Barren flora. In Holt, Perry C., ed.. The distributional history of the biota of the southern Appalachians, Part 11: Flora, pp. 215-248. Klimkiewicz, M. Kathleen 1970. The taxonomic status of the nominal species Microtus pennsylvanicus and Microtus agrestis (Rodentia:Crice- tidae). Mammalia, Tome 34, No. 4:640-665. Kretzoi, M. 1962. Arvicolidae oder Microtidae? Vertebrata Hungarica, Tom. IV. Fasc. 1-2. pp. 171-175. Kurten, Bjorn, and Elaine Anderson 1972. The sediments and fauna of Jaguar Cave: II — The Fauna. Tebiwa, Jour. Idaho State Mus. 10(l):26-36. Latham, R. M. 1950. The food of predaceous animals in northeastern United States. Final report, Pittman-Robertson Proj. 36-R. Report 1, Pa. Game Comm., Harrisburg, Pa. 69 pp. Lensink, Calvin J. 1954. Occurrence of Microtus xanthognathus in Alaska. Jour. Mammal. 35(2):259. Linzey, Alicia V., and Donald W. 1971. Mammals of Great Smoky Mountains National Park. Univ. of Tenn. Press. 114 pp. Lipps, L., and C. E. Ray 1967. The Pleistocene fossiliferous deposit at Ladd’s, Bartow County, Georgia. Bull. Georgia Acad. Sci. 25:113-119. Long, Charles A 1972a. Taxonomic revision of the mammalian genus Microsorex Coues. Trans. Kansas Acad. Sci. 1971/74 (2):181-196. 1972b. Notes on habitat preference and reproduction in pygmy shrews, Microsorex. Canadian Field-Naturalist, 86: 155-160. 1974. Microsorex hoyi and Microsorex thompsoni. Amer. Soc. Mammalogists. Mammalian Species No. 33:1-4. Lynch, John D. 1966. Additional treefrogs (Hylidae) from the North American Pleistocene. Ann. Carnegie Mus. 38( 1 l):265-27l. MacCord, Howard A., Sr. 1973a. The Hirsh Site. Archeol. Soc. Va., Q. Bull. 27{3):151-I57. 1973b. The Hidden Valley Rockshelter, Bath County, Virginia. Archeol. Soc. Va. Q. Bull. 27(4):I98-228. McKeon, John B. 1974. Letter to the editor. Science. 184:88-89. Martin, Robert A. 1974a. Fossil mammals from the Coleman HA fauna, Sumter County. In Webb, S. David, ed.. Pleistocene Mammals of Florida, pp. 35-99. 1974b. Late Pleistocene mammals from the Devil’s Den fauna. Levy County. In Webb, S. David, ed.. Pleistocene Mam- mals of Florida, pp. 1 14-145. Martin, Robert L. 1973. The dentition of Microtus chrotorrhinus (Miller) and related forms. Occas. Papers, Univ. Conn., Biological Sci. Ser. 2(1 2): 183-201. Maxwell, Jean A. and Margaret Bryan Davis 1972. Pollen evidence of Pleistocene and Holocene vegetation on the Allegheny Plateau, Maryland. Quat. Res. 2(4): 506-530. Mayer, W. V. 1941. Variation and systematic position of the flying squirrel of Idaho. Murrelet, 22:30-31. Michels, Joseph W., and Ira F. Smith, eds. 1967. A preliminary report of archaeological investigations at the Sheep Rock Shelter site, Huntingdon, Pennsyl- vania. 2 vols. 943 pp. Miller, G. S. Jr., and G. M, Allen 1928. The American bats of the genera Myotis and Pizonyx. Bull. U. S. Natl. Mus. 144:1-218. Munro, j. A. 1929. Notes on the food habits of certain raptors in British Columbia and Alberta. Condor, XXXI: 1 16. Murray, J. J. 1945. The faunal zones of the Southern Appalachians. The Raven, 16(3-4): 10-22. 1952. A check-list of the birds of Virginia. Va. Soc. Ornithol., pp.1-113. Pagels, John F., and Cathy M. Tate 1975. Small Mammals of the Paddy Knob-Little Back Creek area of Bath County, Virginia. Proc. Va. Jour. Sci. ,26(2): 59., abstract. Paradiso, John L. 1969. Mammals of Maryland. N. Amer. Fauna, No. 66, 193 pp. Parmalee, Paul W. 1967. A Recent cave bone deposit in southwestern Illinois. Bull. Natl. Speleo. Soc. 29 (4): 119-147. Parmalee, Paul W., and John E. Guilday 1966. A Recent record of porcupine from Tennessee. Jour. Tenn. Acad. Sci.,41(3):8l-82. Parmalee, Paul W., Ronald D. Oesch, and J. E. Guilday 1969. Pleistocene and Recent vertebrate faunas from Crank- shaft Cave, Missouri. 111. State Mus. Reports of Investi- gations, No. 14: 1-37. Peterson, O. A. 1926. The fossils of the Frankstown Cave, Blair County, Penn- sylvania. Ann. Carnegie Mus., 16(2): 249-315. Preble, Edward A. 1908. A biological investigation of the Athabaska-Mackenzie region. N. Amer. Fauna No. 27. 574 pp. 1910. A new Microsorex from the vicinity of Washington, D.C., Proc. Biol. Soc. Wash., 23, 24 June: 101-2. Quinn, J. H. 1972. Extinct animals in Arkansas and related C-14 dates circa 3000 years ago. 24th Inter. Geol. Congr. 12: 89-96. Ray, Clayton E. 1965. A new chipmunk, Tamias aristus, from the Pleistocene of Georgia. Jour. Paleont. 39(5): 1016-1022. 1967. Pleistocene mammals from Ladds, Bartow County, Georgia. Bull. Georgia Acad, of Sci. XXV(3):120-150. Ray, Clayton E., Byron N. Cooper, and William S. Ben- NINGHOFF 1967. Fossil mammals and pollen in a late Pleistocene deposit at Saitville, Virginia. Jour. Paleont. 41(3). 86 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 2 Ray, Clayton E., Charles S. Denny, and Meyer Rubin 1970. A peccary, Platygonus compressus LeConte, from drift of Wisconsinan age in northern Pennsylvania. Amer. Jour. Sci., 268( 1 ):78-94. Repenning, Charles A. 1967. Subfamilies and genera of the Soricidae. Geol. Surv. Prof. Paper 565, U. S. Gov. Print. Office, Wash., D. C. 74 pp. Rhoads, Samuel N. 1903. The mammals of Pennsylvania and New Jersey. Pri- vately printed. Philadelphia. 266 pp. Roberts, Harvey A., and Robert C. Early 1952. Mammal survey of southeastern Pennsylvania. Pa. Game Commission, Harrisburg. 70 pp. Saltzman, Barry, and Anandu D. Vernekar 1975. A solution for the Northern Hemisphere climatic zona- tion during a glacial maximum. Quat. Res. 5(3):307-320. Semken, H. a. 1969. Paleoecological implications of micromammals from Peccary Cave, Newton County, Arkansas. Abstr. Pro- gram mtg. Geol. Soc. Amer., 3:27. Shane, Orrin C., Ill, and Paul W. Parmalee [In Vertebrate remains from the Raven Rocks site, Belmont press] County, Ohio. Kent State Univ. Press. Shay, C. Thomas 197 h The Itasca Bison Kill site, an ecological analysis. Minn. Hist. Soc. 133 pp. Smith, Charles R., Jack Giles, M. E. Richmond, Jerry Nagel, and D. W. Yambert 1974. J he mammals of northeastern Tennessee. Jour. Tenn. Acad. Sci. 49(3):88-94. Soper, J. D. 1942. Mammals of Wood Buffalo Park, Northern Alberta and District of Mackenzie. Jour. Mammal. 23(4): 1 19-145. 1948. Mammal notes from the Grande Prairie-Peace River region, Alberta. Jour. Mammal. 29(l):49-64. Todd, W. E. Clyde 1963. Birds of the Labrador Peninsula and adjacent areas. Univ. Toronto Press. 819 pp. Vazquez, A. W. 1956. A new southern record for Muslela erminea cicogiwnii. Jour. Mammal. 37: 113-1 14. Voorhies, M. R. 1974. Pleistocene vertebrates with boreal affinities in the Georgia Piedmont. Quat. Res. 4(l):85-93. Watts, W. A. 1970. The full-glacial vegetation of northwestern Georgia. Ecology. 5l(l):l7-33. Weigel, Robert D. 1974. Investigations in Russell Cave. Part 5: vertebrates from Russell Cave. Publ. Archeol., Natl. Park Serv., U. S. Dept. Inter. 13:81-85. Wetmore, Alexander 1959. Notes on certain grouse of the Pleistocene. Wilson Bull. 71(2): 178-182. 1967. Pleistocene aves from Ladds, Georgia. Bull. Georgia Acad. Sci. XXV(3):15I-I53. Wilder, C. D. Jr., and R. D. Fisher 1972. Occurrence of the golden mouse in southwestern Vir- ginia. Chesapeake Sci. 13(4):326-327. Wrigley, Robert E. 1972. Systematics and biology of the woodland jumping mouse, Napaeozapus insignis. 111. Biol. Monogr. 47. Univ. 111. Press. 1 1 7 pp. Youngman, Phillip M. 1975. Mammals of the Yukon Territory. Natl. Mus. Can., Ottawa 1975; Publ. in Zool., 10:1-192. 1977 GUILDAY, PARMALEE, AND HAMILTON: CLARK’S CAVE BONE DEPOSIT 87 BULLETIN 0/ CARNEGIE MUSEUM OF NATURAL HISTORY ISSN 0145-9058 THE CHACOAN PECCARY CATAGONUS WAGNERI (RUSCONI) RALPH M. WETZEL Biological Sciences Group, University of Connecticut NUMBER 3 PITTSBURGH, 1977 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 3, pages 1-36, figures 1-10, tables 1-8 Issued June 29, 1 977 Price $6.00 a copy ©1977 by the Trustees of Carnegie Institute, all rights reserved. CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE PITTSBURGH, PENNSYLVANIA 15213 CONTENTS Introduction 4 Acknowledgments 7 Specimens Examined 10 Measurements and Symbols 11 Relative Age of Specimens 11 Comments 12 Systematic Account 12 Comparisons 12 Comparisons 16 External Appearance and Measurements 16 Sexual Dimorphism 20 External Nares 20 Rostrum 20 Infraorbital Foramen 21 Zygoma 22 Orbits 22 Size of Brain Case 23 Basicranial Flexure 23 Sinuses 23 Mandible 24 Size of Teeth and Diastemas 24 Cups and Cingula 25 Incisors 25 Canines 27 Discussion and Conclusions 33 References Cited 35 INTRODUCTION A third species of living peccary (Fig. 1), Tayas- suidae, representing a generic addition to the fauna of the Paraguayan Chaco, was announced by Wetzel, Dubos, Martin, and Myers (1975). We assigned this form to the genus Catagonus Ameghino, formerly known only from the Lower and Middle Pleistocene of Argentina, and indicated evidence of its close rela- tionship to the genus Platygonus LeConte from the Upper Pliocene to postglacial times. We postulated that Catagonus occurs in most or all of the Gran Chaco. Since that writing, two skulls of Catagonus, collected in 1936 by Jose Yepes from the Argentine province of Salta, have been identified (Wetzel and Crespo, 1976) and conversations with hunters in the Bolivian Chaco indicate that Catagonus also occurs in that nation. [Also see Olrog et al ( 1 976) ] The Gran Chaco extends from southeastern Bolivia through western Paraguay and into northwestern Argentina. It is an area of thorn steppe or thorn forest, often with dense and spiny undergrowth. In Paraguay, the Chaco covers an area two-thirds the size of Cali- fornia and is traversed by the Trans-Chaco highway that begins north of Asuncion and runs northwesterly to the Bolivian border. Rainfall decreases from 1400 mm per year in the palm savannah of the Paraguay River valley to 400 mm per year along the western border of Paraguay, where grasses and bare sand occur between the more scattered trees and shrubs. The landscape is virtually flat, broken by extensive grass and palm swamps in areas of higher rainfall and by a few permanent and numerous intermittent streams {see Wetzel and Lovett, 1974; Short, 1975). Western Paraguay occupies three-fifths of the nation, but has only 4% of the population. Except for the Mennonite farming communities around Filadelfia, the Paraguayan Chaco contains large ranches, occa- sional small army posts, and much unused land. Sev- eral different derivations of the word Chaco have been postulated, but an appropriate one is that it comes from the Quechuan chacu, meaning an abun- dance of animal life (Weil et al., 1972). Jose de Acosta stated that chaco was a method of hunting by encirclement, and Garcilaso de la Vega used chacu to mean the annual ceremonial hunt of the Incas, in which encirclement was the method of capture ( Barbara G. Beddall, pers. commun.) . Zoogeographical studies that mention the Chaco have been largely without benefit of recent studies of its mammalian fauna. The more recent continental analyses treat the Gran Chaco as follows; Hershkovitz ( 1 972 ) grouped the Chaco with other adjacent faunal areas as a Parana-Paraguay Valley District, a transi- tional zone between the Brazilian and Patagonian Subregions. Fittkau (1969) placed the Chaco on the boundary between the Guiana-Brazilian and the Andean-Patagonian Regions, but fragmented it among three extensive zoogeographic provinces. MUller (1973:143-145), after considering the overlap of northern and southern faunas, found sufficient en- demic animal species and subspecies to designate a separate Chaco Center. He grouped (p. 175) the Chaco with nonforest dispersal centers that he con- sidered to have alternately expanded in arid periods and contracted during moist periods of the Quaternary. Other recent authors who have summarized the evi- dence for this ebb and flow of xeric vs. mesic habitats and their biota during Quaternary glacial-interglacial cycles include Eden (1974), Haffer (1974), Van der Hammen (1974), Short (1975), and Fairbridge (1976). Blair (1976) compared the anurans of the Sonoran of North America, the Chaco, and the Monte. The data he reviewed for the Chaco were restricted to Argentina, and his use of the term Chaco was much broader than the use in this paper. Solbrig (1976) wrote that Cabrera (1971) designated one of his three principal floral divisions of Neotropical South America as the Chaco Dominion, and one of its seven subdivisions as the Chaco Province. It is this latter use, except for restricting the eastern border of the Chaco to the Rio Paraguay, that I follow here. Prior to our report of the Chacoan peccary, Cata- gonus was restricted to two extinct species. Ameghino (1904) named the genus Catagonus when he de- scribed C. metropolitanus, based on a palate in two fragments, from Lower Pleistocene deposits in the city of Buenos Aires. In the same paper he described from the Middle Pleistocene of Buenos Aires Listrio- don bonaerensis, based on isolated teeth, and this species was later moved to the genus Catagonus by Rusconi (1930). LeConte (1848) described Platygonus from the North American Upper Pleistocene, and several species have been described from South American deposits. Among these, Rusconi (1930) described Platygonus (Parachoerus) carlesi from Middle Pleis- tocene deposits near the Rio Duke in Santiago del Estero, western Argentina. Platygonus carlesi wagneri Rusconi (ibid.) was reported in association with pre- Hispanic funeral urns, artifacts, and a large modern mammalian fauna from the vicinity of Melero in Santiago del Estero (Fig. 2). Rusconi (1948) later raised wagneri to the level of species. Both P. carlesi 4 1977 WETZEL: THE CHACOAN PECCARY 5 Fig. 1. Skull of the Chacoan peccary or Tagua, Catcigonus wagneri (Riisconi ). Specimen is conn 16886, collected 10 km W Fortin Teniente Montania, Depto. Boqueron, Paraguay, 21 July 1974. 6 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Fig. 2. Provenance of Recent peccary specimens examined. # Catagoms wagneri. 0 Tayassu pecari. O Tayassu tajacu. 1, type- locality, C. wagneri (Rusconi 1930, 1948). and P. wagneri were based upon complete or nearly complete crania. The Chacoan peccary proved to be conspecific with P. wagneri. Despite the similarities of this species and P. carlesi to North American Platygonus, the molari- form premolars and larger molars were distinct from all the Platygonus examined. P. carlesi and P. wagneri, in fact, were nearer to Catagonus metropolitanus and C. bonaerensis and were considered congeneric (Wetzel et al., 1975). I have since examined Platy- gonus rebuffoi Rusconi ( 1952), based upon a partial mandible, from the Middle Pleistocene of Uruguay 1977 WETZEL: THE CHACOAN PECCARY 7 and find it assignable to the genus Catagomis. Thus all the species placed in the subgenus Parachoerus Rus- coni belong, in my view, to the genus Catagomis. In the Chaco, Catagomis wagneri is distinguished from the other living peccaries by any one of a num- ber of Guarani or Spanish names, e.g., Pagua, Tagua, or Cure-buro. The white-lipped peccary, Tayassu pecari (Link) , is there known as Tagnicate or Tachy- cati, and the collared peccary, T. tajacii (L.), as Tayte-tou or Cure-i. The Recent specimens of C. wagneri are here described in more detail than was possible in the first brief report, and are contrasted with T. pecari and T. tajacu from Paraguay and ad- jacent Argentina. During the course of this study, such repeated comparisons have strongly emphasized the much closer relationship of T. pecari and T. tajacu. I am therefore abandoning Woodbume’s (1968) generic separation into T. pecari and Dicotyles tajacu, and returning to a congeneric treatment of those two species. Detailed comparisons of the crania of T. pecari and T. tajacu may be found in both Rusconi (1929) and Woodburne (1968), and comparisons of those two species with Platygonus compressus in Guilday, Hamilton, and McCrady ( 1971 ) . Such com- parisons are therefore not repeated here. Com- parisons with fossil forms are restricted to those assigned to the genera Catagomis, Platygonus, and, for comparative purposes, samples of Prosthennops Gidley and Mylohyus Cope. The genus Selenogonus Stirton (1947) from the Upper Pliocene or Pleis- tocene of Colombia is not included in the comparisons, and I have avoided a definite stand on the genus Argyrohyus Kxa^itv\c\\ (1959) from the Upper Plio- cene of Argentina. Neither holotype was examined, and the limited information in the descriptions, be- cause of the fragmentary nature of the specimens, dictates restraint in commenting on those important specimens. I use Platygonus marplatensis Reig (1952), rather than Argyrohyiis chapadmalensis (Castellanos) which Kraglievich considered the senior synonym. This is both a convenience and a bias on my part. In the use of Platygonus chapadmalensis (Ameghino) for another specimen (mmcn 246) in the collection at Mar del Plata, I have followed Reig. In June 1975, during a visit to the Museo Argentina de Ciencias Naturales “Bernardino Rivadavia” in Buenos Aires, the holotype of C. bonaerensis was pho- tographed and measured. Unfortunately, the holotype of C. metropolitanus has been lost. Hereinafter, mea- surements of C. metropolitanus are from Rusconi (1930). Rusconi’s holotypes of C. wagneri and C. carlesi have not as yet been located. Measurements in this paper, therefore, are from Rusconi (1930, 1948) . Most measurements for P. compressus from Welsh Cave, Kentucky, are from Guilday et al. (1971). When my measurements of this species are given, the number of specimens is always indicated. Measure- ments of crania, teeth, and metapodials of C. wagneri from the Chaco of Paraguay are listed in Table 1 and may be compared with Rusconi’s measurements of his holotypes and with Guilday’s measurements of P. compressus. ACKNOWLEDGMENTS My wife, Drew S. Wetzel, has been continually in- volved in this study during most of the trips to ex- amine specimens in museums and in preparation of this manuscript. Encouragement and sponsorship of our fieldwork in Paraguay, 1972-1975, have been generously provided in Asuncion by Ing. Hernando Bertoni, Ministro de Agricultura y Ganaderia, Re- publica del Paraguay; Ambassador George W. Landau and his staff of the Embassy of the United States; and Mr. Robert J. Eaton. Peter M. Berrie, ISCES, or- ganized the fieldwork of 1972. Robert L. Martin par- ticipated in the fieldwork of 1973-1975; J. W. Lovett in 1972-1973; Juan Balbuena in 1972 and 1974; Philip Hazelton in 1973 and 1975. In 1975 Robert Dubos, John J. Mayer, Philip N. Brandt, and Juan Guerrero Cruz were also colleagues in the field. Philip Myers collected three of the C. wagneri specimens. These are in the mvz, Berkeley, collection. Of numerous hunters interviewed in the Chaco, P. Erhard Schneider and Hans Enns of Filadelfia, Hugo and Jorge di Stilio of San Nicolas, Argentina, and Alberto and Oscar Arrigoni of Buenos Aires provided experi- enced observations on the habits and abundance of the three species of peccaries. John E. Guilday, Malcolm C. McKenna, Clayton E. Ray, Duane A. Schlitter, Frank C. Whitmore, Jr., and Michael O. Woodburne provided encouragement and advice. Gerald E. Schultz made helpul comments on the Blancan. For advice on location of specimens of fossil tayassuids, thanks are due J. F. Bonaparte, Guillermo del Corro, Paul S. Martin, Larry G. Marshall, Alvaro Mones, and Rosendo Pascual. BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Table 1 . Skeletal measurements, Catagonus wagneri, Chaco of Paraguay. Measurement Y s C O.R. N Greatest length Crania with 309.9 adult dentition 6.7 2.2 298-324 29 Condylobasal length (CBL) 266.6 7.2 2.7 252-280 27 Basal length 255.7 7.4 2.9 241-268 26 Anterior border of orbit to anterior of premaxilla (RL) 199.1 5.2 2.6 191-209 31 Postrostral length (CBL-RL adj. = PRL) 71.4 4.1 5.8 62.6-77.0 27 Anterior border of orbit to margin of infraorbital foramen 86.7 3.6 4.1 81.1-97.4 31 Zygomatic breadth 125.6 5.7 4.5 108-137 25 Depth of suborbital zygoma 30.3 2.3 7.6 26.0-35.2 30 Depth of zygoma: postorbital process of zygoma to preglenoid process 56.0 2.7 4.8 50.9-61.0 24 Vertical diameter of orbit 35.8 1.0 2.8 33.4-37.7 31 Width across canines 60.5 2.4 4.0 55.8-65.7 28 Width across canine buttresses 64.3 2.4 3.8 60.5-69.2 27 Width, maximal, across molar rows 64.4 3.0 4.7 59.0-74.0 27 Width, minimal, between p2’s 27.2 2.0 7.2 22.7-32.4 29 Width, minimal, between M^’s 24.7 1.9 7.6 20.8-29.9 30 Width, minimal, between orbits 77.5 4.1 5.3 70.0-89.0 30 Width across postorbital processes 100.4 3.9 3.9 93-109 29 Length of precanine diastema 20.1 2.3 11.6 13.3-26.3 31 Length of postcanine diastema 23.9 3.2 13.4 18.8-29.1 31 Height of nasal opening 31.2 1.2 3.8 29.0-33.3 23 Height of occiput from ventral border of condyles 100.4 4.2 4.2 93-109 26 Width across occipital condyles 43.6 1.6 3.5 40.146.1 26 Cranial capacity 114.1 7.9 6.9 102-130 22 Length, condyle to tip Mandibles with adult dentition 209.8 5.7 2.7 200-222 20 Height, maximal, posterior 93.0 3.5 3.7 86-99 20 Width, maximal, posterior 109.8 7.2 6.6 99-122 21 Length of postcanine diastema 30.8 4.1 13.4 21.4-36.4 23 Depth below middle of postcanine diastema 23.4 1.4 6.0 20.2-25.3 23 Depth below anterior margin of Mj 31.2 1.8 5.8 25.8-33.6 23 Length of symphysis 56.2 3.2 5.7 47.4-61.8 21 dp2-4^ length Deciduous dentition 33.3 — 1 dp2, length 9.2 — — — 1 width 7.5 — — — 1 dp3, length 11.6 — — 11.1, 12.2 2 width 11.2 — — 10.3, 12.2 2 dp4, length 13.3 — — 13.1-13.7 3 width 12.4 — — 12.1-12.7 3 dP2^, length 34.8 — — 34.1-35.3 4 dP2, length 7.4 — — 6.7-7.S 4 width 4.0 — — 3.94.0 5 dP3, length 9.8 — — 9.6-10.2 5 width 6.4 — — 6.1-6.6 5 dP4, length 18.2 — — 17.7-18.9 5 width 10.4 — — 10.2-10.5 5 1977 WETZEL: THE CHACOAN PECCARY 9 Table 1. Skeletal measurements, Catagonus wagneri, Chaco of Paraguay (continued). Measurement Y s C O.R. N Permanent dentition length 94.8 2.6 2.7 91.0-99.1 23 p2-4^ length 38.2 1.9 5.0 32.1-41.3 26 length 57.9 2.5 4.4 52.7-64.9 27 Upper canine, anteroposterior diameter 15.1 1.0 6.5 13.5-17.3 24 transverse diameter 10.4 0.8 7.7 9.0-12.2 23 p2, length 11.4 0.6 5.6 10.1-12.8 29 width 9.6 0.7 7.1 8.4-11.1 29 p2, length 13.2 0.8 6.0 11.5-14.8 34 width 13.1 0.7 5.2 12.1-14.6 33 P^, length 15.0 0.8 5.6 13.1-17.8 36 width 15.2 0.8 5.3 13.6-16.9 36 m1, length 16.6 1.0 5.9 14.7-18.4 38 width 15.4 0.7 4.5 14.0-16.9 37 m2, length 20.6 0.9 4.5 18.7-22.5 37 width 18.8 1.0 5.1 17.0-20.9 37 m3, length 22.3 1.3 5.9 20.1-25.4 28 width of anterior moiety 18.9 1.0 5.3 16.5-20.6 27 width of posterior moiety 17.4 0.8 4.3 16.1-18.6 28 P2-M3, length 98.7 3.8 3.8 90.5-107.6 16 P2-4. length 37.4 2.0 5.4 32.841.5 16 Mi_3, length 61.9 2.0 3.2 58.0-66.7 21 Lower canine, anteroposterior diameter 15.4 0.9 5.7 13.7-17.0 18 transverse diameter 10.5 0.7 6.8 9.0-11.7 18 P2, length 9.3 0.6 6.4 7.8-10.1 15 width 5.9 0.5 7.7 5.4-6.8 15 P3, length 12.6 0.6 5.1 11.3-14.0 19 width 8.8 0.5 5.4 8. 1-9.8 19 P4, length 16.2 1.0 6.2 14.4-19.4 22 width 12.6 0.6 5.1 11.4-13.8 22 Ml , length 17.0 0.9 5.4 15.2-18.8 28 width 13.0 0.5 3.9 11.9-13.9 28 M2, length 20.5 1.0 4.7 18.7-22.3 28 width 16.5 0.6 3.8 15.1-17.8 27 M3, length 26.2 1.6 6.0 24.3-30.6 21 width of anterior moiety 16.4 0.8 5.1 14.9-18.1 21 width of posterior moiety 15.9 0.6 3.8 14.4-16.8 21 Length of metapodials Metacarpal II 41.5 39.844.4 4 III 67.4 65.1-70.5 5 IV 69.8 68.5-71.4 5 V 52.6 50.3-53.7 5 Metatarsal II 47.2 43.6-51.2 4 III 80.0 79.0-81.5 5 IV 82.4 81.5-83.4 5 V 48.6 46.0-50.4 5 10 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 At this University, Robert E. Dubos measured many of the skulls with me, John J. Mayer noted the sim- ilarity to Platygonus, Theodore A. Heist adapted data for computer analysis, Mary Hubbard and her staff prepared the figures, Adelaide Ellsworth and Robert Pirozok sectioned hairs, and Solomon E. Wollman photographed the skull and toothrows. I am especially indebted to the following institutions and curators who generously permitted examination of specimens under their care. The following abbrevia- tions are used in later accounts to identify the location of specimens: AMNH — The American Museum of Natural History, New York, Malcolm C. McKenna. BMNH — British Museum (Natural History), Lon- don, G. B. Corbet. CM — Carnegie Museum of Natural History, Pitts- burgh, Mary R. Dawson. CONN — The University of Connecticut Museum of Natural History, Storrs. MACN — Museo Argentino de Ciencias Naturales “Bernardino Rivadavia,” Buenos Aires, Guillermo del Corro and Jorge A. Crespo. Mcz — Museum of Comparative Zoology, Harvard University, Cambridge, Barbara Lawrence. MDC — Museo Departmental de Colonia, Uruguay, Bautista Rebuffo. MHN — Museo de Historia Natural, Universidade Federal de Minas Gerais, Belo Horizonte, Jose Silvio Fonseca. MMCN — Museo Municipal de Ciencias Naturales, Mar del Plata, Galileo J. Scaglia. MN — Museu Nacional, Rio de Janeiro, Fernando D. de Avila-Pires and Fausto L. de Souza Cunha. MNHN — Museo Nacional de Historia Natural, Montevideo, Alvaro Mones and Alfredo Ximenez. mu’ — The Museum, Texas Tech University, Lub- bock, Patricia Vickers Rich. Mvz — Museum of Vertebrate Zoology, University of California, Berkeley, J. L. Patton. TMM — Texas Memorial Museum, University of Texas, Austin, Ernest L. Lundelius, Jr. USNM — National Museum of Natural History, Smithsonian Institution, Washington, D.C., Charles O. Handley, Jr. and Clayton E. Ray. Support for this study was provided by the National Geographic Society from 1973 through 1975. Addi- tional funds were supplied by the University of Con- necticut Research Foundation in 1973 and the Carnegie Museum of Natural History in 1975. SPECIMENS EXAMINED Collection sites along the Trans-Chaco highway of Paraguay are given in kilometers (Km) from the beginning of the road at Villa Hayes, northeast of Asuncion. Lower case (km) is used for distances from other points. Numbers preceding abbreviations for museums indicate the number of specimens examined. Numbers following such abbreviations are museum accession numbers. Catagoniis wagneri, Recent — 56 (Fig. 2). ARGENTINA. Salta: 1 macn, unspecified locality; 1 macn, Dragones. PARA- GUAY. Boqueron: 4 conn, 28-40 km N Filadelfia, road to Fortin Teniente Montania; 5 conn, 10-42 km W Ftn. Tte. Montania, road to Mariscal Estigarribia; 1 conn. Km 480; 1 CONN, Km 530; 1 conn, Ftn. Capitan O. Serebriakof. Nueva Asuncion: 12 conn. Km 555-607, vicinity of Tte. Ochoa; 24 CONN, Km 613-667, vicinity of Tte. Enciso; 2 conn. Km 764, Ftn. Sgto. Rodriguez (Villazon). Presidente Hayes; 3 MVZ, Km 275; 1 conn. Km 277. Catagonus spp., fossil. ARGENTINA. Buenos Aires; MMCN 41, nne Mar del Plata (Miramar Form., Ensenadense, Mid-Pleistocene); mmcn 707, Cuidad de San Antonio de iThe symbol TTU-P was inadvertently not used to indicate the paleontological collection of The Museum, Texas Tech University. Areco (“Bonaerense?,” Mid-Pleistocene?); mmcn 972, Ayo. Loberia (Barranca Lobos Form., inf.. Lower Pleistocene). Distrito Federal: macn 2440, L. homierensis Ameghino, holo- type (Bonaerense, Mid-Pleistocene). URUGUAY. Colonia: MDC 1345, P. rebuff oi Rusconi, holotype, 50 km N Colonia del Sacramento (Bonaerense, Mid-Pleistocene). Platygonus spp. ARGENTINA. Buenos Aires: mmcn 25, P. marplatensis Reig, holotype. Barranca de Los Lobos, 1 km NE Baliza Caniu (layer 3, Chapadmalal Form., Montehermo- sense, Upper Pliocene); mmcn 156,P. scagliae Reig, holotype, Chapadmalal (Barranca de Los Lobos Form., Vorhue inf., Uquiense, Lower Pleistocene); mmcn 246, Punta San Andres (Vorhue inf., Uquiense, Lower Pleistocene); mmcn 878, SW Ayo. Loberia (Barranca de Los Lobos Form., Uquiense, Lower Pleistocene); mmcn 1212, Canada Chapar (Vorhue inf.. Lower Pleistocene). BRAZIL. Minas Gerais; 3 mhn, caves (Pleistocene-subRecent). MEXICO. Guanajuato: 1 USNM, P. alemanii Duges, holotype, Moroleon (Pleistocene). UNITED STATES. Arizona: 4 amnh, Papago Springs (Rancholabrean). Florida: 1 usnm, Melbourne (Pleistocene). Idaho: 1 usnm, Hagerman Form. (Hemphillian). Kansas: 9 AMNH, Edson Quarry (Hemphillian). Kentucky: 15 cm, Welsh Cave (Rancholabrean). Maryland: 8 usnm, P. cum- 1977 WETZEL: THE CHACOAN PECCARY 1 1 berlandensis Gidley, incl. holotype, Cumberland Cave (Irvingtonian). Missouri: 7 amnh, Cherokee Cave (Rancho- labrean). Nebraska; 1 amnh. Snake Creek Form. (Hemphil- lian). Tennessee: 5 cm, Guy Wilson Cave ( Rancholabrean). Texas: 2 mu. Carter Quarry (Blancan); 3 tmm, Crosby Co. (Blancan); 1 amnh, Channing area (“Hemphillian-Blancan”). Prosthennops spp. UNITED STATES. California: 1 amnh, Eden (late Hemphillian). Florida: 1 amnh, Mixon Bone Bed (late Hemphillian). Nebraska: 1 amnh. Snake River (Valen- tinian. Lower Pliocene); 3 amnh. Ash Hollow Form, (late Clarendonian, Lower Pliocene). URUGUAY. Colonia: mdc 398, ? Prosthennops nruguayensis Rusconi, holotype, 12 km N Colonia del Sacramento (Pampean Form., Belgranense, Mid- Pleistocene). Mylohyiis spp. UNITED STATES. Arkansas; 1 amnh, Conard Fissure (Irvingtonian). Florida: 1 amnh, Seminole Field (Rancholabrean). Tayassu pecari, Recent — 37 (Fig. 2). ARGENTINA. Chaco: 2 macn, unspecified locality. Entre Rios: 1 bmnh, unspecified locality. Misiones: 3 macn, unspecified locality. Santa Fe: 1 macn, unspecified locality. PARAGUAY. 1 MNHN, unspecified locality. Nueva Asuncion: 3 conn, Tte. Enciso; 1 conn, Garrapatel-i, 7 km SW Km 620; 1 conn. Km 592; 6 conn, Ftn. Sgto. Rodriguez. Presidente Hayes: 13 CONN, Juan de Zalazar; 1 conn, Km 194; 1 mvz. Km 275; 2 usnm, Pto. Pinasco; 1 conn, 85 km w Pozo Colorado, 10 km N road to Ftn. Avalos Sanchez. Tayassu tajacii. Recent — 59 (Fig. 2). ARGENTINA. Chaco: 1 macn, unspecified locality. Formosa: 1 macn, unspecified locality. Jujuy: 1 bmnh, 1 macn, unspecified localities. Misiones: 5 macn, unspecified localities. Salta: 16 macn, Dragones; 1 macn, Urundel. BOLIVIA. Chuquisaca: 1 CONN, 8 km E Santa Rosa. PARAGUAY. Boqueron: 4 CONN, 29 km N Filadelfia, road to Ftn. Tte. Montania. Caaguazu: 2 mcz, Rio Yuqueri. Guatra: 2 bmnh, Villarrica. Itapua: 1 conn, Pto. Pirapo. Nueva Asuncion: 1 conn. Km 750, Ftn. Grab E. A. Garay; 5 conn. Km 764, Ftn. Sgto. Rodriguez; 1 conn. Km 607; 3 conn, Tte. Enciso. Presidente Hayes: 3 conn, Estancia-i, 100 km e Filadelfia; 7 conn, Juan de Zalazar; 3 mvz. Km 275. MEASUREMENTS AND SYMBOLS All linear measurements, given in millimeters, were taken with dial calipers accurate to .1 and, for the larger dimensions, with GPM calipers accurate to 1. The cranial measurements used in Table 1 and else- where include most of those used by Guilday et al. (1971) and Rusconi (1930, 1948) plus the following additions and derivations: Height of nasal opening: Distance from the most anterior point of nasal bone to midline of dorsal sur- face of premaxilla, in a line at right angles to the toothrow. Cranial capacity: An estimate, given in milliliters, of the volume of brain cavity, based upon volume of small beans required to fill the cavity. Rostral length (RL) : Distance from anterior tip of premaxilla to anterior margin of orbit. Rostral length, adjusted (RL adj.) : Distance from anterior tip of premaxilla, along midline axis of skull, to a line between the anterior margins of orbits. RL adj. is derived as the altitude of a triangle in which RL is the hypotenuse and one-half the interorbital width is the base. Postrostral length (PRL): Distance from anterior margin of orbits, along the midline axis, to a line across the posterior margin of occipital condyles. PRL is de- rived as the difference between condylo-basal length and RL adj. Symbols used in the Tables and elsewhere refer to: Y = mean or X; s = standard deviation of the sample (S.D. of Simpson, 1949 and Guilday et al., 1971); C : - coefficient of variation computed as s/Y (V of Simpson and of Guilday); O.R. = observed range; N = number in sample. RELATIVE AGE OF SPECIMENS Following Herring’s ( 1974) method, I rated the de- gree of closure of 22 cranial sutures for 45 C. wagneri, 21 r. pecari, and 20 T. tajacii, all conn specimens from Paraguay. The relative sequence of closure of these sutures within each species was then compared by mean suture fusion scores and by plotting suture closure against age categories. Arbitrary age categories were based upon the following chronology of tooth eruption and wear. These age categories were applied to all three species, but the following sequence of suture closure is for C. wagneri. 0. Immature: Only deciduous teeth in place; no molars present. 1. Young juvenile: Deciduous premolars and first 12 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 3 molar in place. Sutures closed by end of period: inter- parietal and occipito-parietal. 2. Juvenile: Deciduous premolars and first two molars in place. Sutures closed by end of period: inter- frontal and intermaxillary; sutures beginning to close: naso-frontal and occipitals. 3. Young adult: Permanent premolars and first two molars in place. Additional sutures closed by end of period: basispheno-occipital, naso-frontal, and occip- itals; sutures beginning to close: fronto-parietal, inter- nasal, interpremaxillary, jugo-frontal, maxillo-frontal, maxillo-jugal, naso-maxillary, occipito-squamosal, premaxillo-maxillary, premaxillo-nasal. 4. Adult: Third molar in place, but without wear. Additional sutures closed by end of period: fronto- parietal, internasal, interpremaxillary, jugo-frontal, maxillo-frontal, maxillo-jugal, naso-maxillary, occip- ito-squamosal, premaxillo-maxillary; sutures begin- ning to close: alispheno-squamosal, basispheno-pre- sphenoid, jugo-squamosal, maxillo-alisphenoid, pa- rieto-squamosal. 5. Adult: Moderate wear on first and second mo- lars, slight wear on third molar. Additional sutures closed by end of period: alispheno-squamosal, jugo- squamosal, maxillo-alisphenoid, parieto-squamosal, premaxillo-nasal. 6. Adult: Moderate to heavy wear on first two mo- lars, moderate wear on last molar. Additional suture closed before end of period: basispheno-presphenoid. 7. Old adult: Molars worn to basins. comments: The sequence of eruption of premolars and molars proved to be the same in all three species. The age at which teeth erupted is not known for C. wagneri nor T. pecari. Kirkpatrick and Sowls (1962) determined the sequence and time of tooth eruption in 26 captive T. tajacu in Arizona. In that species the first two upper molars were present by 43 (37-50) weeks of age, the permanent upper premolars by 72 (66-83) weeks, and the last upper molars by 83 (74-94) weeks; there was no significant difference between upper and lower dentition. The relative sequence of closure of cranial sutures is indicated in Table 2 for the three peccaries as mean suture fusion scores and as the sequence of closures within each species. As the samples for this compari- son are restricted to conn specimens from Paraguay and all but one specimen, T. tajacu from southeastern Paraguay, are from the Chaco, any possible effect of geographical variation upon sequence of closure has been reduced. This geographical homogeneity and some dissimilarity in sutures chosen for scoring may explain variation between the differences found by Herring (1974) for the species of Tayassu and the dif- ferences found in my study. The major differences found here, however, are between C. wagneri on one hand and the two species of Tayassu on the other. Of 22 sutures rated, 18 differed from Tayassu by more than .1 fusion score and one closure rank or more. Seven of the 22 sutures closed earlier and 10 sutures closed later than in T. pecari and T. tajacu. Fusion scores for internasal and basispheno-presphenoid su- tures differed from the other species by more than .2 but fell between the scores of the two Tayassu. For the remaining three sutures scored, the fusion score in C. wagneri was similar to T. tajacu for occipito-pari- etal and similar to T. pecari for interpremaxillary and basispheno-occipital. SYSTEMATIC ACCOUNT Catagonus Ameghino Catagonns Ameghino, 1904:188; Rusconi, 1930:164; Wetzel etal., 1975. lYPE species: C. nielropolitaniis Ameghino, loc. cit. Holo- lype not at macn, June 1975. Ensenadense, Mid-Pleistocene, Argentina, Buenos Aires. Listriodon Ao«flcre/i.v;.v Ameghino, 1904:186. Holotype: macn 2440. Bonaerense, Mid-Pleistocene, Argentina, Buenos Aires. Ptatygoniis (Ptirachoeriis) carlesi Rusconi, 1930:150. Holo- type not at MACN, June 1975. Bonaerense, Mid-Pleistocene, Argentina, Santiago del Estero, Las Termas, Rio Dulce. Catcigonns (Interchoents) honaeren.sis. — Rusconi, 1930:168, based upon Ameghino, 1904: 186. Platyganus (Parachoent.s) carlesi wagneri Rusconi, 1930:231 and Platygoniis wagneri. — Rusconi, 1948:231. Pre-Hispanic deposits, Argentina, Santiago del Estero, Llajta Mauca. Platygoniis (Parachoeriis) rehiiffoi Rusconi, 1952: 125; Mones and Francis, 1973:77. Holotype: mdc 1345. Bonaerense, Mid-Pleistocene, Uruguay, Colonia, Arroyo de Las Limetas. RANGE OF catagonus: Lower Pleistocene of Argentina and Uruguay to Recent of Gran Chaco. comparisons: The skull of Catagonus (Fig. 1) is similar to Platygoniis and differs from Tayassu (in- cluding Dicotyles) in having extreme development of the rostrum, nasal chamber, and sinuses; brainease 1977 WETZEL: THE CHACOAN PECCARY 13 proportionally and often actually smaller; infraorbital foramen well anterior to zygomatic arch; pronounced articular fossa absent on anterior face of zygomatic arch; zygomatic bar below orbit deeper; orbits more posterior in position, anterior edge of orbit well pos- terior to last molar, and postorbital process of zygo- matic well posterior to preglenoid process; pro- nounced basicranial flexure; teeth hypsodont; and mo- lars with four major cusps. Cranial characters of Catagonus that distinguish it from Platygonus (Fig. 3) are: molars larger, tooth- rows longer; postcanine diastemas shorter; premolars and molars with cuspules; last premolars (Pi molari- form, with four major cusps vs. two cusps; cingula well developed on most premolars and molars, complete on buccal side of all upper premolars and molars; lower first premolar (P2) with a single major cusp (Fig. 4) vs. paired lateral cusps; last upper molar (M^) quad- rangular, lacking a pronounced posterior constriction in transverse width. Excepting the earliest forms of Platygonus, Catagonus also differs in having three pairs of lower incisors and a more gracile skull that lacks the flared zygomata and keeled genium. Catagonus is similar to Prosthennops in that the last premolars are molariform, but differs from the latter in having larger molars with cingula more com- pletely developed, longer toothrows, and shorter post- canine diastemas. A more definitive separation of Prosthennops (with the latest occurrence in the Upper Pliocene) from Catagonus, which was first known from the Lower Pleistocene, awaits a better under- standing of the upper limits of Prosthennops. Catagonus differs from Platygonus (Brasiliochoerus Rusconi, 1930) in having the larger than M“, particularly for the anteroposterior diameter. Brasilio- choerus is unusual among the Pliocene-to-Recent tayassuids of North and South America in having the M® smaller than M- for both anteroposterior and transverse diameters, as shown in Table 3. The re- maining characters presented by Rusconi are not unique to Brasiliochoerus: ( 1 ) The features of molari- form premolars and the P^ with four main cusps and nearly the same size as are shared by Catagonus, Brasiliochoerus, Prosthennops and Mylohyus. (2) A glenoid fossa ventral to the orbit and, although not listed by Rusconi, an anterior opening of the infra- Table 2. Comparison of mean suture fusion scores* and sequence of closure of cranial sutures for the three species of Recent peccaries. Suture fusion scoring follows Herring (1974): 0 = unfused, 1 = less than half fused, 2 = about half fused, 3 = more than half fused, and 4 = completely fused. Cranial suture C. wagneri T. pecari T. tajacu interparietal 3.98 1st 3.46 2nd 3.40 2nd occipito-parietal 3.95 2nd 3.63 1st 3.92 1st interfrontal 3.89 3rd 3.46 2nd 3.40 2nd occipitals 3.82 4th 3.39 7th 3.23 5 th intermaxillary 3.64 5th 3.46 2nd 3.40 2nd naso-fronta! 3.58 6th 3.46 2nd 3.22 6 th occipito-squamosa! 3.50 7th 1.72 21st 2.28 17th premaxillo-maxillary 3.16 8th 3.46 2nd 3.22 6 th internasal 3.11 9th 3.37 8th 2.93 14th interpremaxiilary 3.07 10 th 3.04 14th 2.86 15th naso-maxillary 2.98 11th 3.24 9th 3.22 6th fronto-parietal 2.96 12th 2.83 15th 2.79 16 th basispheno-occipital 2.85 13th 2.80 16th 3.12 11th premaxillo-nasal 2.80 14 th 3.24 9th 3.08 12th maxilio-frontal 2.78 15th 3.12 11th 3.22 6th jugo-frontal 2.73 16th 3.07 13th 3.03 13 th maxillo-jugal 2.49 17th 3.16 11th 3.17 10th maxillo-alisphenoid 1.86 18th 2.33 18th 2.18 19th alispheno-squamosal 1.67 19th 1.82 20th 1.90 21st jugo-squamosal 1.62 20th 1.95 19 th 2.22 18th basispheno-presphenoid 1.62 20th 1.16 22nd 1.99 20th parieto-squamosal 1.58 21st 2.50 17 th 1.90 21st *Original mean suture fusion scores for entire samples: C. wagneri, 2.893; T. pecari, 3.188; T. tajacu, 2.915. The following constants were applied to adjust Tayassu samples to a mean age equivalent to that of Catagonus sample: T. pecari, .9075; T. tajacu, .9923. 14 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Fig. 3. Right maxillary toothrows, P-ML Left, Platygonus compressus, cm 20114, Welsh Cave, Kentucky. Right, Catagonus wagneri, CONN 16886, Chaco of Paraguay. Artist, Mary M. Hubbard. 1977 WETZEL: THE CHACOAN PECCARY Fig. 4. Right cheek teeth, C. wagneri, conn 16886. Left, maxillary P--M^. Right, mandibular P2-M3. 16 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Table 3. Anteroposterior and transverse diameters of m2 and for Brasiliochoenis and Catagonus. Except where stated otherwise, measurements are from Rusconi (1930). m2 m3 Platygoniis (Brasiliochoerus) platensis 16.3/15.5 15.5/13.5 P. (Brasiliochoerus) platensis parodii 17.3/16 16 /14.3 P. (Brasiliochoerus) stenocephalus, Reinhardt (1880;295) 18 /18 17 /16 P. (Brasiliochoerus) stenocephalusl*, Ameghino (1889;575) 16 /16 16 /13 Catagonus bonaerensis 19.6/18 22 /18 C. carlesi 19.8/18.5 22 /20 C. wagneri, Y, holotype and paratype, Rusconi (1948;235) 19.6/19.5 21.5/19.5 C. wagneri, Recent of Paraguay, this study 20.6/18.8 22.3/18.9 *Considered by Rusconi (1930:176, 219) to be “Catagonus (Interchoenis)l” orbital foramen near the middle of the rostrum occur in Catagonus, Mylohyus, Prosthennops, Brasilio- choenis, and most of the remaining species of Platy- gomis. These two charaters are in contrast to Tayassu where the glenoid fossa is posterior to the orbit and the infraorbital foramen lies in the posterior third of the rostrum. Partial exceptions, discussed later, occur in those South American Platygoniis that also dis- play other features suggesting some proximity to the ancestry of Tayassu. (3) Slender, gracile skulls occur in Catagonus, Brasiliochoenis, and some early Platy- gonus as opposed to the more massive skulls with flaring zygomata and angular processes and large canine buttresses of most Platygoniis and Prosthen- nops. (4) Long toothrows and short diastemas occur in both Catagonus and Brasiliochoenis. The general similarity of Catagonus and Brasiliochoenis may be seen by comparing C. carlesi in Rusconi (1930: Plates 3, 4, 5) for the former genus with P. steno- cephalus in Reinhardt (1880: Plate 7) and Winge (1906: Plate 6) for the latter subgenus. Only the vs. M-'^ differences noted above and the larger premolars and molars of C. carlesi appear to separate the two taxa. If Platygoniis is restricted by non-molari- form premolars, Brasiliochoenis must be removed from the genus as Woodburne (1968:32) believed should be done eventually. The question then remain- ing would be the use of Brasiliochoerus as a genus, or as a subgenus of Catagonus, a question that cannot be resolved here. Catagonus wagneri (Rusconi) Platygoniis carlesi wagneri Rusconi, 1930:231. Platygoniis wagneri. — Rusconi, 1948:231. Catagonus wagneri. — Wetzel, Dubos, Martin, and Myers, 1975:379; Wetzel and Crespo, 1976:25. type-locality: Pre-Hispanic deposits, Argentina, Santiago del Estero, Llajta Mauca, 28° 12'S, 63°05'W (Fig. 2). RECENT RANGE AND HABITAT: Semiarid thorn-forest and steppe of the Gran Chaco; specimens reported in this study (Fig. 2) are from the middle to western Chaco of Paraguay and the province of Salta, Argentina. Interviews with hunters indicate that the present range extends into the Chaco of Bolivia and the Argentine provinces of Formosa, Chaco, and northern Santiago del Estero. Olrog, Ojeda, and Barquez (1976) report specimens from Nueva Esperanza, Depto. Pellegrini, in the latter province, as well as information on the tagua in the Chaco of Salta. COMPARISONS EXTERNAL APPEARANCE AND MEASUREMENTS: Compared with the gray color of T. tajacu and the black of T. pecari, C. wagneri is a large brownish-gray peccary with a faint collar of lighter hairs across the shoulders. Hair on ears and legs is longer and paler in color than in Tayassu. The head is larger, and ears, legs, and tail are longer. The larger ear accounts, no doubt, for one of the local names, Cure-buro. Like T. pecari and T. tajacu, C. wagneri has vestigial hooves or dew claws on the reduced second and fifth digits of the forefeet. A single, median dewelaw is present on the posterior side of the hindfeet of T. pecari and T. tajacu but absent in C. wagneri. Dew claws are missing altogether in the extinct P. com- pressiis and P. cumberlandensis (Guilday et ah, 1971:291). Mean measurements of five freshly killed adults and one young adult C. wagneri from the vicinity of Teniente Ochoa and Tte. Enciso in the western Chaco of Paraguay are as follows: Length of head -j- body, 1026 Us 1112, 923; $ $ 1005, 1005, 1087); length of tail, 86.7 (3 . . . , 102; 2 2 88, 70, . . . ); length of hindfoot to tip of longest hoof, 227.6 ( $ S 1977 WETZEL: THE CHACOAN PECCARY 17 222, 235; 9 9 228, 238, 215); height of ear from notch, 119.4 (S S 115, 118; 9 9 120, 120, 124); weight, one 9 , 37 kg. Although most of our Paraguayan specimens of the three species of peccaries are skulls from hunters’ kills, the external measurements of C. wagneri may be compared with the following means for five adult T. pecari (usnm), collected by the Smithsonian Venezuelan Project from the states of Apure and Bolivar, and two adult T. tajacu (conn), from Sgto. Rodriguez, Depto. Nueva Asuncion, Paraguay, re- spectively: Length of head -|- body, 1049.6 and 841; length of tail, 38.2 and 55.0; length of hind foot in- cluding hoof, 220 (N 1) and 188; height of ear from notch, 82.0 and 92.0 One of the foregoing T. pecari weighed 26.9 kg. The greater size and weight of C. wagneri are obvious except for length of head + body. The overlap of this measurement with that of T. pecari reflects the relatively short postcranial di- mensions of C. wagneri. The greater size of the Tagua’s head is responsible for the saying among hunters in the Chaco that when the head is cut off, one is left with only half the animal. Larger cranial size in the Tagua is illustrated by comparing the greatest length of skulls of adult C. wagneri with those of T. pecari, all conn specimens from Paraguay: 309.9 (O.R., 298-324; N = 29) and 271.7 (254- 282;N = 20). The ratio of distal to proximal elements of the limbs of C. wagneri is comparable to Platygonm com- pressus and Mylohyus nasutus, rather than to T ay assn (see Table 4). This type of comparison, made by Guilday et al. (1971:291), assumes that the ratio of length of scapula to humerus or the ratios of lengths of the more distal limb elements to the humerus or femur are all greater in more cursorial mammals. Although limb bones in C. wagneri are shorter than in C. compressus, their measurements are generally nearer those of the latter species than of either Tayassu. Comparative mean lengths of metapodials of C. wagneri (Table 1) and C. compressus (Welsh Cave, Kentucky), respectively, follow: Metacarpal III, 67.4 and 85.5; metacarpal IV, 69.8 and 86.3; metatarsal, 82.4 (IV) and 91.1 In C. wagneri, meta- tarsals II and V are vestigial; phalanges and hooves for these digits are lacking. In Platygonus, metatarsal II is vestigial and V is lacking; in Tayassu, II is complete and V is vestigial (Guilday et al., 1971:291). As in Tayassu, digits II and V of the forefoot of C. wagneri are complete with hooves, phalanges, and articular surfaces on the distal ends of metacarpals. In addition to the evidence of limb proportions, the digits of C. wagneri suggest a species that is more cursorial than Tayassu but less cursorial than Platygonus. hair: The dorsal hairs or bristles, reaching 220 mm, are longer than in other peccaries (Fig. 5 ) . The basal third to half (65-100 mm) of the shaft, strikingly paler than the terminal portion, is indistinctly banded by alternate shades of grayish-tan and oflf-white. This pale portion gradually merges into a band of dark brown ( 15-20 mm), followed by a narrower band of white (8-13 mm), and then by a long (65-80 mm) apex of dark brown to black. The shorter, less nu- merous hairs range from entirely dark brown or black for the shortest (approximately 40 mm) hairs to those with a beginning of a banded pattern. Pronounced apical fraying results in a plumose termination in all but the shortest hairs. In T. pecari, hair has a much _ Table 4. Relative scapula and limb proportions of selected Tayassuidae; measurements in parentheses. Data for C. wagneri (Y CONN 17803 and 18006) are from this study and for all other species, from Guilday, Hamilton, and McCrady, 1971 :291. Skeletal Tayassu Tayassu Platygonus Mylohyus Catagonus element tajacu pecari compressus nasutus wagneri scapula 95% 90% 99% — 99% humerus (171.6/173.9) radius 69% 72% imc 84% 81% humerus (95.6/138) (119/166) (160/206) (182/217) (140.6/173.9) metacarpal IV 36% 38% 45% 47% 40% humerus (49/138) (63.3/166) (92.1/206) (103/217) (70.2/173.9) tibia 94% 94% 99% 105% 101% femur (135/144) (161/172) (198/201) (230/216) (177.0/175.0) metatarsal IV 41% 46% 53% 417o femur (59/144) — (92.4/201) (115/216) (82.6/175.0) 18 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Fig. 5. Dorsal hair of peccaries; left, C. wagneri; middle, T. tajacu; right, T. pecari. V 1977 WETZEL: THE CHACOAN PECCARY 19 T. tajacu Tpecari ZOOp C. wagneri Fig. 6. Hair of the peccaries; cross sections of dorsal hair at middle of shaft. Note the undulating surface and thinner cortical layer for G. wagneri as compared with the smaller diameter, smooth surface and thicker cortical layer for T. tajacu and T. pecari. 20 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 shorter pale basal area, followed by dark brown to black on the remaining three-quarters of the shaft. In T. tajacii, hair lacks the pale basal portion entirely, having instead, a series of distinct contrasting bands of alternating light and dark along the shaft before the terminal dark tip. Hair is also much shorter than that of C. wagneri, not exceeding 150 mm in all specimens examined from Paraguay. The shortest hairs of T. lajacu are white rather than black, as in C. wagneri. Scanning electron micrographs indicate that in C. wagneri the hair surface has undulations or shallow grooves that are absent in both T ay assn. This feature can be seen better in cross sections prepared for light microscopes. Figure 6 illustrates these undulations, the larger diameter, and thinner cortical layer of C. wag- neri. I measured diameter and minimal depth of the cortical layer, between the internal radial ribs, at the base, middle, and tip of hairs of the three peccaries. The thinness of the cortical layer in C. wagneri is illustrated by the means (in microns) and the ratio of minimal depth of the cortical layer to greatest diam- eter of the section as follows: C. wagneri, 17 (O.R., 15-20; N = 14) and .02 (.02-.03; N = 12); T. pecari, 37 (30-50; N = 20) and .06 (.05-. 12; N = 10); r. tajacu, 28 (20-35; N= 11) and .06 (.05-.06; N = 9). SEXUAL DIMORPHISM: Since most specimens of the Chacoan peccary are skulls from hunters’ kills, sexes are chiefly unknown, but it is obvious that the variation of the canines and surrounding bone is not as extreme as in P. compressiis. Coefficients of variation for C. wagneri and for P. compressiis (Guilday et al., 1971 : 293) are respectively: Width across upper canines, 2.86 and 10.65; width across canine buttresses, 3.03 and 12.26. When the sample of adult C. wagneri is divided by a combination of size of canines and width across canines, the group with the larger canines has a wider skull and somewhat shorter postcanine diastema and condylobasal length. The magnitude of variation is reduced by this segregation, a major reduction occur- ring in the zygomatic breadth from s = 10.8 for the unstratified group to s = 4.4 for the large-canine group, and 2.1 for the small-canine group. I am reluctant to associate size of canines with sexual dimorphism in C. wagneri because of the lack of clear- cut separation of “males” and “females.” Since the more gracile skull of C. wagneri lacks the extreme development of canine buttresses and has less varia- tion in canines, the possible effects of both age and any sexual differences are less accentuated than in P. compressiis. EXTERNAL NARES: Height of the nasal opening is greater in Catagoniis than in Tayassii, as follows, with the mean, if given, followed by s: Catagonus metro- politaniis 38; C. carlesi 28; C. wagneri 31.2, 1.2 (O.R. = 29.0-33.3; N = 23); Tayassii pecari 24.8, .80 (23.3-26.6; N = 18); T. tajacu 20.7, .71 (19.7- 22.5; N = 18). In Catagonus wagneri, C. carlesi, Platygonus compressiis, P. ciimberlandensis, and Tayassii pecari, the most posterior lateral margin of the nares (= narial notch) is well posterior to P. In T. tajacu the narial notch is broader, not so deep, and falls above the posterior part of P. In T. pecari the narial notch differs from both T. tajacu and C. wag- neri in being acuminate. C. wagneri differs from both T. pecari and T. tajacu in having a broadly curved midlateral projection, a process of the premaxilla, which divides the narial notch into subequal dorsal and ventral arcs. Where a slight projection occurs in the narial notch of T. pecari and T. tajacu from Para- guay, the projection is formed by the nasal bone rather than the premaxilla. rostrum: The much greater size of the rostrum, in both actual measurement and in proportion to the postrostral length of skull, separates Catagonus and Platygonus from the shorter-nosed members of Tayassii. Note in Table 5, however, that the postrostral to rostral ratios of Platygonus from the Lower Pleisto- cene of Argentina are approximately midway between Tayassii on one hand and Platygonus of North America and Catagonus on the other. The ratio of Platygonus sp. (mhn 305) is nearest to ratios of Tayassii. The lateral profile of the rostrum in Catagonus, Platygonus compressiis, and P. ciimberlandensis is distinctly convex. The rostrum in Tayassu pecari is slightly concave in profile; in T. tajacu, Platygonus scagliae, P. chapadmalensis (mmcn 246), and Platygonus spp. (mhn 305, mmcn 1 2 1 2 ), slightly con- vex or straight. In transverse section, the dorsum of the rostrum is broadly rounded in Catagonus and Platygonus, more sharply rounded in T. tajacu, and flat in T. pecari. Woodburne (1968:28) used the well-defined an- terior portion of the supraorbital canals in T. tajacu as a distinction between that species and T. pecari. In Paraguayan specimens of C. wagneri and T. pecari, the anterior portion of the canals varies from indis- tinct to, in the older specimens, as deep and well de- fined as the canals in T. tajacu. The anterior portions of the canals are also well defined in Platygonus scag- liae, P. chapadmalensis (mmcn 246), Platygonus 1977 WETZEL; THE CHACOAN PECCARY 21 Table 5. Comparison of ratios* of postrostral to rostral length and of position of infraorbital foramen to rostral length for Catagonus, Platygonus, and Tayassu. See text for explanation of abbreviations. PRL / RL adj. Orbit to infraorbital foramen / RL Catagonus wagneri. Recent Catagonus sp., MMCN 41 , .37, s .02, N 15 .47, s .02, N 14 Mid-Pleist., Argentina Platygonus compressus, CM 12885, .36 12886, 201 14, U-Pleist., N. Am. P. cumberlandensis, USNM 8000, 8146, .37 ( . .. ,.37, .38) .44 (.46, .43, .44) 8147, Mid-Pleist., N. Am. Platygonus sp., MMCN 1212, .36 (.38, . . . , .33) .42 (.40, .43, . . . ] L-PIeist., Argentina P. scagliae, MMCN 156, .48 .41 L-Pleist., Argentina P. chapadmalensis, MMCN 246, .48 .41 L-Pleist., Argentina Platygonus sp., MHN 305, .48 .31 Pleist., Brazil .51 .30 Tayassu tajacu. Recent .59, s .02, N 17 .30,s.02, N 19 T. pecari. Recent .60, s .03, N 13 .23, s .01, N 17 *Measurements for C. wagneri are from Table 1. Pertinent measurements for other specimens follow. Measurements in parentheses are for individual specimens; Y, s, and N are given for T. tajacu and T. pecari Orbit to infra- PRL RL adj. orbital foramen RL Catagonus sp. 7 1 195 200 P. compressus 76 ( . . . , 74, 77) 202 ( ... , 202, 202) 90.5 (93, 88, 90) 205.8 (203, 207, 207) P. cumberlandensis 90 (97, . . . , 82) 251 (255, . . . , 246) 111 (103, 119, . . . ) 261 (260, 275, 250) Platygonus sp. 112 232 97 237 P. scagliae 92 178 72 184 P. chapadmalensis 91 190 61 194 Platygonus sp. 97 190 58 194 Tayassu tajacu Ih.l , A. 1 124.8, 6.1 35.1, 2.9 127.5, 6.1 N=17 21 19 22 T. pecari 88.8, 3.7 150.2,6.4 38.5, 3.5 153.0, 5.1 N=13 18 19 19 Sp. (mmcn 1212), and in the lateral views of P. should not, in my view, be used as evidence of relation- stenocephalus (Winge, 1906: Plate 6) , and Catagonus ship of C. carlesi to T. tajacu as Woodburne (1968 : 32) carlesi (Rusconi, 1930: Plate 3). All these specimens from the Pleistocene were adults with evident molar wear. Thus the definite outline or sculpturing of the anterior canals, per se, is not a unique character of T. tajacu or Dicotyles vs. Tayassu, and especially not of older animals. The anterior canals of T. tajacu, being located nearer the summit of the dorsal rostral curve, are more evident from dorsal view. The ante- rior canals of C. wagneri, P. compressus, and T. pecari are located more laterally on the rostrum and are thus more broadly separated than the constricted anterior canals of T. tajacu. The anterior canals shown for C. carlesi by Rusconi (1930: Plates 3, 4) appear to be within the range of variation of C. wagneri, and suggested. INFRAORBITAL FORAMEN (Table 5) : In C. wagneri and C. carlesi, the anterior opening of this foramen is almost halfway between the orbit and the tip of the premaxilla, lying above P^. This position is approxi- mated in Platygonus from North America and in some of the specimens from South America, such as P. scagliae and Platygonus sp. (mmcn 1212). The remaining specimens from South America, P. chap- admalensis (mmcn 246) and Platygonus sp. (mhn 305), are similar in this respect to T. tajacu. In both species of Tayassu the opening is in the posterior third of the rostrum, above P^ or ME In T. tajacu it lies under the anterior part of the zygomatic arch, and in 22 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Table 6. Comparison of suborbital zygomatic depth in Recent peccaries from Paraguay, and specimens from the Pleistocene of Argentina, Brazil, and North America. See text for explanation of abbreviations. Platygonus cumherlandensis, USNM 8000, Suborbital zygomatic depth Ratio of suborbital zygomatic depth / PRL Mid-Pleistocene, North America Platygonus chapadmalensis, MMCN 246, 79.6 .82 Lower Pleistocene, Argentina Platygonus sp., MMCN 1212, 42.8 .47 Lower Pleistocene, Argentina Platygonus scagliae, MMCN 156, 49.6 .44 Lower Pleistocene, Argentina Platygonus sp., MHN 305, 39.1 .42 Pleistocene, Brazil Platygonus compressus, AMNH 42781 & TUC 5-1211 , Upper Pleistocene, 39.0 .40 North America 35.0 (35.0, 35.0) .44 (.44, .45) Catagonus wagneri. Recent, Paraguay Catagonus sp., MMCN 41, 30.3, s 2.3, N 30 .43, s .04, N 15 Mid-Pleistocene, Argentina 25.7 .36 Tayassu pecari. Recent, Paraguay 22.8, s 1.6, N 21 .26,s.02, N 13 Tayassu tajacu. Recent, Paraguay 17.5, s 3.9, N 22 .24, s .05, N 17 T. pecari, even more posteriorly, well under the zygo- matic shelf. The transverse shape of the opening differs in all three Recent species. In C. wagneri the opening is ovoid and oriented vertically along its longest axis; in T. tajacii, the longest axis of the oval is directed dorsolaterally; in T. pecari the opening is narrow and slit-like. zygoma: In both Tayassu pecari and T. tajacu, the ventrolateral face of the maxillary zygomatic process is deeply excavated as the fossa for the dilator naris lateralis muscle (as described for T. tajacu by Wood- burne, 1968) and the fossa extends anteriorly above the opening of the infraorbital foramen. Catagonus wagneri, C. carlesi, and Platygonus compressus differ markedly from Tayassu in having only a shallow, short fossa not extending anteriorly beyond the infraorbital foramen. Vertical depth of the suborbital zygoma is much greater in Catagonus and Platygonus than in Tayassu. This is shown, both as ratios and as measurements, in Table 6. Depth of the suborbital zygoma is, however, less in the more gracile skull of Catagonus than in Platygonus. In C. wagneri and Catagonus sp. (mmcn 41), the suborbital zygoma lacks the distinct con- cavity on the lateral face found in Platygonus. orbits: In Catagonus and Platygonus, the orbits lie posteriorly in the skull; the anterior margin of the orbit is distinctly posterior to the last molar; and the postorbital process of the zygomatic is dorsal to the glenoid fossa. In both Tayassu, the anterior margin of the orbit lies above either the M“ or and the post- orbital process of the zygomatic is well anterior to the preglenoid process. In Catagonus and Platygonus the eyes are thus set posteriorly behind a much longer rostrum, while in both Tayassu, eyes are more ante- riorly positioned behind a shorter rostrum. A mid- horizontal line through the orbits of either Tayassu is dorsal to the anterodorsal tip of the rostrum. In the genera Catagonus and Platygonus, com- pensatory shift in the orbital position took several evolutionary pathways that probably reduced inter- ference of the longer rostrum with vision: (1) The entire frontal region, including the orbits, shifted dorsally well above the dorsal margin of the rostrum, as in Platygonus scagliae, P. chapadmalensis (mmcn 246 ) , and Platygonus sp. (mmcn 1212). (2) Only the orbits became positioned more dorsally, nearer the upper margin of the skull, as in Platygonus cumber- landensis and P. compressus, as pointed out by Guil- day et al. (1971:298). ( 3 ) The long axis of the orbits came to lie at a more oblique angle to the long axis of the skull, as in Catagonus and Platygonus except for P. scagliae and relatives, above. This position of the eyes, along with the basicranial flexure of the 1977 WETZEL: THE CHACOAN PECCARY 23 Table 7. Comparison of cranial capacities of Recent peccaries, Catagonus wagneri, Tayassii pecan, and T. tajacii. All specimens are adults from Paraguay. See text for definitions of abbreviations. Species Y s C O.R. N Cranial capacity, ml. C. wagneri 114.1 7.9 6.9 102-130 22 T. pecari 158.9 5.1 3.2 150-167 9 T. tajacu 95.9 6.2 6.5 84-105 10 Cranial capacity / Condylobasal length C. wagneri .42 .03 6.6 .37-.51 21 T. pecari .67 .03 3.8 .62-.70 9 T. tajacu .50 .04 7.5 .43-.53 8 skull discussed next, would permit forward vision when the head flexed in a feeding position, permitting less interference by the rostrum (see Guilday et al., 1971:304). (4) The eyes shifted to a more lateral position, e.g., as Guilday et al. (ibid: 298) found for Platygonus. SIZE OF BRAIN CASE: Measurements of cranial ca- pacity (Table 7) indicate a proportionally much smaller brain than in Tayassii. The smaller skull of T. pecari has a greater cranial capacity than C. wag- neri, while the O.R. of the cranial capacity of the smallest species, T. tajacu, overlaps with that of the large C. wagneri. The markedly lower ratio of cranial capacity to condylobasal length in C. wagneri is sim- ilar to its low ratio of postrostral length to rostral length, adjusted (Table 5). Both ratios are expres- sions of the long rostrum and short postrostral dimen- sion in C. wagneri. It will be noted in Tables 1 and 5 that the postrostral lengths of the crania of Catagonus and P. compressus are much shorter than that mea- surement in T. pecari, and that the ratios of PRL/RL adj. in Catagonus and Platygonus are significantly lower than those in Tayassii. It is therefore to be ex- pected that when the cranial capacities are estimated for fossil Platygonus and Catagonus, they will prove to have had proportionally smaller brains than Tayassii. This would have conferred a distinct advantage upon Tayassu, although Guilday et al. (1971:309, 311) doubt that Tayassu and Platygonus could have been competitors in North America. It is difficult to imagine, however, that Catagonus and Tayassu were not com- petitors in the shifting ecotone between forest and forest-edge habitats in South America. In all but dusty, open habitats, the greater success of Tayassu must have been assured. BASICRANIAL FLEXURE: Distinct basicranial flexure, absent in T. pecari and T. tajacu, is remarkable in Catagonus wagneri, Catagonus sp. (mmcn 41 ) , Platy- goniis scagliae, Platygonus sp. (mmcn 1212), P. ciiinberlandensis, and P. compressus. In these species the basioccipital region is directed more anterodor- sally, and the basisphenoid, with an even more dorsal tilt, approaches the vertical. In Recent C. wagneri, the ventral surface of the basisphenoid is 65°-75° from horizontal as defined by the premolar-molar toothrow, compared to approximately 20° in Tayassu. Guilday et al. (1971:304) associated this flexure and the oblique long axis of the orbits in P. compressus with evolutionary adaptation for life in more open habitats. Such modifications would permit a horizontal direction of the main axis of sight even when the head was flexed in a grazing attitude. This is a reasonable sug- gestion, especially when correlated with the extreme rostral development in Platygonus and Catagonus, which would require cranial flexure for reduced im- pairment of vision in both grazing and nongrazing attitudes. sinuses: Extreme development of the sinuses, a distinctive feature of the skull of Platygonus (see Guilday et al., 1971), occurs also in C. wagneri. A pair of prominent suborbital sinuses project dorso- laterally to the pterygoid processes and posteriorly to the level of the tympanic bullae. Air passages connect the dorsal side of these sinuses with a posterior exten- sion of the nasal chamber. The extreme posterior mar- gin of the nasal chamber reaches the level, lateral to the midventral line, of the anterior part of the basi- sphenoid. The well-developed maxillary sinuses, the posterior extension of the frontal sinuses into the dor- sal part of the parietal bones, the suborbital sinuses. 24 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Table 8. Comparison of mandibles, Catagoniis, Platygonus, and Tayassu. Where available, Y, s, and N, or Y and N are given. Mandi- bular length is from condyle to anterior tip. Measurements of P. compressus are from Guilday et al. (1971 :293). Species Length, Maximal Depth at Depth at Height, ramus mandible height ramus postcanine diastema ant. margin Ml L., mandible C. wagneri. Recent, Paraguay 209.8, 5.7 93.0, 3.5 23.4, 1.4 31.2, 1.8 .44, .02 N = 20 20 23 23 19 Catagonus sp., MMCN 41, 202 30.6 35.8 Mid-Pleist., Argentina P. compressus, Welsh Cave, 218.1, 6.0 94.6, 5.0 31.2, 1.5 39.0, 2.7 .43* U-Pleist., Kentucky N = 12 10 18 18 P. cumberlandensis, USNM 8147, 278.3 113 36.7 48.2 .41 8921-3, Mid-Pleist., Maryland N =4 1 3 3 1 Platygonus sp., MMCN 1212, 270 40.9 50.1 L-Pleist., Argentina P. chapadmalensis, MMCN 246, 243 40.5 49.8 L-Pleist., Argentina P. scagliae, MMCN 156, 211 31.4 41.7 L-Pleist., Argentina Tayassu pecari. Recent, Paraguay 189.0,4.3 88.4, 3.3 29.2, 2.1 38.4, 2.9 .47, .02 and northern Argentina N = 14 14 15 15 13 T. tajacu. Recent, Paraguay and 152.4, 8.8 72.6,4.0 26.6, 1.4 32.2, 1.4 .47, .02 northern Argentina N = 19 16 18 18 16 *ratio of Y’s. and the posterior extension of the nasal chamber re- sult in the brain case of C. wagneri being nearly sur- rounded by air chambers. Finch, Whitmore, and Sims (1972:18) commented on the cul-de-sacs associated with the extreme development of sinuses in Platygonus compressus: “The passage of air through the nasal passage was thus very tortuous, which may have been an advantage in a dust-laden atmosphere.” It should be added that such extreme development of the sinus- nasal system in Catagoniis and Platygonus could serve as both a dust trap and a well-developed olfactory system, with the latter requiring the former in a dusty atmosphere. Tayassu pecari and T. tajacu, although sharing the familial development of sinuses about the dorsal and anteroventral part of the brain case, lack this extreme development of sinuses. The size of the cribiform plates, much smaller in Tayassu, also sug- gests a lesser dependence upon olfaction than in Platygonus and Catagoniis. mandible: As with the cranium, the mandible of C. wagneri is somewhat shorter than it is in the small- est Platygonus, and much longer than in the largest Tayassu (see Table 8). The body of the mandible is so slender that the depth is actually less than in the smallest peccary, T. tajacu. In both Tayassu, the mandible is proportionally deeper and has a higher ratio of height of ramus to total length than in Cata~ gonus, P. compressus, and P. cumberlandensis. In Tayassu, the posterior margin of the mandible bulges distinctly beyond the condyles, but in Cata- gonus and Platygonus, it projects only slightly if at all. In the more massive skull of Platygonus, the angular process of the mandible flares laterally, in contrast to Catagoniis and Tayassu, where the angular portion is deflected medially in a gentle curve, or is vertical. Like Tayassu, C. wagneri lacks the distinctive keel on the mandibular symphysis found in P. cumberlandensis and P. compressus. A slight keel is observable on the holotype of P. scagliae and on P. chapadmalensis (mmcn 246). Of the specimens of North American Platygonus examined, only those from the Upper Pliocene (Hemphillian) lacked a keel: Edson Quarry, Kansas (amnh). Snake Creek Formation, Nebraska (amnh), and Hagerman Formation, Idaho (usnm 13798). This keel and the lateral flare of the angular processes seem to be specialized features of the Middle to Upper Pleistocene radiation of Platygonus in North America, not occurring in the more gracile skulls of earlier Platygonus or the more conservative Catagonus. SIZE OF TEETH AND DIASTEMAS: Despite the smaller 1977 WETZEL: THE CHACOAN PECCARY 25 skull of C. wagneri, the teeth are larger than in many species of Platygonm and are proportionally larger than in any member of that genus. The maxillary toothrows of C. wagneri and P. conipressus are com- pared in Figure 3. This longer toothrow is accom- modated in the shorter jaw of C. wagneri through a reduction in length of diastemas. This is illustrated in Figure 7, where the maxillary postcanine diastema is plotted against the length of The premolars of C. wagneri are frequently at oblique angles to the main axis of the toothrow. I presume this is an effect of crowding caused by phylogenetic reduction of length of jaw not entirely compensated by reduction in diastemal space. CUSPS AND cingula: The teeth of Catagonus are similar to Platygonus in being functionally lophodont and more hypsodont than in Tayassu. As shown in Figures 3 and 4, the premolars and molars of C. wagneri have numerous small cusps in addition to the major cusps that are characteristic of Platygonus. The cingula are well developed and surround the anterior, buccal, and posterior margins of Similar cusp- ules and well developed cingula occur in the holotype of C. wagneri (Rusconi, 1930: Plate 16) and on P--1VF of C. metropolitanus (ibid.: Plates 8 and 9). Wear on the teeth of C. carlesi (ibid.: Plate 5) prevents in- spection for cingula and cuspules. Although teeth of the holotype of C. bonaerensis are also worn, I was able to see pronounced cingula on the anterior faces of and Mo,.? and cuspules at the boundaries be- tween the anterior and posterior moieties of P4 and Mo. In North American Platygonus, cingula on the buccal side of the molars are less well developed than in Catagonus. It should be noted, however, that in the temporal sequence of Platygonus examined, the buccal cingula of molars were more evident in the Blancan and Irvingtonian specimens than in the later Rancholabrean. Buccal cingula of the molars occur in the holotypes of Platygonus marplatensis and P. scagliae, in P. chapadmalensis (mmcn 246), and the large Playtgonus sp. (mmcn 1212). These specimens from the Upper Pliocene to Lower Pleisto- cene of Argentina, therefore, are intermediate in this character between Catagonus and North American Platygonus. In both Catagonus and Platygonus, cin- gula of the P-‘^ are well developed on the anterior and posterior faces and, for Catagonus, the buccal face and for Platygonus, the lingual face. In the mandible, no pronounced cingula occur on P0.4 of Catagonus or P. conipressus, but are well developed on the anterior, posterior, and buccal surfaces of P.4,4 of P. cumber- landensis. Tayassu differs markedly from both Catago- nus and Platygonus in lacking such well-developed cingula on premolars and molars. Where cingula do occur, they are restricted to portions of the anterior and posterior faces. incisors: The incisors of C. wagneri are longer and the maxillary incisors more procumbent than are those of the other living peccaries. Across the poste- rior base of each incisor is a cingulum, absent in T. pecari and present on only the upper incisors of T. tajacu. A minute median cuspule, rather than a cingu- lum, is present on the unworn posterior face of the lower incisors of Paraguayan Tayassu. The unworn P of C. wagneri differs from that of Tayassu in having a rounded tip, crenulated on its posterior margin, and a cavity in the occlusal (posterior) surface. A slight ridge partially divides the cavity into medial and lateral halves. The P of Tayassu has a pointed tip without crenulations and without the cavity or median ridge on the posterior surface. The size of the P, as measured transversely at the alveolus to avoid the effects of wear, does not differ greatly among Catagonus, Tayassu, and Platygonus from the Lower to Mid-Pleistocene of South America. However, I“’s in North American Platygonus from the Middle to Upper Pleistocene are smaller. The lat- ter teeth are compared with Catagonus, with Y and s given where available: C. wagneri, 6.1 and .37 (O.R., 5. 5-8. 6; N = 15); Catagonus sp. (mmcn 41), 6.8; P. cumberlandensis, 5.2 (4.9, 5.5; N = 2); P. ale- manii, Papago Springs, 5.5 (5. 2-5. 9; N 4); P. com- pressus, Welsh Cave, 4.6 (3. 3-5. 6; N = 5). The statement of Guilday et al.. (1971: 305 ) that the P of Platygonus is “. . . reduced to a small peg . . .” reflects the small size of this tooth in some of the specimens from Welsh Cave. As in Tayassu, all specimens of C. wagneri from Paraguay have three pairs of lower incisors, whereas 1.4 is rarely present in North American Pleistocene Platygonus. One (cm 2634) of 15 undamaged man- dibles of P. conipressus from Welsh Cave has alveoli for I4. One (amnh 45719) of seven undamaged mandibles of the same species from Cherokee Cave, Missouri, has an alveolus for the left 1,4. Of four com- plete mandibles of P. cumberlandensis, one (usnm 8147) has vestigial alveoli for I4. The following intact mandibles of Platygonus from the North American Upper Pliocene have alveoli for I4: One (usnm 13798) from the Hagerman Formation and six of seven specimens (amnh) from Edson Ouarry, Kan- sas. The small sample from the Pleistocene of Argen- 26 BULLETIN CARNEGIE MUSEUM OF NATURAE HISTORY NO. 3 66 62 — 58 — E E o E S 30 26 — 22 — 18 — Prosthennops crassigenis AMNH (late Clarendonian) \ Cumberland Cave (Irvingtonian) PLATYGONUS MMCN 156, Chapadmalal \ (Uquiense) Welsh Cave (Rancholobrean) VIMCN 1212, Canada Chapar (Uquiense) Prosthennops (Hemphillian) Papago Springs (Rancho- labrean) MU 6415, Blanco ^ lane an) Edson Quarry (Hemphillian) MMCN 246, Punta San Andres • (Uquiense) A (Bonaerense and CATAGONUS ® Ensenadense) MMCN 41, Mar del Plata ®( Ensenadense) Paraguay (Recent) (pre-Hispanic Recent) C C 44 48 52 56 60 64 68 il-3 Length of M ^ (mm) 1977 WETZEL; THE CHACOAN PECCARY 27 tina suggests a more conservative character; Catago- nus sp. (mmcn 41) and Platygonus sp. (mmcn 1212) have I3 on both right and left; Catagonus sp. (mmcn 707) and P. scagliae have 1,^ or an alveolus present on only the right side; P. chapadmalensis (mmcn 246) has no alveolus for the right I3 and probably none for the left. canines: The canines of Catagonus and Platygonus are similar in being longer from alveolus to tip, and, in proportion to that length, more slender than the shorter, broader canines of Tayassu. The base of the canine in C. wagnei and P. compressus is smaller than in T. pecari and, as would be expected, in those pecca- ries with larger skulls such as P. cumberlandensis and C. metropolitanus. Anteroposterior and trans- verse diameters of the upper canine at its alveolus follow: Platygonus alemanii, Papago Springs, 14.2 and 10.4 (N = 4); Catagonus wagneri, Paraguay, 15.1 and 10.4 (N = 24 and 23); C. wagneri, holo- type, 15.5 and 10; C. carlesi, holotype, 15.5 and 11; C. metropolitanus, 21 and 14; P. cumberlandensis, 20.0 and 14.3 (N = 3); Tayassu pecari, 16.3 and 11.0 (N = 9 and 10); and T. tajacii, 14.0 and 9.7 (N== 17). P“: The first upper premolar of C. wagneri is multi- cuspate and ovoid, with the anteroposterior diameter greater than the transverse. In Platygonus, the P^ is bicuspate and has equal diameters; in Tayassu, the P“ is roughly triangular in shape, its apex being an ante- rior lobe formed by the single large cusp. The antero- posterior diameter of Catagonus’ P^, although larger in proportion to its transverse diameter than in Platy- gonus, is approximately the same as in those Platy- gonus with large crania: C. wagneri, Paraguay, 11.4 (O.R., 10.1-12.8; N 29); Platygonus sp. (mmcn 1212), 11.3; P. cumberlandensis, holotype and para- types, 11.4 (N = 5); P. texanus (mu 6415), 10.9; P. bicalcaratus (tmm 31 175-12), 1 1.3. Po: The first lower permanent premolar of Catago- nus is proportionally narrow and has a single tall major cusp preceded by a low cuspule and followed by poste- rior cuspules (Fig. 4). As in Tayassu, the Po lacks the laterally paired major cusps of Platygonus, but such cusps are present in the deciduous Po of C. wagneri. p®: The second upper premolar of C. wagneri has three major cusps, a variably larger number of cusp- ules, and definite cingula on the anterior, buccal, and posterior margins. It shares the character of three major cusps with Tayassu, and differs from Platygonus whose P'^ has two major cusps, transversely arranged, and a complete cingulum on the lingual instead of the buccal margin. Although the P^ of C. wagneri is much larger than in Tayassu, Prosthennops, Platygonus compressus, P. scagliae, P. chapadmalensis, and Platy- gonus sp. of Edson Quarry, the O.R.’s of both diam- eters overlap with the diameters of Platygonus sp. (mmcn 1212), P. cumberlandensis, P. texanus (mu 6415), and P. bicalcaratus (tmm 31175-12). The P^ of Prosthennops differs in being narrower; those of most Platygonus are wider than long; and those of C. wagneri, approximately equal in both diameters. This range is indicated by the following examples of ratios of transverse to anteroposterior diameter: Prosthen- nops edensis, .94; Prosthennops sp., Mixon Bone Beds, .92; P, crassigenis, .86 (N = 2); Platygonus sp. (mmcn 1212), 1.09; P. chapadmalensis (mmcn 246), 1.09; P. compressus, Welsh Cave, 1.11 (from Y’s of Guilday et al., 1971); P. cumberlandensis, 1.13 (O.R., 1.09-1.16; N = 5); P. texanus (MU 6415), 1.13; Catagonus wagneri, Paraguay, 1.00 (.88-1.11; N - 32). P3: Although similar to Po, the second lower pre- molar of C. wagneri differs in being larger and in hav- ing the central, major cusp partially separated into lateral subdivisions. This is suggestive of the two trans- versely oriented cusps of Platygonus and Tayassu. The P.3 differs, like that of Tayassu, in lacking a cingulum, as contrasted with Platygonus in which a cingulum is continuous from the anterior and buccal to posterior face. Also like Tayassu but differing from Platygonus, the main cusp is preceded and followed by smaller cusps. The unworn anterior cusp of C. Fig. 7. Comparison of Recent and fossil Catagonus with Platygonus and Prosthennops. A single line represents the observed range; a bar represents two standard errors of the mean of sample. Measurements are from this study and the literature, as follows; Catagonus: A, C. carlesi, holotype, Prov. Santiago del Estero (Rusconi 1930); B, composite, of C. bonaerensis and diastema of C. metropoli- tanus, Prov. Buenos Aires (ibid.); C, C. wagneri, holotype and paratype, Prov. Santiago del Estero (Rusconi 1948); mmcn 41, Catagonus sp., Prov. Buenos Aires; Paraguay, C. wagneri. Recent (N = 27 X and 31 Y). Platygonus, reading from top down ; Cumberland Cave, Maryland, P. cumberlandensis, holotype and paratypes (N = 4); mmcn 1212, Platygonus sp., Prov. Buenos Aires; mmcn 156, P. scagliae, holotype, Prov. Buenos Aires; Welsh Cave, Kentucky, P. compressus (Guilday et al. 1971); Papago Springs, Arizona, P. alemanii (N = 3); MU 6415, P. texanus, Blanco Beds, Texas; Edson Quarry, Kansas, Platygonus sp. (N = 3); mmcn 246, P. chapad- malensis, Prov. Buenos Aires. Prosthennops: P. crassigenis, Ainsworth area, Nebraska; Prosthennops sp., Mixon Bone Bed, Florida. greatest antero-posterior length (mm) 28 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 14 A CATAGONUS B Chaco a. TayassjL . Tayas^ tajacu pecan Edson Quarry MHW 305® Welsh Cave • MHN 313 PLATYGONUS Cumberland Calve Blanco^MMCN 1212 (Uquiense) Ch erokee Cave >MMCN 246 (Uquiense) • R scagfiae ( Uquiense) 10 14 15 16 greatest transverse width (mm) 1977 WETZEL: THE CHACOAN PECCARY 29 wagneri is crenulated, and the posterior mass consists of a series of at least three cuspules. In Catagonus both diameters of the P3 are larger than those of Tayassu, most Prosthennops, Platygonus sp. of Edson Quarry, P. compressus of Welsh and Cherokee Caves, P. scagUae, and P. chapadmalensis (mmcn 246). A second group of Platygonus and allies with large crania, from the Upper Pliocene and Lower Pleisto- cene, have large, broad Pa’s, with transverse diameters above the O.R.’s of Catagonus, while their anteropos- terior diameters overlap the O.R.’s of the latter genus. The transverse diameters and, as it was used by Kraglievich (1959) to distinguish Argyrohyus, the ratios of transverse to anteroposterior diameters are contrasted between Catagonus and this second group: Catagonus sp. (mmcn 707), 8.9 and .63; C. wagneri, holotype, 9 and .66; C. wagneri, Paraguay, 8.8 (8.1- 9.8; N = 19) and .69 (.62-. 74; N = 24) vs. Platy- gonus texanus, 12.4 (11.3,13.5, Meade 1945) and .85 (.82, .89); Argyrohyus chapadmalensis, 10.7 and .88 (Kraglievich, 1959); Platygonus marplatensis, 10.2 and .85 (ibid.); Platygonus sp. (mmcn 1212), 10.05 and .82. These same ratios for the first group of Platygonus ranged from .77 {Platygonus from Edson Quarry) and .78 (P. compressus from Welsh Cave, from Y’s of Guilday et al., 1971) to .82 (P. cumber- landensis and P. chapadmalensis) . The Prosthennops specimens (amnh) differ by having very slender Pa and low diameter ratios: P. niobrarensis, .62 (.60, .65) and P. crassigenis, .61. PL The last permanent premolar of the genus Catagonus is molariform, nearly as large as the first molar, and has four major cusps. This is in contrast to Platygonus, which has much smaller P^’s with only two major cusps (Figs. 3 and 4). In Paraguayan C. wagneri (Fig. 8) the anteroposterior length of the P^ (Y, 15.0; N = 36) exceeds that of even the large Platygonus, mmcn 1212 (1 1.4) of the Lower Pleisto- cene of Argentina, as well as the large Blancan species reported by Meade (1945:528), P. bicalcaratus (11.0, 11.4, 12.0) and P. texanus (14.0). Although in Tayassu and the Prosthennops examined, the P^’s are more molariform than in Platygonus, they do not have four equal major cusps nor do they attain antero- posterior diameter approaching that of the ME In Tayassu, the P^ has only three major cusps. If the supernumerary teeth in the holotype of P. (Brasilio- choerus) stenocephalus (Winge, 1906: Plate 6) are molars, the P^ apparently has four cusps and ap- proaches the length of the molar. The greater size of the P4 of Catagonus as com- pared to Platygonus and relatives is illustrated by the following series of anteroposterior diameters: C. wagneri, Paraguay, 16.2 (14.4-19.4; N = 22); Platygonus texanus (MU 6415), 1 4.5; Platygonus sp. (mmcn 1212), 14.2; P. bicalcaratus (tmm 31175-12), 13.9; P. marplatensis, 13.0 (Reig, 1952: 122); Argyro- hyus chapadmalensis, 12.5 (Kraglievich, 1959:226); Platygonus cumberlandensis, 12.2 (N = 4); P. chapadmalensis (mmcn 246), 10.8; P. compressus, Welsh Cave, 10.8. Although all the foregoing speci- mens had smaller anteroposterior diameters than Catagonus, some of them had transverse diameters approaching that genus. The ratios of transverse to anteroposterior diameter reflect this character, as pointed out by Kraglievich (1959) for Argyrohyus: Catagonus bonaerensis, . 78; C. rebuffoi, .72; C. wagneri, Paraguay, .77, s = .04 (.66-. 82; N — 25) vs. Platygonus marplatensis, 1.00 (Kraglievich, 1959); P. cumberlandensis, .97 (.94-1.00; N = 4); P. cha- padmalensis (mmcn 246), .95; Argyrohyus chapad- malensis, .93 (Kraglievich, 1959); Platygonus tex- anus (mu 6415), .92; P. bicalcaratus (tmm 31178- 13), .91. mL The anterior and posterior cingula are inter- rupted at mid-point by cuspules. A variable number of three to four cuspules lie between these anterior and posterior cuspules. As in the other maxillary cheek teeth, a cingulum on the buccal side is continuous with cingula of the anterior and posterior faces. Both diam- Fig. 8. Comparison of Recent and fossil peccaries; lines and bars as in Fig. 7. Measurements are from this study and from the literature as follows: Catagonus: A, C. nietropolitaiuis and B, C. bonaerensis, holotypes, Prov. Buenos Aires, Mid-Pleistocene (Rusconi 1930); C, C. wagneri, holotype and topotype, Prov. Santiago del Estero, pre-Hispanic (Rusconi 1948); D, C. carlesi, holotype, Prov. Santiago del Estero, Mid-Pleistocene (Rusconi 1930); Chaco, C. wagneri, Paraguay, Recent (N = 36). Platygonus, reading from top down: mhn 313, Platygonus sp., Caves of Minas Gerais, Pleistocene; Edson Quarry, Kansas, Platygonus sp.. Upper Pliocene (N = 5); Cumberland Cave, Maryland, P. cumberlandensis, holotype and paratypes, Mid-Pleistocene (N = 5); Blanco Beds, Te.xas, P. bicalcaratus, Blancan (Meade 1945); mmcn 1212, Platygonus sp., Prov. Buenos Aires, Lower Pleistocene; mhn 305, Platygonus sp.. Caves of Minas Gerais, Pleistocene; Cherokee Cave, Missouri, P. compressus. Upper Pleistocene (Simpson 1949); Welsh Cave, Kentucky, P. compressus. Upper Pleistocene (Guilday et al. 1971); mmcn 246, P. chapadmalensis, Prov. Buenos Aires, Lower Pleistocene; P. scagliae, holotype, MMCN 156, Prov. Buenos Aires, Lower Pleistocene. Tayassu: T.pecari, Paraguay, Recent (N = 22); T. tajacu, Paraguay, Recent (N= 1 8). 30 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 eters of the first upper molars of Catagonus are larger than those of Tayassu, Platygonus compressus, and P. scagUae. The transverse diameter of Catagonus is greater than that of P. bicalcaratus (mu 6415) and P. chapadmalensis (mmcn 246), but smaller than that of P. texaniis (tmm 31175-12). MU Cuspules occur near the midlines of the ante- rior and posterior cingula. Medial cuspules lie in the valley between the anterior and posterior pairs of major cusps. Unlike the M, of Tayassu and Platygo- nus, there is a definite cuspule on the lingual side of the valley between the two pairs of major cusps. In un- worn Ml this is flanked by a lesser cuspule in the buccal position. The anteroposterior diameter of the Ml of Catagonus is larger than that of Tayassu or Platygonus. The transverse diameter overlaps that of P. texanus (mu 6415), P. bicalcaratus (tmm 31178-13), P. cumberlandensis, and Platygonus sp. (mmcn 1212). M-; Pronounced cuspules occur medially at the anterior and posterior borders of the second upper molar. As with the M\ medial cuspules lie next to the valley between the two moieties. Cingula bind the buccal and, except for the interruption by cuspules, the anterior and posterior margins of the tooth. Both diameters of the M- of C. wagneri are larger than in Tayassu, P. compressus, P. scagliae, and P. chapad- malensis (mmcn 246). The O.R.’s overlap those of P. texanus (mu 6415), P. bicalcaratus (tmm 31175-12), P. cumberlandensis, and Platygonus sp. (mmcn 1212). The entire sample is dwarfed by the transverse diameters of P. henningi Rusconi (1930), 21.3; and Catagonus sp. (mmcn 707), 22.6. Mo: The second lower molar is similar to Mi except that it is larger in all dimensions and has more pro- nounced cuspules. In Catagonus, both diameters of this tooth are greater than those of Tayassu, P. com- pressus, P. chapadmalensis (mmcn 246), P. scagliae, and Prosthennops. The O.R.’s of both dimensions overlap those of P. cumberlandensis, P. texanus (mu 6415, 6416) and P. bicalcaratus (tmm 31178-13, 31197-2). Transverse diameter of the Mo in the large Platygonus specimen (mmcn 1212) is 19.8, and that of Catagonus sp. (mmcn 707), 20.6 as compared to C. wagneri, Y = 16.5 (O.R., 15.1-17.8; N = 27). M^: The last upper molar of Catagonus is quad- rangular in outline, lacking the distinct posterior con- striction in transverse width that occurs in Tayassu, Mylohyus, and most Platygonus and Prosthennops. Figures 3 and 4 show the four major cusps, the cuspules, and truncated appearance of the M^. Some ratios of greatest transverse diameter of the posterior moiety to anterior moiety follow: C. wagneri, Para- guay, .92, s = .03 (O.R., .87-.96; N = 14); Catago- nus sp. (mmcn 41 ), Mid-Pleistocene, Argentina, .96; Platygonus bicalcaratus (tmm 31175-12), Blancan, .89; P. scagliae, holotype. Lower Pleistocene, Argen- tina, .88; P. compressus. Upper Pleistocene, Missouri, .85 (Simpson 1949:29; ratio of means of Wp to Wa); P. cumberlandensis, Mid-Pleistocene, Maryland, .88 (.79-. 96; N = 3); Prosthennops crassigenis, late Clarendonian, Nebraska, .86 (.81, .90); P. edensis, late Hemphillian, California, .93; Prosthennops sp., late Hemphillian, Florida, .86; Tayassu tajacu, Re- cent, Paraguay, .88 (.86-.95; N = 4); P. pecari, Recent, Paraguay, .88 (.82-.97;N = 9). The appearance of posterior taper is accentuated in Mylohyus, Tayassu, and most Platygonus by the pres- ence of a posterior lobe or heel as contrasted with the truncated M^ of Catagonus, Prosthennops, and Platy- gonus bicalcaratus. The M^’s of Catagonus and P. bicalcaratus (tmm 31 175-12)^ are quite similar, both having a well-developed cingulum on the posterior margin The chief differences in C. wagneri are the less pronounced posterior taper and the presence of medial cuspules, including one dividing the posterior cingulum rather than the medial ridge that divides this cingulum in P. bicalcaratus. Both diameters of the M® of C. wagneri are larger than those of Tayassu, Platygonus compressus, P. scagliae, P. chapadmalensis, Platygonus spp. (mhn 305, 313), and Prosthennops crassigenis. The O.R.’s ^Hibbard and Riggs (1949, Bull. Geol. Soc. Amer., 60:829-860), followed by Dalquest ( 1975, Occ. Papers No. 30, The Museum, Texas Tech Univ.), considered P. texanus to be conspecific with P. bicalcaratus. It is convenient for discussion purposes here to retain the two species, following Meade (1945), rather than speak of “texanus-like P. bicalcaratus" with a distinct heel on the M-^ (MU 6415) and “typical P. bical- caratus’’ with a truncated M3 (TMM 31175-12). Fig. 9. Canonical analyses, fossil and Recent peccaries, 21 variables as follows; condylobasal length, rostral length, greatest width ^ across maxillary toothrow, palatal widths inside P-'s and M“’s, height of cranium at condyles, length of postcanine diastema, length of P--M'\ length of P-"‘, length of M'"\ and anteroposterior and transverse diameters of individual maxillary teeth from P'* to M^. Except for P. compressus, Welsh Cave (Guilday et al. 1971), measurements are from this study. Note; Platygonus sp., mmcn lAb—P. chapadmalensis. 1977 WETZEL: THE CHACOAN PECCARY 31 -10- -5 0 T. tajacu TAYASSU T pecan' MHN 313 10 15 20 25 CATAGONUS .Caves of Minas Gerais, Pleistocene- SubRecent Rcompressus Guilday etal.(l97l) Roncho^_ labreo/^ ( > sp., MMCN 41 • /R scagliae, Uquiense Ensenadense MHN 305 Recent -10 -5 Platygonus sp., MMCN 246, Uquiense lAMNH 42781 R a lemon a PLATY GONUS Rancholobrean •MMCN 1212 Platygonus sp. Uquiense P, cumber/andensis Irvingtonian O 10 Catagonus wagneri 32 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 C >s JC CL a E -0^5® UJ-O-X- 8 § 55 ^ Jo .9 ♦- ^ S a ® 00 S Si CL SOw «• c 13^ L. o o o JC ® >\ o tlli ^a,? (B T — I — r T m 1 — \ — r o I IT) o Platygonus cumberlondensis Irvingtonian 1977 WETZEL: THE CHACOAN PECCARY 33 of the anteroposterior diameter overlap that of Pros- thennops edensis, Prosthennops sp. from Mixon Bone Beds Platygoniis spp. (mmcn 1212, mhn 309), P. bicalcaratus, and P. marplatensis. The M® of C. wag- neri is much smaller in both dimensions than Cata- gonus sp. (mmcn 707), in transverse diameter than P. bicalcaratus and P. marplatensis, and in antero- posterior diameter than P. texanus (mu 6415) . M3: The last lower molar differs from the truncated upper molar in having a large, median posterior cusp. This cusp (hypoconulid) in Catagomis, as well as in Platygonus, has an undivided point and is a prominent feature of the M3, secondary in height only to the pair of major cusps in each moiety. The equivalent posterior lobe in Tayassu is highly variable, with nu- merous cuspules. In Catagomis, a cuspule lies on the buccal flank of the posterior cusp. The cingulum on the anterior face of the M3 is interrupted by a median cuspule and the overlapping posterior cuspule of the M2. A chain of medial cuspules extends from the valley between the anterior pair of major cusps to the large posterior cusp. The transverse diameters of the posterior and an- terior moieties of M3 in C. wagneri are nearly equal (Fig. 4), as compared with proportionally lesser widths of the posterior moiety in other specimens ex- amined. This greater posterior taper is illustrated by the following ratios of transverse diameter of poste- rior moiety to anterior moiety: C. wagneri. Recent, Paraguay, .98 (O.R., .95-1.0; N = 10); Platygonus compressus, Rancholabrean, Missouri, .96 (Simpson, 1949:30; ratio of means of Wp to Wa); P. cumber- landensis, Irvingtonian, Maryland, .86; P. bicalcara- tus, Blancan, .91 (N = 2); P. texanus, Blancan, .90; P. marplatensis, holotype, .89; Prosthennops sp., late Hemphillian, Florida, .90; P. crassigenis, late Claren- donian, Nebraska, .86; P. niobrarensis, Valentinian, Nebraska, .79; Tayassu pecari, Recent, Paraguay, .89 (.84-.94; N = 10); T. tajacu. Recent, Paraguay, .91 (.84-.97;N= 10). DISCUSSION AND CONCLUSIONS The close relationship of the genera Catagomis and Platygonus indicated in the foregoing comparisons is also supported by canonical analyses. Multivariant comparisons were made using width and length of mandibular teeth plus length of mandibular diastemas, maxillary teeth measurements, and all the measure- ments or combinations listed for Figures 9 and 10. In Figure 9, the specimens of Catagomis are posi- tioned adjacent to the middle of the grouping of Platygonus specimens from the Pleistocene of both South and North America. This is also shown in Fig- ure 10, except that the closeness of two large speci- mens of Platygonus to Catagomis is exaggerated. Some of the large dimensions of the teeth of Platy- gonus sp. (mmcn 1212) and P. texanus are similar to those of Catagomis, but when cranial dimensions are included in the multivariant comparisons, as in Figure 9, this relationship is not so close. Note also that in Figure 8 the dimensions of a single critical tooth, P^, of the Blancan specimens and mmcn 1212, are simi- lar to each other and group with the rest of Platygonus, well separated from Catagomis. The specimens of Platygonus from the South Amer- ican Pleistocene, although separable from Catagomis by both multivariant and univariant comparisons, show some blending of Platygonus and Catagomis characters not noted in the North American Platy- gonus. P. scagUae, P. chapadmalensis, and Platygonus sp. (mmcn 1212), all from the Lower Pleistocene (Uquiense), have typical Platygonus P'* (see Fig. 8), but show some similarity to Catagomis in cingula on the anterior and buccal sides of P“ and M^’^. In the relationship of the maxillary postcanine diastema to length of M^ P. chapadmalensis is also intermediate between Catagomis and Platygonus (see Fig. 7) . All multivariant comparisons, including those shown in Figures 9 and 10, positioned Tayassu pecari closely with T. tajacu. Tayassu is widely separated Fig. 10. Canonical analyses, fossil and Recent peccaries, 26 variables as follows: lengths Pj-Mj, P)_4, M1.3, anteroposterior and trans- verse diameters of individual teeth from P^ to and Pa.to M3. Measurements from the literature are indicated in figure; measurements from this study are as follows: Catagomis: C. wagneri, Paraguay, Recent; Catagomis sp., mmcn 41, Prov. Buenos Aires, Mid- Pleistocene. Platygonus, reading from top down: Platygonus sp., mmcn 1212, Prov. Buenos Aires, Lower Pleistocene; Platygonus sp., Edson Quarry, Kansas, Upper Pliocene; P. scagliae, holotype, mmcn 156, Prov. Buenos Aires, Lower Pleistocene; P. cliapad- nialensis, mmcn 246, Prov. Buenos Aires, Lower Pleistocene; P. cumberlandensis, holotype and paratypes, Cumberland Cave, Mary- land, Mid-Pleistocene. Prosthennops: P. crassigenis, Ash Hollow Formation, Florida, Lower Pliocene. Tayassu: T. pecari and T. tajacu, Paraguay, Recent. 34 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 3 from Catagonus in all these comparisons, but near some Prosthennops (Fig. 10) and two Platygonus, MHN 3 1 3 (Fig. 9 ) and mhn 305 (Table 5 ) . The latter specimens are from caves in the region of Lagoa Santa, Minas Gerais. These and other mnh specimens strongly suggest a phylogenetic continuum from Prosthennops-Platygonus to modern Tayassii pecari. As none of the Lagoa Santa specimens at mhn have M“’s larger than M^’s, as does Platygonus (Bmsilio- choerus) stenocephalus (see Table 3), I have not fol- lowed Paula Couto’s (1970:7) application of that name to the large specimens in that collection. Woodburne (1968:30-32) hypothesized that a common ancestor of T. pecari, P. (Brasiliochoerus) stenocephalus, and Mylohyus would be Prosthennops niobrarensis Colbert (Valentinian, Lower Pliocene of Nebraska). The additional specimens from Minas Gerais at mhn, not seen by Woodburne, suggest that T. pecari originated in South America, probably either in the forested highlands of Brazil or during a period of isolation in a forest center in the Amazonian basin. Woodburne also believed there was an affinity of T. tajacii with the Platygonus compressus group and con- sidered the similarity of their common ancestry to certain peccaries in the Frick Collection (amnh) — the specimens from Edson Quarry, Kansas (Woodburne, pers. commun.). In the present study some compari- sons place these Edson Quarry specimens in the genus Platygonus. In other comparisons they fall in an inter- mediate position with Platygonus, Catagonus, and Tayassu, and seem to support Woodburne’s idea, but with T. pecari and T. tajacu more closely related to each other than to any other species. It is probable that T. tajacu evolved in South America from a proto- T. pecari. Both the more northerly extension of the range of T. tajacu and its much smaller size, which would have made available to them many more suit- able dens in which the species could take refuge, sug- gest that the selective force was cold stress upon a population isolated in montane forests. In tracing ori- gins of T. pecari separate from T. tajacu, Woodburne placed chief emphasis upon difference in origin of two muscles of mastication, but ignored the many common characters in which they differ from other genera. I would not challenge his excellent list of salient differ- ences (ibid.: 28-29), except for the minor note that exceptions to his differences, numbers 9, 19, 22, 26, and 27, have been encountered in peccaries from Venezuela and Paraguay. And, of course, I would use his list of characters as specific rather than generic differences. The early history of Catagonus cannot be traced with the data at hand, but its close relationship to Platygonus seems apparent. In many respects, C. wag- neri is more conservative than P. cumberlandensis and P. compressus of the Middle to Upper Pleistocene — in retention of dewclaws on the forefeet and the third pair of lower incisors, in failure to develop the flaring zygomata, angular processes and orbits, the canine buttresses, and the symphyseal keel. In these features and in the more gracile skull, C. wagneri more nearly resembles Platygonus of the Upper Plio- cene. Also, the molariform premolars of C. wagneri are somewhat like those of Prosthennops of the Pliocene. The more primitive features of C. wagneri suggest that its ancestry was not subjected to as severe selec- tion toward a specialized cursorial existence in open habitats as has been suggested for Platygonus by Guilday et al. (1971). North American Platygonus was preyed upon by large, cursorial carnivores that were absent in South American grasslands. This may have provided a selective force developing a more cursorial group in North America (Guilday, pers. commun.). Another effect of greater predation pres- sure upon North American Platygonus might have been the selection of males with larger canines for defense of the herd. Catagonus, not subjected to this degree of selective predator pressure upon defenders of the herd, would retain a more ancestral, minimal sexual dimorphism. In support of this thesis, the variation in size of canines and canine buttresses in Platygonus is greater by five-fold than in C. wagneri. Despite its more conservative evolution, C. wagneri retains those features of Platygonus that are probably adaptive for a cursorial life in an open, arid habitat: Greater size than the other modern peccaries, elon- gated limbs, basicranial flexure associated with eye position and long rostrum, loss of some external dew- claws, and extreme development of olfactory cham- bers and sinuses. The close relationship of Catagonus and Platygonus does not necessarily indicate a phylogenetic continu- ity: A better understanding of the limits of Prosthen- nops and Platygonus is needed. If the limit of the genus Platygonus is restricted by the presence of non-molariform premolars, Platygonus and Catago- nus must be considered to be sister groups, with Prosthennops containing a common ancestor. Catagonus wagneri has survived as a component of the Chaco dispersal center (as used by Muller, 1973) along with other mammalian species such as the three- 1977 WETZEL: THE CHACOAN PECCARY 35 banded armadillo Toly pent es matacus, the greater pichiciego Burmeisteria retusa, the cricetid rodent Pseudoryzomys wavrini, and the mara Dolichotis salinicola. The growing literature on climatic and biotic fluctuations in South America and their possi- ble correlation with the glacial cycles of the world has been recently summarized by Muller (1973), Haffer (1974), and Short (1975). Little is known of the history of the Chaco in the Pleistocene except that it also was subjected to xeric-mesic cycles that extended into postglacial times (Short, 1975:171). It is possi- ble that in the Chaco, Catagonus and other arid- adapted plants and animals were afforded suitable habitats that were continuous with other nonforest centers during aridity cycles. Muller found, for exam- ple, strong affinities between the Chaco and the Bra- zilian Caatinga centers. His Pampa center has more affinities with the Chaco center than he presents, an example being the range of one of his indicator spe- cies, the hairy armadillo Chaetophractus villosus. This species is not confined to the Pampa center as recorded by Cabrera (1958:214) but, as we have found in our studies in Paraguay, occupies the Chaco center as well. Such a speculative picture of Pleistocene ebb and flow of Catagonus populations should be tempered with an accounting for the isolation of C. wagneri in a single zoogeographic center. Some features suggest that the Chacoan peccary, evolving from larger an- cestors of the Pleistocene, has become stunted. The large skull vs. the proportionally smaller body and the large, crowded premolars and molars vs. the short diastema and gracile skull may seem to support this hypothesis. Although our knowledge of this newly dis- covered living relict is presently inadequate for this task, it is tempting to invoke either, or both, genetic drift in a restricted population (an islandic type of speciation) or severe selection toward a smaller-bod- ied Catagonus in a minimally adequate habitat. It is reasonable to suppose that in an earlier, more mesic cycle, the habitat for Catagonus was much more re- stricted than now. It is also reasonable to consider that its habitat may have been a rather unsatisfactory compromise between minimally adequate tempera- ture, when the pampean grasslands to the south were no longer warm enough for peccaries, and the nearest available semi-open habitat — the thorn scrub of the Chaco. In conclusion, Catagonus wagneri is a diurnal, cur- sorial animal, better equipped for far vision and speed for escape from predators than are the nocturnal Tayassu. C. wagneri is also an “olfactory” animal, uti- lizing an elaborate sinus system as a dust trap in the semi-arid environment to which it is restricted. Both preliminary stomach analyses and tooth structure in- dicate that C. wagneri is a browser, as compared with the more omnivorous Tayassu. T. pecari and T. tajacu have survival advantages over C. wagneri in having larger brains, either actually or proportionally, prob- ably greater flexibility of diet, and shorter rostra af- fording greater flexibility of head positions without interference with forward vision. The three species of living peccaries meet in the Gran Chaco during what may be an interim period — the present — between a more arid cycle that favored thorn forest, steppe, and C. wagneri, and a moist cycle that will increasingly favor more mesic forests and Tayassu. References Cited Ameghino, F. 1904. Nuevas especies de mamiferos cretaceos y terciarios de la Republica Argentina (continuacion). An. Soc. Cient. Argent., 58: 182-192. Blair, W. F. 1976. Adaptation of anurans to equivalent desert scrub of North and South America. In Goodall, D. W., Evolu- tion of desert biota. Austin, University of Texas Press, 197-222. Cabrera, A. 1958. Catalog© de los mamiferos de America del Sur. Rev. Mus. Argent. Cienc. Nat. ‘Bernardino Rivadavia,’ Zool. (1957) 4(1): 1-307. Cabrera, A. L. 1971. Fitogeografia de la Republica Argentina. Boln. Soc. Argent. Bot., 14: 1-42, cited by Solbrig, 1976. Colbert, E. H. 1938. Pliocene peccaries from the Pacific coast region of North America. Carnegie Inst. Washington, Publ. No. 487: 243-269. Eden, M. J. 1974. Paleoclimatic influences and the development of savanna in southern Venezuela. Jour. Biogeogr., 1:95-109. Fairbridge, R. W. 1976. Shellfish-eating Preceramic Indians in coastal Brazil. Science, 191:353-359. 36 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 3 Finch, W. I., F. C. Whitmore, Jr., and J. D. Sims 1972. Stratigraphy, morphology, and paleoecology of a fos- sil peccary herd from western Kentucky. U.S. Geol. Survey Prof. Paper, 790, 25 pp. Fittkau, E. J. 1969. The fauna of South America. In Fittkau, E. J., J. Hies, H. Klinge, G. H. Schwabe, and H. Sioli, Biogeogra- phy and ecology in South America, vol. 2. The Hague, Dr. W. Junk Publishers, 624-658. Guilday, j. E., H. W. Hamilton, and A. D. McCrady 1971. The Welsh Cave peccaries (Platygonus) and asso- ciated fauna, Kentucky Pleistocene. Ann. Carnegie Mus., 43:249-320. Haffer, j. 1974. Avian speciation in tropical South America. Cam- bridge, Mass., The Nuttall Ornithological Club, Piibl. No. 14, 390 pp. Herring, S. W. 1974. A biometric study of suture fusion and skull growth in peccaries. Anat. Embryol., 146: 167-180. Hershkovitz, P. 1972. The Recent mammals of the Neotropical Region: a zoogeographic and ecological review. In Keast, A., F. C. Erk, and B. Glass, Evolution, mammals, and southern continents. Albany, State University of New York Press, 311-431. Kirkpatrick, R. D., and L. K. Sowls 1962. Age determination of the collared peccary by the tooth-replacement pattern. Jour. Wildlife Manage- ment, 26:214-217. Kraglievich, j. L. 1959. Rectification acerca de los supuestos “molares hu- manos fosiles” de Miramar (Prov. de Buenos Aires). Rev. Inst. Antropol., 1:223-236. LeConte, j. L. 1848. On Platygonus Compressus: a new fossil pachyderm. Mem. Amer. Acad. Arts Sci., new ser., 3:257-274. Meade, G. E. 1945. The Blanco fauna. Contrib. Geol. 1944, Univ. Texas Publ. 4401:509-540. Mones, A., AND J. C. Francis 1973. Lista de los vertebrados fosiles del Uruguay, II. Mammalia. Comun. Paleont. Mus. Hist. Nat. Monte- video, 1:39-97. Muller, P. 1973. The dispersal centres of terrestrial vertebrates in the Neotropical realm. The Hague, Dr, W. Junk Pub- lishers, 244 pp. Olrog, C. C., R. A. Ojeda, and R. M. Barquez 1976. Catagonus wagneri (Rusconi) en el Noroeste Argen- tine. Neotropica, 22:53-56. Paula Couto, C. de 1970. Paleontologia da regiao de Lagoa Santa, Minas Gerais, Brasil. Bol. Mus. Hist. Nat., Univ. Fed. Minas Gerais, Geol. No. 1, 21 pp. Reig, O. a. 1952. Descripcion previa de nuevos ungulados y marsupi- ales fosiles del Pliocene y del Eocuartario Argentines. Rev. Mus. Munic. Cienc. Nat. Tradic., Mar del Plata, 1:119-129. Reinhardt, J. 1880. De i de brasilianske Knoglehuler fundne Navlesvin- Arter. Videnskabelige Meddelelser fra den naturhis- toriske Forening i Kjobenhavn, 1879-1880:271-301. Rusconi, C. 1929. Anatomi'a craneodental de los tayassuinos vivientes. An. Soc. Cient. Argent., 107:66-82, 177-242. 1930. Las especies fosiles argentinas de pecaries (Tayassui- dae) y sus relaciones con las del Brasil y Norte Amer- ica. An. Mus. Nac. Hist. Nat. “Bernardino Riva- davia,” 36: 121-241, 18 pis. 1948. Restos de platigonos y malformaciones oseas proce- dentes de los tumulos indigenas de Santiago del Estero. Rev. Mus. Hist. Nat. Mendoza, 2:231-239. 1952. Pecaries extinguidos del Uruguay. Rev. Mus. Hist. Nat. Mendoza, 6: 123-127, 2 pis. Short, L. L. 1975. A zoogeographic analysis of the South American Chaco avifauna. Bull. Amer. Mus. Nat. Hist., 154:163-352. Simpson, G. G. 1949. A fossil deposit in a cave in St. Louis. Amer. Mus. Novitates, 1408:1-46. SOLBRIG, O. T. 1976. The origin and floristic affinities of the South Amer- ican temperate desert and semidesert regions. In Goodall, D. W., Evolution of desert biota. Austin, University of Texas Press, 7-49. Stirton, R. A. 1947. A rodent and a peccary from the Cenozoic of Colom- bia. Colombia Inst. Geol. Nac., Compilacion de los Estudios Geologicos Oficiales en Colombia, 7:317- 324, 1 pi. Van der Hammen, T. 1974. The Pleistocene changes in vegetation and climate in tropical South America. Jour. Biogeogr., 1:3-26. Weil, T. E., J. K. Black, H. I. Blutstein, D. S. McMorris, F. P. Munson, and C. Townsend 1972. Area handbook for Paraguay. Washington, U.S. Gov- ernment Printing Office, 316 pp. Wetzel, R. M. and J. A. Crespo 1976. Existencia de una tercera especie de pecari (Fam. Tayassuidae, Mammalia) en Argentina. Rev. Mus. Argent. Cienc. Nat. “Bernardina Rivadavia,” Zool. (1975) 12:25-26. Wetzel, R. M. and J. W. Lovett 1974. A collection of mammals from the Chaco of Para- guay. Univ. Conn. Occas. Papers, Biol. Sci. Ser., 2:203-216. Wetzel, R. M., R. E. Dubos, R. L. Martin, and P. Myers 1975. Catagonus, an “extinct” peccary, alive in Paraguay. Science, 189:379-381. WiNGE, H. 1906. Jordfundne og nulevende Hovdyr (Ungiilata) fra Lagoa Santa, Minas Geraes, Brasilien. Copenhagen, Museo Lundii, 3 : 1-239. WOODBURNE, M. O. 1968. The cranial myology and osteology of Dicotyles tajacii, the collared peccary, and its bearing on classi- fication. Mem. S. Calif. Acad. Sci., 7:1-48. BULLETIN of CARNEGIE MUSEUM OF NATURAL HISTORY REEVALUATION OF EARLY MIOCENE NORTH AMERICAN MOROPUS (PERISSODACTYLA, CHALICOTHERIIDAE, SCHIZOTHERIINAE) MARGERY CHALIFOUX COOMBS Department of Zoology, University of Massachusetts Amherst, Massachusetts 01002 NUMBER 4 PITTSBURGH, 1978 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 4, pages 1-62, figures 1-28, tables 1-6 Issued I August 1978 Price: $5.00 a copy Craig C. Black, Director Editorial Staff; Hugh H. Genoways, Editor', Duane A. Schlitter, Associate Editor, Stephen L. Williams, Associate Editor, Teresa M. Bona, Technical Assistant . ® 1978 by the Trustees of Carnegie Institute, all rights reserved. CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE PITTSBURGH, PENNSYLVANIA 15213 CONTENTS Abstract 5 Introduction 5 Abbreviations and Usage 5 Stratigraphic Distribution 6 Systematics 10 Subfamily Schizotheriinae 10 Included Genera 10 Known Distribution 10 Revised Diagnosis 10 Discussion 10 Genus Moropus 11 Included Species 11 Known Distribution 11 Diagnosis 11 Discussion 11 Moropus eUilus Marsh 12 Lectotype 12 Paralectotype 12 Hypodigm 13 Diagnosis 13 Discussion 13 Description and Comments 13 Moropus distans Marsh 45 Lectotype 45 Paralectotype 45 Hypodigm 45 Diagnosis 47 Discussion 47 Description and Comments 47 Moropus oregonensis (Leidy) 48 Holotype 48 Hypodigm 48 Diagnosis 48 Discussion 48 Description and Comments 48 Moropus holla ndi Peterson 49 Holotype 49 Hypodigm 49 Additional Materia! 49 Diagnosis 49 Discussion 50 Description and Comments 50 Conclusions and Zoogeography 59 Acknowledgments 60 Literature Cited 60 ABSTRACT The subfamily Schizotheriinae. as represented by the genera Schizotherium, Borissiakia, Moropiis, Phyllotillon, and Ancy- lotlieriiim. is distinguished from the Chalicotheriinae by its more elongated, higher crowned molar teeth and less progressive foot structure. Of these genera, only Moropus ever reached North America, arriving in the medial or late Arikareean. Four late Arikareean/early Hemingfordian species of Moropus are recog- nized. Moropus distans . the type species, is represented by a few postcranial remains of small size. Based on comparison with other Moropus species, it is closely related to, and may be syn- onymous with, Moropus oregonensis . a species also from the late Arikareean of Oregon but represented only by dental re- mains. Moropus elatus and Moropus hollandi of the Great Plains include individuals near the large end of Moropus size range and are more closely related to one another than either is to M. distans or M. oregonensis. M. elatus is quite well known from numerous individuals and is an important basis for studies of intraspecific variation and interspecific and intergeneric com- parisons. M. hollandi is unique among Moropus species in hav- ing lost the trapezium from the manus. Moropus senex, repre- sented by a single phalanx from Oregon . is considered a nomen duhiutn. Moropus is known to have ranged from Oregon to Flor- ida in the late Arikareean and also occurs in the Aquitanian of France. Fragmentary, scattered remains are the worst barrier to an understanding of the early evolutionary history of this genus. INTRODUCTION Excavations by Carnegie Museum of Natural History expeditions at the Agate Spring Quarries in Sioux County, northwestern Nebraska, between 1904 and 1908 provided the major stimulus for the first comprehensive review of North American chalicotheres. In 1914, Holland and Peterson pub- lished a complete description of material of Moro- piis elatus collected by the Carnegie Museum of Natural History at Agate and included a compari- son with known remains of Moropus hollandi and other species of Moropus . No broad rediscussion of Moropus has appeared since. In 1970, I began a restudy of the Schizotheriinae (Coombs, 1973), with special emphasis on Hemingfordian and later North American relatives of M. elatus . This later material had been only peripherally studied by Holland and Peterson, and much new material had appeared. Some of the results of this study have been pub- lished (Coombs, 1974, 1975, 1976), but before a complete reevaluation of the later species will be meaningful, it is necessary to rediscuss the early Miocene North American species of Moropus using newer material and data. Planned funetional and zoogeographic considerations of Moropus also ne- cessitate this reevaluation. ABBREVIATIONS AND USAGE The following museum abbreviations are used with specimen numbers: AM — Pratt Museum. Amherst College, Amherst; AMNH — Department of Vertebrate Paleontology, American Museum of Natural History. New York; CM — Carnegie Museum of Natural History. Pittsburgh; F:AM — Frick American Mammals. American Museum of Natural History. New York; FMNH — Field Museum of Natural History. Chicago; UCMP — University of California. Museum of Paleontology, Berkeley; YPM — Peabody Museum of Natural History, Yale University, New Haven. Other abbreviations used are Me and Mt followed by a Roman numeral to indicate, respectively, metacarpal and metatarsal. The anatomical terms generally follow usage by Butler (1965). except where an element was not discussed by him; the termi- nology utilized by Holland and Peterson ( 1914) is often obsolete. Directional terminology applied to the skeleton is as follows: A. Dorsal-ventral as applied to the skull, vertebrae, and pel- vis; B. Anterior-posterior as applied to skull (including tooth- rows). vertebrae, pelvis, scapula, and long bones of fore and hind limbs; C. Dorsal-volar ( = anterior-posterior, dorsal-ventral, dorsal- plantar of several other authors) for carpals, metacarpals. tarsals, metatarsals, and phalanges; D. Radial-ulnar ( = medial-lateral) for carpals, metacarpals, and phalanges of the manus; E. Tibial-fibular ( = medial-lateral) for tarsals. metatarsals, and phalanges of the pes; and F. Labial-lingual in reference to the teeth. The three phalanges of each digit are named, respectively, prox- imal. medial, and ungual (the latter is the fissured claw); this designation is to avoid the confusion with digit numbers (I-V), which results when the phalanges are also given numbers. Fused proximal and medial phalanges of digit II of the manus and oc- casionally also of digit II of the pes are termed duplexes follow- ing the usage of Holland and Peterson (1914). 5 6 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 STRATIGRAPHIC DISTRIBUTION Skinner's report (1968) correlating chalicotheriid occurrences in North America provided an impor- tant basis on which the present study could build. Nevertheless, new chalicothere finds, discovery of additional specimens in old collections, further data concerning the location or correlation of old local- ities. and the greater knowledge acquired through further taxonomic study all make necessary a re- discussion of stratigraphic distribution. A map of the localities discussed is provided in Fig. 1. Lo- calities yielding chalicotheres of medial Hemingfor- dian or later age and of uncertain age will be dis- cussed in later papers. Although chalicothere remains can give some idea of the correlation of deposits, they are usually too rare and too incom- plete to be of much practical value in biostratigra- phy. John Day Fonnatkm . — Skinner ( 1968: 12-13) sug- gested that Moropus distans and M. sene.w species established by Marsh from the John Day Basin, could be from the Rattlesnake Formation or Mas- call Formation in the region, as well as from the John Day Formation. Some additional data, which suggest that at least some of Marsh's material is from the John Day Formation, are available. Specimen labels on Moropus material from the John Day Basin give the following general locality information; DYPM 12193a. holotype of Moropus distans — “Bridge Creek beds'' (Marsh specimens included in the hypodigm have the same collecting data); 2) YPM 12194, holotype of M. senex — “Day- ville?, Oregon''; 3) YPM 10030. holotype of M. or- egonensis — “Upper John Day beds. Bridge Creek. Wasco County, Oregon," collected by Condon (ad- ditional specimens from the same locality cata- logued as YPM 10030a, 10030b); 4) AMNH 7259, referred by Holland and Peterson (1914) to M. or- egonensis — “near Antelope Springs, Oregon." Al- though none of this data is precise enough to sug- gest exact locations, at least the “Bridge Creek beds" included important early collecting areas in the John Day Formation. The size and morphology of the specimens is near that hypothesized for early Miocene Moropus immigrants to North America and may suggest a relatively early age for the John Day chalicotneres. However, the designation of “Upper John Day beds" for the holotype of M. oregoncnsis signifies no more than that the speci- men probably came from the John Day Formation. As Hay ( I J63). Rensberger ( 1971 , 1973), and Fisher and Rensberger ( 1972) have pointed out, the bound- aries as designated by Merriam (1901) between members of the formation are not synchronous everywhere in the area. Fisher and Rensberger ( 1972) recently advocated a four-fold division of the John Day Formation; the divisions between mem- bers cross biostratigraphic boundaries, especially the division between the Turtle Cove Member and Kimberly Member. Fisher and Rensberger (1972:9) also pointed out the difficulties of using adherent matrix as a criterion for assigning formerly collect- ed museum specimens to members of the John Day Formation. Further information concerning distribution of chalicothere remains in the John Day Formation may be forthcoming. In the John Day stratigraphic collection being studied by Rensberger is a small chalicothere phalanx. Rensberger, who intends to describe the occurrence along with the remainder of his perissodactyl collection, stated (personal communication, 1972) that the specimen came from a stratigraphic level, which represents an early or possibly medial Arikareean age. Although some of the earlier collected Moropus material from the John Day Basin may be younger than this, John Day chalicotheres are clearly among the earliest known representatives of Moropus in North Amer- ica. Harrison Formation and Upper Harrison For- mation ( = Marsland Formation sensu McKenna, 1965), Nebraska and Wyoming. — The highest fre- quency of Moropus specimens, including material of both M. elatus and M. hollandi , occurs in these two formations. Recently, Robert M. Hunt of the University of Nebraska has restudied the type areas of both formations and traced the contact between them to important fossil localities. One of the most far-reaching results of his study has been the con- clusion that the Harrison-Upper Harrison contact lies below the Agate Spring Quarries and that there- fore the well-known fossil assemblage from these quarries comes from the Upper Harrison Formation (Hunt, personal communication, 1976). By Hunt's reinterpretation, the fauna associated with the Har- rison Formation is relatively sparsely known, but it does include isolated specimens of ehalicotheres probably referable ioM. elatus. Chalicotheres from the Agate Spring Quarries are in the present paper attributed to the Upper Harrison Formation, in ac- cordance with Hunt's findings. 1978 COOMBS— REEVALUATION OF MOROPUS 1 Fig. I. — Map of Arikareean-early Hemingfordian Moropus localities discussed in this paper: 1) Caliente Formation. California; 2) Dayville, Oregon (type of "M. senex")\ 3) Bridge Creek, near Mitchell, Oregon (types of M. Jistans and M. oref>onensis)\ 4) near Antelope Springs, Oregon; 5) Lemhi River Valley, 46 mi southeast of Salmon. Idaho; 6) 7 mi south of Chugwater. Platte Co.. Wyoming; 7) Big Muddy Creek. Wyoming; 8) Jay Em. Goshen Co.. Wyoming; 9) near Van Tassell. Niobrara Co., Wyoming; 10) western Sioux County along Niobrara River, Nebraska (type of M. holltincli): II) Agate Spring Quarries and American Museum-Cook Quarry, Sioux Co., Nebraska; 12) Morava Ranch Quarry and surrounding area. Box Butte Co., Nebraska (type of M. elatusl)\ 13) Buda local fauna, Alachua Co., Florida. One important problem in discussions of Moro- pus elatus Marsh has concerned the geographic and stratigraphic position of its type locality in relation to strata exposed in the Agate Spring Quarries. Skinner (1968, Table 1, footnote a) correctly stated that Marsh’s type of M. elatus, if from the locality suggested by Holland and Peterson (1914:226), could have come from one of several formations exposed near the mouth of Whistle Creek in Sioux County. Nebraska. However, Matthew ( 1929:520, footnote 1) had differed from Holland and Peter- son’s estimation of the locality in concluding that the type of M. elatus was actually found some 18 mi east of Agate, Nebraska, a site well east of that proposed by Holland and Peterson. Matthew’s reason for proposing that the type lo- cality of Moropus elatus was farther east was based on some surface exploration and digging conducted by Harold J. Cook and communicated by letter to Osborn in 1917 and 1918 and to Matthew in 1927. Cook had concluded that the Agate Quarry vicinity was too distant to have been easily visited from the stage road presumed to have been followed by Marsh’s collector Hank Clifford en route between Sidney. Nebraska, and the Red Cloud Agency. Cook had discovered a fossil pocket, which ap- peared to have been previously worked, relatively close to the stage road and about 18 mi east of Ag- ate. Cook collected at this spot some chalicothere material, which he sent to the American Museum for study. Subsequently the specimens were ex- amined by Richard S. Lull and compared with the type material of M. elatus at Yale University. Lull concluded (personal communication to Osborn. 1917) that Cook’s material was similar in color, preservation, and general appearance to the type. 8 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 although no actual break contacts could be found. Unfortunately, I have been unable to locate Cook’s specimens for further comparison. Cook asserted (letter, 1927, quoted in Matthew, 1929:520, footnote 1) that the pocket in which he found the bones “is located in the lower part of the Upper Harrison beds, whereas the Agate Springs Fossil Quarries and the Moropus obtained there are some sixty feet below the top of the Lower Harri- son beds.’’ On this basis Matthew (same footnote) stated that the Upper Harrison species M. hollandi was probably synonymous with M. eUitus and that the Agate Quarry material should be referable to a different species, M. cooki Barbour. Such a conclu- sion is not. however, borne out by specimen mor- phology. Since Cook’s excavation and possible discovery of Clifford’s locality, the area between 16 and 20 mi east of Agate along the Niobrara River has been studied by parties of the Frick Laboratory and of the University of Nebraska. The area is complexly channeled and several formations are represented, including the type section of the Runningwater For- mation (Cook. 1965). Cook’s description of his old diggings stated, “It is really the nearest command- ing hill to that old road crossing [Niobrara crossing], so it seems a particularly probable place for him [Clifford] to have found the type specimens [of Moropus elatiis] figured and described by Marsh” (letter of 21 November 1917. to Osborn, bracketed material mine). Cook described the ma- trix as “mostly very hard around the bones, so that collecting is rather slow .... As the slope is largely grassed over, it is small wonder that bones at this spot should have missed recent attention by people living around there .... The color of the matrix varies a good deal locally, and the deposits are quite ’pockety,’ due to channel beds, and shifting stream depositions” (H. J. Cook, letter to Osborn, 20 De- cember 1917). These descriptions coincide very closely with the location and lithology of Morava Ranch Quarry, so it seems likely that Cook’s lo- cality was at or very near the site of the later quar- ry, excavated by Ted Galusha in 1940 and by me in 1975. It may also be at or near the original Clifford locality. Moropus remains from Morava Ranch Quarry agree quite well in color (bluish gray) and preservation with the type material of M. elatus. Detailed location of the Morava Ranch Quarry is available in FrickiAmerican Museum records. As mapped in Cook (1965:3), it lies in a Harrison For- mation channel, but this assessment may require reevaluation as study of the area progresses. Mor- phologic evidence (below) suggests that Agate and Morava Ranch Quarry Moropus are referable to M. elatus and that Moropus hollandi is a separate but closely related species. The Cook Collection at the American Museum of Natural History contains a number of chalicothere remains, primarily teeth and foot elements, from the Agate Quarries and from north and east of Ag- ate. Locality data for some of this material is not sufficient for placement in the Harrison or Upper Harrison Formation. Some material is from the American Museum-Cook Quarry, 2 mi north of the Agate Quarries. This latter locality was discussed in detail by Hunt (1972:35-37) and on the basis of data presented by him is almost certainly the place where American Museum parties collected an ex- cellent skull, AMNH 10,645, referred to M. elatus. Hunt (personal communication, 1976) now places the American Museum-Cook Quarry in the base of the Upper Harrison Formation. AMNH 13,765, an upper jaw with P^-M^ was collected by Olcott in 1907 at the “top [of] Lower Harrison beds. Van Tassel, Wyoming” (specimen field label by Olcott). This specimen yields very lit- tle information of specific taxonomic value and is regarded here as Moropus indet. Locality data are not precise enough to be certain that it comes from the Harrison Formation. A thick section of the Upper Harrison Formation is exposed along the Niobrara River near the Ne- braska-Wyoming state line. It was here that Peter- son (1907u, 1909) designated the type section of the Upper Harrison beds. McKenna (1965) considered Peterson’s section to be the type section of the Marsland Formation and used “the term Marsland [proposed by Schultz, 1938] as an objective syn- onym (at the rank of a formation) of the term Upper Harrison beds” (1965:10, bracketed material mine). In the Upper Harrison faunal list from this area, Peterson (1907a:56) included Moropus lelatus Marsh, later in the paper (1907a:60) identified as CM 1424, found “near the base of the Upper Har- rison beds on the Niobrara in Sioux County, Ne- braska, in 1901.” Peterson later (1913) made this specimen the type of a new species, M. hollandi. Holland and Peterson (1914:232) further specified the locality as “near Wyoming state line,” and the Carnegie Museum catalogue lists further, “near Vantassel.” No additional locality data is available. Hunt (personal communication) has found addition- al Moropus material at Harper Quarry, just above 1978 COOMBS— REEVALUATION OE MOROPUS 9 the Harrison-Upper Harrison contact near the state line. Additionally. Skinner (1968) mentioned deposits yielding chalicotheres from the Jay Em district, Goshen County, east-central Wyoming, and from 7 mi south of Chugwater. Platte County, south- eastern Wyoming. He considered them biostrati- graphically equivalent to the Upper Harrison. Fur- ther, Riggs collected Moropus material, now at the Field Museum of Natural History, from beds pre- sumably also Upper Harrison equivalents near Jay Em, Wyoming. Where diagnostic specimens are present, the Jay Em and Chugwater chalicotheres seem to be referable to M. hollandi. It is not pres- ently clear to what extent the difference between M. elatus and M. hollandi represents an age and/or ecologic distinction between the faunas containing each. Correspondingly, the boundary between Ari- kareean and Hemingfordian Moro/7«5 -containing faunas needs redefinition. Other Arikareean or early Hemingfordian occur- rences. — Among the specimens collected by Am- herst College expeditions of 1907-1908 at Big Mud- dy Creek, Wyoming, were several chalicothere phalanges. Locality and stratigraphic data with the material are inexact, and the Amherst College cata- logue lists Muddy Creek specimens from "Lower Rosebud,” "Lower Harrison,” and "Upper Har- rison” beds. The basis on which these distinctions were made is unclear. Loomis (1909, 1911) pub- lished on turtles and camels from Muddy Creek and concluded that they were from "Upper Harrison beds.” The type of Testndo brevisterna , he stated, was found near a skeleton of Merychyus minimus and was thought to be "Upper Harrison” on that basis (Loomis, 191 1). However, Schultz and Fal- kenbach (1947) mentioned no representatives of Merychyus from Muddy Creek; they later (1949) listed the Monroe Creek genera Mesoreodon and Merycoides from Muddy Creek. McKenna and Love (1972) reevaluated Oxydactylus gibbi and Pro- tome ryx leonardi, camel species named by Loomis (191 1) from Muddy Creek. They referred both spe- cies to Miotylopus gibbi (Loomis) and considered it probable that the type specimens came from be- neath the Harrison Formation. The Muddy Creek area and Yale and Amherst College collections de- rived from it need serious restudy. Monroe Creek equivalent beds clearly seem to be represented, but Upper Harrison equivalents may be present also. The chalicothere phalanges from Muddy Creek are generally similar to and within the size range of Agate Quarry Moropus elatus. They are probably not useful for correlation, but they may be among the earliest known chalicotheriids in North Ameri- ca. Patton (1967:8) mentioned the presence of a chal- icothere in the Buda local fauna, near Newberry, Alachua County, Florida. Skinner ( 1968) gave a personal communication from Patton, which further identified the remains, whereas Patton and Webb (1970) described it as a "dwarf version of the large chalicothere . . . common in western faunas.” Pat- ton and Taylor (1971:128) considered the fauna to be earliest Hemingfordian, but Rich and Patton (1975:695) subsequently suggested an Arikareean age. Patton intends to discuss the chalicothere at a future date. It should be noted in regard to the Buda specimen that specimens of small chalicotheres are also known from the John Day Formation in Ore- gon. Repenning and Vedder ( 196LC-237) reported a chalicotheriid in an Arikareean assemblage from continental deposits in the Caliente Formation of the eastern Caliente Range, California, between their sections 2 and 3 (1961:C-236) and about 3,000 ft above the base of their section. Repenning (per- sonal communication) identified the remains as the proximal one-half of a rather small duplex bone (fused proximal and median phalanges), abraded and badly gnawed by rodents, and a ?chalicothere tooth fragment. Listed in the same fauna (I96LC- 237) were Oxydactylus brachyodontus [7=Para- tylopus cameloides (Wortman)] and Parablasto- meryx aff. P. falkenbachi Frick, both restricted elsewhere to the late Arikareean and early Hem- ingfordian. Geology of this area has recently been detailed by Woodburne (1975). Skinner ( 1968:18) mentioned a Moropus phalanx, F:AM 54,900, collected at the upper end of the Lemhi River Valley, about 46 mi southeast of the town of Salmon, Idaho. The age of the deposits was unknown but considered possibly biostratigraphi- cally equivalent to the Marsland Formation (sensu McKenna, 1965) of the Great Plains. Morphology of the phalanx is not useful taxonomically and does not contribute any further information to a corre- lation. 10 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 SYSTEMATICS Class Mammalia Order Perissodactyla Superfamily Chalicotherioidea Gill, 1872 Family Chalicotheriidae Gill, 1872 Subfamily Schizotheriinae Holland and Peterson, 1914 Included Genera Schizotherimn Gervais, 1876; Ancylotlieriiim Gaudry, 1862; Moropus Marsh, 1877; Phyllotillon Pilgrim, 1910; and Borrissiakia Butler, 1965. Known Distribution Oligocene-Miocene of Eurasia, Miocene of North America, Pleistocene of Africa (usage of Miocene as in Berggren and Van Couvering, 1974). Revised Diagnosis Dental formula 10/1-3, C 0/0, P 3/3, M 3/3 (an- terior dentition not verified in all genera; see Coombs, in press); molar teeth higher crowned than in the Chalicotheriinae, with molars slightly to strongly elongated; protoloph on upper molars com- plete on unworn teeth; ectoloph on upper molars less strongly slanted than in the Chalicotheriinae (paracone and metacone in labial half of tooth); metastylid always separate from metaconid on low- er molars; jaw symphysis typically shorter and di- astema relatively longer than in the Chalicotheri- inae; Mt III longest metatarsal or Mt III and Mt IV subequal in length; Me III longest metacarpal; fore- limbs versus hindlimbs not so disproportionate in length as in the Chalicotheriinae; proximal pha- langes more symmetrical than in the Chalicotheri- inae and ungual phalanges less transversely com- pressed. Discussion The present-day classification of the Chalicothe- rioidea, which gives the early (primarily Eocene) chalicotheres ( = Eomoropidae) equal rank with the later, unmistakably clawed group ( = Chalicotheri- idae), originated with Matthew (1929). Within the latter group he distinguished lineages having high- crowned and low-crowned molar teeth, now des- ignated the Schizotheriinae and Chalicotheriinae. He did not, however, name his subgroups or give full family status to the Eomoropidae and Chali- cotheriidae. These refinements evolved gradually in the works of later authors. The accepted subfamilies Chalicotheriinae and Schizotheriinae show better than any other simple taxonomic scheme the evolutionary relationships within the Chalicotheriidae. Particularly, they ex- press the strong morphological and ecological di- vergence of Chalicotheriiim and closely related Nestoritherium from the remainder of the Chali- cotheriidae. These Chalicotheriinae, while remain- ing generally conservative in their dental evolution, attained the most specialized limb structure, evi- dent in even the earliest known representatives (Chalicotheriiun pilgrimi, C. rusingense). Among these unusual limb features are hindlimbs much shorter than forelimbs, loss of trapezium and Me V, lunate displaced almost entirely onto the mag- num, increase in length from Me II through Me IV and from Mt II through Mt IV, astragalus articulat- ing with both navicular and cuboid, asymmetrical proximal phalanges with metapodial facet parallel to dorsal surface, and ungual phalanges strongly compressed transversely. Improved knowledge of the Chalicotheriinae is available in Butler (1965) and Schaefer and Zapfe (1971). Schizotheriines remain more conservatively perissodactyl-like in their foot structure but evolve more rapidly in their dentitions by increasing the crown height and length of mo- lars, probably as a selective response to coarser diet. Schizotheriines are often associated with sa- vannah faunas, whereas Chalicotheriinae are more often found with dense woodland assemblages. Problems exist, however, in making a clear dis- tinction between Chalicotheriinae and Schizotheri- inae. Relations of the primitive Oligocene schizo- theriine genus Schizotherium to the Chalicotheriinae are unclear. Schizotherium lacks the foot special- izations of chalicotheriines, but its molars are not so elongated or high crowned as in other schizo- theriines, although several species trend in this di- rection. More complete knowledge of Schizother- iiim should settle some of these questions. Another small difficulty arises because certain schizotheri- ines parallel the Chalicotheriinae in a few aspects of foot structure, although no schizotheriine attains the derived state seen in even the most primitive known chalicotheriines. Several schizotheriine spe- cies independently lost the trapezium; some Ancy- lotherium lost Me V; the tall astragalus Borissi- akia articulated with the cuboid as well as with the navicular; metatarsal length was gradually reduced in schizotheriine evolution (see Coombs. 1974). On the other hand, a few primitive species of Chali- 1978 COOMBS— REEVALUATION OF MOROPUS 11 cotherium have a metastylid on lower molars, whereas most other chalicotheriines have a small metastylid or none. Parallelism and problematic primitive genera and species are not uncommon phenomena in taxonomic work, however, and in this case provide no special barrier to the accep- tance of two chalicotheriid subfamilies. The Chalicotheriidae made its first known ap- pearance in North America during the Arikareean Land Mammal Age. Only the Schizotheriinae ever appeared in North America. All North American chalicotheriids could have been derived from a sin- gle or several Arikareean immigrants belonging to the genus Moropus . Genus Moropus Marsh Moropus Marsh, 1877:249. Type species . — Moropus distans Marsh, 1877. Included Species Moropus distans; M. oregonensis (Leidy, 1873); M. elatus Marsh, 1877; M. Iwllandi Peterson, 1913; M. matthewi Holland and Peterson, 1914; M. rner- riami Holland and Peterson, 1914; and additional presently undescribed species. Known Distribution Arikareean-Barstovian and possibly later faunas of North America (see Skinner, 1968, and Coombs, 1973, for later generically questionable material); Aquitanian-? of Eurasia. Diagnosis 1) Dental formula I 0/3, C 0/0, P 3/3, M 3/3; 2) molars intermediate in proportionate length and crown height between those of Schizotherium and Ancylotheriunr, 3) absence of a) crochet, b) labial rib on ectoloph between mesostyle and metastyle, and c) accessory cuspules on M^ posterolingual to hypocone; 4) no hypoconulid on M3; 5) frontals and parietals of skull without dorsal expansion; 6) Me V present; 7) scaphoid never contacting Me II, even during extreme carpal flexion; 8) volar process on lunate better developed than in Ancylotherium: 9) no dorsal flattening of metacarpals; 10) asymmet- rical astragalus; 1 1) astragalus articulating distally only with the navicular; 12) metatarsals proportion- ately shorter than in Schizotherium and Borissiakia; 13) Mt III and Mt IV subequal in length; 14) Mt IV having ectocuneiform facet; and 15) proximal and medial phalanges of digit II of the manus fused to form a duplex in all except very young individuals. Discussion Moropus differs from Schizotherium in charac- ters 2, 4, 12, and 15; from Borissiakia in characters 3b, 3c, 11, 12, and 15; from Phyllotillon in charac- ters 3a and 3b (but a crochet and posterior ectoloph rib are present only rarely \n Phyllotillon)', and from Ancylotherium in characters 1 (incisors only), 2, 3a, 3b, 6, 7, 8, 9, 13, and 14. Moropus also differs from Borissiakia, Ancylotherium , and Schizotherium tur- gaicum in the retention of a trapezium in the manus in all species except Moropus hollandi ; this bone was lost independently in several schizotheriine lin- eages. Moropus, Wkt Ancylotherium , has a tenden- cy to fuse the proximal and medial phalanges of digit II of the pes to form a smaller, more symmet- rical duplex than that belonging to the manus. The species included in each of the Old World genera discussed above follows the treatment by Coombs (1974). Two new schizotheriine genera have recently been named — Huanghotherium Tung et al. (1975) and Gansuodon Wu and Chen (1976). Both genera are based on upper molars and differ from Moropus in the following characters: larger size; greater crown height; smaller hypocone; more prominent, ridge-like cingulum lingual to the protocone. Gan- suodon has a large crochet on M'^, and Huangho- therium has a trace of one on M^. All of the pre- ceding differences from Moropus are shared similarities among Gansuodon, Huanghotherium , and Ancylotherium. Ancylotherium pentelicum , as figured by Thenius ( 1953), is almost as large as the type of Gansuodon and resembles it closely. The large M® of Gansuodon compared to M^ is probably not a valid difference from Ancylotherium, for M^ and M® of the type of Gansuodon may not be the same individual. Huanghotherium was distin- guished especially by its tall ectoloph. whose height on M^ exceeds the width of the tooth. Ancylother- ium also has very high-crowned molars, but possi- bly not so tall as in Huanghotherium ; further com- parison is necessary. Wide intraspecific ranges of variation are common in schizotheriines, and the Chinese material may in future best be viewed as species within Ancylotherium . It is not clear how Gansuodon and Huanghotherium might be related to probable Ancylotherium postcranials from China already figured by Bohlin (1936; an astragalus) and Colbert (1934, Figs. 13a, d, f; an Mt II and pha- langes). North American schizotheriines from the Arika- 12 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 2. — Mounted skeleton of Moropus elatus, AMNH 14,375, as reconstructed by Osborn (1919). Courtesy of the American Museum of Natural History. reean onward form a distinct radiation, probably separate from Old World forms. A European Aqui- tanian representative of Moropus is also recognized (Moropus sp., Coombs, 1974). Both Schizotherium and Borissiakia are easily distinguished from Mor- opus, but there are problems in distinguishing Old World representatives of Moropus from Phyllotillon (known from limited material from the Burdigalian of Baluchistan and possibly later from Europe) and from Ancylotherium (Metaschizotherium). Moropus elatus Marsh Moropus elatus Marsh, 1877:250. Moropus cooki Barbour, 1908:215. Moropus petersoni Holland, 1908:810. Moropus parvus Barbour, 1909:222. Moropus elatus: Peterson, l907/?:733; Holland and Peterson, 1914:222. Moropus petersoni: Holland and Peterson, 1914:226. Moropus cooki: Osborn, 1917:131. Moropus cookei: Osborn, 1919:251, Fig. I. Moropus cooki: Matthew, 1929:520; von Koenigswald, 1932:22. Moropus elatus: von Koenigswald, 1932:22. Moropus elatus, M. petersoni, M. parvus, M. cooki: Colbert, 1935:13; Belyaeva, 1954:49. Moropus elatus: Coombs, 1974:275; Coombs, 1975;55. Lectotype Mt II of YPM 13,081 (Eig. 18), collected by Hank Clifford in Nebraska, probably near the Niobrara River (see discussion of locality above), ?Harrison Eormation, ?late Arikareean. Paralectotypes YPM 24,63 la-d, part of duplex of digit II manus, proximal phalanx, tuber of calcaneum, and proxi- mal end of Mt III (Fig. 19). The paralectotypes were figured as part of the type by Holland and Peterson (1914:223-224). 1978 COOMBS— REEVALUATION OE MOROPUS 13 Hypodigm YPM 24,632, cuneiform, patella, and distal end of Me II (figured as part of the type collection by Holland and Peterson but not mentioned by Marsh, 1877); a large quantity of material, including com- plete skeletons, from the Agate Spring Quarries, Sioux County, Nebraska, in the Carnegie Museum of Natural History. American Museum of Natural History, University of Nebraska State Museum, and other museums; extensive dental and postcra- nial material (but no complete skeletons) from Mo- rava Ranch Quarry, 18 mi east of Agate, Box Butte County, Nebraska, in the Prick Collection of the American Museum and the Pratt Museum, Amherst College; and additional fragmentary specimens from various localities in northern Sioux County, Nebraska. All referred specimens are from the Har- rison Formation or Upper Harrison Formation (see above). Diagnosis 1) Chalicotheres at the middle to large end of M or opus size range; 2) upper molars more propor- tionately elongate than in M. oregonensis', 3) labial metaloph origin on unworn M^ very near mesostyle; 4) no lingual cingula on lower molars; 5) trapezium present and well developed; 6) calcaneum with strong extension of narrow ectal facet onto tuber calcis; 7) navicular facet on tibial surface of cuboid without any proximal extension; 8) Mt II with ec- tocuneiform facet having primarily fibular orienta- tion and an oblique tibial ridge present on the prox- imal part of the shaft; and 9) proximal and medial phalanges of digit II of the pes only occasionally fused (approximately 10% of cases). Discussion Of the Moropus species discussed in this paper, M or opus elatus differs from M. distaus in charac- ters 1 , 7, and probably 9, from M. oregonensis in characters 1 and 2, and from M. hollandi in char- acter 5. Further, M. elatus has, on average, pro- portionately longer metatarsals than are known for M. hollandi (Table 6). Most of the other diagnostic features mentioned differentiate M. elatus from more advanced Moropus species like M. merrianii. Reasons for synonymizing Moropus petersoni Holland, 1908 (including its junior synonym M. par- vus Barbour, 1909) with M. elatus were detailed by Coombs (1975). Individuals previously referred to M. petersoni are probably females of M. elatus. Morphologic comparisons between the type of M. elatus and Agate material (below) provide evidence that Moropus cooki Barbour, 1908, is also a junior synonym of M. elatus . 1 have not been able to locate the large scapula (CM 1776) described by Holland and Peterson (1914;230, 332, Fig. 77) as Moropus (?) nuiximus . This specimen, from the Agate Spring Quarries, is as figured different from all known scapulae of Mor- opus. despite the large amount of material and vari- ation known. It is here thought to be either an aberrant specimen of M. elatus or does not belong to a chalicothere. M . (?) nuiximus must be consid- ered a noinen duhium . Moropus elatus is the most completely known of all chalicothere species. Agate material collected by the American Museum of Natural History after Holland and Peterson’s (1914) monograph, as well as more recently collected specimens from other localities, provide additional information concern- ing intraspecific variation and other aspects of anat- omy. For example, an M. elatus edentulous pre- maxilla, AMNH 1 1,321, verifies the ruminant-like vegetation cropping mechanism in this species (Coombs, 1978). Such new data are discussed in the present paper with a brief update of morphology already detailed by Holland and Peterson. This re- view will be especially useful for comparison with other Moropus species. No attempt at muscular re- construction or detailed functional analysis is made here; these topics are reserved for a separate paper. Description and Comments Dentition Interpreting specific taxonomic relationship and phylogeny on the basis of dentitions is very difficult within the genus Moropus . This difficulty is unfor- tunate, because in most cases teeth are better rep- resented in the fossil record than limb elements. Teeth show a great amount of variation in mor- phology and relative proportions within a single population, in many cases combined with the lack of substantive changes in the teeth between popu- lations, even over a long period of time. Yet changes in the limbs of the same animals have been much greater. It is often impossible to identify an isolated tooth or even several associated teeth be- low the generic level. All things considered, how- ever, it does appear that a few regular changes in dentition occurred as North American chalico- theres evolved. Some of these changes are in pro- 14 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 3. — Top: right lateral view of skull of AMNH 11,322. Bottom: lateral view of left mandibular ramus of AMNH 14,427 (coronoid process broken). Both are Moropiis elatiis from the Agate Spring Quarries, Sioux Co., Nebraska. portions, others in morphology, often in one part of the toothrow but not in another. Particularly im- portant may be subtle changes in tooth wear (that is, which areas of a tooth are worn first and along what plane). At the same time, tooth wear is often a complicating factor because dentitions that would be similar at the same stage of wear may appear quite different at other stages of wear. Eor the most part treatment of North American chalicothere den- titions has been inadequate, even when specimens were well known (see Holland and Peterson, 1914:245-250), but several European authors (Bor- issiak, 1946; Butler, 1965; and others) have given good accounts of variation among teeth of a single species. Possibly the lack of early treatment of den- titions of North American forms was a result of the problems created by intraspecific variation. The ter- minology used in this section is diagrammed in Fig. 4; upper and lower teeth of M. elatus are shown in Fig. 3 and 5. Upper Molars M^. — This tooth is the least worn of the molars in any given toothrow and is easily recognized by the shape of its posterior part. It is approximately the same size as M^, sometimes slightly longer or shorter; M^ is in some cases relatively shortened because of wear. Both teeth are much larger than M' (see Table 1). In the progression from M* to M^ the anterior V of the W-shaped ectoloph becomes progressively longer compared to the posterior V. The change in proportions is particularly visible be- tween M^ and M®, for in the latter the ectoloph pos- terior to the mesostyle is not V-shaped but curves 1978 COOMBS— REEVALUATION OF MOROPUS 15 Fig. 4. — Diagrammatic Table 1. — Lengths® and widths'^ (in mm) of upper cheek teeth of Moropus elatus. M. hollandi . and M. oregonensis. Measurements CM 2I03(R)® AM 9923(D® AMNH 1 1.322(R)® AMNH 14,427(R)‘-- FMNH PI2094(L)“ YPM I0,030a.b‘- YPM 10,030^' Labial length P^ 21.3 21.9 20.1 20.5 18.8 Width P^ 21.0 20.6 20.3 16.2 16.0 Labial length P^ 23.1 23.2 23.2 21.1 19.4 17.4 Lingual length P^ 21.8 21.7 21.4 18.0 18.8 15.3 Width P® 28.3 29.1 26.9 26.0 24.4 20.3 Labial length P* 26.0 25.8 23.2 25.3 22.5 19.1 18.2 Lingual length 24.0 24.2 22.3 22.2 20.3 16.5 17.7 Width P^ 32.6 32.1 29.9 29.8 26.6 23.1 23.5 Labial length M‘ 41.2 35.8 35.4 37.4 35.7 25.9 Width M‘ 35.2 37.6 35.9 32.9 31.7 21.4 Labial length M^ 55.9 45.7 53.0 55.3 49.6 36.5 Width M^ 42.9 44.7 42.0 40.6 40.8 26.2 Labial length M^ 59.8 54.0 51.6 54.4 48.3 36.5 Width M^ 48.6 51.0 42.6 44.1 43.2 33.8 Length premolar row 68.5 69.3 62.7 62.4 58.1 Length molar row 144.3 131.5 128.3 136.2 130.1 Length premolar row/length molar row 0.47 0.53 0.49 0.46 0.45 ® Lengths are maximum anterior to posterior dimensions. '' Widths are maximum labial to lingual dimensions, including mesostyle on molars. '■ M. elatus from the Agate Spring Quarries and Morava Ranch Quarry (AM 9923). M. hollandi from near Jay Em, Wyoming. ® Af. oregonensis: YPM 10,030a = left F-P, right M-*; YPM 10,030b = left M‘-M^; YPM 10,030 = holotype right F. 16 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Fig. 5.— Occlusal views of cheek teeth of M. elatiis: A) left P^-M^ of AMNH 11.322; B) left of AM 9564; C) left of AM 9922; D) left P2- AMNH 14.427. A and D from Agate Spring Quarries. Sioux Co.. Nebraska. B and C from Morava Ranch Quarry. Box Butte Co., Nebraska. 1978 COOMBS— REEVALUATION OF MOROPUS 17 posterolingually. The ectoloph of is never worn posterior to the metacone. Correlated with the shortening of the posterior ectoloph, M'^ is trans- versely narrow in its posterior part. All of these features are common to other schizotheriine chali- cotheres. On M® both parastyle and mesostyle are very strongly developed, whereas the metastyle is com- paratively weak. The paracone is usually taller than, and labial to, the metacone. On many specimens a weak labial rib passes from the base of the ectoloph to the tip of the paracone; there is in no case such a rib passing to the metacone. Development of the protoloph varies among specimens, and its struc- ture is rapidly obscured by wear. In the least worn specimens a protoconule is developed on the pro- toloph just lingual to the paracone. In most unworn specimens the protoloph is continuous all the way to the protocone but somewhat depressed lingual to the protoconule. In others, slightly more worn, the protoloph is incomplete and there is no connection between protocone and protoconule. Where the protoloph is complete, it curves posterolingually from the protoconule to join the protocone’s ante- rior margin. The protocone is a broad-based, coni- cal cusp with a blunt-pointed tip, which lies barely in the anterior half of the tooth, posterior to the paracone and protoconule. The tip of the protocone is worn only after the lingual surface of the ectoloph is well worn and in the majority of specimens is untouched. Even where the protocone is not still connected to the protoloph, a weak ridge passes anteriorly from its tip. The anatomy of the metaloph of M^ is fairly uni- form, the primary differences resulting from vary- ing degrees of wear. No metaconule is developed in any of the specimens, and the hypocone forms the lingual end of the metaloph, its tip intimately connected to and about the same height as the crest. In a slightly worn specimen the labial origin of the metaloph from the ectoloph is close to the meso- style (less than 10 mm). In an older individual, at- trition has worn the mesostyle to increase the ap- parent angle of the fold and has moved the apparent metaloph origin posteriorly so that there is a greater distance between metaloph origin and mesostyle. In the most worn specimens, the original tip of the metacone has been worn off and the apparent meta- cone is positioned posterior to the original one. This change is because of the attrition of the ectoloph anterior to the metacone but not posterior to it; the ectoloph crest then passes directly posteriorly or even posterolabially rather than posterolingually from the repositioned metacone. Bordered by the lophs and cusps of M'^ are three valleys, two of which increase in size somewhat during wear. The largest of these is the central val- ley, lingual to the ectoloph between paracone and metacone, anterior to the metaloph and skirting the posterior base of the protocone. This furrow is open lingually between protocone and metaloph. where it forms a notch in the tooth’s lingual outline; in worn specimens there may also he an anterior open- ing across the worn-down protoloph. Between the protoloph and anterior cingulum is a shallower val- ley. A third small valley, or postfossette, which de- creases in size as wear progresses, is bordered by metaloph and ectoloph between hypocone and metastyle. No crochet, as described by Butler (1965:178) for Chalicotheriuin rusingense , is pres- ent in any specimen, although there is sometimes the hint of a very small crista. Development of cingula varies, but in each spec- imen a very weak labial cingulum is present. The anterior cingulum is strong and forms a border for the anterior valley, curving posterolingually to skirt the protocone. Near the protocone it is variable. It may be notched anterior to the protocone to create an anterolingual exit for the central valley where the protoloph is worn down or very low; it may thicken and almost merge with the lingual wall of the protocone, sending off a branch to merge with the protoloph or with the small anterior ridge from the protocone tip; or it may retain its ridge-like cin- gulum character as it skirts the protocone lingually. In all cases it has some (though variable) ridge-like character along the posterolingual part of the pro- tocone; it may then continue weakly to merge with the hypocone or disappear at the large lingual exit of the central valley lit is never so thickened as in Ancylotherium (Ancylotherium ) pentelicum]. A posterior moderate to well-developed cingulum connects hypocone and metastyle and partially or completely closes off the posterior opening of the postfossette [similar io Ancylotherium (Metaschizo- therium) fraasi but not as high as in A. (A.) pen- telicum]. In those specimens where roots are visible there are three present — two small labial roots and a large anteroposteriorly elongated lingual one. The lingual root sends a connecting crest to the small postero- labial root. and M\ — Both these teeth are similar mor- phologically to M'^. In most specimens they are too 18 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 worn to give good evidence of structural details, but occasionally (CM I707A, see Holland and Pe- terson, I914:pl. 51) an unworn M' is preserved in association with a deciduous dentition. Where M‘ is unworn it appears larger than it does in a mature toothrow between and M^, where anterior and posterior margins of the tooth have been removed by wear. Prior to wear it is clear that a protoconule and complete protoloph are present, although these are quickly worn away, but the protoloph does not necessarily continue directly onto the protocone. Heavier wear on M’ and M^ reduces the sharp- ness of the cusps and increases the relative size of the parastyle in comparison to that of M^. Because of wear on the ectoloph posterior to the metacone on these teeth, the size of the postfossette increases rather than decreases with tooth use. There is a slight notch on the metaloph between metacone and hypocone, and the hypocone is taller relative to the metaloph than on M^. Cingula vary in much the same way as on M^, and root development is also similar. On both M* and M^ there is a labial ectoloph rib opposite the paracone but none opposite the metacone. Upper Premolars P^ is worn very early in the life of the animal, and its unworn structure is therefore difficult to deter- mine. Holland and Peterson (1914:245) described this tooth in fairly great detail, but their terminology is difficult to understand; the terms here follow those of Butler (1965: 183) for Chalicotherium rusin- gense (see Fig. 5). In the least worn specimens (for example F:AM 54,449, AM 9,922 — Fig. 5C) the ec- toloph is nearly straight (not W-shaped). The para- cone is in line with the metacone, and there is a very low parastyle. Lingual to the paracone and connected to it by a short crest is a cuspule, prob- ably the protoconule, which is not visible in more worn specimens, though traces of it remain. The single lingual cusp, or protocone, is close to the posterior border of the tooth. Because of its loca- tion and because it is connected by a strong crest (metaloph?) to the ectoloph posterior to the para- cone, it looks more like a hypocone than a proto- cone. Labially the cingulum is very weak, but be- tween the parastyle and protocone a prominent cingulum is present on the anterior and lingual edges of the tooth. A small posterior cingulum clos- es posteriorly the valley between protocone and metacone. The most obvious variation among specimens of this species is the occasional in- creased development of the protoloph lingual to the protoconule and consequent broadening of the an- terior part of the tooth (particularly well shown by CM 2103, the Agate Quarry specimen figured by Holland and Peterson, 1914:pl. 49). In this speci- men the protoloph is in some places as tall as the protocone but is separated from the protocone by a deep, broad valley. At its lower, lingual end the protoloph merges with the cingulum, which skirts the protocone lingually. P^ and P"* are very similar to one another, but P^ is larger and less worn. Its ectoloph has a rudimen- tary W-shape, formed by differences in height of the various points rather than by changes in cur- vature or slant of the labial wall, which is still quite straight and flat. The paracone is slightly taller than the metacone, and a medium-sized parastyle is de- veloped. Most of the early wear on the tooth is on the lingual side of the ectoloph but does not obscure relations with the lingual side of the tooth. The pro- tocone, the only lingual cusp, is large and crescent- shaped and is the last part of the tooth to be worn. It is connected to the ectoloph by two crests, of which the protoloph is taller in some specimens, the metaloph in others. Both lophs are curved and merge into ectoloph and protocone at either end. Enclosed by ectoloph, protoloph, metaloph, and protocone is a deep central valley. Forster-Cooper (1920) noted a weaker metaloph than protoloph on P and P of Phyllotillon and used this character to differentiate Phyllotillon from Chalicotherium . Variation of this character within Moropus elatns suggests that such a distinction between Phyllotil- lon and Moropus does not apply. As in P^, the cingulum is the most variable feature of P, although in all specimens the labial cingulum is extremely weak. In CM 2103 the lingual cingulum is very strong and the tooth is therefore anteropos- teriorly expanded in its lingual part. The anterolin- gual cingulum is here taller than the posterolingual one. Lingual to the protocone the cingulum loses its ridge-like nature at two points where it nearly merges with the lingual wall of the protocone. Be- tween these points it is separated from the proto- cone by a small pit. In other specimens (for exam- ple, AMNH 11,322, Fig. 5) the cingula are in general much weaker, and in others the posterior cingulum is more strongly developed than the an- terior. The lingual cingulum may also remain com- pletely distinct from the protocone. The smaller has a less W-shaped ectoloph, a weaker parastyle, and a consistently weaker ante- 1978 COOMBS— REEVALUATION OE MOROPUS 19 rior than posterior cingulum. Variations among specimens of P'^ seem generally linked in the same individual to those of and P^. Lower Molars Lower teeth were given less attention by Holland and Peterson (1914) than upper teeth, but at the same time these teeth provide fewer features of taxonomic value. M3, the least worn lower molar, is approximately the same size as (slightly larger than) M2; both are larger than Mj. The trigonid is shorter but not narrower than the talonid; each con- sists of a V-shaped crescent of straight to slightly curved lophids. Trigonid and talonid join at the metaconid-metastylid. The metaconid, the tallest point on the trigonid, is sloped gradually anteriorly so that its base partly closes the lingual opening of the trigonid basin. In an unworn specimen, the pro- toconid, slightly anterior to the metaconid, is nearly as tall as the metaconid, but in a worn specimen it is considerably lower. The lowest cuspid of the tooth, the paraconid, does not rise above the level of the paralophid; usually the paralophid is the most curved of the four lophids. On the talonid, the highest point is the metastylid, which in unworn specimens is completely separated from the metaconid at its tip. In more worn teeth the two cuspids are more continuous but still can be distinguished by the grooves between them on both labial and lingual sides. Of approximately equal height but lower than the metastylid are the hypoconid and entoconid, the former slightly an- terior to the latter but both well posterior to the metastylid. The labial part of M3 is more rapidly worn down than the lingual part. Protoconid and hypoconid are worn by the ectoloph of the corresponding upper tooth near parastyle and mesostyle respectively and are eventually abraded to below the level of the lingual cuspids, even the paraconid. Trigonid and talonid basins gradually disappear. As wear pro- gresses the lingual cuspids also become more worn, particularly the metaconid and metastylid, which cut across the lingual end of the protoloph of the upper molar. Finally the entire crown is worn flat. Cingula on M3 and the other lower molars vary among individuals. They are in all specimens strongest posteriorly, anterolabially, and on the la- bial side between the bases of protoconid and hy- poconid. In many of the specimens no further cin- gula are visible, but in some the cingulum forms a well-defined ridge all around the labial base of the tooth. In general, the Morava Ranch Quarry spec- imens seem to have slightly greater cingulum de- velopment than those from the Agate Quarries. On all specimens there is practically no development of a lingual cingulum except at the base of the para- conid, where a weak one may be visible. Where known in an unworn condition, M, and M2 are very similar to M3, but metaconid and meta- stylid are less deeply separated. Lower Premolars P4 is very similar to the molars but differs in sev- eral ways. Most importantly, the talonid is reduced in length and height so that it is about the same length as the trigonid; its posterior part is particu- larly shortened so that the entoconid lies directly lingual to the hypoconid. Both of these cuspids are lower compared to metaconid-metastylid than are their counterparts on the molar teeth. Metaconid and metastylid are barely distinct at their tips and rapidly become confluent during wear. On the tri- gonid the paraconid is so low or so rapidly worn away that it is not visible. On the labial surface of the tooth the groove between trigonid and talonid slants strongly posteriorly toward its base, thus showing an increase in slant over that in M,, which in turn has more slant than on M2 or M3. There is a strong ridge-like cingulum all along the labial base of the talonid, ascending at its anterior end opposite the base of the protoconid. P3 continues the trend away from molar mor- phology begun by P4. Here the talonid is shorter than the trigonid as well as lower in height, but it exceeds the trigonid in transverse width. The tal- onid has only a very small basin which is lost very early during wear to form an oblique flat wear sur- face. Both posterior cuspids, particuarly the ento- conid, are very low and are transversely aligned. The trigonid, in contrast to that of P4, has a well- developed paraconid that is approximately as tall as the protoconid and is not rapidly worn away. No separate metastylid is developed. The groove be- tween trigonid and talonid is oblique but shallow, and there is a well-defined cingulum along the labial base of the talonid. On P2 only the protoconid forms a well-developed cuspid, and the talonid is rudimentary. Neither tri- gonid nor talonid basin remains, for all cuspids are in a straight antero-posterior line. Anterior to the protoconid, at the anterior edge of the tooth, is a distinct but small paraconid. Posterior to the pro- toconid, at the end of a crest, is the low, indistinct 20 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 6. — (A and B) Ventral views (anterior at top) of atlases of A) M. elatus. AMNH 14,378, and B) M. hollandi. FMNH PI3000. (C and D) Occlusal views of deciduous teeth of M. ehitus: C) right dp2-dp4 of AMNH 86,099 and D) right dp^-dp"* of AMNH 94,232. A. C. and D from Agate Spring Quarries, Sioux Co., Nebraska; B from vicinity of Jay Em, Goshen Co,, Wyoming. 1978 COOMBS— REEVALUATION OF MOROPVS 21 hypoconid. On this tooth there is a weak lingual cingulum opposite trigonid and talonid, and on this cingulum there may be occasional traces of a meta- conid and/or entoconid. Compared to the trigonid, the talonid is very short. As Holland and Peterson (1914:243) mentioned, the border of the mandible descends sharply anterior to P2, and there is a long diastema between this tooth and the lower incisors. Incisors Of the anterior teeth, only the three lower inci- sors are present. Their morphology has been dis- cussed in association with an edentulous premax- illa, AMNH 11,321, from the Agate Quarries (Coombs, 1978). Moropns elatiis had a cropping mechanism very similar to that of modern cervids and bovids, except that no lower canine was pres- ent. There is some evidence (AMNH 1 1 ,322) that in aberrant cases a procumbent lower canine or de- ciduous incisors might have been retained in line with the incisor row. Deciduous Teeth Upper teeth of M. elatiis juveniles are known from at least five specimens (CM 1707A, CM 1743, CM 1747, CM 1738, AMNH 94,232). Both dp'' and dp"* strongly resemble the permanent molars but are smaller and more symmetrical and quadrate than M'; the size difference between dp'* and M’ is great- er than that between dp® and dp“*. Because of wear on dp®, the protoconule is not visible and the pro- toloph is incomplete in all known specimens. On dp^ there is, in the least worn specimens, a visible protoconule and a very low (already almost worn away) continuation of the protoloph to the proto- cone. A tiny anterolingual cuspule is present on the protocone of dp^ in some cases (CM 1743). Wear on the metaloph causes the hypocone to be taller than the rest of the loph. On all specimens of dp® and dp"*, the labial cingulum is absent. The elongated, subtriangular dp^ resembles the permanent P® more than it does the other deciduous teeth. As in P®, dp® is broadest in its posterior part , due here to strong development of the hypocone, an isolated conical cusp; any metaloph has been rapidly worn away by wear cutting down the lingual face of the ectoloph. The ectoloph has the parastyle in a straight line with the barely distinguishable paracone and metacone; strong development of the parastyle causes the anteroposterior elongation of the tooth. There is a small protocone anterolingual to the paracone and connected to it by a short crest. Table 2. — Maximum anterior-posterior lengths (in mm) of lower decidnous teeth (pins M, and M2 where preserved on these jaws) of Moropns eluliis and Moropns sp. (from St. Gerand Le Puy. France; Coombs, 1974). Measurements CM CM 1759 AMNH 86,098 AMNH 86.099 Moropns sp. Length dpa — 14.4 — 14.3 14.0 Length dpn 26.4 28.2 27.2 27.6 broken Length dp4 29.2 29.2 29.4 — 20.5 Length dpg-dp^ 67.4 70.8 — 69.5" 54.2 Length M, 35.3 35.1 — ~ — Length M2 — 46.7 — — — “ Approximate value. Posterior and lingual cingula are strong, but there is no anterior cingulum. A relatively large number of mandibles with de- ciduous teeth are available for this species. Dp4. though considerably smaller than M,, is strongly molariform. Its metaconid and metastylid are clear- ly separate at the tips, and in unworn specimens a small paraconid tip is visible; this anterior cuspid is rapidly worn away, however, as the trigonid of dp4 is compressed against the talonid of dps. Both trigonid and talonid basins are well defined, the lat- ter slightly larger and deeper than the former. Dps, although more molariform than its permanent coun- terpart, differs in several respects from dp4. partic- ularly in the trigonid. Although the tooth retains the double crescent shape of the molars and both tri- gonid and talonid basins are present, the former is very shallow. Compared to the talonid, the trigonid is very narrow, and its paralophid is elongated so that the paraconid lies well anterior to the proto- conid. On dpa the paraconid is almost as strongly developed as the protoconid and is not worn down rapidly. Thus the talonid of dp:j is similar to that of dp4 and the permanent molars, while the trigonid more closely resembles the permanent Pg. Meta- conid and metastylid are separated at their extreme tips, and the cingula are like those of molars. The small dp2 is in most specimens strongly worn (AMNH 86,099 retains it in good condition, see Fig. 6). It is not at all molariform and strongly resembles the permanent Pg. Paraconid, protoconid (the tallest cuspid), and hypoconid are in a straight anteropos- terior line. A well-defined crest joins the paraconid and protoconid. Although neither metaconid nor entoconid is visible, there is a trace of a talonid basin on dpg that is not present on Pg. BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Skulls (see Fig. 3) Holland and Peterson (1914) gave some details of the cranium of CM 2103, a well-preserved (although dorsoventrally crushed) specimen of a large indi- vidual from the Agate Quarries. Some of the sub- sequently recovered material in the American Mu- seum of Natural History shows remarkable preservation of the basicranium, and an M. elatiis skull (AM 9923) has also recently been excavated from Morava Ranch Quarry. Comparison of all these skulls with later North American specimens should be profitable when all the material is fully prepared. There are a number of similarities among M. elatus skulls to those of Cluilicotherium (Filhol, 1891; Colbert 1934) and Borissiakia (Borissiak, 1946), which illustrates the basic unity of cranial structure among all chalicotheriids. Among general chalicothere features are the arrangement of basi- cranial foramina, the extension of the flat glenoid fossa onto the ventral surface of the zygomatic arch, the partial but not complete posterior closure of the orbit, and the presence of a long medial ven- tral crest from the internal nares through the pre- sphenoid. The lateral view of the skull of AMNH 14,373 figured by Coombs (1978) shows uncrushed skull proportions of M. elatus and reconstructs the eden- tulous premaxilla in position. The gently curved dorsal surface of the skull of M. elatus lacks any suggestion of doming or elaboration. The stylomas- toid foramen, anterior to the base of the paroccipital process and posterior to the enclosed auditory tube, was not mentioned by Holland and Peterson. The condylar foramen of Holland and Peterson is more usually termed the hypoglossal foramen (see Sisson and Grossman. 1953:72). In all cases the bullae are large and rounded, but the paroccipital processes, although well developed, are variable. In some cases (CM 2103) they are concave anteriorly and convex posteriorly, but in other specimens the ex- act reverse is true. Of particular note in M. elatus is the consistent position of the internal nares and the infraorbital foramina in all specimens examined. In adult individuals the palatines extend posterior to M'^, and the internal nares are therefore posterior to M^ also, at least no farther forward than the hy- pocone of M'^ Infraorbital foramina lie on the snout opposite the anterior part of M^. Mandibles Holland and Peterson compared lower jaws of M. elatus to those of £'(/////,v and Ancylotherium . With a few minor differences, mandibles of M. elatus are very similar to those of other Moropus species. Ba- sic features of the mandible are 1) a narrow anterior part with a short symphysis ending opposite or an- terior to Pa, 2) a long (about as long as and often longer than the premolar row), curved diastema be- tween I3 and P2 with the mental foramen (sometimes two) lying slightly posterior to its midpoint, 3) base of Pa raised somewhat above the level of the dia- stema, 4) labial and lingual sides of jaw primarily flat but slightly convex below cheek teeth, 5) base of ascending ramus broad, rugose, 6) coronoid pro- cess curved posteriorly with long anteroposterior extent, 7) articular condyle flat dorsally with dorsal and posterior parts meeting at almost a right angle, 8) temporal fossa somewhat and pterygoid fossa more strongly excavated. 9) dental foramen large, and 10) angle narrow, flat, not projecting notably below the remainder of the jaw margin. Vertebrae and Ribs Vertebrae and ribs of M. elatus were described in great detail by Holland and Peterson (1914) and will not be described here. Trends in Moropus cer- vical vertebrae from II-VII include a very gradual shortening of the centrum, elongation, narrowing, and increased anterior slant of the neural spine, a decrease in the ventral keel, and a slight increase in width but simplification of the transverse pro- cess. Between cervical VI and VII the vertebrar- terial canal is lost. General chalicothere characters of the cervical vertebrae are elongated centra, oblique anterior and posterior facets on the centra, and a strongly developed ventral keel on anterior vertebrae. Twenty-one dorsal vertebrae are present in M. elatus. Of these, 15 bear ribs (XV has an anterior but no posterior caput facet): I-XV represent tho- racics and XVI-XXl lumbars. Some trends repre- sented along the course of the dorsal vertebrae are the following: 1) the neural spine increases in length from l-III, decreases from IV-XIII, then increases again into the lumbar series from XIII-XIX and decreases again from XIX on; 2) the neural spine slants progressively more posteriorly from VI- XIII, but then decreases its angle from the vertical to stand almost erect at the last thoracic (XV), and then is nearly straight or very slightly anterior or posterior slanted through the lumbar series; 3) the ventral keel decreases along the most anterior dor- sals; 4) there is an increase in the transverse process and separation of a metapophysis from dorsal V 1978 COOMBS— REEVALUATION OE MOROPUS 23 Fig. 7. — Lateral views at scaled size of M. elatiis scapulae of a small individual (?female), CM 1700, on left, and a large individual (?male), CM 1604, on right. Modified from Holland and Peterson (1914, Fig. 76 and pi. 65). onward; 5) from VI-XV the exit notch for the spinal nerve forms a progressively larger bony circle; 6) the neural spine becomes less laterally compressed from IV-XII but more compressed from XII onward; 7) the prezygapophyses migrate onto well- developed metapophyses from XVI onward, where- as the postzygapophyses of these vertebrae are di- rected more laterally; 8) the centrum becomes less triangular and more oval and dorsoventrally com- pressed from XIX-XXI. Well-preserved sacra show fusion of five verte- brae; the transverse processes of the anterior four are enlarged and strongly fused together to form a strong attachment to the ilium. The first sacral is especially broad. The neural spines, which are tightly fused together, become progressively more posteriorly slanted, whereas the size of the centrum decreases as you follow the sacral series posterior- ly. The metapophyses on sacral I bear an articular surface for the posteriormost lumbar, but posterior to sacral 1 they decrease in size and disappear. No caudal vertebrae are known for any Moropiis species outside M. elatus, but several belonging to M. elatiis are preserved, some fused to the sacrum. Known specimens all suggest a small, short tail. Forelimb Scapula. — All M. elatiis scapulae are quite sim- ilar in general structure. Care must be taken in as- sessing morphological differences, for here in par- ticular many structural variations are the result of differential allometric growth in animals of different absolute size, either different sexes or growth stages. An example of the differences between scapulae of the same species is shown in Fig. 7, individuals referred by Holland and Peterson to different species. In the larger specimen the tuber spinae is heavier and more posteriorly curved and the upper posterior border of the infraspinous fossa is more rounded, thick, and rugose — adaptations to greater weight bearing (see Osborn, 1929:740-741) in the larger heavier animal. Scapulae have the following general features: 1 ) a circular to oval glenoid fossa with a moderately developed coracoid process just anterior to it; 2) an indented anterior border below the spine; 3) ante- 24 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 8. — Dorsal (A) and volar (B) views (excluding trapezium, pisiform) of right carpus of AMNH 14,378, M. elatiis, from the Agate Spring Quarries, Sioux Co., Nebraska. Labels in Fig. 9. rior border of blade slightly more convexly rounded than posterior border; 4) upper (proximal) end slightly squared to gently rounded; 5) well-devel- oped spine with posteriorly reflected tuber spinae; 6) supraspinous and infraspinous fossae of subequal size; 7) subscapular surface gently concave with a slightly convex ridge in the anterior or middle part. M. elatiis has a more strongly posteriorly reflected tuber spinae than is known in other Moropus spe- cies. Humerus (Fig. 13A). — The humerus of Moropus , unlike the femur, is not easily confused with humeri of contemporary perissodactyls (for example, of rhinoceroses). Compared to the probable weight of the animal and the size of the proximal articulation and greater tuberosity, the shaft is very long. Even more unusual is the comparatively strong develop- ment of the entepicondyle, otherwise reduced or absent in all but the most primitive ungulates. The distal trochlea is transversely very broad and is not deeply grooved; the anterior part of the trochlea (capitLilum), with which the radius articulates, is particularly wide. Holland and Peterson (19I4;334- 335) mentioned the following important features of the humerus of M. elatus: large, rounded head; well, but not unusually, developed greater tuber- osity; relatively shallow but broad bicipital groove; prominent deltoid ridge extending well down the shaft and flexed posteriorly; prominent ectocondy- lar ridge (but not as strong as in most other peris- sodactyls); deep olecranon fossa; shallow coronoid fossa; plus the very important characters of the dis- tal end mentioned above. Radius-ulna . — Although a number of radii and ulnae are available, there are few consistent differ- ences from other Moropus specimens, which can be regarded as of taxonomic value. In many cases only a single radius or ulna (often broken) is known; in other cases the two bones are solidly fused, usu- ally at the distal end and sometimes also for part of 1978 COOMBS— REEVALUATION OE MOROPUS 25 the shaft. There does not seem to be any quarry or locality correlation between groups in which more or less fusion occurs; rather, fusion may be corre- lated with the age of the animal. Whether or not strong fusion occurred does not seem to be func- tionally critical, for in any case the facets between the two bones at their proximal ends would allow little or no movement between them, certainly no rotation of the radius. General features of the ulna include rather short but heavy olecranon process, prominent anconeal process, deep semilunar notch with strong expan- sion near the radius facet, radius facet distal and at an acute angle to semilunar notch and forming a deep articular fossa, shaft subtriangular in section, distal end narrow with cuneiform facet flat and ad- jacent (across a 90-degree arc) to the pisiform facet on the posterior (volar) surface. Some radius fea- tures include proximal and distal ends broader than shaft articular surface for humerus with broad, slightly concave ulnar (lateral) part and narrower, flatter radial (medial) part, coronoid process weak, facet for ulna with proximally oriented central tongue and posteriorly (volar) oriented radial and ulnar parts (somewhat variable), distal end with dorsal groove between scaphoid and lunate facets, lunate facet more concave and of greater dorsovolar extent than scaphoid facet. Cm pals. — Carpal elements of M. elatus were de- scribed and figured by Holland and Peterson (1914:337-339), but additional material has given a more accurate view of the absence of significant variation among specimens and the general useful- ness of manus and pes elements in taxonomy. Therefore the most salient features of the carpals are worth rediscussing here as a basis for taxonomic and functional work to follow (see Eigs. 8-12). On the scaphoid the weakly convex articular sur- face for the radius covers all of the proximal surface except for a rugose protuberance at the ulnovolar angle (Eig. lOA). The radiovolar surface bears a small but well-defined, distally slanting, convex fac- et for the trapezium (Eig. 8B). Adjacent and distal to the trapezium facet is the saddle-shaped trape- zoid facet, a large facet, which reaches distally onto the radial side of the distal process (Eig. lOB, pal- 26 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Dorso-radial process Fig. 10. — Proximal (A and C with radial edge at right) and distal (B and D with radial edge at left) views of right cuneiform, lunate, scaphoid (A, B) and right unciform, magnum, trapezoid (C, D) of M. elatiis. AMNH 14,378, from the Agate Spring Quarries, Sioux Co., Nebraska. Abbreviations for facets in Figs. 10-12, 22, and 24-26 are as follows; R, radius; U, ulna; P, pisiform; S, scaphoid; L, lunate; Cn, cuneiform; Un, unciform; M, magnum; Td, trapezoid; Tm, trapezium; II, Me II; III, Me III; IV, Me IV; V, Me V. mar process of Holland and Peterson). Of the two facets for the lunate on the ulnar surface (Fig. I lA), the proximal oval one stands on a low platform and is particularly prominent at its volar edge. The more distal lunate facet occupies the ulnar surface of the distal process and curves from the magnum facet in a proximoulnar direction so that in its most proxi- mal part it borders the volar edge of the trapezoid facet. Between proximal and distal lunate facets, the ulnar surface of the trapezoid is depressed. Be- cause the large distal process is squared, the facet for the magnum on its distal end is flat to slightly convex, bordered on one side by the trapezoid facet and on the other by the distal lunate facet. On the lunate, the proximal articulation, which is for the radius, is of suboval shape and convex in all directions, especially strongly so in the dorsal to volar direction (Fig. lOA). Very prominent in a proximal view is the volar process near the distal edge, which has a facet (for the magnum) on its distal but not on its proximal surface (Fig. lOB). Of the two scaphoid facets on the radial surface (Fig. I IB), the proximal one is large and adjoins the ra- dius facet across a prominent keel; separated by a groove from the proximal facet, the distal scaphoid facet varies slightly in position in different speci- mens, sometimes passing along the radiodistal sur- face of much of the volar process but barely dis- tinguishable from the magnum facet except on the volar process. On the most distal part of the radial surface of the lunate, the magnum facet curves onto the distal surface of the volar process, where it 1978 COOMBS— REEVALUATION OF MOROPUS 27 Fig. !!. — A) Ulnar view of right scaphoid, trapezoid, and Me II (dorsal edge at right), and B) radial view of right lunate, magnum, and Me III (dorsal edge at left) of M. elatus, AMNH 14.378, from the Agate Spring Quarries, Sioux Co., Nebraska. See Fig. 10 for facet abbreviations. Unlabeled line points to volar process of lunate. forms a concave, cup-shaped facet separated by a ridge from the distal lunate facet (Fig. lOB). All along its most distal and ulnar edges the magnum facet is divided by a beak-like ridge from the unci- form facet (Figs. lOB, 12A). The unciform facet has two parts, the dorsal part flat to convex and with an almost entirely distal orientation, the volar ad- joining part slightly concave and curving onto the ulnar side of the bone, where it is weakly disting- uishable from the cuneiform facet. Flat to weakly convex and on a low platform, the cuneiform facet forms a D-shaped tongue onto the ulnar surface of the lunate (Fig. 12A). Several cuneiforms from the Agate Quarries and Morava Ranch Quarry were compared with YPM 24,632, which is part of the type collection of M. elatus and closely resembles the more recently col- lected specimens. This bone is very deep in the dor- sal to volar direction but is transversely narrow. At the dorsoradial angle and volar edge the cuneiform tapers to a blunt point, giving the bone a wedge- shaped appearance in proximal view (Fig. lOA). The dorsal and ulnar surfaces of the bone are rough, but not strongly rugose, and on the distal part of the ulnar surface is a blunt roughened protuberance for the attachment of muscles or ligaments (Figs. 8A, lOB). There is another, smaller protuberance near the proximal edge of the narrow volar surface. Both pisiform and ulna articulate with the proximal surface of the cuneiform (Fig. lOA). The relative amounts of the proximal surface occupied by each varies in the specimens examined, but in each case the more dorsal facet, for the ulna, is larger. A very weak diagonal ridge separates the two facets, and the pisiform facet is raised slightly above the ulna facet. Both facets are wedge-shaped, the slightly concave ulna facet narrowest in the volar direction, the weakly convex pisiform facet narrowest dor- sally. On the distal surface of the cuneiform (Fig. lOB), the unciform facet is concave in the dorsal to volar direction and closer to the dorsal than to the volar edge of the bone. There is only one, tongue- like facet for the lunate, at the distal edge of the radial surface (Fig. 12B), and adjoining the unci- form facet at a right angle. No second, more prox- imal lunate facet is observable. On pisiform specimens of M. elatus, the cunei- form facet is parallel to the long axis of the pisiform and the adjacent ulna facet perpendicular to it. Both surfaces curve in tongues away from their straight 28 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 process Fig. 12. — A) Ulnar view of right lunate, magnum, and Me III (dorsal edge at right). B) radial view of right cuneiform, unciform, and Me IV (dorsal edge at left), and C) proximal view of right Me IV, Me III, and Me II (dorsal edge below, radial edge at right) of M. elatiis . AMNH 14,378, from the Agate Spring Quarries, Sioux Co., Nebraska. See Fig. 10 for facet abbreviations. adjoining edge, each slightly concave, but the ulna facet more so. When placed in articulation with the cuneiform, the pisiform does not occupy the entire pisiform facet on the cuneiform and so may have been capable of some dorsovolar movement against the cuneiform. The part of the pisiform on which the articular facets lie is expanded in all directions, but volar to that region the pisiform is laterally con- stricted; still farther in the volar direction the pisi- form is expanded and rugose for muscle attach- ment, its end slightly curved in the radial direction. The unciform (hamate) has two prominent pro- cesses, a dorsoradial one and an ulnovolar one. Of these, the former is near the proximal edge of the bone and the latter near the distal edge (Figs. IOC, D). There are two facets on the proximal surface (Fig. IOC), the more distal and radial for the lunate and extending onto the above-mentioned dorsora- 1978 COOMBS— REEVALUATION OE MOROPUS 29 dial process. On its radial side (Fig. 12B), this pri- marily concave facet adjoins the much smaller mag- num facet at a right angle; its volar edge borders the cuneiform facet along a blunt ridge. The cunei- form facet is strongly convex in the dorsal to volar direction and curves onto the volar side of the bone (Fig. 8B). This facet is much larger than the cor- responding facet on the cuneiform, so that during flexion of the carpus the unciform was able to move considerably with the result that the volar surface of its cuneiform facet was then in contact with the cuneiform. No facets are present on the roughened ulnar surface of the unciform, for Me V does not articulate with the unciform. On the radial side of the unciform are two facets (Fig. 12B). The proxi- mal one, for the magnum, has a trapezoidal shape and flat surface and lies on a very weak, distally slanted platform, with the radial surface of the un- ciform roughened and depressed on either side of it. Generally speaking, the platform for the magnum facet is more pronounced in specimens from Mo- rava Ranch Quarry than in those from the Agate Quarries. There is also variably present in M. elatiis (absent in Fig. 12) a narrow volar extension of the magnum facet adjoining the lunate facet. The trap- ezoidal magnum facet indistinctly borders the facet for Me III, which in turn indistinctly borders the Me IV facet near the distal edge of the radial suiface of the unciform (visibility and location of Me 111- Mc IV facet boundaries vary somewhat among specimens). The weak ridge, whieh appears to sep- arate the Me III facet from the Me IV facet, is ac- tually shared by both facets. The Me III facet varies in its dorsal to volar extent but in any case does not reach as far in the volar direction as does the Me IV facet, which occupies the distal surface of the ulnovolar process (Fig. lOD). Although the magnum is one of the largest bones in the carpus, onjy a small part of it is visible in dorsal view; it is most strongly developed in its vo- lar part. The largest of the four facets visible from the radial side (Fig. 1 IB) is that for Me II, irregular in shape and slightly concave. There are two facets for the trapezoid, a larger dorsal one between scaphoid and Me II facets and passing onto the proximal surface of the magnum, and a smaller oval-shaped volar one also near the proximal edge (it borders proximally parts of both scaphoid and lunate facets). Of the two large curved facets on the proximal surface of the magnum (Fig. IOC), the more dorsal one articulates with the distal process of the scaphoid. Most of the volar part of the prox- imal surface is occupied by a convex, ball-shaped lunate facet, which fits into the cup of the volar process of the lunate. The dorsal part of the lunate facet curves onto the ulnar surface of the magnum and is separated by only a weak ridge from the un- ciform facet. Compared with adjoining facets, the unciform facet is small and flat, lying on a low plat- form (Fig. 12A). It attains the distal edge of the ulnar side of the magnum where it joins the Me 111 facet at right angles. The latter facet, on the distal surface (Fig. lOD), is a very elongated facet, pass- ing in the distovolar direction for nearly the entire extent of the well-developed volar hook. The Me III facet is irregularly concave along most of its length and may be divided by a dorsal to volar di- rected ridge into moieties of unequal size. In both proximal and dorsal views (Figs. IOC. 8A). the trapezoid has a triangular appearance, in the former case the apex being the volar angle, in the latter the distal angle which fits into the proxi- mal groove on Me II. The entire proximal surface (Fig. IOC) is occupied by the scaphoid facet, con- cave transversely and convex in the dorsal to volar direction. Curving onto the volar surface of the trapezoid (Fig. 8B), this facet merges into the facet for the trapezium, which occupies most of the volar suiTace and is continuous with trapezium facets on the scaphoid and Me II. The scaphoid facet also curves onto the ulnar surface of the trapezoid (Fig. 1 1 A), where it borders the two facets for the mag- num. The larger dorsal magnum facet has a slight distal slant and a subtrapezoidal shape. Volar to this facet is a rough-surfaced depression, larger than the dorsal facet itself. Only at the extreme proximovolar edge of the dorsal magnum facet, bor- dering the scaphoid facet, is a small flattened area representing the volar articular facet for the mag- num. Separated by a distinct ridge from the dorsal magnum facet is the facet for Me II. This distal facet, like its counterpart on Me II. has a V-shape (Fig. lOD), and its radial edge meets the trapezium facet at an acute to right angle on the volar surface of the trapezoid. A number of trapezium specimens of M. elatus are known. This small bone contacts facets on the scaphoid, trapezoid, and Me II. Metacarpals . — The “dorsal surface” of Me II described by Holland and Peterson (1914) is ac- tually a dorsoradial surface (with radiovolar and dorsoulnar processes at either end. see Fig. 12C) when Me II is placed in articulation with Me 111. For this reason, dorsal of Holland and Peterson is 30 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 here called dorsoradial. a term also adopted by Bor- issiak (1946) and Belyaeva (1954). In dorsoradial view (Fig. 8A) the proximal surface of Me II is M-shaped, broadened by radiovolar and dorsoulnar processes that extend to either side of the shaft. Most of the radiovolar process is occupied by a well-developed trapezium facet (Fig. 8B), which is weakly concave and widest in its radial part. The trapezium facet is contiguous to, but separated by, a sharp crest from the concave trapezoid facet, which forms the middle V of the above-mentioned M (Fig. 12C). An ulnar crest, which joins the one between trapezium and trapezoid facets to form the apex of a triangle enclosing the trapezoid facet (Fig. 12C), separates the trapezoid facet from the mag- num facet. The magnum facet is the last (ulnar) limb of the M and overhangs the Me 111 facet (Fig. 1 1 A), which forms the proximal part of a deep concavity on the ulnar side of the bone. On the volar side of the dorsoulnar process the Me III facet curves dis- tally, so that when Me 11 and Me III are articulated. Me II covers both dorsoradial and proximoradial parts of Me 111. The shaft of Me 11, almost circular in section, is more massive than that of either Me III or Me IV. The distal end is rotated clockwise in relation to the proximal end, having an effect in turn on the orientation of the fused proximal and medial phalanges (bearing the large claw) of this digit. Gen- eral features of the distal end of Me II are 1) the strongly developed ulnar tubercle proximal to the phalangeal articular surface, 2) the unequal devel- opment of the sesamoid facets, the ulnar facet being wider and extending farther distally than the radial facet, 3) the weakness of the keel between the two sesamoid facets, and 4) the presence of a foramen on the radial surface Just proximal to the articular facets. YPM 24,632, a distal end of Me 11, which was part of Marsh’s type collection of M. elalus. shows no special differentiating features from Agate and Morava Ranch Quarry material. Most of the proximal part of the dorsal surface of Me 111 is rugose (Fig. 8A), except for a subtrian- gular area, which articulates with the dorsoulnar process of Me IT This facet is continuous with the Me II facet on the proximal surface (Fig. I IB, de- spite the implication by Holland and Peterson, 1914:350, 352, that the two are separate facets). The proximally oriented part of the Me 11 facet extends in a dorsal to volar strip across the radial side of the proximal surface. It is delineated only by a faint ridge from the broader magnum facet, which oc- cupies the middle part of the Me 111 proximal sur- Tabile 3. — Maximum lengths (in mm) and ratios of maximum lengths of metacarpals in individuals of selected chalicothere species. Specimens Me 11 Me III Me IV Me 11/ Me 111 Me IV/ Me III Moropus elatiis AMNH 14.378 198 vO O O 0.86 0.87“ CM 1700 164“ 188 166 0.87“ 0.88 CM 1604'’ 230 274 221 0.84 0.81 Moropus hollandi CM 1424 174 — — — — FMNH P13000 178 206 190 0.86 0.92 Schizotheriuni prise um Specimen in Paris'' 1 15 130 126 0.88 0.97 Ancylotheriiun (A.) penteliciim AMNH 32504“ 234“ 279 275 0.84“ 0.98 Cluilicotheriiiin grande CM 2298'- 163 198 212 0.82 1.07 “ Approximate measurement. " Measurements taken from Holland and Peterson, 1914. ‘ An uncatalogued specimen, possibly a composite, from the Phosphorites of Quercy. located in Museum National d'Histoire Naturelle. Paris. ‘‘ Cast of an uncatalogued specimen from Pikermi in the Museum National d’Histoire Naturelle, Paris. Cast of specimen from Sansan (Gers), France. face (Fig. 12C). Along its dorsal to volar axis the magnum facet is concave in its dorsal part and con- vex in its volar part, extending farther in a volar direction than any other part of the proximal sur- face. A sharp crest separates the magnum facet from that for the unciform, which adjoins it in the ulnar direction. This latter facet is subtriangular with a volar apex and overhangs the facet(s) for Me IV, separated by another sharp crest (Fig. 12A). Despite Holland and Peterson’s (1914:353) obser- vation of a double Me IV facet, divided into two by a sinus, there is often a single, undivided facet. In any case the Me IV facet(s) occupies (-y) the prox- imal surface of a large concavity on the ulnar sur- face of Me HI. The shaft of Me 111 is subquadrate in section and the distal end is only slightly asym- metrical. A blunt ridge continues distally from the Me II facet along the dorsoradial angle for about half the length of the shaft. Me IV is oriented so that what seems to be a flat dorsal surface (Fig. 8A) actually has a dorsoulnar orientation when the bone is articulated with Me 111, a situation the opposite of that for Me 11. Of the two Me 111 facets (Fig. I2B, or parts of a single. 1978 COOMBS— REEVALUATION OF MOROPUS 31 undivided facet), the largest and most convex is the dorsal one, whereas the smaller volar facet has a less proximal and more completely radial orienta- tion. Separated from the Me III facet(s) by a weak to somewhat pronounced rib, the unciform facet occupies almost all of the proximal surface of Me IV and is subtriangular with a volar apex (Fig. 12C). An important feature of Me IV is its articular sur- face for Me V, the reduced metacarpal that articu- lates with no other bone. This articulation is on the ulnovolar angle of Me IV (Fig. 8B) and is separated from the proximal edge by paired tubercles (one ulnar, one volar). The Me V facet is on the distal volar surface of the ulnar tubercle. The shaft of Me IV becomes more oval and less triangular in section distally. Me IV is less symmetrical than Me III, but more so than Me II, and at its distal end the radial tubercle is larger than and proximal to the ulnar one, whereas the ulnar of the two sesamoid facets is smaller. In M. elatus Me V has a two-part Me IV facet on the dorsoradial side of its proximal end and a trans- versely compressed shaft. Small sesamoid and pha- langeal facets are present on the distal end. which reaches the level distally of the proximal end of sesamoid facets in Me IV. In articulation with Me IV, Me V was somewhat volar to the other meta- carpals and was strongly divergent at its distal tip from the remainder of the manus. Manus summary. — A complete analysis of the manus is best taken as part of a broad functional study of Moropus (Coombs, in preparation), but a few comments are useful. Despite the lack of ro- tation ability at the wrist (see radius-ulna com- ments), carpal facets suggest that the wrist was quite flexible from side to side and that strong flex- ion was possible along two planes, at the ulnocarpal joint and between the proximal and distal rows of carpals. Manus flexion in M. elatus was not so pro- nounced, however, as in Ancylotherium {Ancylo- therium) pentelicum (Schaub, 1943), where the tra- pezium was lost and the scaphoid could contact Me II during extreme flexion. Weight was borne pri- marily by the distal ends of the metacarpals, with participation by the sesamoids and the proximo- volar ends of proximal phalanges. Except for the tips of the ungual phalanges, the hooked digits were probably held clear of the ground during walking by hyperextension of the proximal phalanx and flexion of the other phalanges (see digit II of manus in Fig. 2). Me III and Me IV are almost the same length, but weight was concentrated on the more symmet- rical Me III. The torsion and other unusual features of the relatively short Me II are associated with the bearing of the large hooked claw by this digit. Hindlimb Innominate . — Complete pelvis specimens are otherwise rare in Moropus but are well known in M. elatus. Chalicothere features not shown so strongly in other perissodactyl innominates are the longer and narrower pelvic proportions, primarily because of expansion of the ischium and pubis pos- terior to the acetabulum and the fairly long shaft of the ilium. Articulation with the femur and relations to the sacrum and lumbar vertebrae suggest that the ilium was nearly vertical in the living animal, al- though probably not so sharply vertical as it was mounted by Osborn (AMNH 14,375, Fig. 2). On the medial surface of the ilium, the articular surface for the sacrum is heavy and rather long. The iliac crest is expanded but not strongly so. Aside from the rugosities along the iliac crest, there is an obvious rugose area of muscle attachment (for the rectus femoris) just anterior and dorsal to the acetabulum. The acetabulum is deeply excavated with a heavy dorsal overhang, which aided in supporting the body over the femur. Femur (Fig. 13B). — Holland and Peterson (1914), while describing the femur of M. elatus. compared and differentiated it from the femur of titanotheres. Actually the greatest problem in the identification of Moropus femora is in comparison with those of rhinoceroses found in beds of the same age, usually in greater numbers. Features, which usually differ- entiate a Moropus femur, are 1) the symmetrical or nearly symmetrical (less symmetrical in older indi- viduals). slightly oblique patellar facet (that in rhinos is more asymmetrical). 2) the asymmetrical development of the distal condyles, the medial con- dyle extending farther posteriorly (condyles of con- temporary rhino femora are more nearly symmet- rical), 3) the lesser development of the third trochanter, and 4) the very small but persistent fo- vea capitis on the head for attachment of the liga- mentum teres (i\\c fovea on rhinoceros femora is usually larger). The large number of known M. elatus femora are essentially similar in morphology, although there is some individual variation in the proportional width of the shaft. In larger femora of M. elatus. the shaft is relatively broader. Weight-bearing adaptations such as this, in a species where size variation is as great as in M. elatus . suggest that one must be care- 32 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 3cm Fig. 13. — Anterior views of A) distal end and shaft of left humerus. B) right femur, C) patella, and D) left tibia of M. elatus, AMNH 14.378, from the Agate Spring Quarries. Sioux Co.. Nebraska. Abbreviations: D.r., deltoid ridge; C.f., coronoid fossa; Cap, capitulum; Ent, entepicondyle; Ect, ectocondylar ridge; I , greater trochanter (broken); 2, lesser trochanter; 3, third trochanter; Pat, patellar facet; Cnm, cnemial crest; Mm, medial malleolus; F, fibida articulations. fill in drawing specific boundaries on the basis of proportions alone. In addition to the characters cit- ed above. M. elatus femora share well-developed greater and lesser trochanters and a fairly deep tro- chanteric fossa. Distal to the third trochanter, the shaft narrows quite abruptly and then broadens close to the distal end. Tibia and fibula . — Among known Moropits spe- cies, the fibula has been preserved only in M. ela- tus', it will not be discussed further here. Tibiae vary little in proportions and morphology, and there is therefore little to add to the description of Holland and Peterson (1914:363). Some of the most promi- nent features are (Eig. 13D): I) a heavy cnemial crest, which is rugose proximally and extends a good distance distally along the shaft, 2) two broad proximal articular surfaces for the condyles of the femur, of which the lateral surface is slightly more curved, 3) a V-shaped or double (spreading apart anteriorly) spine between the two femur articular surfaces, with the medial side of the spine sharper and taller, 4) well-developed articular surfaces both proximolaterally and distolaterally for the fibula and a well-defined ridge for the interosseous ligament connecting the proximal and distal facets, and 5) a strong medial malleolus. There is no evidence that the medial malleolus contacted the calcaneum dur- ing extreme flexion as Holland and Peterson sug- gested. Patella (Pig. 13C). — Patellae are not known for all Moropus species, but there is some evidence of differences among species. Patellae of M. elatus 1978 COOMBS— REEVALUATION OE MOROPUS 33 Table 4. — Measurements (in mm) of astragali of selected chali- cothere specimens, including proportions of tihial height/maxi- mum width. Taxa and specimens Maximum width Height of tibial side" Height/ width Moropus elatin' , AMNH 14378'' 89.9 74.7 0.83 Moropus elatus, CM I701'> 76.7 68.8 0.86 Moropus elatus , AM 9722^- 89.7 68.8 0.77 Moropus elatus . AM 972y 92.9 74.5 0.80 Moropus elatus, AM 9724‘- 90.8 71.2 0.78 Moropus elatus , Range of six Frick Morava Ranch Quarry specimens 81.1-98.0 62.5-77.2 0.77-0.82 Moropus hollandi , CM 1424" 84.1 64.0 0.76 Moropus hollandi, F:AM 54,902g 103.5 74.4 0.72 Moropus merriami , UCMP 1 1 ,605 100.7 64.9 0.64 Moropus merriami, UCMP 19.404 102.6 62.2 0.61 Schizotherium turgaicund' 43.5 38.5 0.89 Schizotlierium turgaicund 47 41 0.87 Borissiakia betpakdalensis^ 92 74 0.80 Borissiakia betpakdalensis^ 68 58 0.85 Ancylotherium {Metaschizotherium ) fruasd 83.4 56.0 0.71 Chalicotherium rusingense*' 66.0 35.7 0.54 Chalicotherium salinum , AMNH 19,436 88.6 37.7 0.43 Chalicotherium grande , CM 2299 (cast) 100.7 41.2 0.41 Chalicotherium grande*' 100.5 46.2 0.46 Chalicotherium gold fuss i*' 98.7 44.7 0.45 ® Height measurement includes neck and edge of navicular facet. ’’ Fj'om Agate Spring Quarries, Sioux County, Nebraska. From Morava Ranch Quarry, Box Butte County, Nebraska. Holotype. right side. Taken from Belyaeva, 1954:57-58. 'Taken from Borissiak. 1946. ® Taken from von Koenigswald, 1932. '' Taken from Butler, 1965. have 1 ) the facet for femur not so wide transversely as tall, 2) the trochlear ridge between the two sides of the femur facet blunt, with curved sides of the facet at obtuse angles to one another, 3) the anterior patellar surface rugose with some but not strong development of a proximal ridge for the insertion of knee extensors, and 4) a distal tongue developed at the patellar apex (except in one Morava Ranch Quarry specimen). It is important to note that YPM 24,632, a patella, which was part of Marsh's type collection of M. elatiis and first figured by Holland and Peterson (1914:223), is like the majority of pa- tellae of M. elatus. Tarsals (Pigs. 14-17). — In all astragali of M. ehi- tus a small neck distal to the trochlea is present (Fig. 14C), on whose distal surface the large navic- ular facet lies. The neck is most pronounced on the tibial side of the bone, in some specimens very pro- nounced, but it may be reduced at its fibular edge (distal to the middle part of the astragalar trochlea). Over half the width of the astragalus is occupied by the fibular side of the trochlea, which is more grad- ually slanted than the tibial side. The fibular edge of the trochlea does not ascend proximally quite so far as does the tibial edge, and distally it hangs free from the rest of the bone (but does not extend far- ther distally than the navicular facet). On the volar surface of the astragalus (Fig. 14B), the most ob- vious feature is the deeply concave ectal facet for the calcaneum, which extends more than half the transverse width of the astragalus as well as over half the height. The ectal facet is rather uniformly concave from proximal to distal throughout its width. A sharp proximal ridge separates the ectal facet from the dorsal surface of the trochlea. Dis- tally, across a blunt ridge along the fibular half of its width, the ectal facet borders the smaller, weak- ly convex calcaneal facet, also for the calcaneum. The latter facet is slanted strongly in a distal direc- tion and away from the ectal facet. Both ectal and calcaneal facets are separated by a rugose area con- taining nutrient foramina from a third facet for the calcaneum, the sustentatucular facet, on the tibio- distal side of the astragalus’ volar surface. The sus- tentacLilar facet is generally oval in shape, weakly convex, and lies flat on the plane of the volar sur- face. On the distal surface of the astragalus the sin- gle facet for the navicular is distinctly pear-shaped with a narrow fibular end. Weakly convex trans- versely and at its fibulovolar angle, the facet passes slightly onto the fibular and volar surfaces of the astragalus, but along its entire dorsal edge a distinct ridge separates it from the dorsal surface. The tuber calcis of the calcaneum is long and massive, with a subtriangular cross-section. Its dor- sal surface is narrower than its rugose volar surface. 34 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 I'ig. 14. — A) Dorsal view right calcaneum, B) volar view right astragalus, C) dorsal view right tarsus, and D) volar view right tarsus of AMNH 14,378 (missing mesocuneiform and entocuneiform), M. elatiis. from the Agate Spring Quarries. Sioux Co.. Nebraska. Tarsus elements identified in Fig. 15; facets on astragalus and calcaneum abbreviated as follows: E, ectal; S, sustentacular; C, calcaneal. “r 1^,1 Sustentacular • proces s Lesser process 1978 COOMBS— REEVALUATION OE MOROPUS 35 Fig. !5. — Labeled drawing of dorsal and volar views of tarsus of AMNH 14,378 shown in Fig. 14. and the end of the tuber, where there is a pro- nounced rugose area for muscle attachment partic- ularly characteristic of this species (Eig. 14D), is thicker than the shaft. Of the facets for the astrag- alus (Fig. 14A), the ectal facet is relatively narrow but extends well onto the dorsal surface of the tu- ber. Here it is only weakly convex, but distal to the tuber it forms a protruding right-angle bend corre- sponding to the deeply concave ectal facet on the astragalus. On some specimens there appears to be a facet for the tibia at the edge of the ectal facet on the tuber (mentioned by Holland and Peterson, 1914:366, although no particular evidence for it ex- ists on the tibia). In those specimens where such a facet exists, the ectal facet, rather than being flat on the tuber, is raised on a platform so that the tibia facet is continuous with the astragalar trochlea when the astragalus and calcaneum are articulated [also noted in Ancylotherinm (A .) pentelicum]. The oval, slanted sustentacular facet occupies almost the entire sustentacular process and is separated from the ectal facet by a groove for the interosseous ligament. The third facet, corresponding to the cal- caneal facet on the astragalus, is distal to the ectal facet on what Holland and Peterson termed the “lesser process," which is well developed. On the distal end of the calcaneum is the pear-shaped cu- boid facet, flat on its fibular side and slightly con- cave on its tapering tibial side. There is no facet on the narrow distal surface of the sustentacular pro- cess. The paralectotype of M. elatiis, YPM 24,631, includes a broken calcaneum. Although retaining only the tuber calcis, it can without question be connected with the other specimens here described. Important common features are the strong rugose protuberance at the end of the tuber calcis and the shape of the ectal facet as it extends onto the tuber. On the cuboid, the calcaneum facet is pear- shaped (Fig. 16E), corresponding exactly to the dis- tal facet on the calcaneum. It occupies the entire proximal surface and is set off by a ridge from the remainder of the bone; articulation with the calca- 36 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 16. — Tibial (A - C. proximal edge at top) and proximal (D - F. dorsal edge at bottom) views of cuboids of M. Iiollandi (A, D. left cuboid). FMNH P13000, from near Jay Em, Goshen Co., Wyoming; M. elatiis (B, E, left cuboid), AMNFl 14,378, from the Agate Spring Quarries, Sioux Co., Nebraska; and M. disfans (C, F, right cuboid), YPM 24,628, from Bridge Creek, Oregon. Distal views (dorsal edge at bottom) of naviculars of G) M. Iiollandi, FMNH P13000. and H) M. eiatus . AMNH 14,378. Facet abbreviations used in Figs. 16-19 and 27-28 are as follows; Ca, calcaneum; Cu. cuboid; N. navicular; Ec, ectocuneiform; Me. mesocuneiform; II, Mt II; III, Mt III; IV, Mt IV. neiim is slightly oblique. The dorsal, fibular, and most of the volar surfaces of the cuboid are gently curved and relatively smooth, but at the tibiovolar angle is developed a strongly rugose process for muscle attachment, which overhangs Mt IV (Eigs. 14D, I6B, 17C). The tibial side of the cuboid (Eigs. 16B, 17C) has no articular surface in its most prox- imal and volar parts, where it is separated from the navicular by a small space. The navicular facet lies in a dorsal to volar strip, which occupies one-third of the height of the tibial surface and lies about halfway between the proximal and distal edges of the bone. This facet is slanted to form the proximal part of a V of which the ectocuneiform facet is the distal part. Occupying the most distal part of the tibial surface, the ectocuneiform facet consists of dorsal and volar divisions, which may or may not be separate. The distal surface of the cuboid, except for the tibiovolar process, is occupied by the large quadrilateral facet for Mt IV, weakly concave ex- cept near its tibiovolar angle where it is convex. The proximal facet of the navicular is transverse- ly concave, particularly at its tibiovolar angle, where it follows the curve of the corresponding fac- et on the astragalus (Eigs. 14C, 17B). Three other facets are visible on the distal surface (Eigs. 16H, 1978 COOMBS— REEVALUATION OE MOROPUS 37 Fig. 17. — Views of right tarsals and metatarsals of M. elcitiis, AMNH 14,378, from the Agate Spring Quarries, Sioux Co., Nebraska; A) Tibia! view of navicular, ectocuneiform, and Mt 111 (dorsal edge at left); B) fibular view of navicular, ectocuneiform, and Mt III (dorsal edge at right); C) tibial view of cuboid and Mt IV (dorsal edge at left); D) proximal views of Mt IV, Mt III, and Mt II (dorsal edge at bottom). Facet abbreviations as in Fig. 16. 17A, B). Of these, the cuboid facet (fibularmost) does not abut the astragalus facet and is like a short dorsal to volar bar (like the corresponding facet on the cuboid). The ectocuneiform facet is about two times as large as that for the mesocuneiform (the most tibial of the distal facets) and has an irregular surface (primarily convex in a dorsal to volar direc- tion) and shape (broadest dorsally and tapering in the volar direction). A shallow groove near the vo- lar edge of the ectocuneiform facet to accommodate a weak ridge of the ectocuneiform prevents dorsal to volar movement of the navicular and ectocunei- form against one another. The irregularly curved, weakly convex mesocuneiform facet is slanted to approach on its tibial side, but not adjoin, the facet for the astragalus. Weak ridges between the three distal facets of the navicular are their only distin- guishing boundaries. 38 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Table 5. — Maximum lengths (in mm) of the three metatarsals within single chalicotheriid specimens. Specimens or taxa Length Mt 11 Length Mt 111 Length Mt IV Mt 11/ Mt 111 Mt IV/ Mt 111 AMNH 14378" 133.7 146.4 148.3 0.91 1.01 CM 1701" 109.9 124.9 126.8 0.88 1.01 CM 1706a" 134 156 157 0.86 1.01 F:AM 54903'’ Schizotlierium — 115.7 129.8 — 1.12 priscum'^ Ancylotherium (A.) 95 118 1 16 0.81 0.98 pentelicum'^ Chalicotherium 127.0 171.3 147.2 0.74 0.86 grande'' 65.8 77.7 107.7 0.85 1.39 ® Specimens of Moropus elatiis from the Agate Spring Quarries, measurements of CM 1706a from Holland and Peterson. 1914. Moropus cf. M. hoUandi . '■ Specimen (possibly a composite) from the Phosphorites of Quercy in the Museum National d’Histoire Naturelle, Paris. AMNH 10.564. cast of specimen from Pikermi in the Museum National d'Histoire Naturelle. Paris. '■ CM 2299, cast of specimen from Sansan. On the proximal surface of the ectocuneiform, the shield-shaped navicular facet is broadest dor- sally and occupies all except the volar angle of the bone, where there is a well-developed, rugose volar process (Figs. 14D, 17B). This process closely ap- proaches the larger tibiovolar process of the cuboid when ectocuneiform and cuboid are articulated. There are some differences among ectocuneiform specimens in proportions and in the development of a single or double cuboid facet on the fibular surface. An important feature on the dorsal part of the fibular surface (Fig. 17B), distal to the cuboid facet(s), is a rounded facet for Mt IV, very distinct and continuous with the cuboid facet(s). On the tib- ial side of the ectocuneiform lie two articular sur- faces (Fig. 17A). The proximal one, for the meso- cuneiform, is small and flat. It occupies the middle third of the dorsal to volar depth of the ectocunei- form and is delineated from the proximal surface of the bone by a sharp ridge. Distal to this facet and occupying the entire depth of the ectocuneiform is the band-like, slightly convex facet for Mt II. Thus the ectocuneiform of M. elatus contacts all three metatarsals. On its distal surface is an irregularly shaped and generally flat articular surface for Mt III, occupying all except the most volar part of the bone. The mesocuneiform is a small bone with a slightly concave navicular facet covering most of its prox- imal side. The slightly convex facet for Mt II oc- cupies the volar two-thirds of the distal surface. There is also a small articular facet on the tibial side of the mesocuneiform, probably for an entocunei- form. No entocuneiform has been shown for any Moropus species, but apparently a small remnant of the bone was present. It did not articulate with Mt II, and there is no sign of an articulation for it on any naviculars examined. The presence of an entocuneiform in M. elatus , as well as in Ancylo- therium (A.) pentelicum where a facet for it is also known on a mesocuneiform specimen (a specimen in the Museum National d’Histoire Naturelle. Paris, casted as AMNH 10,564), is in disagreement with a statement by Radinsky (1963:7) that the entocu- neiform was lost in the Chalicotheriidae. The en- tocuneiform facet has only a small dorsal to volar extent in M. elatus . Metatarsals (Figs. 14-15, 17-19, 27). — Metatar- sal proportions provide an important means of sep- arating certain Moropus species, for there was ap- parently an increase over time in width relative to length. Proportions and other morphological fea- tures are fairly consistent within a single species, even despite variations in absolute size (Table 6). Thus where intraspecific variation can be evaluat- ed, a metatarsal or other foot element can provide a very useful characterization of a species. For this reason a metatarsal (YPM 13,081, Mt II) was cho- sen as M. elatus lectotype from among the type material mentioned by Marsh (1877). Mammalian feet are often considered too plastic for detailed taxonomic work, but at least in Moropus such a restriction does not seem to apply, possibly because of rapid evolution of and strong selection on the manus and pes while the dentition remained rela- tively plastic within species and constant between species. YPM 13,081 , the Mt II chosen as the lectotype, corresponds closely in proportions and morphology to specimens from both the Agate Quarries and Morava Ranch Quarry (Fig. 18). There is some vari- ation among Mt 11 specimens in the size and shape of the concave mesocuneiform facet, which occu- pies the proximal surface (Fig. 17D). In some (YPM 13,081 and the Morava Ranch Quarry specimens) the facet has a greater dorsal to volar extent and accompanies the ectocuneiform facet to the volar edge, separated from the latter facet by a blunt pro- jection. In others (primarily Agate Quarry speci- mens) the mesocuneiform facet does not extend so far in the volar direction and is more separated in its volar part from the ectocuneiform facet. The sin- 1978 COOMBS— REEVALUATION OE MOROPUS 39 Table 6. — Length and width (in mm) of metatarsals of selected schizotheriine species, with proportions of width versus length. Species and specimens Greatest length Greatest distal width Minimum shaft width Length/distal width Length/shaft width Moropus elatiis AMNH U31%^ 133.7 Mt I! 41.6 26.6 3.2 5.0 Moropus elatus CM 1701 109.9 30.4 18.9 3.6 5.9 Moropus elatus YPM 13.081 (lectotype) 136.2 36.4 26.2 3.7 5.2 Moropus elatus F:AM 54,447'> 137.1 38.3 24.5 3.6 l5 Moropus elatus ( range three Frick specimens'’; 109.3-139.5 30.4-37.5 20.9-22.3 3. 4-3. 7 5.1-6.25 mean three Frick specimens'’) 120.0 33.6 21.6 3.6 5.5 Schizotheriuin priscum (see specimen in Table 5) 95 19 15 5.0 6.3 Ancylotherium (A.) pentelicum (see specimen in Table 5) 127.0 52.3 38.1 2.4 3.3 Moropus elatus AMNH 14,378" 146.4 Mt III 44.3 28.9 3.3 5.1 Moropus hollandi FMNH PI 3000 130.7 39.4 28.4 3.3 4.6 Moropus cf. M. hollandi F:AM 54,903 115.7 43.3 30.4 2.7 3.8 Schizotherium priscum (see specimen in Table 5) 118 24 23.2 4.9 5.1 Ancylotherium (A.) pentelicum (see Table 5) 171.3 61.9 47.1 2.8 3.6 Moropus elatus AMNH 14.378" 148.3 Mt IV 43.9 28.2 3.4 5.3 Moropus elatus CM 1701" 126.8 34.9 23.2 3.6 5.5 Moropus elatus AM 9955" 158.7 51.2 36.0 3.1 4.4 Moropus elatus (range five Frick specimens": 123.9-164.8 35.7-45.2 26.0-31.9 3. 5-3. 8 4. 8-5. 4 mean five Frick specimens") 149 41 29 3.6 5.1 Moropus cf. M. hollandi F:AM 54.903 129.8 38.5 27.5’- 3.4 4.T Moropus sp. (St.-Gerand) (see Coombs, 1974) 88.9 28.0 19.8 3.2 4.5 Schizotherium priscum (see specimen in Table 5) 116 22 28 5.3 4.7 Ancylotherium (A.) pentelicum (see Table 5) 147.2 59.3 45.3 2.5 3.2 From the Agate Spring Quarries. From Morava Ranch Quarry. Approximate measurement. gle ectocLineiform facet at the fibular edge of the proximal surface has a primarily fibular, but slightly proximal, orientation (Fig. 18D). In some speci- mens the area connecting the dorsal and volar parts of this facet is slightly constricted. Distal to the ec- tocuneiform facet and adjacent to it is a small Mt III facet, whose dorsal and volar parts are con- nected only by a very narrow band at the facet’s proximal edge. A dorsofibular tubercle at the prox- imal end of all the Mt II specimens helps to prevent 40 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Fig. 18. — Dorsal views of A) left Mt II of F:AM 54,447. M. ekitus. from Morava Ranch Quarry, Box Butte Co., Nebraska; B) right Mt II of YPM 13.081, lectotype of M. ehitns\ C) left Mt II of AMNH 14,378, M. elatus. from the Agate Spring Quarries, Sioux Co., Nebraska; D) Fibular view of left Mt II of AMNH 14,378. Facet abbreviations as in Fig. 16. any dorsal to volar movement of Mt II against the ectocimeiform or Mt III. A depressed area on the fibular surface bordering the Mt III facet is suc- ceeded distally by a strongly rugose area. Very characteristic of this bone is an oblique longitudinal ridge for muscle insertion along the tibiodorsal an- gle of the proximal part of the shaft; this ridge is more distinct in some specimens than in others. The distal end of Mt II is slightly asymmetrical. The proximal end of an Mt III of YPM 24,631, a paralectotype of M. elatus (see Fig. 19A, B), is in- distinguishable from other referred Mt III speci- mens but probably belonged to a larger individual than the Mt II considered as the lectotype. On each Mt III specimen, the ectocuneiform facet, on the proximal surface, is rather flat and slanted distally toward the tibial side (Figs. 17D, 19A). The shape of the proximal end of the bone is subtriangular with the ectocuneiform facet occupying practically all of this surface. On the tibial side of the bone (Figs. 17A, 19A) the small articular surface for Mt III is bipartite with dorsal and volar parts connected by a very narrow strip of articular surface (the con- nection may be absent in some specimens). Some- times the dorsal part of the facet is larger, some- times the volar, and sometimes the parts are subequal. The two facets for Mt IV, on the fibular side of the proximal end (Figs. 17B, 19B), are large compared to the Mt II facet. In some specimens the dorsal and volar facets are distinctly separated by an intervening sharp depression; in others they are directly adjacent though not confluent. In all the specimens examined, the dorsal Mt IV facet is eas- ily the larger (about twice as large as the volar facet), and both facets are flat to slightly concave and slanted slightly toward the center of the fibular 1978 COOMBS— REEVALUATION OF MOROPUS 41 Fig. 19. — Tibial (A) and fibular (B) views of proximal end of Mt III of YPM 24,631, a paralectotype of M. elutns. Dorsal (C) and volar (D) views of rare fused proximal and medial phalanges of digit II of the pes of M. elaius, F:AM 54,444, from Morava Ranch Quarry, Box Butte Co.. Nebraska. Dorsal (E) and volar (F) views of unfused proximal and medial phalanges of digit II of the pes of M. elaius, AMNH 14,378. from the Agate Spring Quarries, Sioux County, Nebraska. Facet abbreviations on Mt III as in Fig. 16. surface (probably restricting dorsal to volar move- ment between Mt III and Mt IV). Distal to the ar- ticular surfaces, rugosities are present on the tibial, volar, and fibular surfaces of the shaft. At the distal end of Mt III the phalangeal articulation is roundly convex and laterally symmetrical, separated dor- sally by a depression from the shaft of the bone. There are strong lateral tubercles and a prominent but not sharp keel between the sesamoid facets. In all examined specimens of Mt IV the proximal head is of quadrilateral shape, slightly wider trans- versely than deep. The proximal surface (Fig. I7D) is occupied almost entirely by the flat to slightly concave cuboid facet, which is separated by the facets on the tibial side of the bone by a sharp dorsal to volar ridge. The proximal end of the tibial surface has two oval-shaped facets for Mt III (Fig. 17C), of which the dorsal is about twice as large as the volar. The two articular platforms are separated from each other by a groove, and the area just distal to the facets is very rugose. Part of the dorsal facet faces in a dorsal as well as a tibial direction. When Mt IV 42 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Fig. 20. — Dorsal (A, C) and ulnar (B, D) views of the duplex of digit II of the manus of a presumed male (A, B) of M. elatiis, AMNH 14.427. left side, and presumed female (C, D) of M. elcitiis, AMNH 14,425, right side, both from the Agate Spring Quarries, Sioux Co,. Nebraska. is articulated with Mt HI, the contact between the two bones occupies almost all of the volar facet on Mt IV but only the distal three-fourths of the dorsal facet. The remainder of the dorsal facet articulates with the distal fibular edge of the ectocuneiform (Figs. I7C, D). There are no facets on the fibular surface of Mt IV, but the fibular-volar angle of the proximal end is enlarged and appears to overhang the shaft because of some dorsal to volar compres- sion of the shaft’s fibular side. The proximal head and distal end of Mt IV are each rotated in a dif- ferent direction relative to the proximal part of the shaft axis. On the slightly asymmetrical distal end, the tibial lateral tubercle is stronger than the fibular one, and the tibial side of the phalanx articulation extends farther distally. Pes summary. — Movement in the pes of M. ela- tiis is more restricted than that in the manus, and the pes as a whole is less divergent than the manus from that of other tridactyl perissodactyls. Flexion and extension of the pes occurs primarily at the joint between the tibia and astragalus; most of the joints among tarsals and between tarsals and meta- tarsals are restricted in size and have stops espe- cially to prevent dorsal to volar flexion. A small degree of side to side movement is possible, how- ever. The relatively large number of associated speci- mens permit the observation that Mt III and Mt IV were subequal in length in M. elatus and that both were longer than Mt II (Table 5). Consistent with this emphasis on both Mt III and Mt IV, the ecto- cuneiform articulates with Mt IV as well as Mt III so that weight is spread on both digits (the ecto- cuneiform also has some articulation with Mt II). In Ancylotlieriiim (Ancylotheriiim) penteliciim (specimen in Museum National d’Histoire Natu- relle, Paris, figured by Gaudry, 1862), Mt III is clearly the longest digit, and there is no contact between the ectocuneiform and Mt IV. Mt IV in M. elatus was also capable of some movement in the fibular direction relative to the rest of the pes; dur- ing such movement its contact with the ectocunei- form was lost and its ectocuneiform facet then ar- ticulated with Mt III. Metatarsals are much shorter than metacarpals of the same animal. Phalanges of the pes are similar in structure to those of the manus and were presumably capable of roughly the same movements. They are, how- ever, smaller, and there is no special enlargement of the phalanges of digit II of the pes. Proximal Phalanges In M. elatus, morphology of phalanges does not differ markedly from digit to digit, except for the distinctive phalanges belonging to digit II of the manus and pes. Thus an isolated phalanx cannot easily be referred to a particular digit or even to the manus or pes, especially when one considers the great size variation among individuals (Coombs, 1975). Nevertheless, certain generalizations con- cerning the phalanges are possible. Digit II manus (Fig. 20). — A large, fused proxi- mal and medial phalanx of digit II of the manus is a characteristic feature of Moropus . Described as a duplex by Holland and Peterson (1914:357-359), such a bone is present, though broken, among the paralectotypes of M. elatus (YPM 24,631, figured 1978 COOMBS— REEVALUATION OF MOROPUS 43 by Holland and Peterson, 1914:223). In juvenile in- dividuals the two bones were not yet fused or were incompletely fused, but in at least one old individual (UCMP 14,377, originally AMNH 14,377, Fig. 21) the ungual phalanx is also fused to the duplex. The angle of fusion in the latter specimen may suggest the usual hooked position of these phalanges during life. Characteristic of the duplex of the manus is the strongly dorsal orientation of the facet for Me II. The Me II facet is roughly heart-shaped, without any particular asymmetry, and is gently concave. Its ulnar edge is slightly sharper than its radial one. Two volar tubercles (radial and ulnar), separated by a notch, extend proximally beyond the Me II facet. The shaft of the duplex is deepest in those speci- mens where fusion has proceeded most completely; in some specimens the junction of the original two bones has become obscured. Distally the duplex is expanded in the dorsal to volar direction by the curved facet for the ungual phalanx. This facet is shaped like a pulley, with a broad, deep groove be- tween the two flanges, and curves from the dorsal edge of the distal end onto the volar surface of the duplex. Dorsally the groove comes to an end before the disappearance of its lateral flanges so that a stop for dorsal movement of the ungual phalanx is de- veloped. On the volar side, the flanges of the artic- ular facet curve apart and become sharper before they disappear. When a number of duplexes from digit II of the manus are available, as in the Cook Collection (N = 9) and other collections from the Agate Quar- ries in the American Museum of Natural History, it becomes clear that two size groups are present. The smaller duplexes (Figs. 20C, D) are slightly less asymmetrical than the larger ones (Figs. 20A, B), and their proximal facet is not quite so flattened at its radial edge. The possibility that such smaller du- plexes might belong to a digit other than digit II of the manus must be considered, especially in view of the occasional fusion of phalanges belonging to digit II of the pes. However, the difference in size in this case seems to be due to sex dimorphism, not to phalangeal fusion in additional digits. The mean length of the smaller manus duplexes in the Cook Collection (four specimens) is about 80% of that of the larger duplexes (six specimens). This ratio cor- responds closely to the male/female size ratio cal- culated from the mean sizes of large and small radii and tibiae given by Coombs ( 1975). Rare duplexes, which do belong to digit II of the pes (see below), are only about two-thirds the size of the large ma- nus duplexes and correspond closely to the com- bined lengths of unfused phalanges of digit II of the pes. Small or intermediate sized duplexes in schizo- theriine species other than M. elatus must be evaluated carefully to determine whether they be- long to the manus or the pes. Digits III and IV of manus . — In M. elatus much of the great expansion of digit II of the manus is the result of expansion of the ungual phalanx and the part of the duplex representing the medial phalanx. The proximal phalanges of digits 111 and IV of the manus are not much smaller than the proximal part of the duplex. Because digit III and digit IV proximal phalanges are about the same size, it is difficult to assign each to the correct digit even though they differ from each other in morphology. One kind of proximal phalanx 1) is more symmetrical than the other sort, narrowing rapidly in the distal direction, 2) has metacarpal facets on its dorsal surface but with more proximal orientation than on the duplex, 3) has flattened articular areas on the dorsal surface just distal and adjacent to the metapodial facet (pre- sumably for stops against hyperextension), 4) has some development of variable volar intermediate tubercles on the shaft, increasing the depth of the bone, and 5) has a slightly asymmetrical distal facet (a shallow, curved pulley). A phalanx of YPM 24,632 (not mentioned by Marsh, 1877, but figured by Holland and Peterson, 1914:224) is among the second sort of proximal pha- lanx. Features include I) transversely expanded metacarpal facet, 2) some lateral torsion of the shaft (more than in previous sort), 3) pronounced flat- tened area on dorsal surface just distal to the meta- carpal facet and particularly noticeable on the ulnar (?) side, and 5) distal facet asymmetrical (one side of pulley longer and less curved). Digit II of pes (Figs. 19C-F). — On a single indi- vidual the proximal phalanges of the pes are slightly smaller than those of the manus and are not so strongly expanded at their proximal ends. Digit II of the pes is represented by a comparatively large number of specimens from the Agate Quarries and Morava Ranch Quarry. In all of the specimens the bone is flattened in the dorsal to volar direction compared to most other proximal phalanges, and the articular facet for Mt II has a strongly dorsal orientation (a similarity to the facet for Me II on the duplex of the manus). The Mt II facet extends onto the proximovolar tubercles, between which there is a pronounced notch. This notch is generally characteristic of the phalanx in its sharp indentation 44 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 process Duplex facet B Subungual process Fig. 21. — A) Unusual, completely fused proximal, medial, and ungual phalanges of digit II of the manus belonging to UCMP 14,377, M. elattis. B) Radial view of a left ungual phalanx of digit II of the manus of a presumed female of A/, elatiis , AMNH 14,425. C) Ulnar view of a right ungual phalanx of digit II of the manus of a presumed male of M. elatus. AMNH 14.427. All from the Agate Spring Quarries, Sioux Co., Nebraska. D) Dorsal view of fused proximal and medial phalanges of digit 11 of the pes of YPM 12,194, type of “A/, senex." from Dayville. Oregon, and small width. A distal facet for the medial pha- lanx. grooved but with an irregular surface, curves at the distovolar angle of the phalanx. The articu- lation between proximal and medial phalanges is so close that there could have been little or no move- ment between them (Eigs. I9E, E). Holland and Peterson (1914:375-376) noted the occasional pres- ence among the Carnegie material of fused proximal and medial phalanges of digit II of the pes. Their observation is borne out by the presence among the Morava Ranch material (E:AM 54444, Eigs. 19C, D) of another such example. A duplex for digit II of the pes is best distinguished from that of digit II of the manus by its smaller size and by its articular surface for the ungual phalanx being less deeply grooved and with less strongly flared volar edges. 1978 COOMBS— REEVALUATION OF MOROPUS 45 Such fusion in the pes was quite uncommon in M. elatiis, probably occurring in no more than one out of ten individuals. Digits 111 and IV of pes. — Proximal phalanges of these digits are not easily distinguishable from one another. Both are longer and thicker than unfused proximal phalanges of digit II of the pes. One group of these phalanges (? of digit III) is slightly shorter and more expanded proximally than is the other group. In ail of the specimens the articular surface for the metatarsal has more proximal orientation than is on the proximal phalanx of digit II. This dif- ference is consistent with comparisons of digits III and IV of the manus against digit II, except that in the phalanges of the pes flat stopping surfaces at the distal edge of the metatarsal facet are small or absent. This difference probably corresponds to the smaller development on the metatarsals of the tu- bercles against which the flattened areas of the pha- langes may rest during hyperextension. Both pha- langes are more symmetrical than their counterparts in the manus. Medial Phalanges Unfused medial phalanges of manus and pes are not easily distinguished in this species. All have a curved concave proximal facet (for the proximal phalanx), divided into two halves by a ridge, and a pulley-like, curved distal facet for the ungual pha- lanx. Medial phalanges of digits HI and/or IV of the manus are more likely than others to have asym- metrical development of the proximal facet. The medial phalanx of digit II of the pes, even where it has not fused to the proximal phalanx, has a right angle bend in its proximal facet which would pre- vent any movement of the two phalanges against one another. Ungual Phalanges (Figs. 2 IB, C) Except for the large ungual of digit 11 of the ma- nus, it is practically impossible to refer ungual pha- langes to a given digit, except that in a given indi- vidual, unguals of the manus are usually larger than those of the pes. The large ungual phalanx of digit II of the manus is taller and more transversely compressed than other unguals of this species, and there is a thick, tall dorsal process onto which the duplex facet ex- tends. A large, rugose subungual process prevents the duplex facet from reaching the volar surface of the bone. Although the duplex facet covers a large arc at the proximal end of the ungual phalanx, it does not follow an evenly rounded curve but has a sharp angle between its proximal oriented and volar oriented parts; this angle corresponds to one on the distal facet of the duplex and effectively prevents most movement between the two bones. Dorsally the claws are quite sharp, and there is either equal fission between the dorsal and volar edges of the claws or greater fission dorsally. An intermediate sized ungual, not so tall nor so strongly compressed transversely as the larger one, belongs to digit 111 or IV of the manus and has both a large dorsal process and subungual process. As in the larger phalanx, the claws are sharp and clo- ven equally above and below. The other ungual of the manus compares closely with the unguals of the pes. Of the three small pes unguals one has a larger dorsal process than the others. Generally the pes unguals are relatively broad transversely, and the two claw sides are separated by a small space. The two sides may be slightly asymmetrical and are usually cloven more deeply above than below. The subungual process is much reduced. Sesamoids A number of sesamoids are known in M. elatiis. These articulate with the distovolar surfaces of metacarpals and metatarsals, but it is difficult to determine, which sesamoids go with which digits. Moropiis distans Marsh Moropus distans Marsh, 1877:249. Moropiis distans: Peterson, l907/>:734; Holland and Peterson, 1914:221; von Koenigswald, 1932:22; Colbert, 1935:13; Be- lyaeva, I954;49 (locality incorrectly cited as Nebraska by von Koenigswald, Colbert, and Belyaeva). Lectotype YPM 12,193a, coossified proximal and medial phalanges (duplex) of digit II of pes from the “Bridge Creek beds” in the John Day Basin, Ore- gon, ?Arikareean (Fig. 22F). Paralectotype Medial phalanx of YPM 24,627 mentioned by Marsh ( 1877). Hypodigm Cuboid of YPM 24,628, first mentioned and fig- ured by Holland and Peterson (1914:220; see Figs. 16C, F); distal end of Me II of YPM 24,628. Para- lectotype and hypodigm were collected with the lec- 46 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 1978 COOMBS— REEVALUATION OF MOROPUS 47 totype and were originally given the same catalogue number; however, they are probably not the same individual and are renumbered in this paper. Diagnosis A small chalicothere, smaller than any known specimen of M. elatus\ cuboid having more proxi- mal extension of navicular facet than in M. elatus; duplex of digit II of pes flatter in dorsal to volar direction than rare pes duplexes of M. elatns, fusion to form this duplex possibly more common than in M. elatns . Discussion Although Moropus distans is not so well known as M. elatns, Holland and Peterson (1914:217) es- tablished it as the type species of Moropus because Marsh (1877) had mentioned it first. Additional ma- terial must be obtained before an estimate of the size range in this species can be made. Small av- erage size may be a primitive feature in the genus Moropus (see also Moropus sp. from St.-Gerand- le-Puy, Coombs, 1974). The distribution of M. dis- tans cannot be evaluated without fuller knowledge of other small, early North American chalicotheres, for the species is presently known only from Ore- gon. Attention is directed to Moropus oregonensis (Leidy, 1873) from the same region as M. distans. Both species were described before it became known that the teeth and footbones of chalicotheres actually belong to the same kind of animal (Filhol, 1891). M. distans may be synonymous with M. or- egonensis, which includes only teeth, but conspe- cificity cannot be established. If evidence of syn- onymy should become available, M. oregonensis (discussed below) would become the type species of Moropus . Marsh (1877:250) also named Moropus senex for a duplex, YPM 12,194 (Fig. 2 ID), from near Day- ville, Oregon. This phalanx, which belongs to digit II of the pes, is described below but has no specific diagnostic features. Moropus senex is therefore considered to be a no men dnbinm. The type spec- imen resembles rare pes duplexes of M. elatns but is considered for the present to represent Moropus indet. Description and Comments YPM 12,193a, the small duplex here designated as the lectotype of Moropus distans (Fig. 22F), ap- pears to belong to digit II of the pes but is slightly flatter, especially in the metapodial facet, than du- plexes of the same digit in M. elatns and "M. se- nex." Its distal end is not known, and thus definite identification as a duplex of the pes cannot be made, despite its small size. In addition to its unusually small size compared to other Moropus specimens, this duplex may be significant in the mere presence of fusion of proximal and medial phalanges in a pes digit, for such fusion is rare in M. elatns (no more than 10% of known specimens). Although it is haz- ardous to generalize on the basis of limited material, the presence of such a duplex among the very few specimens of M. distans may suggest that devel- opment of a pes duplex was common in this species. Formation of a duplex for pes digit 11 is apparently universal in certain Moropus species, for example M. merriami . Fusion of phalanges to form a duplex of pes digit II is also of interest in the single specimen of "M. senex," YPM 12,194 (Fig. 21D). This specimen clearly belongs to digit II of the pes by virtue of its shallow distal facet without flared lateral edges. It is thus not a manus duplex, unless an unusual one, of a small sized Moropus species. Morphological resemblances to an M. elatns duplex of pes digit II from Morava Ranch Quarry (F:AM 54,444, Figs. 19C, D), including a size similarity, make tempting a synonymy of "M. senex" with M. elatns. YPM 12,194 does not seem to belong to M. distans , but the rarity of phalangeal fusion in the pes of M. ela- tns and our lack of knowledge of postcranials of M. oregonensis make any taxonomic alignment of "M. senex" premature. YPM 24,621 , the medial phalanx paralectotype of M. distans, presents no features of special taxo- nomic interest, except for small size, nor does the Fig. 22. — A) Holotype of M. oregonensis , YPM 10,030, a right maxillary fragment with P"* and partial M'. B-E) Hypodigm of M. oregonensis: left P^-P'* (B) and right (C) of YPM 10,030a; left M‘-M^ (D) of YPM 10.030b; and right M^ (E) of AMNH 7259. E) Lectotype of M. distans. YPM 12,193a, coossified proximal and medial phalanges of digit II of the pes. A-D and E are from the “Bridge Creek Beds” in the John Day Basin, Oregon; E is from “near Antelope Springs,” Oregon. G and H) Dorsal (G) and volar (H) views of a left carpus of M. hollandi , EMNH P13000 (minus pisiform), from near Jay Em. Goshen Co., Wyoming. Bones correspond to those of right carpus of M. elatns labeled in Eig. 9; facet abbreviations as given in Fig. 10. 48 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Me II distal end of YPM 24.628. Both specimens generally resemble their counterparts in M. elatus except that the two sesamoid facets on the M. dis- tans Me II are slightly less unequal in size than they are in M. elatus. The cuboid of YPM 24,628 is small, with a max- imum dorsal to volar thickness of 47 mm, compared with 75 mm for AMNH 14,378 (a medium-sized cu- boid of M. elatus). Morphologically, the specimen is close to cuboids of M. elatus. M. hollaudi , and M. inerriami. but the tibiovolar rugose process is proportionately somewhat smaller than in other Moropus specimens (Figs. I6C, F). Unfortunately the volar part of the tibial side of the bone is dam- aged. so that the extent of navicular and ectocu- neiform facets are not completely clear. Neverthe- less. the navicular facet appears to extend proximally to reach the proximal edge of the tibial side and thus adjoins the facet for the calcaneum (Fig. 16C). Such proximal extension of the navic- ular facet is seen in M. inerriami (UCMP 78.727) but not in M. elatus or M. Iwllandi . The ectocu- neiform facet on the M. distans cuboid, though bro- ken at its volar edge, is apparently single rather than double; M. elatus has variably a single or double facet, but more typically a double one. Moropus oregonensis (Leidy) Lopliiodon oregonen.sis Leidy. IS73. pi. 2. fig. I (also described without name pp. 219-220). Lopliiodon oregonensis: Sinclair, 1901:702. Moropus oregonensis: Holland and Peterson, 1914:219; Colbert. 19.35:13; Belyaeva 1954:49. Holotype YPM 10,030, right maxillary fragment with worn P"* and partial M' from Bridge Creek area, John Day Basin, Oregon (specimen label states "upper John Day beds”); ?Arikareean; see Fig. 22A. Hypodigm Left P'^-P^ and right M'^ of YPM 10,030a, which may be part of the same individual as the type but were not figured by Leidy (Figs. 22B, C); YPM 10.030b, left M’-M^ collected at the same locality as type (Fig. 22D); AMNH 7259, M^(Fig. 22E) from "near Antelope Springs. Oregon.” Diagnosis Species known only from dental remains, smaller than the smallest known representatives of M. ela- tus', metaloph on unworn upper molars (especially M'b with labial origin very close to mesostyle, even closer than in M. elatus', upper molars not strongly elongated; P'* and P“* without strong anterolingual or posterolingLial cingula and with wear along trans- verse lophs proceeding to protocone before lophs are completely worn away (especially along meta- loph). Discussion Like Moropus distans, Moropus oregonensis shows a number of similarities to Moropus merria- nii and some of its undescribed Hemingfordian rel- atives, possibly more resemblance than to Moropus elatus. Although of small size, the known teeth of M. oregonensis are slightly larger than might be ex- pected to belong to the same individuals as the small postcranials of M. distans . Nevertheless these two species may be synonymous, for the pos- sibility of sexually dimorphic size differences must be considered. The labial origin of the metaloph so close to the mesostyle on unworn M® is a character shared with Seliizotherium (see 5. priseuin in Coombs, 1976); it is apparently primitive within the genus Moropus and is lessened in more advanced species. Description and Comments /V/^ (Figs. 22C, E). — The two available speci- mens. YPM 10,030a and AMNH 7259, are of similar size. At a glance they appear to have quite different morphology, but most of the points of variance are easily explained by wear, for AMNH 7259 had only recently erupted. In the unworn tooth the origin of the metaloph is very close to the mesostyle, where- as in the worn tooth the origin has moved poster- olingually. All parts of the metaloph are of approx- imately equal height, but there is a slightly separated summit, representing the hypocone, at the lingual end of the metaloph. The anterior part of the tooth, including the relatively large proto- conule and complete protoloph, is little worn even on the more worn specimen, YPM 10,030a. In nei- ther specimen does a branch of the anterolingual cingulum join the protoloph to ascend to the tip of the protocone. In YPM 10,030a, the anterolingual cingulum remains separate, ridge-like and uniform as it skirts the protocone lingually; posteriorly it enters the central valley as a weak ridge or fold but does not close off the lingual opening of the central valley. In AMNH 7259 the cingulum is less uniform and approaches the wall of the protocone more closely but does not fuse with it. The postfossette 1978 COOMBS— REEVALUATION OF MOROPUS 49 is partly closed off by the posterior cingulum, and a labial rib is present opposite the paracone. On AMNH 7259 a very small crista (see Butler, 1965: 178, for definition) is present. Principal differ- ences from M. elatus include the smaller degree of wear on the protoloph and the morphology of the anterolingual cingulum. The origin of the metaloph is closer to the mesostyle in an unworn specimen than in any other known Moropus species, even M. elatus . and M' (Fig. 22D). — In addition to its smaller size, M' is more quadrate (less elongated compared to width) than M^. Both teeth closely resemble M^. but on a branch of the anterolingual cingulum appears to join the protoloph (but weakly), in con- trast to M'f The parastyle and mesostyle, particu- larly on M^, are slightly weaker than they are in M. elatus. Posterior to the protocone the lingual cin- gulum is only weakly visible. and P^ (Figs. 22 A, B). — Known upper pre- molars are somewhat worn. They show many sim- ilarities to specimens of M. elatus . but in their pat- tern of wear come closer to some of the small Hemingfordian relatives of M. nierriaini. It is clear, despite wear, that the two transverse lophs were complete and that a protoconule was present. Wear reaches the protocone earlier in the life of each tooth than it does in P* and P^ of M. elatus , for in these specimens the protocone has been worn near- ly flat and lies quite below the level of the protoloph at the same time that the protoloph is still complete- ly or at least partially visible. Apparently wear pro- ceeds from the ectoloph along the metaloph to the tip of the protocone fairly early during the life of the animal, whereas wear along the protoloph is much slower. Wear on the anterior part of the ec- toloph slightly reduces the parastyle in comparison to that often seen in even more worn specimens of M. elatus. Anterolingual and posterolingual cingula are not at all expanded but form a uniform ridge skirting the protocone, fusing with it at only one point. Moropus hollandi Peterson Moropus elatus: Peterson, 1907a: 60. Moropus hollandi Peterson. 1913:673. Moropus hollandi: Holland and Peterson. 1914:232. Moropus elatus: Matthew, 1929:520. Moropus hollandi: von Koenigswald, 1932:22; Colbert, 1935:13; Belyaeva, 1954:49. Holotype CM 1424, a partial skeleton of a medium-sized individual, including radius-ulna, scaphoid, trape- zoid. magnum, unciform. Me 11-V. femur, two tib- iae, two astragali, two calcanea, duplex of digit 11 of manus, proximal phalanx, and medial phalanx, from Peterson’s ( 1 901a ) Upper Harrison beds near Nebraska-Wyoming state line along Niobrara Riv- er. Early Hemingfordian. Many of the elements of the holotype of M. hollandi were figured by Holland and Peterson (1914) and. except for the scaphoid and trapezoid (Figs. 26E-H), are not repeated in the present paper. Hypodigm F:AM 54902a- 1. two scaphoids (Fig. 26C). two lunates (Fig. 26D), distal end of radius, distal end of radius-ulna, cervical vertebra VI 1, two proximal phalanges, astragalus (Fig. 23C). sesamoid, proxi- mal end of ulna, and cuboid, from Upper Harrison equivalent beds (Skinner, 1968) 7 mi south of Chug- water, Wyoming; FMNH P13000, an atlas (Fig. 6B), pelvis, almost complete left manus (Figs. 22G, H. 24, 25, 26A, B), duplex and ungual phalanx of digit II of manus (Figs. 28D-F), and navicular, cu- boid, and Mt 111 of left pes (Figs. 16A, D, G, 27B, D-F) from “Jay Em Creek, 2 mi east of ranch. Lusk, Wyoming;’’ and FMNH PI 2094, a skull (Figs. 23A, B) from the “east wall of Jay Em Creek, a tributary of Rawhide Creek, near Lusk, Wyo- ming.’’ The Field Museum specimens, collected by Riggs near Jay Em, probably also come from Upper Harrison equivalent beds. All the material referred to F;AM 54,902 is larger than the holotype, and some of it probably belongs to a single individual. Additional Material F;AM 54,903, a Mt III and Mt IV from Upper Harrison equivalent beds (Skinner, 1968) near Jay Em, Wyoming, is here referred to Moropus cf. M. hollandi. Its uncertain reference is because of dif- ferences, whose significance is unclear, between its Mt III (Figs. 27C, 28A-C) and that of FMNH PI 3000. The latter specimen is more definitely re- ferable to M. hollandi by virtue of its known manus morphology. Diagnosis Moropus species resembling M. elatus in most respects, including size, dental morphology, and absence of proximal extension of navicular facet on the cuboid. Differs from M. elatus in the generally shorter proportions of Mt 111 relative to minimum shaft width (Table 6) and from all known Moropus species in the absence or strong reduction of a tra- 50 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 pezium in the manus. Trapezium facets absent on scaphoid. Me II, and probably on trapezoid. Lu- nate, cuneiform, and trapezoid commonly com- pressed in the dorsal to volar direction relative to their width. Discussion The Frick and Field Museum specimens cited in the hypodigm are of great importance in affirming M. hollandi as a separate species. Three additional scaphoids, one trapezoid, and one Me II, all of which closely resemble the holotype, and new metatarsal specimens reinforce the idea that the dif- ferences of the holotype of M. hollandi from M. elatus are not mere individual peculiarities. Ab- sence of a trapezium, along with the dorsal to volar compression of several carpals, suggests a degree of superflexion of the manus unusual within Mor- opus . The large number of similarities between M. hol- landi and M. elatus emphasize the close affinities between these two species. At the same time they create difficulties in identifying specimens, espe- cially in those many cases where diagnostic ele- ments of M. hollandi are not included. Because of difficulties in identification, it is still unclear wheth- er the two species are separated by a temporal boundary or whether their occurrences reflect eco- logical differences in the assemblages in which they are represented. Additional specimens may some- day show a morphological continuity in the evolu- tion of M. hollandi from a population of M. elatus. Certainly some aspects of M. elatus morphology, when taken at one end of their range of variation (for example, shorter proportions of some metatar- sals) approach conditions seen in M. hollandi. On the other hand, no M. elatus specimen shows the scaphoid, trapezoid, or Me II characteristics con- sidered diagnostic of M. hollandi. Absence of a trapezium occurs in Moropus, so far as is known, only in M. hollandi. It is paralleled, however, in Ancylotherium (Ancylotheriiun) pen- telicum. where there is an even greater degree of mobility of the scaphoid on the trapezoid. There the scaphoid contacts Me II during extreme flexion as was reported by Schaub (1943; although Schaub concluded, probably erroneously, that the trape- zium in Ancylotherium was fused with Me II). Schizotherium turgaicum and Borissiakia betpak- dalensis apparently have also independently lost the trapezium (see Borissiak, 1946; Belyaeva, 1954). Slight proportionate shortening of the metatarsals of M. hollandi compared to M. elatus is consistent with the gradual shortening over time of metatarsals among a number of schizotheriine species (see Coombs, 1974). The Mt IV proportions of F:AM 54,903 (Moropus cf. M. hollandi) most closely ap- proach those of the small Mt IV of Moropus sp. from Aquitanian deposits of St.-Gerand-le-Puy, Al- lier, France (Table 6). Present evidence suggests that M. hollandi had a size range rather similar to that of M. elatus. The two best preserved individuals (the holotype and FMNH PI 3000) are of modest size, however, but material in F;AM 54,902 suggests that larger indi- viduals were also included in the species. The ho- lotype and FMNH PI 3000 probably represent fe- males, and the morphology of their duplex phalanges of the manus also supports this conclu- sion (see discussion of M. elatus duplexes). Description and Comments Skull and Upper Dentition (FMNH P12094, Figs. 23A, B) The single skull with upper teeth known for M. hollandi falls within the range of morphology of M. elatus and shows no special differentiating features. There is no sign of doming or skull elaboration, and certain landmarks, like the internal nares and in- fraorbital foramen, are in the same position as they are in M. elatus. The upper teeth also cannot be differentiated from those of M. elatus, although the length of the toothrow (P^-M'fi falls near the lower end of the M. elatus size range and the premolar row may be proportionately shorter. The teeth in this specimen are very little worn, so that the pro- toconule and protoloph on M^ are still complete. The labial origin of the metaloph on M^ is still quite close to (less than 10 mm from) the lingual base of the mesostyle. The base of the protocone is not folded in the central valley. Premolars of FMNH P12094 are most similar to those specimens of M. elatus in which the protoloph of P^ is not strongly developed and the lingual cingula of P^ and P"* are relatively small. On P^ and P"* the protoloph is taller than the metaloph, which is more worn. Despite the progressing wear on protoloph and especially meta- loph. the protocone is still completely unworn, a similarity in wear pattern to M. elatus as compared to M. oregonensis . Vertebrae and Pelvis The atlas and pelvis (FMNH P13000) cannot be easily differentiated from specimens of M. elatus. 1978 COOMBS— REEVALUATION OE MOROPUS 51 Fig. 23. — Lateral view of skull (A) and occlusal view of right P^-M^ (B) of M. hollandi . FMNH P12.094, from near Jay Em. Goshen Co., Wyoming. Volar view of astragalus (C) of F:AM 54.902g. referred to M. hollandi , from 7 mi south of Chugwater. Wyoming. For facet abbreviations on astragalus see Fig. 14. However, the proportions of the atlas are slightly different (see Eig. 6), for the anterior to posterior length of the ventral arch is small relative to the transverse width of the articular surfaces for the occipital condyles. F:AM 54,902e, a cervical ver- tebra VII, corresponds generally with specimens of M. elatiis but is also rather broad, particularly an- teriorly. Forelimb CM 1424, the holotype of M. hollandi, includes much of one forefoot; additional forelimb material belonging to F:AM 54,902 and FMNH PI 3000 makes the forefoot the best known part of the skel- eton of M. hollandi . The fused radius-ulna of CM 1424 does not differ from specimens of M. elatns in any important way but is strongly fused into a single unit for more of its length than are most M. elatns specimens. There is a slight difference between F:AM 54,902c and F:AM 54,902d, both of which include the distal ends of radii, in that the former has a more strongly compressed and dorsal to volar concave distal facet for the lunate. The less com- pressed condition (F:AM 54,902d) is more similar to that in M. elatns, but the compressed condition corresponds more closely to the morphology of the lunate of F:AM 54902b (Fig. 26). The carpus provides the principal characters by which M. hollandi is distinguished from M. elatns. The scaphoid differs from that of M. elatns in the apparent lack (or at least very strong reduction) of a trapezium facet, so prominent on the radiovolar 52 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Volar Volar hook Ulno-volar process Fig. 24. — Proximal (A. C. radial edge at left) and distal (B. D. radial edge at right) views of left scaphoid, lunate, and cuneiform (A. B) and left trapezoid, magnum, and unciform (C, D) of M. hollandi. FMNFl PI3000. from near Jay Em, Goshen Co.. Wyoming. Facet abbreviations as in Fig. 10. surface of the scaphoid in other Moropus species. In M. hollandi , the same area is occupied by a ridge separating the rough radiovolar surface from the smooth saddle-shaped trapezoid facet distal to it (Figs. 22H, 26A. C. E). The trapezoid facet (Fig. 24B) is very similar to its counterpart in M. elatiis, as is the proximal lunate facet (Figs. 26B, F). The distal lunate facet is also similar but does not extend so far in the proximal and volar directions; how- ever, it does reach the edge of the trapezoid facet in its most proximal part. As in M. elatns , the distal process is squared with a relatively flat magnum facet on its distal surface, bordered by trapezoid and distal lunate facets (Fig. 24B). Two lunate specimens are slightly compressed in the dorsal to volar direction, with proximal and dis- tal scaphoid facets curtailed at their volar edges (Fig. 25A). In F:AM 54,902b (Fig. 26D) the volar process is also rather short, and the facet for the radius is very strongly convex in a dorsal to volar direction and abbreviated at its volar edge. A third, larger, specimen (F:AM 54,902k) has a large volar process and does not seem to be compressed. The single cuneiform specimen, like the lunate of the same individual (FMNH P13000), shows a ten- dency toward compression in a dorsal to volar di- rection. Such abbreviation on the cuneiform is mostly in the lunate (Fig. 25C) and unciform (Fig. 24B) facets; the unciform facet is more concave than in specimens of M. elatns and is set off by a volar ridge. The unciform (two specimens) is basically similar 1978 COOMBS— REEVALUATION OF MOROPUS 53 Fig. 25. — A) Radial view of left lunate, magnum, and Me III (dorsal edge at right), B) ulnar view of left lunate, magnum, and Me III (dorsal edge at left), C) radial view of left cuneiform, unciform, and Me IV (dorsal edge at left), and D) proximal view of left Me 11, Me 111. and Me IV (dorsal edge below, radial at left) of M. hollandi, FMNH PI3000, from near Jay Em, Goshen Co.. Wyoming. Facet abbreviations as in Fig. 10. to that of M. elatus. On the radial surface (Fig. 25C), the platformed magnum facet extends farther distally, however, and is separated from the Me III facet by a clearly defined ridge. Distinctness and slant of the magnum facet of these specimens re- sembles those in Morava Ranch Quarry specimens of M. elatus slightly more than those from Agate. At its proximal edge the magnum facet has a narrow proximal strip of facet passing in a volar direction from the main body of the facet and bordering the radial edge of the lunate facet for some distance; this addition to the magnum facet is present in some but not all M. elatus specimens and presumably al- lows tighter articulation between magnum and un- ciform. As in M. elatus, the boundary between Me III and Me IV facets is indistinct. Because of the great distal extent of the magnum facet, that for Me III is a bit smaller than the one in M. elatus', there appears to be another tiny area of contact with Me III, only rarely present in M. elatus, on the radial side of the ulno-volar process. The two known magnum specimens resemble material of M. elatus very closely, except that the more distal extension of the contact between mag- num and unciform is reflected in a flat and rather inconspicuous distal continuation of the unciform facet on the magnum (Fig. 25D). The trapezoid differs from that of M. elatus in several points, the most important being the ab- sence of a trapezium facet. The scaphoid facet of CM 1424 (Fig. 26G), like that of M. elatus, curves onto the radiovolar surface of the trapezoid, but 54 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Fig. 26. — A) Radiovolar view of left scaphoid, trapezoid, and Me II, and B) ulnar view (dorsal edge at left) of left scaphoid, trapezoid, and Me 11 of M. hollandi . FMNH P13000, from near Jay Em, Goshen Co., Wyoming. C) Radiovolar view of left scaphoid of M. hoUundi . F:AM 54.902a, and D) proximal view (volar edge below) of right lunate of M. hollandi. F:AM 54,902h, both from 7 mi south of Chugwater, Wyoming. E) Radiovolar, and F) ulnar views of left scaphoid, and G) radiovolar, and H) ulnar views of left trapezoid of holotype of M. hollandi. CM 1424, from near the Wyoming state line along the Niobrara River, Sioux Co., Nebraska. Facet abbreviations as in Fig. 10. instead of merging with a facet for a trapezium, it passes distally for almost the height of the trape- zoid. It does not, however, quite adjoin the Me II facet at the distal end of the trapezoid but is shaped like a tongue with its volar and distal edges espe- cially well defined. When the scaphoid is moved along the trapezoid, its concave distal surface fol- lows the contour of this facet all the way to the distal edge of the trapezoid. This range of move- ment suggests that the entire radiovolar facet was indeed for the scaphoid. The scaphoid facet in EMNH P13000 (Eig. 26A), although resembling that in M. elatus in not being clearly abbreviated at its distal edge, was probably also only for the scaph- oid. Scaphoid and Me II in this specimen show no trace of trapezium facets, and the scaphoid during extreme flexion of the carpus reaches close to the trapezoid’s distal edge. It is probable, considering the absence of trapezium facets on scaphoid, trap- ezoid, and Me II, that no trapezium was present in M. hollandi . This possibility was mentioned by Hol- land and Peterson (1914:233), but their evidence for it was not so strong, for they thought that a trape- zium facet was present on the trapezoid. Trape- zoids of M. hollandi also are more abbreviated at their volar edges than are those of M. elatus . 1978 COOMBS— REEVALUATION OF MOROPUS 55 Me II (two specimens) shows basic similarity in articular facets to Me II of M. elatus , including the following: I) the concave nature of the trapezoid facet, which is set off by well-defined ridges from all other facets (Fig. 25D); 2) the lack of strong proximal orientation of the volar part of the mag- num facet and of any sharp ridge separating it from the trapezoid facet (Fig. 25D); 3) Me III facet over- hung for almost all of its extent by the magnum facet (Fig. 26B). These features are not shared by certain other North American schizotheriine spe- cies. The radiovolar angle of the proximal end of Me II differs from all known Moropus species in the absence of a trapezium facet and the weakly developed radiovolar process (Fig. 25D). There is no difference from M. elatus in the shaft or distal end. The proximal end of Me III is narrower in relation to its depth than in other Moropus species, and this narrowness is reflected in the shape of the proximal facets. The proximal part of the Me II facet is flatter than that in M. elatus and slants more dorsally (Fig. 25A); as in M. elatus it is continuous with the dorsal part of the Me II facet. Magnum and unciform fac- ets (Fig. 25D) are both narrower than, but otherwise similar to, the same facets in M. elatus . On the ulnar side of Me III, weakly overhung by the un- ciform facet, are two Me IV facets in the holotype, the dorsal facet larger and separated from the volar one by a depression of greater width than any seen on specimens of M. elatus. However, the joining of these two facets in FMNH PI 3000 (Fig. 25B) sug- gests that variability of the Me IV facet(s) persisted in M. Iwllandi . Shaft and distal end of Me IV are similar to those in M. elatus . Me IV is quite similar to its counterpart in M. elatus. There are two separate articular facets for Me III (Fig. 25C); the dorsal facet differs in the holotype from that in M. elatus in being not much larger than the volar facet, in being concave, and in having an orientation far more radial than prox- imal. These unusual features are not shared, how- ever, by FMNH PI 3000. An important feature of the Lilnovolar surface of Me IV is the weaker de- velopment than in M. elatus of ulnar and volar tu- bercles (Fig. 22H). There is, however, the same depression between the tubercles that occurs in M. elatus, and an Me V facet is clearly developed and separated by a space from the proximal edge of the bone. Only the proximal end of Me V is preserved in CM 1424. The preserved part has a single cup- shaped facet for Me IV on its dorsoradial side. Dis- tally the bone does not seem to diverge as far from Me IV as does Me V of M. elatus. Hindlimb A left femur and two damaged tibiae are pre- served in the holotype, but these cannot be differ- entiated from specimens of M. elatus. The holotype also preserves both astragali, and F:AM 54,902g (Fig. 23C) represents an additional specimen. In each, there is a distinct, short distal neck, similar to that in M. elatus. In dorsal aspect the three bones are similar to one another and to M. elatus in 1) the greater transverse width of the fib- ular side of the trochlea and its more gradual slope compared with the tibial side, 2) the somewhat greater proximal extension of the tibial side, and 3) the distal extension of the free part of the fibular side to the level of the distal articular facet. In the holotype the relief on the volar surface of the as- tragalus is considerably less than in F:AM 54,902g and in M. elatus. Its ectal facet is quite shallowly and uniformly concave, and the strong ridge sepa- rating the ectal facet proximally from the trochlea is straight (as in M. elatus) rather than curved. There is barely any eminence between the confluent ectal and calcaneal facets, and the sustentacular facet, weakly convex, is abruptly truncated at its almost straight proximofibular edge. Between the sustentacular and ectal facets is a broad but rela- tively shallow depression. In F:AM 54,902g the vo- lar facets, while resembling in shape those of the holotype, have a very strong relief. The large ectal facet is especially deep in its distal part and is di- vided from the also slanted, contiguous calcaneal facet by a very sharp ridge. A deep but narrow groove separates the ectal from the sustentacular facet. A difference from both the holotype and M. elatus is the curved rather than straight proximal crest between the ectal facet and trochlea. All three astragali have a navicular facet on the distal surface very similar to that in M. elatus. In astragalus pro- portions there is a distinct dichotomy between M. elatus and M. Iwllandi on the one hand and all known later North American specimens (and also later Eurasian schizotheriines) on the other. M. elatus and M. Iwllandi are similar in that the pro- portion of tibial height/transverse width in all spec- imens (Table 4) falls between 0.72 and 0.81; in con- trast. the later North American specimens have a lower value (range 0.61-0.66 for a total of four spec- imens). The decrease in the ratio in the later spec- 56 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 Fig. 27. — Dorsal views of A) right Mt 111 of M. elatus. AMNH 14.378, from the Agate Spring Quarries. Sioux Co., Nebraska. B) left Mt 111 of M. Iiollandi, FMNH P13000, from near Jay Em, Goshen Co., Wyoming, and C) left Mt III of Moropiis cf. M. hollandi , F;AM 54.903. from near Jay Em. Goshen Co., Wyoming, Proximal (D. dorsal edge below), tibial (E). and volar (F) views of Mt III of M. hollandi, FMNH P13000. Facet abbreviations as in Fig. 16; X = broken surface; unlabeled line to Mt IV facet. imens is related not only to genuine broadening of the bones relative to height but also to reductions in the neck (particularly in Moropiis merriami) and a decrease in height of the tibial side of the trochlea relative to the fibular side. On the two broken calcanea of the holotype the tuber calcis is relatively shorter than in M. elatus, and there is only a slight volar prominence at the end of the tuber. As might be expected from holo- type astragali (above), the ectal facet is not so strongly convex as in M. elatus and apparently does not extend so far onto the tuber. No facet for the tibia can be discerned in this individual. The sus- tentacular process is broken off, but the well-de- veloped “lesser process” remains, separated by a sharp ridge from the cuboid facet on the distal sur- face. This latter facet is pear-shaped as in M. elatus but almost completely flat. Two known cuboids resemble specimens of M. elatus and differ from M. distans and M. merriami 1978 COOMBS— REEVALUATION OE MOROPVS 57 Fig. 28. — Tibial (A) and fibular (B) views of Mt III and proximal view (C) of Mt III and Mt IV (dorsal edge below) of F:AM 54,903, Moropus cf. M. hollandi . Dorsal (D) and ulnar (E) views of left duplex, and ungual phalanx (F) of digit II of the manus of FMNFI P13000, a presumed female specimen of M. hollandi. All specimens from near Jay Em, Goshen Co,, Wyoming, Facet abbreviations as in Fig. 16. in the failure of the navicular facet to reach the proximal edge of the bone or to adjoin the calca- neum facet (Eig. 16A). Navicular and cuboid are thus separated at their proximal and volar edges by a small space. In this specimen dorsal and volar parts of the ectocuneiform facet are continuous and show no constriction between them, whereas in M. elatus the two parts of the facet are sometimes sep- arate. The dorsal part of the ectocuneiform facet extends distally to adjoin the Mt IV facet, but the volar part of the facet is well separated from the distal end of the cuboid. In this individual the tib- iovolar process is not remarkably developed (Eig. 16A). A navicular of EMNH P13000 (Eig. 16G) falls easily within the range of morphology of M. elatus and cannot be differentiated in any special way. No metatarsals are known from the type of M. hollandi, but EMNH P13000 preserves a Mt III 58 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 4 (Figs. 27B, D-F), which can be compared with Mt III and Mt IV of F:AM 54,903 (Moropiis cf. M. holUifuli , see Figs. 27C, 28A-C). Table 6 gives the length versus width proportions of metatarsals of M. hollandi and M. elatus. FMNH P13000 (Mt III) falls within the range of M. elatus in length/maxi- mum distal width but has a lower length/minimum shaft width than most specimens of M. elatus. Ap- parently, broadening of the metatarsal shaft relative to length has proceeded in M. hollandi despite con- tinued resemblance to M. elatus in many other re- spects. Mt III and Mt IV of F:AM 54,903 fall below the range of M. elatus in both proportions. Mt III of FMNH P13000 and F:AM 54,903 complement one another in preservation in that the proximal end of the former is broken at the dorsofibular angle, whereas that of the latter lacks the facets for Mt II and most of the volar part of the facet for Mt IV. The SLibtriangular, tibially slanted proximal facet, for the ectocLineiform (Fig. 27D), resembles its counterpart in M. elatus, as does the small bipartite Mt II facet (Fig. 27E), where it is preserved in FMNH PI 3000. In this specimen dorsal and volar parts of the Mt II facet are connected by a thin proximal articular platform. Facets for Mt IV differ between the two Mt 111 specimens. In FMNH PI 3000 the shape of the volar part of the facet sug- gests that it was discrete and well separated from the dorsal Mt IV facet (Fig. 27F). In F:AM 54,903 the two parts of the Mt IV facet are not only adja- cent but are confluent at their most proximal edges (Fig. 28B). Further, the two parts of the facet in this latter specimen are more closely equal in size than in M. elatus: the situation in FMNH P13000 is not clear. Both specimens show more dorsal to volar compression of the shaft, especially toward the distal end, than occurs inM. elatus . This change seems to be concomitant with the shortening and broadening of the shaft. Mt IV of F:AM 54,903 is slenderer and distinctly longer than Mt 111 of the same specimen, propor- tionately more so than in M. elatus. The greatest difference between this bone and specimens of M. elatus is the presence of one continuous facet rather than two facets on the tibia! side of the proximal end. corresponding with facets on Mt III above. The two parts of the facet are of almost equal size (the dorsal part is slightly larger), and the connec- tion between them begins at the crest dividing them from the cuboid facet and continues distally over half of the facet. In addition to this partial dorsal- volar division, there is a functional division into proximal and distal parts, separated only by a bare- ly visible ridge. The more proximal part of the facet, with a proximal to tibial orientation, articulates with the ectocuneiform (Fig. 28C). while the larger, more distal parts of the facet articulate with Mt HI. There is strong development of the fibulovolar angle of the proximal end of this Mt IV. although not so strong as in M. elatus, and there is a weak longitudinal ridge on the fibular side of the shaft. The distal end of this specimen is clearly asymmetrical. Phalanges A single large duplex of digit II of the manus be- longing to the holotype and two similar duplexes of FMNH P13000 (Figs. 28D, E) suggest by their size and morphology that both of these specimens are females, a conclusion consistent with the small size of these individuals compared to known males of M. elatus . The specimens are quite similar to those of female M. elatus in having a less flattened, more symmetrical Me II facet than in M. elatus males. These duplexes are compressed transversely but thickened in the dorsal to volar direction, more so than is usual in M. elatus. Additional proximal phalanges are available but add little to characterization of M. hollandi. Most of these appear to belong to digits III or IV of the manus, but one specimen, belonging to F:AM 54.902f. may belong to digit III or IV of the pes despite its unexpectedly large size. This specimen has a pronounced thickening of its volar part, re- markable because in this and other features it is extremely similar to F:AM 54,91 1 from Flint Hill Quarry in the Batesland Formation of South Da- kota. The presence of M. hollandi in the early Hem- ingfordian fauna of the Batesland Formation is un- clear, however, for other Moropus material from Flint Hill Quarry collected by the University of Cal- ifornia shows no special resemblance to known spec- imens of M. hollandi. M. hollandi is not known from the roughly contemporaneous fauna of the Runningwater Formation of Nebraska. Small medial and ungual phalanges of M. hollandi show no features of particular taxonomic interest. A large ungual phalanx belonging to FMNH PI 3000 (Fig. 28F) closely resembles the large ungual pha- lanx of digit 11 of the manus in M. elatus but is slenderer. Its dorsal and subungual processes are large, but the facet for the duplex is not deeply ex- cavated and has only a weak keel dividing its two halves. 1978 COOMBS— REEVALUATION OF MOROPUS 59 CONCLUSIONS AND ZOOGEOGRAPHY The genus Moropus is derived relative to Schizo- therium in its higher crowned and more elongated molar teeth, in the absence of a hypoconulid on M;,, in the proportionate shortening of metatarsals, and in fusion of proximal and medial phalanges of digit II of the manus to form a duplex bone. Its closest relatives among the Schizotheriinae are Phyllotil- lon, Ancylotheriunu and a yet unpublished genus of North American Miocene schizotheriine. These genera share the above characters with Moropus, but they are additionally derived in a number of others, for example, the presence in Ancylotlierium and sometimes in Phyllotillon of a crochet and pos- terior labial rib on upper molars. The four Moropus species discussed in this paper are all relatively primitive representatives of Mor- opus , little removed from the Eurasian/North American common stock of the genus (see Coombs, 1974). M. distans , the type species, is poorly known but includes Moropus individuals of relatively small size. Because the few available remains of M. dis- tans include a duplex of digit 11 of the pes. it is possible that phalangeal fusion to form a pes duplex was more common in M. distans than in M. elatus. M. distans also differs from M. elatus but resembles Moropus merriami in the more proximal extent of its navicular facet on the cuboid. Moropus orego- nensis is known only from upper premolars and molars, which are smaller than comparable teeth of M. elatus and differ in the manner of wear on upper premolars. M. distans may be synonymous with M. oregonensis , but no elements clearly linking the two species have yet been found. Moropus senex, known from a single specimen, shows no useful fea- tures for differentiating species and is therefore considered a nonien dubiuni . Moropus elatus is the best known Moropus spe- cies and is the most useful basis for intrageneric and intergeneric comparisons. Male M. elatus are at the extreme high end of Moropus size range, and the molar teeth of M. elatus are slightly more elon- gated and high crowned than those of M. orego- nensis . Yet M. elatus has otherwise diverged little from its common ancestry with M. distans and M. oregonensis. Moropus hollandi closely resembles M. elatus but differs in the loss or strong reduction of the trapezium in the manus and the dorsal to volar compression of certain other carpal elements. Loss of the trapezium occurs independently in Schizotherium turgaicuin, Borissiakia betpakdalen- sis and Ancylotheriiun {Ancylotlierium) pentelicum but within Moropus is characteristic only of M. hol- landi, where it presumably allowed increased flexion of the carpus. M. hollandi also has proportionately shortened its metatarsals relative to most of M. elatus, consistent with a general trend toward such shortening within the Schizotheriinae over time. On the basis of the above given similarities and differ- ences. Moropus elatus and M. hollandi seem to be closely related, whereas M. distans and M. orego- nensis may be closer to M. merriami. Moropus mer- riami and its kin will be more thoroughly discussed in a separate paper, as will Moropus matthewi . It is difficult to discuss the zoogeography of early Moropus species on present scanty evidence. M. distans and M. oregonensis are known only from the late Arikareean of Oregon, M. elatus from the late Arikareean/early Hemingfordian of Nebraska, and M. hollandi from the early Hemingfordian of Nebraska and Wyoming. Contemporary materials from other areas are few and for the most part too fragmentary for taxonomic treatment. Scattered re- mains from Oregon to Florida do suggest a broad geographic range and probably a more complex ear- ly Miocene evolutionary story than can currently be reconstructed. Moropus is much more common in the fluvial channel fills in the base of the Upper Harrison Formation than in any other rock unit (Hunt, personal communication). This high rate of occurrence of Moropus elatus in particular could be a result of its relatively greater abundance or could be related to dietary, herding, or other habits that made them more liable to preservation. On the other hand, the conditions of deposition of the Up- per Harrison Formation could have been unusually conducive to chalicothere entombment. The earliest documented occurrences of Moropus in North America are late Arikareean. At this time or somewhat earlier the genus apparently immi- grated from Eurasia across the Bering land bridge or migrated southward as part of a hitherto unsam- pled northern Holarctic fauna. An Eurasian origin is supported by the absence of chalicotheres in North American Oligocene deposits and their presence in the Oligocene of Eurasia (Oreinotherium bilobatum from the Oligocene of Saskatchewan is a bronto- there, not a chalicothere. as Skinner, 1968. also as- serted), the presence of small Moropus sp. very similar to M. elatus in the European Aquitanian (Coombs, 1974), and the late Arikareean or early 60 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Hemingfordian appearance in North America of other immigrants from Eurasia (see, for example, the amphicyonids Cynelos and Ysengrinia in Hunt, 1972, and the primitive deer Blastomeryx cited along with other genera by Wilson, 1967, 1968). Floral geography may also be correlated with the early Miocene appearance of Moropiis in North America. Wolfe (19696:85; 1972:230) suggested that the probably early Miocene Collawash flora from Oregon indicated that floristically the warm tem- perate vegetation at middle latitudes represented a Mixed Mesophytic forest for the first time. At that time the temperate, moist forests of both Oregon and Alaska contained a number of common species and are considered part of the same floristic prov- ince (Wolfe, 1969( Southeastern Geological Society. 13th Field Trip (H. K. Brooks and J. R. Underwood, Jr., eds.). Patton, T. H., and Beryl E. Taylor. 1971. The Synthetoceratinae (Mammalia, Tylopoda, Proto- ceratidae). Bull. Amer. Mus. Nat. Hist.. 145:119-218. Patton, T. H., and S. D. Webb. 1970. Fossil vertebrate deposits in Florida. The Plaster Jack- et. 14:1-18. Peterson, O. A. I907«. The Miocene beds of western Nebraska and their ver- tebrate faunae. Ann. Carnegie Mus.. 4:21-72. 19076. Preliminary notes on some American chalicotheres. Amer. Nat., 41:733-752. 1909. A revision of the Entelodontidae. Mem. Carnegie Mus., 4:41-156. 1913. A new species of Moropus (M . hollandi) from the base of the middle Miocene of western Nebraska. Science, n.s., 38:673. Pilgrim, G. E. 1910. Notices of new mammalian genera and species from the Tertiaries of India. Rec. Geol. Surv. India, 40:63- 71. Radinsky, L. B. 1963. The perissodactyl hallux. Amer. Mus. Novitates, 2145:1-8. Rensberger. J. M. 1971 . Entoptychine pocket gophers ( Mammalia, Geomy- oidea) of the early Miocene John Day Formation, Or- egon. Univ. California Publ. Geol. Sci.. 90:1-163. 1973. Pleurolicine rodents (Geomyoidea) of the John Day Formation, Oregon, and their relationships to taxa from the early and middle Miocene, South Dakota. Univ. California Publ, Geol. Sci., 102:1-95. Repenning, C. A., and J. G. Vedder. 1961. Continental vertebrates and their stratigraphic corre- lation with marine mollusks. eastern Caliente Range. California. U.S. Geol. Surv., Prof. Pap. 424:C235- C239. 62 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 4 Rich, T. H. V., and T. H. Patton. 1975. First record of a fossil hedgehog from Florida (Erina- ceidae. Mamalia). .1. Mamm., 56:692-696. Schaefer. FI., and H. Zapfe. 1971. Chalicotherium grande Blainville und Chalicotherimn goldfussi Kaup. Odontologische und osteologische Unterschiede. Verhandl. Naturf. Ges. Basel, 81:157- 199. SCHAUB. S. 1943. Die Vorderextremitiit von Anvylotherimn penlelicnm Gaudry und Lartet. Schweizerischen Palaeont. Ab- hand!.. 64:1-36. Schultz. C. B. 1938. The Miocene of western Nebraska. Amer. .1. Sci., ser. 5. 35:441-444. Schultz. C. B., and C. H. Falkenbach. 1947. Merychyinae. a subfamily of oreodonts. Bull. Amer. Mus. Nat. Hist., 88:157-286. 1949. Promerycochoerinae. a new subfamily of oreodonts. Bull. Amer. Mus. Nat. Hist.. 93:69-198. Sinclair. W. ,1. 1901. The discovery of a new fossil tapir in Oregon. J. Geol., 9:702-707. Sisson. S., and J. G. Grossman. 1953. The anatomy of the domestic animals. W. B. Saunders Co., Philadelphia, 972 pp. (reprinted from 1938). Skinner. M. F. 1968. A Pliocene chalicothere from Nebraska and the distri- bution of chalicotheres in the late Tertiary of North America. Amer. Mtis. Novitates. 2346:1-24. Thenilis, E. 1953. Studien fiber fossile Vertebraten Griechenlands. Hi. Das Maxillargebiss von Ancylotlierium penlelicnm Gaudry und Lartet. Ann. Geol. Pays Helleniques. 5:97-106. Tung. Yung-sheng. Wan-po Huang, and Zhu-ding Qiu. 1975. [Hipparion fauna in An-lo, Hohsien. Shansi.] Vertebr. Palasiat., 13:34-47. VON Koenigswald. G. H. R. 1932. Metaxchizotheriinn fraasi n.g. n. sp., ein neuer chali- cotheriide aus dem Obermiocan von Steinheim A. Al- buch. Palaeontographica. Beitr. Naturgesch. Vorzeit, Suppl., 8: 1-24. Wilson. R. W. 1967. Fossil mammals in Tertiary correlations. Pp. 590-606, in Essays in Paleontology and Stratigraphy (C. Tei- chert and E. L. Yochelson. eds.). Dept. Geol., Univ. Kansas Spec. Publ. 2, Univ. Kansas Press, Lawrence. 1968. Insectivores. rodents, and intercontinental correlation of the Miocene. Internat. Geol. Congr. 23, 10:19-25. Wolfe, J. A. I969u. Neogene fioristic and vegetational history of the Pa- cific northwest. Madrono. 20:83-110. 19696. Paleogene floras from the Gulf of Alaska region. U. S. Geol. Surv.. Open-file Rept., Ill pp. 1972. An interpretation of Alaska Tertiary floras. Pp. 201- 233, in Floristics and Paleofloristics of Asia and east- ern North America (A. Graham, ed.), Elsevier Publ. Co., Amsterdam. Wolfe, J. A. and E. B. Leopold. 1967. Neogene and early Quarternary vegetation of north- western North America and northeastern Asia. Pp. 193-206, in The Bering Land Bridge (D. M. Hopkins, ed.), Stanford Univ. Press. WOODBURNE. M. O. 1969. Systematics, biogeography, and evolution of Cynorca and Dyseoliyns (Tayassuidae). Bull. Amer. Mus. Nat. Hist., 141:271-356. 1975. Cenozoic stratigraphy of the transverse ranges and ad- jacent areas, southern California. Geol. Soc. Amer. Spec. Papers. 162:1-91. Wu. Wenyli, and Guanfeng Chen. 1976. [A new schizotheriine genus from the Neogene of Ping- liang, Gansu.] Vertebr. Palasiat.. 14:194-197. y !■ f f- ijK' JL. 5 .is. ' & « i WEIGHTS OF 151 SPECIES OF PENNSYLVANIA BIRDS ANALYZED BY MONTH, AGE, AND SEX "'vV' MARY H.> CLENCH ROBERT C. LEBERMAN ■j. NUMBER 5 PITTSBURGH, 1978 /■t i .4' 2 9 t I BULLETIN of CARNEGIE MUSEUM OF NATURAL HISTORY WEIGHTS OF 151 SPECIES OF PENNSYLVANIA BIRDS ANALYZED BY MONTH, AGE, AND SEX MARY H. CLENCH Associate Curator, Section of Birds ROBERT C. LEBERMAN Naturalist , Powdennill Nature Reserve NUMBER 5 PITTSBURGH, 1978 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 5, pages 1-85 Issued 20 September 1978 Price $5.00 a copy Craig C. Black. Director Editorial Staff. — Hugh H. Genoways, Publications Editor; Duane A. Schlitter, Associate Publi- cations Editor; Stephen L. Williams, Associate Publications Editor; Teresa M. Bona, Technical As- sistant. 1978 by the Trustees of Carnegie Institute, all rights reserved. CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE PITTSBURGH, PENNSYLVANIA 15213 INTRODUCTION In 1961 Carnegie Museum of Natural History ini- tiated a bird-banding program at its field research station, Powdermill Nature Reserve. The Reserve is located in the Ligonier Valley, which at its lati- tude is the westernmost valley of the Allegheny Mountains. The banding station is approximately three miles south of Rector, Westmoreland County, Pennsylvania (40°10'N, 79°16'W). The Powdermill program has concentrated pri- marily on the study of small land birds, by using 30- and 36-mm mesh mist nets; larger birds are some- times captured in traps or in nets of larger mesh. During the banding procedure the following data are routinely recorded: band number, species, age and sex (when possible), date, time of day, wing chord measurement, fat index, and weight. Since the au- tumn of 1966, many of the passerines also have been examined for the degree of skull pneumatiza- tion, to determine age. As the Powdermill program has become better known to the ornithological community, we have received many requests for data from our files, par- ticularly for weights. The types of studies prompt- ing these requests have been remarkably diverse. For example, bird weights are used by theoretical ecologists concerned with biomass calculations; by laboratory physiologists needing comparative data from normal wild birds; and, of course, by orni- thologists and banders for taxonomic, zoogeograph- ic, and many other types of studies. Weight data also have some unexpected appli- cations. For instance, we have been asked on two different occasions to supply British ornithologists with weights of common North American species. In both cases vagrant birds of North American or- igin had appeared in England and it was suspected that they might have reached those shores by “as- sisted passage” rather than having flown (or been blown) there on their own. Both vagrants had been weighed when they were found in England, but nor- mal weights for migrating individuals of their spe- cies could not be located in the literature for com- parison. Another unusual request for a bird weight came from an experimental engineer. During a test flight of a new jet airplane, a small bird had collided with the plane's windshield. The glass broke, the plane went out of control and crashed, with the pi- lots ejecting safely. The company that had devel- oped the glass for the windshield asked us for the weight of the bird species involved. That figure was critical for their calculations on the force that had broken the glass, in order to strengthen the wind- shield in the next model. Although weights of many eastern North Amer- ican species can be found in the literature, most represent single birds or small samples, often with- out full supporting data (date, locality, age, sex). Seldom have they represented sufficiently large samples to be useful to researchers. In addition, some of the published weights have been incidental to the main study and hence are difficult to locate from the paper’s title. Weight data are also widely scattered through the literature. An incomplete manuscript bibliography of bird weights and their evolutionary interpretation compiled by G. A. Clark, Jr. in 1976 contains almost 600 titles gleaned from both ornithological and non-ornithological sources. Therefore, seeing a need for a compilation of adequate samples of the weights of North Amer- ican birds, we present here an analysis of much of the data from the Powdermill program accumulated through the spring of 1974: 97,762 weights from 151 species. All of the weights in this study are from living, banded birds. The birds are weighed, while being restrained in a sock or plastic cone, on a triple-beam balance. The scales are counterbalanced against the sock or cone and hence the weight of the bird alone is read directly, to the nearest 0.1 gram. The sex of the bird is recorded as M (male), E (female), or U (unknown). In species that are not sexually dimor- phic in plumage, the sex of some individuals has been determined during the breeding season by the presence of a brood patch or cloacal protuberance, changes in eye or mouth color, etc. The age designations are the same as those cur- rently in use by the Bird Banding Laboratory, U.S. Eish & Wildlife Service: HY = a bird in its hatch- ing year, up to 31 December in its first year of life; SY = second year, the calendar year after the bird was hatched; ASY = after second year; TY = third year; ATY = after third year; AHY = after hatching year, but precise age not known; and U = unknown, age not determined. Among the birds aged specifically SY through ATY, some have been thus recorded through plumage characteristics or skull pneumatization, but most are birds that were originally banded as HY and subsequently rehan- dled. Those recorded as age unknown are birds banded after the breeding season has begun and 3 4 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 before 31 December: primarily non-passerines that cannot be aged by skull pneumatization or plumage, passerines in which the skull may pneumatize early, and certain other autumn-banded passerines han- dled before l%6 when we first began to examine skulls routinely. A few birds were not recorded by age or sex inadvertently, or because they were in anomalous plumage. Nomenclature follows The A.O.U. Check-list of North American Birds (American Ornithologists’ Union, 1957) and Supplements (Auk 90: 411^19, 1973 and Auk 93: 875-879, 1976). Included in the list of species is one composite, “Traill’s Flycatch- er.’’ This name includes the Alder and Willow fly- catchers, which cannot safely be distinguished in the hand. Also in the list are several hybrid types such as “Brewster’s Warbler’’ and “Lawrence’s Warbler.’’ The sample sizes range from one (eight species) to 7,715 (Dark-eyed Junco). We have included the very small samples because even though they rep- resent only one or two birds, they may well be the only, or among the very few, published weights for those species (e.g.. Least Bittern, Boreal Chicka- dee). I’he weights are analyzed by age/sex class by month. In the smaller samples the data are arranged linearly as follows: Sample size in parentheses (N); age/sex class; month; weight in grams (means of two or more weights italicized). The larger samples are arranged in table group- ings: (N) sample size Mean weight to 0. 1 grams, italicized [I standard deviation] in samples of 5 or more Minimum weight in sample weight range within sample Maximum weight in sample Note that the 97,762 weights here recorded are not from 97,762 different birds. Most of the birds are, of course, weighed only once — at the time of banding. But the study also includes many thou- sands of birds that have been weighed several times: migrants that are recaptured within a few hours or days of banding, and resident birds that are repeatedly captured throughout their lifetime. Each recapture weight has been included in its ap- propriate category. For example, if a Black-capped Chickadee were banded and weighed as an HY-U in September, then were recaptured the following June and could be sexed as a male from its cloaca! pro- tuberance, its recapture weight would be entered in the SY-M category for June, and the original weight in the HY-M September category. Known age birds were carried through their third year (TY) in detail, and then grouped as ATY thereafter. The reader will notice that the standard deviation values for many samples are large. These reflect several factors: weights can vary greatly during the breeding season (a female with a mature egg will weigh well above the mean), and the contents of the stomach and gut have a strong effect on the overall live weight (a thrush that has just filled its stomach with wild cherries, with pits, will weigh considerably more than a newly arrived migrant of the same species that has not yet fed). Whether or not a bird has recently defecated also has some ef- fect, but most birds do so during the handling pro- cess and before they are weighed. The most impor- tant variable that affects weight is the quantity of fat deposits within the body and under the skin. To illustrate the effect of fat, we cite the extreme case of a recently handled migrant Cape May Warbler: the bird was banded as an HY-F in September with no visible fat (index of 0) and weighing 9.2 grams (10.0 is the mean for that category). It was recap- tured several weeks later, with maximum fat (index 3) and weighing 15.8 grams — a 72% gain over the original weight. Resident birds, in contrast, show little weight and fat index variation. A Tufted Tit- mouse that has been weighed 16 times over a ten- year period has only varied between 19.3 and 21.4 grams, always with a fat index of 0 or 1. Although we are well aware that the weight of a bird is biologically more meaningful when related to the fat index of that bird, we did not attempt to include fat indices in this analysis. With a possible 21 different age/sex classes, analyzed by the 12 months of the year, some species could yield a table with 252 groups. Breaking the age/sex classes fur- ther into the four fat index classifications (0-3) would be too unwieldy for a study of this sort. De- tailed analysis of weight variation is also more meaningful when correlated with each individual’s body size, but similarly, such a fragmentation would be inappropriate for this study. The Powder- mill banding records are, however, being entered into an electronic data processing system and anal- yses of weight by fat index and/or body size class could be made in the future. We are also, of course, continually adding data to our files at the rate of over 10,000 per year. Requests for special analyses may be possible to fill: address the Section of Birds, Carnegie Museum of Natural History, Pittsburgh, Pennsylvania 15213. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 5 ACKNOWLEDGMENTS A study of this magnitude has had many contrib- utors. Eirst and foremost we thank A. C. Lloyd for his many years of devoted service as a volunteer in the Powdermill banding program. We also thank the dozens of people, too many to name individually, who have helped in a variety of ways with the band- ing since 1961 , and the several clerks who have kept our banding records. This analysis was largely ac- complished by two research assistants; Dr. Gail Schiffer, who began the data compilation and sta- tistical work, and Mr. G. Thomas Bancroft, who saw the task to completion and ran most of the anal- yses on the University of Pittsburgh computer. Marilyn Niedermeier volunteered to copy the re- sults into a uniform format for typing. She was as- sisted in proof reading by Thomas E. Herman and Miriam A. Stern. Cynthia Kraus typed the manu- script. The assistance of Dr. Schiffer and Mr. Ban- croft was supported by the Powdermill Research Eund through the generosity of Mrs. Cordelia S. May. 6 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Pied-bi!led Grebe PodUymhus podiceps (I) HY-U, Sept., 429.4; (1) AHY-U, Sept., 341.1; (1) U-U, Oct., 229.0. Green Heron Butorides striatus (1) HY-U, July, 186.8; (1) HY-U, Sept., 171.2; (1) HY-U, Oct., 210.6; (1) AHY-U, June, 236.7; (1) SY-U, June, 215.2. Least Bittern Ixobiychiis exilis n - 1 (1) AHY-U. May, 115.5. Goshawk Accipiter gentilis n = 1 (1) HY-U, Nov., 444.5. Sharp-shinned Hawk Accipiter striatus n - 12 (1) HY-F, Oct., 159.2; (1) HY-M, Sept., 91.0; (1) HY-U, Sept., 93.0; (1) HY-U, Nov., 115.8; (1) AHY-M, Apr., 101.5; (1) SY-F, Apr., 184.1; (2) SY-M. Apr., 92.2, 97.3 {94.8)\ (1) SY-U, Apr., 102.5; (1) ASY-M, Mar., 88.5; (1) ASY-U, Apr., 100.9; (1) ASY-U, Nov., 101.8. Broad-winged Hawk Buteo platypterus n = 4 (3) AHY-U. Apr., 311.2, 452.2, 472.2 (411. 9)\ (1) AHY-U, May, 360.7. Ruffed Grouse Boiuisa umbellus n = 3 (1) HY-U (poult), July, 171,7; (1) AHY-F, July, 516.3; (1) AHY-M, Oct., 506.8. Bobwhite Colinus virginianus n = 1 (1) AHY-M, May, 162.7. Virginia Rail Rallus limicola n — 2 (1) AHY-U, Apr., 79.5; (1) AHY-U, May, 90.2. Sora Porzana Carolina n = 3 (2) HY-U, Sept., 77.8, 84.2 (81 .0)\ (1) AHY-U, May, 85.0. American Woodcock Philohela minor n = 34 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (4) 160.6 (2) 183.0 139.3 179.4 179.4 186.6 HY-M (1) 141.0 (2) 142.4 (2) 155.4 (1) 159.0 127.2 135.4 157.6 175.5 HY-U (1) 142.2 (1) 191.5 AHY-F (!) 200.0 (1) 200.5 (2) 191.7 184.7 (1) 187.6 198.7 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 7 Age/Sex AHY-M AHY-U U-U Jan. Feb. Mar. Apr. May (1) (2) (4) 134.3 123.2 151.2 116.1 135.1 130.4 183.1 (2) (1) 130.6 III.8 127.3 133.9 June July Aug. Sept. Oct. (1) 152.4 Nov. (2) 153.9 144.7 163.1 (2) 169.4 134.4 204.4 Common Snipe (I) HY-U, Oct., 12! Spotted Sandpiper Age/Sex Jan. AHY-U U-U Solitary Sandpiper Pectoral Sandpiper (1) HY-U, Sept., 44.8. Mourning Dove Capella galUnago Feb. Mar. .0; (1) AHY-U, Apr., 118.6; (1) 1 U-U, Sept ,, 110.8; (1 ) U-U Ac tit is macularia Apr. May June July Aug. Sept. (6) (32) (2) 40.6 39.9 39.9 [3.53] [4.95] 33.8 35.5 32.8 46.0 44.6 53.1 (1) (2) (3) 39.6 37.0 54.6 33.7 46.1 40.2 59.8 Tringa ' solitaria -U, May , 46.2, 52.6, 53.4 (50.7); (1) U-U, Sept., 60.7 Oct. n Nov. Calidris melanotos Zenaichi macroura Age/Sex Jan. Feb. Mar. Apr. May June July AHY-F (2) (1) (3) 140.3 130.8 128.1 138.1 119.7 142.5 132.5 AHY-M (1) (2) (1) (3) (1) 126.5 138.0 135.0 136.7 134.0 135.2 128.7 140.9 143.3 AHY-U (1) (1) 122.1 160.0 Yellow-billed Cuckoo Coccyzus americanus Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. n Nov. HY-U Aug. Sept. Oct. (2) (2) (1) 42.4 50.8 74.3 41.3 50.2 43.5 51.4 n Nov. Dec. = 5 = 46 Dec. = 6 = I = 16 Dec. 90 Dec. 8 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex AHY-F Jan. Feb. Mar. Apr. May June (1) 55.7 July AHY-M (1) 53.6 AHY-U (40) 64.5 [8.62] 52.0 81.5 (27) 70.8 [8.04] 51.0 84.6 Aug. Sept. Oct. Nov. Dec. (5) (3) 60.1 55.1 [6.79] 50.0 50.5 61.2 68.0 U-U (2) (6) 55.3 56.6 54.2 [5.78] 56.4 48.7 63.6 Black-billed Cuckoo Coccyziis erythropthalmus n = 71 Age/Sex HY-U AHY-F AHY-M AHY-U U-U Jan. Feb. Mar. Apr. May June (1) 64.4 (1) 53.4 (35) (9) 51.8 53.4 [6.15] [7.79] 40.9 44.2 63.2 64.9 July Aug. Sept. (5) (1) (3) 39.1 44.4 46.3 [2.25] 44.3 35.1 40.5 (1) 52.8 49.8 (1) (1) 49.9 43.4 (4) (1) 45.7 47.8. 41.5 55.5 Oct. Nov. Dec. (2) 51.6 46.4 56.8 (6) 52.6 [7.60] 41.8 61.2 Screech Owl Otus asio rt 9 (1) HY-U, Sept., 140.1; (2) AHY-U, Mar., 154.3, 170.3 (/62.3); (1) AHY-U, Apr., 160.8; (1) AHY-U. May. 165.9; (3) U-U, Sept.. 153.3, 157.9, 166.6 (I59.3)\ (I) U-U, Nov., 176.0. Saw-whet Owl AegoUus acadiciis n - 7 (2) HY-U, Nov., 74.5, 88.4 (8I.4)\ (1) AHY-U, Mar., 72.6; (1) AHY-U, Apr., 90.8; (I) AHY-U, Oct., 81.7; (1) U-U, Oct., 97.6; (1) U-U, Nov., 87.3. Whip-poor-will Caprimulgus vociferns (I) HY-F, Sept., 51.4; (1) AHY-M, May, 56.5; (1) AHY-M, Aug., 54.6; (1) U-U, Sept., 55.4. Chimney Swift Cliaetura pelagica n = 2 (2) U-U, Sept., 19.5, 21.3 (20.4). 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 9 Ruby-throated Hummingbird Archilochus coinhris n= I3i2 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (6) 3.1 [0.30] 2.8 3.6 (120) 2.2 [0.28] 2.6 4.0 (130) 3.4 [0.37] 2.6 4.4 HY-M (8) 3.1 [0.20] 2.8 3.4 (147) 3.0 [0.28] 2.4 4.0 (161) 3.2 [0.40] 2.4 4.7 HY-U (1) 2.2 (22) 3.2 [0.43] 2.6 4.2 (13) 3.2 [0.22] 2.9 3.6 AHY-F (167) 3.3 [0.28] 2.7 4.7 (19) 3.3 [0.26] 2.9 3.8 (32) 2.2 [0.40] 2.7 4.8 (112) 2.2 [0.36] 2.7 4.5 (89) 2.5 [0.38] 2.7 4.8 AHY-M (3) J.l 3.0 3.2 (135) 3.0 [0.34] 2.4 4.1 (11) 2.7 [0.24] 2.5 2.8 (5) 2.8 [0.00] 2.7 2.8 (43) 3.2 [0.28] 2.5 3.9 (5) 3.2 [0.14] 3.0 3.4 U-F (46) 2.2 [0.30] 2.8 4.0 (24) 3.5 [0.40] 2.8 4.4 u-u (7) 3.0 [0.14] 2.8 3.3 (6) 3.5 [0.53] 3.0 4.3 Belted Kingfisher Megaceryle alcyon n = 26 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (1) 133.5 AHY-F (3) 149.7 137.0 156.3 (1) 135.9 AHY-M (1) 137.4 (1) 136.5 U-F (1) 140.9 (4) 134.9 124.6 140.8 (6) 156.0 [16.37] 139.5 186.5 10 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex U-M Jan. Feb. Mar. Apr. May June July Common Flicker (Yellow-shafted Flicker) Age/Sex HY-F HY-M HY-U AHY-F AHY-M Jan. Feb. Mar. Apr. (I) 148.3 U-F U-M Pileated Woodpecker Colaptes auratus May June July (2) (30) (7) (1) (1) 127.4 127.0 129.3 110. 1 126.3 117.4 [9.72] [6.19] 137.5 106.2 119.0 154.6 135.6 (3) (38) (13) (1) (2) 129.1 1 36. 1 132.2 137.0 134.1 113.9 [10.12] [6.20] 125.1 149.2 117.7 120.7 143.1 160.2 140.8 Dryocopiis pileatiis Red-bellied Woodpecker (1) SY-M, Jan., 80.3; (1) ASY-F, May, 74.5. Aug. Sept. Oct. Nov. (1) (5) (2) 136.8 135.5 171.7 [11.56] 170.8 126.5 172.6 155.0 n Aug. Sept. Oct. Nov (1) (1) (1) 138.5 113.3 141.7 (2) (1) 129.8 141.4 125.2 134.3 (2) 127.8 125.2 1.30.5 (12) (12) 129.8 135.1 [10.74] [12.98] 111.5 110.1 143.3 163.7 (8) (15) (11) (2) 127.3 136.3 140.3 134.4 [6.45] [9.57] [9.75] 127.0 120.6 122.2 124.1 141.8 138.1 159.0 152.2 (1) AHY-F, Apr., 269.6; (2) AHY-F, May, 250.3, 275.1 (262.7); (1) U-U, Nov., 248.1. Melanerpes carol inns Yellow-bellied Sapsucker Age/Sex Jan. Feb. HY-F Mar. Apr. Sphyrapicns variiis May June July Aug. Sept. (4) 51.8 45.2 59.3 Oct. (6) 50.3 [1.98] 47.3 53.0 n Nov. Dec. 167 Dec. = 4 = 60 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-M (4) 49.7 45.6 52.4 (5) 52.6 [2.91] 48.4 55.8 HY-U (2) 48.0 45.5 50.4 (3) 48.0 44.0 50.6 (1) 51. 1 AHY-F (4) 51.4 46.5 57.1 (1) 47.7 (1) 52.0 (1) 46.2 AHY-M (14) 50.5 [5.75] 40.7 62.2 (2) 51.8 51.2 52.3 AHY-U (1) 5L6 SY-F (1) 44.2 SY-M (3) 54.6 48.5 59.2 ASY-F (4) 49.5 46.2 54.7 ASY-M (3) 45.3 45.1 45.6 Hairy Woodpecker Picoides villosiis n = 51 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) 56.2 (1) 65.1 HY-M (2) 65.2 62.2 68.1 (3) 73.5 73.1 74.0 (3) 69.1 67.6 70.4 HY-U (I) 72.0 (I) 67.0 AHY-F (1) 65.9 (2) 61.3 60.6 62.0 (3) 60.6 59.3 62.5 (3) 63.5 61.9 65.6 (1) 63.2 AHY-M (1) 74.6 (1) 69.9 (7) 67.5 [5.88] 60.8 75.5 (5) 73.6 [3.91] 69.0 79.6 (1) 64.1 (1) 70.2 12 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. U-F (1) 63.2 U-M (3) (6) (2) 71.2 69.0 71.5 65.0 [2.10] 71.2 74.7 66.6 71.8 72.8 U-U (1) 69.1 Downy Woodpecker Picoides piihesc, ens n = 560 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) (5) (7) (16) (15) (7) 24.1 24.5 26.4 26.7 27.2 27.6 [0.99] [2.09] [1.14] [1.16] [1.14] 23.6 23.6 24.0 24.2 26.5 25.7 29.3 28.3 28.8 29.3 HY-M (6) (11) (21) (20) (14) 25.7 26.4 27.4 27.6 26.9 [1.28] [0.71] [1.37] [1.37] [2.45] 24.3 25.2 25.0 24.9 23.4 27.5 27.7 31.3 29.6 32.1 HY-U (3) (18) (22) (6) (3) 25.1 25.0 26.0 27.0 27.2 23.1 [1.36] [1.18] [1.74] 26.3 26.2 22.8 23.7 24.6 27.6 11.0 28.6 29.3 AHY-F (6) (2) (18) (38) (15) (2) (2) (4) (4) (11) (12) (3) 27.1 28.2 28.0 27.2 27.3 25.8 25.4 26.2 26.8 27.3 26.8 28.7 [1.871 27.1 [1.76] [1.51] [1.54] 25.4 24.6 25.0 25.0 [1.81] [1.61] 26.4 24.1 29 2 25.5 24.1 25.2 26.1 26.2 26.7 30.1 25.4 24.8 30.7 29.2 30.9 31.6 30.6 30.3 30.1 AHY-M (3) (2) (25) (44) (34) (7) (6) (7) (18) (10) 26.6 25.9 27.4 27.0 27.4 25.7 26.9 26.3 26.9 28.1 25.1 25.7 [1.60] [1.34] [1.89] [1.29] [1.04] [1.51] [1.47] [1.69] 27.9 26.1 25.1 24.4 20.7 23.4 25.6 24.6 23.6 25.6 31.5 29.8 30.3 26.8 28.1 28.2 30.0 31.3 AHY-U (1) 25.2 SY-F (1) (1) 26.8 28.9 SY-M (2) (3) 25.2 29.0 21.9 11. A 28.5 31.5 U-F (2) (12) (19) (5) (2) 27.4 26.7 27.7 26.5 27.0 25.9 [1.33] [1.51] [1.41] 25.9 29.0 24.0 26.0 24.3 28.1 28.9 32 2 28.0 U-M (17) (29) (14) (3) 26.9 26.0 27.1 26.2 [1.88] [2.47] [2.91] 23.8 23.5 21.8 21.2 28.5 31.4 30.9 30.3 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 13 Age/Sex Jan. U-U Feb. Mar. Apr. May June July Aug. Sept. (1) 28.9 Oct. Nov. Eastern Kingbird Tyrannus tymnniis n Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. HY-U (10) 39.3 [3.14] 34.5 44.4 (5) 41.7 [3.19] 37.8 45.4 AHY-U (9) 39.8 [1.16] 37.4 40.8 (3) 40.7 39.0 43.4 U-U (1) 42.0 Great Crested Flycatcher Myiarchus crinitus Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. HY-U (2) (1) 32.4 36.8 32.1 32.8 AHY-U (9) (9) (1) (8) (5) 33.1 34.5 34.3 33.5 33.1 [2.38] [2.71] [3.67] [2.00] 29.7 30.8 28.0 31.2 36.9 38.6 39.5 35.7 U-U (4) (1) 33.6 33.1 32 2 35.0 Eastern Phoebe Sayornis phoehe n Age/Sex HY-U AHY-F AHY M Mar. Apr. May June July Aug. Sept. Oct. (6) (15) (32) (57) (41) 18.3 17.8 18.6 19.5 19.6 [1.46] [1.62] [1.47] [1.59] [2.61] 16.8 14.0 15.5 16.2 11.3 20.6 20.3 21.6 22.8 23.5 (3) (1) (!) 19. ! 19.4 19.5 18.1 20.5 (3) (1) (1) (1) 24.0 23.2 22.4 20.0 23.2 24.4 (28) (63) (15) (2) (35) (18) 19.8 19.8 20.1 20.8 19.8 19.4 [1.82] [1.70] [1.02] 20.6 [1.78] [0.68] 16.6 16.1 18.3 21.0 16.0 18.0 23.6 24.3 21.6 23 2 20.5 Dec. = 28 Dec. = 40 Dec. 485 Dec. AHY-U 14 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. U-U (1) (1) (39) (118) (3) 16.5 18.4 19.6 19.7 20.8 [1.80] [1.81] 19.6 13.3 11.4 21.6 23.0 23.8 Yellow-bellied Flycatcher Empidona. X davi\’enrri.s n = 505 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-M (1) 12.2 HY-U (99) (226) (10) II.2 II.2 II. 9 [0.89] [0.90] [1.44] 8.9 9.2 10.0 13.2 14.3 13.8 AHY-F (1) 12.7 AHY-M (2) (1) 12.2 9.2 11.9 12.5 AHY-U (106) (5) (14) (14) (1) 11.5 11.6 12. 1 II. 9 II.8 [0.90] [0.97] [1.27] 11.45] 9.5 10.5 10.2 9.6 13.9 12.9 14.8 15.5 U-U (1) (4) (20) 11.8 11.8 1 1.2 10.8 [0.92] 13.6 9.7 12.9 Acadian Flycatcher Empidona.x virescens n = 102 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (3) (21) (15) (1) 12.7 13.0 13.0 12.5 12.1 [1.03] [1.01] 13.7 11.5 11.1 15.4 14.7 AHY-F (1) 13.2 AHY-U (52) (3) (4) (1) 12.9 12.7 12.2 13.5 [1.111 12.5 9.8 10.4 12.9 14.4 16.1 u-u (I) 12.4 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 15 “Traill’s” Flycatcher Empidonax traillii and E. ulnorum n Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (73) (54) (3) 12.7 14.5 14.7 [1.00] [2.07] 12.6 10.5 10.2 15.8 15.0 19.5 AHY-F (1) (2) 12.4 11.8 11.7 12.0 AHY-M (1) (1) 11.6 13.5 AHY-U 010) (14) (3) (28) (3) 13.5 12.9 13.3 13.4 13.9 [1.33] [0.89] 11.1 [0.80] 13.2 9.2 11.4 16.0 11.2 14.8 16.9 14.4 14.8 U-U (9) 13.2 [1.69] 10.8 15.7 Least Flycatcher Empidonax minimus Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (6) (397) (461) (12) 9.2 10. 1 10.3 10.7 [0.61] [0.93] [0.90] [0.71] 8.3 8.2 7.6 8.1 10.0 14.1 15.6 11.8 AHY-F (10) (4) (4) (7) 10.3 9.7 9.3 9.3 [0.97] 9.0 8.6 [0.56] 8.9 10.1 9.9 8.7 12.6 10.4 AHY-U (6) (257) (3) (14) (35) (44) (1) 10.6 10.3 10.6 10.8 10.3 10.3 8.6 [0.72] [0.85] 10.2 [0.61] [0.93] [1.15] 9.5 8.2 10.9 9.8 8.6 8.4 11.4 12.3 1 1.6 13.4 14.9 U-U (9) (18) (1) 9.8 10. 1 13.2 [0.54] [0.71] 8.8 8.5 10.5 11.0 Eastern Wood Pewee Contopns virens Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (1) (92) (212) (15) 12.9 13.0 13.3 14.2 [0.89] [1.09] [1.01] 10.6 11.0 12.6 15.2 18.7 16.5 302 Dec. 1289 Dec. 455 Dec. 16 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. AHY-F (2) (1) 15.2 14.0 14.8 15.6 AHY-U (38) (11) (6) 14.5 13.1 14.0 [0.95] [0.70] [0.73] 12.8 12.2 12.8 16.5 14.3 14.8 (30) 13.8 [1.07] 12.2 16.4 (35) 14.5 [1.27] 12.3 18.2 (2) 15.7 13.3 18.1 U-U (1) 12.5 (8) 12.8 [1.23] 10.4 14.1 (1) 14.0 Olive-sided Flycatcher Nuttallornis borealis n = 43 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-U (7) 30.4 [2.01] 28.6 34.6 (10) 31.3 [1.91] 27.8 34.0 AHY-U (8) 33.0 [5.16] 27.9 42.2 (6) 32.2 [2.94] 29.6 36.5 (4) 32.7 31.3 33.9 U-U (2) 27.5 26.7 28.3 (4) 32.0 31.3 33.0 (2) 32.0 29.9 34.0 Tree Swallow Iridoprocne hicolor n = 28 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. AHY-F (2) 20.0 19.5 20.5 (4) 18.6 15.6 20.9 AHY-M (8) 21.5 [1.55] 19.0 23.3 (10) 19.3 [1.28] 17.4 21.3 AHY-U (1) 22.6 (3) 19.9 17.1 22 2 Bank Swallow Riparia riparia n = 22 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec AHY-U (6) 14.6 [0.53] 13.8 15.3 (16) 13.0 [1.41] 9.9 15.3 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 17 Rough-winged Swallow Age/Sex Jan. Feb, AHY-F AHY-M Mar. AHY-U Barn Swallow Age/Sex Jan. HY-U AHY-F AHY-M AHY-U ASY-F ATY-F ATY-M Feb. Mar. Apr. May (4) 15.6 14.2 17.2 (1) 16.2 (2) 16.4 16.0 16.7 (7) (33) 16.6 15.8 [1.02] [1.52] 15.3 10.3 18.0 18.3 Apr. (7) 19.0 [1.66] 16.8 21.1 (5) 20.2 [1.95] 17.8 22.5 (7) 18.9 [1.78] 15.6 20.7 Stelgidopteryx rnficollis June July Aug. Sept. Oct. n = 47 Nov. Dec. Hiruncio msticu May June July Aug. Sept. (5) (7) (7) 19.0 18.7 19.6 1.61] [0.84] [1.34] 17.7 17.7 17.3 21.1 20.0 21.6 Oct. n = 178 Nov. Dec. (39) 18. 1 [1,93] 13.4 23.4 (54) 18.6 [1-55] 15.5 21.7 (44) !8.6 [1.86] 13.7 22.8 (1) 20.0 (I) 20.9 (1) 18.3 Cliff Swallow Petrochelidon pyrrhonota n = 3 (1) HY-U, Aug., 22.6; (1) AHY-U, Apr., 25.5; (!) AHY-U, May, 18.5. Blue Jay Cyanocitta cristata n = 515 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-U (3) (12) (27) (8) (3) 85.3 85.6 87.3 87.0 89.2 80.3 [6.65] [5.22] [5.32] 86.8 87.9 74.7 77.7 79.7 92.1 97.5 96.6 94,4 18 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May AHY-F AHY-U (5) (9) (12) (8) (18) 89.2 86.0 88.0 85.6 83.8 [4.73] [6.45] [3.74] [5.06] [6.55] 82.2 77.1 81.8 79.5 73.1 94.6 96.0 95.4 96.0 97.5 SY-F (1) 79.1 SY-U (13) (5) (27) (43) (112) 85.4 79.7 85.3 86.7 85.2 [7.70] [6.45] [5.57] [6.24] [6.24] 69.0 74.4 75.6 76.1 72.8 97.1 90.9 101.2 97.7 99.7 ASY-F (1) (1) 87.4 86.3 ASY-M (1) 79.4 ASY-U (9) (7) (26) (41) (62) 90.2 86.4 89.5 88.5 86.6 [7.23] [7.68] [5.74] [6.55] [6.15] 81.1 11. A 76.3 76.2 74.6 104.2 95.9 101.4 101.1 102.9 u-u June July Aug. Sept. Oct. Nov. Dec. (1) 90.4 (12) (16) (7) (6) 90.3 88.4 89.6 91.3 [4.14] [8.47] [4.43] [6.01] 83.5 64.1 85.5 81.7 96.2 99.0 97.2 99.5 (2) (3) (2) 82.4 88.1 90.8 80.5 82.7 90.5 84.4 94.3 91.1 (1) 90.9 (1) (1) (1) (4) (1) 12.3 109.4 101.3 92.8 77.5 83.9 98.0 (1) (2) 84.4 89.8 88.5 91.2 Common Crow (I) HY-U, Aug., 411.5. Corviis hrachyrhynchos n = 1 Black-capped Chickadee Pams atricapiUus n = 3307 Age/Sex HY-F HY-M HY-U AHY-F Jan. Feb. Mar. Apr. (2) 10.2 10.1 10.3 (3) (2) (3) (2) 10.9 10.9 11.2 10.8 10.9 10.7 10.7 10.4 11.0 11.1 11.6 11.1 May June July Aug. (1) 10.5 (1) 10.9 (3) 10.3 9.9 10.8 (6) 10.8 [0.54] 10.2 11.3 (3) (70) (118) 9.7 10.3 10.7 8.9 [0.71] [0.77] 10.4 8.2 8.9 (1) 9.6 11.9 12.6 (1) 10.6 Sept. Oct. (3) (3) 10.6 10.4 10.0 9.8 11.3 10.8 (5) (6) 11.2 II.2 [0.55] [0.58] 10.8 10.6 12.2 12.2 (199) (726) 10.7 10.5 [0.78] [0.75] 8.9 8.3 12.5 13.5 Nov. Dec. (5) 10.4 [0.32] 10.0 10.8 (2) 11.5 11.4 11.6 (280) (5) 10.7 10.5 [0.75] [0.65] 8.7 9.8 12.7 1 1.4 (1) 11.6 (1) II.O AHY-M 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 19 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. AHY-U (84) (55) (158) (231) (28) (4) (8) (15) (51) (45) (4) 10.8 10.9 II.O 10.9 10.8 10.2 10.6 10.6 10.5 10.8 10.5 [0.79] [0.84] [0.80] [0.84] [0.80] 9.2 [0.52] [0.56] [0.79] [0.79] 9.9 9.5 9.2 9.2 9.2 9.1 10.8 9.6 10.0 8.9 9.5 11.2 13.0 12.6 13.0 13.2 12.4 11.4 11.8 12.8 12.5 SY-F (8) (3) (2) (3) (1) 10.2 10.7 10.7 10.3 10.3 [0.52] 10.2 10.4 9.7 9.3 11.3 10.9 10.8 11.0 SY-M (9) (3) (5) (1) (1) (3) (3) (6) 11.7 11.6 II. 4 12.0 12.3 1 1.2 10.8 11.3 [0.33] 11.3 [0.45] 10.9 10.0 [0.47] 1 1.2 12 2 10.8 11.6 11.2 10.6 12.2 12.0 11.9 SY-U (38) (7) (83) (34) (7) (3) (10) (18) (13) (1) 10.9 10.7 II.O 11.2 II.3 10.7 10.8 10.9 10.9 11.3 [0.92] [0.48] [0.72] [1.05] [0.33] 10.5 [0.77] [0.91] [0.72] 8.9 9.8 9.5 9.2 10.8 10.9 9.2 9.0 9.9 12.5 11.3 12.6 12.9 11.9 12.0 12.2 12.0 ASY-F (2) (2) (5) (1) (2) 10.2 II. I 11.2 11.4 10.3 9.9 10.7 [0.55] 10.0 10.5 11.5 10.6 10.6 1 1.8 ASY-M (5) (2) (3) (1) (1) (2) 11.4 10.9 ll.l 10.6 II.O 10.5 [0.23] 10.3 10.7 10.2 11.2 11.5 1 1.5 10.9 11.7 ASY-U (20) (8) (42) (24) (1) (1) (2) (3) (7) (2) II. I II. I ll.l ll.l II.O 1 1.2 9.6 9.8 10.5 II.O [0.64] [1.02] [0.77] [0.86] 9.5 8.2 [0.70] 10.9 9.9 9.5 9.4 9.6 9.7 10.9 9.4 11.0 12.4 12.0 12.6 12.7 11.5 TY-F (1) (2) (1) 10.9 10.5 10. 1 10.3 10.6 TY-M (1) (5) (4) (1) (1) (2) 11.6 II. 4 II. 4 II. 8 11.7 10.9 [0.47] 10.5 10.4 10.8 11.9 11.3 11.9 TY-U (9) (18) (8) (2) (1) (1) (2) (4) (5) 11.6 11.5 II. 3 10.9 8.8 12.3 10.6 II.O 10.4 [0.83] [0.65] [0.89] 10.7 10.5 10.3 [1.00] 10.1 10.6 10.3 11.1 10.7 11.6 9.0 12.7 12.8 13.1 11.6 ATY-F (1) (1) (!) (1) (1) 10.8 10.4 10.8 10.9 10.0 ATY-M (2) (5) (1) (1) (2) (1) 11.3 II. 3 II. 5 10.3 10.9 10.5 11.2 [0.38] 10.4 11.3 10.8 11.4 11.8 20 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. ATY-U (15) (7) (57) (34) (3) (2) (6) (9) (13) (3) 10.5 10.8 II. 0 10.8 II.4 10.9 10.8 10.3 lO.O 10.9 [0.76] [0.70] [0.81] [0.79] 11.3 9.8 [0.44] [0.92] [0.70] 10.4 9.3 9.9 9.3 9.5 11.5 12.1 10.1 8.9 8.8 11.2 11.8 12.0 13.0 12.5 11.4 11.6 11.1 U-U (3) (13) (64) (290) (134) (36) 10.5 10.4 10.5 10.6 10.9 10.9 10.1 [0.83] [0.83] [0.71] [0.79] [0.67] 10.8 9.1 9.0 8.8 8.6 9.5 12.3 12.7 13.6 12.7 12.3 Carolina Chickadee Pa riis ca rol i n ensi-s n = 15 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-U (1) (1) (5) (1) 9.8 9.5 9.0 9.6 [0.22] 8.7 9.3 AHY-U SY-U (1) (1) lO.l 9.8 (2) (2) (1) 9.2 9.4 9.9 9.1 9.4 9.4 9.5 Boreal Chickadee (1) AHY-U, Oct., 9.7. Parus hudsonicus n = 1 Tufted Titmouse (Eastern Tufted Titmouse) Pams bicolor n = 1314 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) (1) (1) (2) (1) 20.5 23.3 20.4 22.7 20.8 22.5 22.8 HY-M (1) (1) (1) 22.3 23.0 23.7 HY-U (7) (38) (50) (101) (176) (98) (7) 21.4 20.4 21.5 21.8 21.7 21.9 22.2 [1.28] [1.44] [1.58] [1.55] [1.68] [1.42] [1.37] 19.2 17.6 18.7 18.3 16.7 18.5 20.0 23.5 24.0 25.0 25.8 27.5 24.6 23.6 AHY-F AHY-M (1) (5) (4) (7) 20. 1 20.5 21.2 22.0 [0.31] 19.6 [1.33] 20.1 22 2 20.3 20.9 24.6 (1) 22.5 (2) 21.1 20.4 21.8 (33) (29) (109) (74) (17) (1) (1) (10) (18) 22.4 21.7 22.0 21.5 21.6 19.6 19.4 21.7 21.2 [1.60] [1.50] [1.80] [1.78] [1.35] [1.37] [1.34] 18.7 19.3 18.0 17.2 19.2 19.5 18.9 25.5 24.3 27.0 25.5 23.5 24.0 24.8 (1) 23.8 AHY-U 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS Age/Sex Jan. Feh. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. SY-F (1) (1) (5) (3) (3) (2) 20.7 20.9 21.1 21.5 23.0 20.7 [0.80] 20.7 21.1 20.4 19.9 22.1 24.6 21.0 22 1 SY-M (1) (1) 22.2 21.7 SY-U (23) (8) (53) (44) (15) (6) (14) (31) (1) 22.4 22.3 22.7 22.7 22.2 21.7 22.3 21.7 21.9 [1.22] [0.95] [1.67] [1.37] [0.96] [1.80] [1.60] [1.40] 20.3 20.3 17.7 19.0 20.7 19.4 19.5 17.4 24.9 23.3 26.0 25.3 23.9 24.0 24.6 24.7 ASY-F (!) 21.6 ASY-M (1) (1) 22.1 23.0 ASY-U (2) (2) (18) (12) (8) (!) (9) (10) (4) 23.3 22.6 23.0 22.4 22.2 23.1 21.6 22.1 21.2 23.1 22.4 [1.45] [1.62] [1.39] [1.16] [1.56] 20.6 23.5 22.7 19.7 19.6 20.8 19.5 19.9 ->2 1 25 2 25.5 24.6 23.7 24.2 TY-U (6) (3) (13) (9) (7) (1) (9) (3) (1) 22.9 21.4 21.7 22.9 22.6 22.2 21.5 21.6 22.3 [4.00] 21.2 [2.32] [1.36] [0.83] [1.47] 20.2 17.4 21.7 18.3 19.8 21.6 18.7 24.0 29.3 26.0 24.1 24 2 23.5 ATY-F (1) (1) 20.4 21.7 ATY-M (2) (3) (1) 22.3 22.6 22.1 21.4 22.0 23.2 23.0 ATY-U (5) (5) (26) (17) (4) (2) (12) 22.7 22.3 22.5 22.3 21.3 21.8 21.7 [1.62] [1.27] [1.14] [1.74] 20.4 21.0 [1.22] 20.7 21.3 20.5 20.5 22.4 22.6 20.3 25.0 24.5 25.1 27.2 24.5 U-F (1) (3) 21.1 20.7 20.2 21.2 U-U (1) (2) (29) (22) (34) (11) 23.3 21.2 21.7 22.2 21.4 22.7 20.3 [1.52] [1.87] [1.24] [1.30] 22.0 19.5 19.3 18.5 21.3 24.9 28.9 24.2 24.8 White-breasted Nuthatch Sitta caroiuiensis n = 296 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) (1) (7) (2) 20.7 18.4 20.7 20.3 1'^ BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex HY-U AHY-F AHY-U SY-F SY-U ASY-F ASY-U TY-U ATY-F ATY-U U-F U-U Red-bre; Age/Sex HY-M Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. (2) (3) (7) (6) (1) 19.8 21.9 20.8 21.2 21.7 19.5 20.2 [0.53] [1.36] 20.1 23.2 19.9 19.2 21.2 22.8 (4) (5) (16) (7) (1) (1) (8) (5) (1) 21.0 20.6 21.0 21.4 21.1 20.7 20.8 20.7 20.4 19.4 [1.24] [0.96] [1.30] [1.33] [0.55] 22.0 19.4 19.4 19.0 19.0 19.8 22.3 22.8 23.2 22.4 21 .2 (6) (4) (38) (26) (3) (2) (2) (1) (11) (11) (1) 22.0 22.5 21.3 20.4 21.1 20.3 20.0 17.2 21.2 21.6 21.1 [0.991 20.1 [1.42] [1.30] 20.7 19.5 19.7 [0.74] [0.70] 20.7 24.6 18.8 18.5 21.4 21.1 20.4 20.2 20.2 23.0 24.4 22.8 22.4 22.7 (1) (1) (2) (1) 19.8 21.3 21.8 20.6 2 1 2 22 3 (1) (6) (2) (3) (1) 21.8 21.7 21.2 21.5 23.0 [0.52] 20.0 20.1 20.8 22.4 22.2 22.2 (2) (1) (3) (2) (3) (4) 2 1.1 21.9 21.8 20.2 21.0 21.5 20.2 20.6 19.3 20.4 20.3 21.9 22.7 21.0 22.2 22.6 (1) (9) (5) (1) (4) 20.1 21.7 21.3 21.3 22.6 [2.06] [0.50] 20.7 19.2 20.5 23.9 26.7 21.8 (1) 21.7 (1) (1) (1) (1) 20.5 19.6 20.4 20.6 (3) (3) (1) 21.0 21.1 21.3 20.3 20.0 22.2 22. 1 (2) (8) (1) 21.3 20.6 21.7 20.1 [0.92] 22.4 19.2 22.0 (5) (11) (15) (6) 21.0 21.6 21.4 22.2 [0.92] [0.87] [0.72] [0.65] 19.8 19.7 20.0 21.6 22.3 22.6 22.8 23.3 .sted Nuthatch Silta canadensis n = 7 Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. (1) 12.1 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 23 Age/Sex Jan. HY-U AHY-F Brown Creeper Age/Sex Jan. HY-U AHY-F AHY-U SY-U ASY-U ATY-U U-U House Wren Age/Sex Jan. HY-U AHY-F AHY-M Feb. Feb. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. (2) 10.5 10. 1 10.8 (2) (2) 10.6 II. 2 10.4 10.7 10.8 11.6 Certhia familiaris n = Mar. Apr. May June July Aug. Sept. Oct. Nov. (1) (3) (2) (19) (15) 7.5 7.6 8.6 8.2 7.9 7.1 8.2 [0.43] [0.65] 8.4 9.0 7.5 7.0 9.1 8.9 (1) (1) 8.0 8.0 (19) (26) (19) (6) 8.3 8.5 8.3 8.6 [0.46] [0.66] [0.56] [0.66] 7.4 7.2 7.2 7.4 9.1 9.7 9.3 9.3 (1) (1) 7.9 8.1 (1) 8.2 (1) 8.7 (3) (25) (8) 8.2 8.3 8.3 8.1 [0.61] [0.51] 8.4 7.4 7.7 9.9 9.3 Troglodytes aedon n = Mar. Apr. May June July Aug. Sept. Oct. Nov. (2) (76) (178) (345) (107) (1) 10.4 10.3 10.7 Il.l II.4 12.5 9.7 [0.48] [0.75] [0.84] [0.74] 11.0 8.5 8.0 9.2 9.6 11.4 13.3 14.5 13.2 (6) (2) (5) (6) 10.5 12.2 11.5 11.4 [0.58] 11.1 [0.64] [0.69] 9.9 13.2 10.9 10.7 11.5 12.5 12.3 (5) (1) (2) (1) 10.9 10.9 10.7 10.8 [0.60] 10.5 10.4 10.8 11.9 Dec. 152 Dec. 1055 Dec. 24 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. AHY-U SY-U ASY-F ASY-M ASY-U ATY-F ATY-U U-M u-u Winter Wren Age/Sex Jan. HY-U AHY-U U-U Bewick's Wren Age/Sex Jan. AHY-U Feb. Mar. Apr. May June (24) (87) (2) 10.7 10.6 //./ [0.68] [0.831 11.0 9.7 8.9 11.2 12.2 14.2 (3) 10.6 10.1 11.2 .1 LI 1 y Aug. Sept. (13) (30) (29) ll.O 11.2 11.3 [0.26] [0.79] [0.78] 10.5 9.9 9.7 11.5 12.6 13.8 (1) 9.7 Oct. Nov. Dec. (13) 1 1.2 [0.95] 9.7 13.1 Feb. Feb. (1) (3) 13.2 11.8 11.2 12.1 (3) 10.8 10.2 11.5 (1) 11.8 (1) 10.8 (4) 10.6 9.5 11.4 (1) 11.7 (2) 1 1.2 11.0 1 1.4 (1) 10.2 (1) 10.9 (1) 1 1.0 (3) (22) (64) (8) 10.6 11.0 Il.l II. I 10.2 [0.67] [0.86] [1.05] 11.0 9.9 9.4 10.1 12.5 13.0 12.9 Troglodytes troglodytes n = Mar. Apr. May June July Aug. Sept. Oct. Nov. (14) (85) (18) 8.6 8.9 9.1 [0.92] [0.84] [0.70] 7.2 7,1 8.1 10.1 10.7 10.3 (1) (8) (1) (2) (28) (1) 9.3 8.7 9.3 9.2 9.0 8.3 [0.52] 8.6 [0,78] 7.7 9.8 7.5 9.5 10.1 (9) (4) 8.7 9.0 [0.84] 7.8 7.7 10.5 10.1 Thryomanes hewickii n Mar. Apr. May June July Aug. Sept. Oct. Nov. (1) (1) (1) 9.2 11.3 II.3 171 Dec. = 3 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 25 Carolina Wren Thryothoni; >■ hulovivianns n = 43 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (8) 18.7 [1.08] 17.3 20.3 (13) 18.6 [1.02] 16.6 20.2 (5) 17.6 [1.46] 15.6 19.2 (3) 19.7 17.7 21.6 (1) 20.3 AHY-U (1) 17.5 (3) 19.6 18.2 22.3 (3) 19.2 15.9 21.0 (1) (1) 16.8 22.1 (1) 16.8 SY-U (3) 19.3 19.1 19.6 Long-billed Marsh Wren C istot horns palnstris n = 35 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (10) 11.9 [1.25] 9.9 13.7 (15) 11.2 [1,15] 9.5 13.6 AHY-U (3) 12.7 11.5 13.5 (4) 1 1.3 10. 1 12.1 U-U (2) 9.8 9.5 10.1 (1) 9.4 Short-billed Marsh Wre n Cistothorus platensis n = 3 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (2) 9.2 8.9 9.5 (1) 8.4 Mockingbird Minins polyglottos n _ T Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec AHY-U (2) 49.8 47.7 52.0 Gray Catbird Dnnietella carolinensis n = 3530 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY.F (2) J6.3 35.0 37.5 HY-M 39.0 26 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (2) (154) (333) (1028) (271) 31.4 35.4 36.4 38.6 41.2 30.7 [1.96] [2.16] [2.85] [3.40] 32.0 29.3 30.0 31.5 32.8 41.5 43.4 48.3 52.0 AHY-F (14) (19) (13) (19) (6) (3) 35.0 36.9 37.8 36.3 39.6 38.5 [2.55] [2.40] [2.74] [2.45] [2.49] 32.2 30.9 33.4 33.9 33.3 35.4 46.0 40.9 40.7 42.2 41.7 42.8 AHY-M (12) (17) (5) (3) (1) 35.7 35.4 35.9 37.1 40.1 [4.40] [1.14] [0.80] 35.5 30.0 33.2 35.0 39.8 48.4 37.2 37.1 AHY-U (11) (711) (72) (49) (44) (157) (59) 34.5 34.8 34.7 36.2 37.6 40.0 43.5 [2.37] [2.54] [2.28] [1.87] [2.68] [3.04] [4.37] 32.1 26.7 26.6 32.1 32.1 30.5 34.3 40.7 45.2 40.6 40.2 43.7 47.7 56.5 SY-F (2) (3) (1) 35.3 33.6 41.7 30.3 31.2 40.3 35.7 SY-M (3) (1) (2) 32.1 34.6 35.5 30.1 34.2 34.5 36.8 SY-U (14) (2) (3) (6) (1) 34.4 37.3 39.1 41.3 45.5 [2.13] 33.5 38.4 [3.37] 31.3 41.1 39.6 37.1 39.7 45.5 ASY-F (1) (1) (2) (1) (1) 32.5 43.5 37.3 34.3 37.3 34.0 40.6 ASY-M (5) (2) (3) (1) 34.7 35.0 34.5 42.4 [1.10] 34.3 33.1 33.5 35.6 36.1 36.1 ASY-U (18) (1) (3) (2) (1) 35.4 38.2 39.1 39.7 48.1 [1.60] 34.0 38.2 33.1 41.8 41.2 38.1 TY-F (2) (1) (1) 36.2 32.1 38.2 33.9 38.5 TY-M (3) (1) (1) 31.7 35.9 39.7 31.0 32.1 Nov. (3) 57.8 36.8 39.6 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 27 Age/Sex Jan. TY-U ATY-F ATY-M ATY-U U-M u-u Brown Thrasher Age/Sex Jan. HY-U AHY-F AHY-M AHY-U ASY-U U-U Feb. Feb. Apr. May June July Aug. Sept. Oct. (1) (2) (1) (2) (3) 35.9 33.3 35.4 39.8 39.0 31.1 37.5 38.7 35.5 42.0 39.6 (2) 35.3 34.4 36.2 (8) (1) 34.3 38.3 [1.18] 33.4 37.0 (10) (2) (3) (7) 36.0 33.4 37.3 39.6 [1.53] 33.3 35.3 [1.61] 33.5 33.5 41.1 36.5 38.1 (1) 41.1 38.7 (7) (320) (62) 37.7 39.3 40.7 [1.74] [2.91] [3.14] 35.9 32.3 31.5 39.7 47.9 49.0 To.xostonia rufuin Apr. May June July Aug. Sept. Oct. (10) (17) (47) (22) 66.4 68.9 69.8 72.5 [4.25] [5.17] [4.01] [6.75] 58.2 60.0 62.0 55.1 74.1 78.6 80.6 92 2 (2) (9) (2) (2) (1) 68.4 67.8 67.7 66.0 71.4 67.1 [2.97] 67.2 66.0 69.7 62.1 68.2 66.0 71.6 (2) (2) 72.8 67.8 62.5 66.2 83.0 69.3 (61) (40) (2) (3) (15) (2) 66.6 67.1 66.7 66.3 75.6 63.5 [5.39] [5.82] 65.9 61.0 [6.26] 58.1 57.6 58.5 67.4 70.1 66.3 68.9 79.2 79.2 89.0 (I) 66.3 (22) (11) 73.5 71.2 [5.01] [4.04] 62.3 65.8 83.4 78.5 Dec. 273 Dec. 28 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 American Age/Sex HY-F HY-M HY-U AHY-F AHY-M AHY-U SY-F SY-U ASY-F ASY-M TY-U ATY-F ATY-M Robin Jan. Feb. (I) 81.0 Turdus migratohus Mar. Apr. May June July (3) (9) (61) 75.8 76.1 74.9 68.3 [3.87] [4.97] 85.4 67.3 54.2 81.4 84.2 (26) (85) (37) (8) (7) 77.8 77.2 79.8 78.8 76.1 [5.55] [7.06] [5.44] [4.63] [3.48] 70.8 63.5 69.5 70.4 70.9 94.4 102.6 88.5 86.1 80.4 (37) (76) (57) (9) 77.2 77.1 75.6 76.6 [5.59] [6.34] [5.55] [4.94] 70.6 66.0 66.6 67.3 94.3 94.4 93.9 82.5 (16) (15) (5) 76.6 77.3 76.4 [4.40] [7.52] [6.97] 68.1 66.4 69.6 85.3 94.0 86.4 (1) 88.9 (3) (1) 75.4 68.6 71.0 78.4 (1) (3) (1) (1) 77.1 80.8 79.5 85.9 78.3 85.0 (4) (7) (4) 75.6 70.4 77.1 71.3 [4.52] 75.0 80.1 64.2 78.7 78.1 (1) 77.4 (2) 75.2 71.7 78.7 (1) (6) (3) (1) 68.8 73.0 76.7 74.5 [4.81] 70.9 65.1 84.1 79.8 n = 745 Aug. Sept. Oct. Nov. (1) (1) (12) (2) 82.0 82.8 78.8 70.7 [6.16] 66.7 66.7 74.6 88.8 (1) (16) (3) 74.0 79.8 75.6 [5.86] 72.1 70.5 81.6 94.5 (51) (7) (41) (19) 75.7 79.3 77.3 82.7 [3.55] [5.94] [4.66] [5.06] 68.8 69.2 66.2 74.4 82.0 86.1 87.7 90.5 (2) 81.5 80.3 82.7 (1) (4) 71.6 81.9 79.5 84.6 (I) 70.2 (1) 76.0 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 29 Age/Sex U-F Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. (7) 81.8 [5.16] 75.7 90.0 Nov. U M (4) 82.2 79.7 83.7 (4) 85.5 77.4 93.2 u-u (8) 79.8 [3.93] 73.5 85.3 (51) 79.1 [5.32] 67.7 94.4 (16) 82.3 [7.90] 65.6 97.5 Wood Thrush Hylocichla iniisleliiui n = Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. HY-U (17) 47.8 [3.67] 39.3 52.6 (28) 50.4 [2.77] 42.8 54.2 (116) 50.6 [3.03] 40.1 58.4 (9) 52.7 [4.83] 44.0 60.6 AHY-F (8) 49.1 [1.38] 46.6 50.7 (6) 53.3 [2.55] 51.1 57.7 (7) 48.6 [3.08] 42.8 53.1 (1) 51.4 AHY-M (1) 45.4 (14) 45.4 [1.50] 42.1 47.4 (7) 46.0 [2.06] 42.3 48.5 (5) 48.0 [0.72] 47.2 49.0 (1) 50.1 (1) 45.6 AHY-U (5) 46.6 [4.58] 40.0 52.2 (123) 46.9 [3.41] 39.2 56.4 (9) 48.7 [3.06] 46.2 55.5 (7) 48.7 [1.78] 46.8 52 2 (4) 49.3 45.0 53.3 (7) 49.6 [2.56] 46.9 54.2 (1) 54.2 SY-F (5) 50.3 [3.13] 45.0 53.3 (3) 49.7 49.3 50.3 (1) 46.2 SY-M (6) 48.1 [3.50] 43.6 51.3 (2) 47.7 46.7 48.7 (1) 48.6 SY-U (5) 45.9 [2.98] 42.5 (34) 46.6 [2.90] 40.5 (2) 47.4 45.6 49.1 (2) 45.4 43.1 47.7 50.7 55.1 (1) 46.8 Dec. 478 Dec. ASY-F 30 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. ASY-M ASY-U TY-U ATY-F ATY-M ATY-U U-M U-U Hermit Thrush Age/Sex Jan. HY-U AHY-U SY-U U-U Feb. Feb. Mar. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec (1) (7) (1) (1) 45.1 44.2 47.0 48.3 [1.93] 41.3 46.5 (1) (1) 45.1 44.1 (2) (3) 47.8 45.7 47.0 45.5 48.5 46.0 (1) (1) 44.7 50.4 (1) (3) 50.6 47.4 45.6 48.6 (1) 42.5 (1) 49.2 (14) 50.0 [2.91] 45.7 55.2 Catharus guttatus n = Apr. May June July Aug. Sept. Oct. Nov. (6) (440) (30) 29.1 30.6 32.3 [1.91] [2.31] [2.64] 26.7 24.6 28,9 31.6 39.8 39.6 (3) (4) (1) (29) (1) 28.7 30.4 30.0 31.4 34.4 26.6 26.6 [2.14] 30.3 33.6 27.6 37.4 (5) (5) 31.0 31.5 [2.12] [1.25] 27.9 30.0 33.4 33.2 (22) 30.7 [1.67] 27.5 34.0 Swainson's Thrush Catharus ustulatus n = 2174 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (1) (3) (1036) (112) 31.6 30.5 30.8 33.0 26.4 [2.54] [3.81] 33.3 21.7 26.5 46.7 48.2 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 31 Age/Sex Jan. Feb. Mar. AHY-F AHY-U SY-U U-U Gray-cheeked Thrush Age/Sex Jan. Feb. Mar. HY-U AHY-U SY-U U-U Apr. May June July Aug. Sept. Oct. (1) (8) 31. 1 29.6 [1.39] 28.0 32.1 (3) (249) (4) (2) (14) (119) (10) 29.6 29.8 32.8 29.7 30.8 32.4 34.4 27.6 [2.68] 29.6 29.6 [1.70] [2.96] [3.54] 31.1 21.9 37.0 29.8 27.4 26.2 28.7 39.3 33.5 42.1 39.1 (1) (82) 25.1 29.7 [2.70] 25.2 39.0 (7) (475) (47) 29.2 30.8 35.1 [0.98] [2.82] [5.15] 28.1 24.7 26.5 30.5 50.7 47.7 Catharus minimus Apr. May June July Aug. Sept. Oct. (2) (159) (97) 33.1 32.1 36.5 31.8 [2.58] [5.37] 34.4 25.0 25.4 40.8 52.6 (22) (37) (9) 33.1 33.6 38.2 [3.34] [3.91] [5.98] 27.0 27.1 30.2 39.6 45.1 45.0 (2) 34.4 33.2 35.6 (115) 31.6 [2.77] 26.4 39.8 (21) 35.3 [5.65] 27.2 50.5 Nov. Dec. n = 464 Nov. Dec. Veery Age/Sex HY-U Jan. Feb. Mar. Apr. Cathoms fuscescens May June July Aug. (ID 30.8 [1.63] 28.6 34.1 (2) (1) 30.7 29.1 30.3 31.0 n = 116 Sept. Oct. Nov. Dec (33) 33.9 [4.06] 28.7 48.2 AHY-F 32 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex AHY-U SY-U u-u Jan. Feb. Mar. Apr. May June July Aug. Sept. (38) (4) (6) 31.0 28.8 34.8 [2.68] 26.2 [5.84] 26.3 31.7 29.2 35.8 41.7 (15) 31.0 [2.651 27.9 37.8 (6) 31.8 [3.14] 27.9 36.3 Oct. Nov. Eastern Bluebird Age/Sex Jan. Feb. Mar. HY-M HY-U AHY-F AHY-M AHY-U (4) 29.9 26.9 31.6 (4) 29.7 27.5 32.0 (1) 30.1 U-F Blue-gray Gnatcatcher Age/Sex Jan. Feb. Mar. HY-M HY-U AHY-F Apr. (1) 29.9 Sialia sialis May June July (1) 30.4 Aug. Sept, n Oct. Nov. (1) 29.4 (1) 40.0 (1) (1) 30.2 30.5 (1) (2) (1) 28.9 29.7 29.3 28.4 30.9 PoUoptila caerulea Apr. May June (42) (27) (4) 5.8 6.4 6.5 [0.41] [0.87] 6.1 4.8 5.4 7.0 6.8 8.9 (36) (19) (3) 6.0 6.1 6.3 [0.37] [0.28] 5.9 5.2 5.7 6.6 6.8 6.7 July Aug. Sept (1) 5.7 (10) (36) (4) 6.0 6.0 6.0 [0.22] [0.26] 5.9 5.6 5.5 6.2 6.3 6.6 (4) (15) (1) 5.9 6.0 5.9 5.2 [0.40] 6.4 5.3 6.7 (2) (3) 6.3 6.0 6.1 5.3 6.4 7.0 (2) 31.8 30.7 32.8 n = Oct. Nov. Dec. 21 Dec. 235 Dec. AHY-M 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 33 Age/Sex AHY-U SY-M ASY-M TY-M ATY-M U-U Jan. Feb. Mar. Apr. May June July Aug. (2) (1) (1) (1) (6) 5.6 6.3 6.2 6.2 5.8 5.5 [0.67] 5.6 5.0 6.7 (1) 5.8 (1) (1) (1) 6.0 6.2 6.2 Sept. Oct. Nov. Dec. (I) 6.0 (2) 6.0 5.9 6.1 (1) 5.8 (1) 6.1 (I) 5.9 (4) 6.2 5.0 6.7 (3) 6.2 6.1 6.4 Golden-crowned Kinglet Age/Sex Jan. Feb. HY-F HY-M Mar. U-M Apr. AHY-F (4) (47) 5.6 5.8 5.5 [0.40] 5.8 5.1 7.1 AHY-M (4) (21) (104) 5.9 6.0 6.1 5.9 [0.40] [0.44] 6.0 5.2 4.9 7.2 7.2 SY-M (1) 6.1 U-F Reguliis saUapa May June July Aug. Sept. (I) 5.8 Oct. (55) 6.3 [0.581 5.2 7.8 (63) 6.4 [0.50] 5.6 7.6 (8) 6.4 [0.31] 6.1 6.9 n = (S43 Nov. (51) 6.2 [0.441 5.1 7.3 (65) 6.4 [0.43] 5.4 7.4 Dec. (8) (5) 6.5 6.2 [0.54] [0.46] 5.5 5.5 7.4 6.8 (26) (59) (3) 6.2 6.5 5.8 [0.37] [0.56] 5.8 5.4 4.5 5.9 6.7 7.8 (35) (83) 6.5 6.4 [0.53] [0.56] 5.6 5.3 7.6 7.7 34 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Ruby-crowned Kinglet Regains colendnlci n = 3184 Age/Sex Jan. HY-F HY-M HY-U AHY-F AHY-M AHY-U U-F U-M u-u Cedar Waxwing Age/Sex Jan. HY-F Feb. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. (108) (290) (11) 6.3 6.3 6.8 [0.63] [0.56] [0.38] 4.9 5.2 6.2 8.5 8.3 7.5 (9) (202) (27) 6.7 6.8 7.1 [0.74] [0.62] [0.67] 5.6 5.5 5.9 8.1 8.6 9.5 (6) (1) 6.2 [0.45] 5.6 6.8 6.7 (118) (338) (33) 6.1 6.3 6.3 [0.37] [0.47] [0.47] 5.3 5.2 5.3 7.0 7.8 7.3 (3) (518) (51) (7) 6.6 6.8 7.1 6.7 6.5 [0.53] [0.50] [0.43] 6.7 5.4 5.9 6.0 8.9 8.6 l.A (3) (2) 6.3 6.3 6.2 5.5 6.4 7.1 (39) (460) (106) 6.4 6.5 6.6 [0.72] [0.60] [0.54] 5.3 5.1 5.1 8.0 8.9 8.3 (5) (610) (229) 6.8 6.9 7.1 [0.63] [0.61] [0.54] 6.2 5.0 5.8 1.1 9.7 8.9 (7) 6.8 [0.59] 6.1 7.8 Bomhycilla cedrorum n = 3318 May June July Aug. Sept. Oct. Nov. Dec. (1) 31.5 (7) (506) (675) (248) (89) (3) 29.8 31.6 32.6 33.6 33.6 33.6 [1.58] [2.26] [2.14] [2.12] [2.48] 28.5 27.1 20.0 21.3 26.7 28.5 39.7 31.3 38.3 38.5 44.7 39.7 HY-U 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 35 Age/Sex AHY-F Jan. Feb. Mar. Apr. May (20) 32.4 [2.57] 28.6 38.8 AHY-M (18) 30.2 [2.66] 25.9 35.5 AHY-U (5) 31.2 [1.13] 29.6 32.7 (852) 32.0 [2.62] 22.7 41.1 ASY-F (1) 30.5 ASY-U (1) 27.3 June July Aug. Sept. Oct. Nov. (13) (9) (124) (23) 33.1 33.1 33.1 33.3 [2.29] [1.56] [2.28] [2.38] 30.0 29.8 28.0 29.4 37.6 35.4 40.2 39.9 (2) (1) (29) (8) 29.7 26.7 31.5 29.4 28.8 [2.90] [2.00] 30.5 27.2 25.5 39.6 32.6 (191) (19) (216) (136) (28) (80) 32.4 29.8 31.8 32.2 33.2 34.8 [2.94] [1.95] [2.68] [2.40] [1.86] [3.15] 25 2 25.7 24.6 26.3 30.5 29.4 43.0 34.3 43.0 40.7 37.2 43.0 Dec. (I) 34.0 U-U Starling Age/Sex Jan. Feb. Mar. Apr. HY-F HY-M (1) 30.3 Sturnus vulgaris May June July Aug. Sept. (9) (1) 33.5 32.3 [1.77] 30.7 35.9 n Oct. Nov. (2) (1) 78.8 63.8 77.5 80.0 (3) 79.4 76.1 84.4 HY-U (1) 52.4 AHY-F (11) (9) (3) 74.9 79.3 81.5 [6.09] [8.14] 75.0 63.9 67.5 91.1 83.2 91.2 AHY-M (6) (3) (5) 79.4 80.6 78.1 [6.00] 77.1 [4.68] 69.7 83.2 72.8 86.3 83.2 SY-F (1) 72.8 SY-M (2) (1) (1) 75.2 75.0 79.4 75.0 75.3 (1) 77.0 ASY-F 36 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. ASY-M U-U (I) 82.2 White-eyed Vireo Age/Sex Jan. Feb. HY-U AHY-F AHY-U ASY-F ASY-M Yellow-throated Vireo Age/Sex Jan. Feb. HY-U AHY-F AHY-U U-U Solitary Vireo Age/Sex Jan. Feb. HY-U AHY-U U-U Mar. Apr. May June July Aug. Sept. (I) 82.5 Vireo griseiis Apr. May June July Aug. Sept. (5) (4) 11.3 11.8 [0.42] 10.7 10.7 12.8 11.8 (1) (1) 11.9 12.7 (9) (49) (5) (4) (3) (3) 10.5 II.2 II.7 II. 1 12.0 II.8 [0.52] [0.71] [0.55] 10.6 11.5 10.3 9.6 10.0 10.9 12.3 12.4 13.0 11.2 13.1 12.3 (1) 11.4 (1) (1) 12.4 12.5 Vireo flavifrons Mar. Apr. May June July Aug. Sept. (3) 18.3 16.7 19.4 (1) 19.2 (3) (I) (1) 15.7 17.2 16. 1 15.0 17.0 (1) 18.8 Mar. Apr. (1) 15.1 Vireo solitarius May June July (3) 16.7 15.6 18.4 Aug. Sept. (42) 16.9 [1.30] 14.3 20.5 (6) 16.5 [1.14] 14.7 17.7 Oct. Nov. (1) 77.0 n = Oct. Nov. (1) 12.5 n Oct. Nov. n = Oct. Nov. (49) 16.5 [0.93] 14.0 19.5 (8) 16.9 [0.64] 16.0 18.2 (4) (1) 16.0 16.0 14.5 18.0 Dec. 88 Dec. 10 Dec. 114 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 37 Red-eyed Vireti Vireo olivaceus n = Age/Sex HY-F HY-M HY-U AHY-F AHY-M AHY-U SY-M SY-U ASY-F ASY-M ASY-U TY-F ATY-F Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. (1) 16.7 (4) (1) 16.6 15.8 16.0 17.3 (15) (211) (401) (17) 17.1 16.7 17.3 18.9 [1.01] [1.39] [1.63] [2.11] 15.5 11,5 12.2 17.0 18.6 23.1 25.7 23.7 (16) (24) (64) (51) (7) 15.7 17.5 16.4 16.7 17.3 [0.921 [1.42] [0.92] [1.14] [0.76] 14.6 15.6 14.2 13,5 15.7 17.7 21.5 18.4 20.7 17.8 (15) (25) (15) (18) (2) 16.5 16.7 16.7 16.9 16.8 [1.46] [0.80] [0.86] [1.32] 16.1 14.2 15.5 14.6 15.1 17,5 19.6 18.3 18.2 20.0 (550) (95) (67) (193) (92) (1) 16.5 16.6 16.7 17.2 17.8 20.4 [1.32] [1.02] [0.89] [1.14] [1.64] 12.6 13.7 14.0 15.0 15.3 22.7 18.7 18.6 20.8 25.1 (1) (1) (1) 15.7 16.3 16.2 (6) (1) 15.8 15.6 [0.95] 14,3 17.1 (7) (1) (3) (6) 16.2 17.2 15.5 16.8 [0.60] 15.0 [0.92] 15.4 16.1 15.4 17.3 18.3 (2) (1) (1) (3) 18.9 17.3 17.4 16.6 17.9 15.8 19,9 17.7 (26) (3) (3) (ID (1) 15.9 17.2 16.6 17.0 17.3 [0.71] 16.6 16.3 [0.70] 14.9 18.1 16.9 15.9 18.2 (1) 18.1 17.1 (5) (1) (4) (12) 17.0 16.2 16.0 16.4 [2.21] 13.9 [1.08] 15.5 17.6 13.4 20.9 18.1 Nov. 2055 Dec. 38 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. ATY-M Feb. Mar. Apr. May June (18) (2) 16.2 15.9 [0.80] 15.4 14.5 16.3 18.4 July Aug. (7) 16.8 [0.38] 16.2 17.2 Sept. Oct. Nov. Dec. ATY-U (18) (5) !6.3 16.4 [1.00] [0.78] 15.1 15.4 18.6 17.6 (14) 17.4 [0.87] 16.2 19.5 (2) 15.3 12.0 18.6 U-U (3) 16.6 16.4 16.8 Philadelphia Vireo Vireo pluladelpht icus n = 246 Age/Sex Jan. HY-U Feh. Mar. Apr. May June July Aug. Sept. (117) 12.0 [1.02] 9.4 15.3 Oct. (13) 12.6 [1.42] 11.2 15.4 Nov. Dec. AHY-U (79) 12.2 [1.14] 10.3 16.1 (16) 12.2 [0.90] 10.7 14.3 (6) 12.7 [0.94] 11.0 13.7 U-U (15) II. 8 [1.00] 10.4 13.7 Warbling Vireo Vireo gilviis n = 20 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U 14.8 [0.99] 13.6 16.1 AHY-U (14) 13.6 [1.19] 11.4 17.0 U-U (1* 14.1 Black-and-white Warbler Age/Sex Jan. Feh. Mar. Apr. HY-F Mniotilta varia May June July Aug. Sept. (2) (13) (19) 9.5 10. 1 10.6 9.4 [0.55] [1.23] 9.6 9.3 8.3 11.1 14.0 Oct. n = 180 Nov. Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 39 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. HY-M (2) (17) (26) 10. 1 10.1 10.5 9.9 [0.77] [0.83] 10.2 7.9 8.7 11.1 11.8 HY-U (1) (5) (2) 11.4 10.6 9.3 [0.75] 8.4 9.8 10.2 11.5 AHY-F (1) (14) (3) (10) (12) (3) 9.9 10.8 10.3 10.9 10.3 10.3 [0.93] 9.7 [0.56] [0.70] 10.0 9.0 11.2 10.2 9.3 10.9 12.7 12.2 11.1 AHY-M (5) (9) (2) (2) (5) 10.7 10.4 11.6 11.9 11.9 [0.88] [0.95] 11.2 11.6 [2.80] 10.1 9.1 12.0 12.1 8.8 12.2 12.1 15.2 SY-F (1) 10.2 SY-M (1) 11.2 U-F (1) (1) 10.3 9.3 U-U (18) (5) 10.4 10.4 [0.69] [0.68] 9.2 9.5 11.7 11.3 Prothonotary Warbler Protonotarici citrea Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. AHY-M (I) 12.9 Nov. n Nov. Worm-eating Warbler Hehuitheros vermivonts Age/Sex Jan. Feb. Mar. Apr. May June July HY-U AHY-U (2) (6) 13.0 12.8 12.8 [0.67] 13.1 11.6 13.4 U-U n = Aug. Sept. Oct. Nov. (2) 12.6 12.2 12.9 (1) 12.2 Dec. Dec. 11 Dec. 40 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Golden-winged Warbler Vennivora chrysoptera Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. HY-F (2) (11) (4) 8.7 8.3 8.8 7.1 [0.75] 8.1 10.3 7.0 9.4 9.8 HY-M (4) (10) (1) 9.8 8.8 9.0 9.6 [0.54] 10.4 8.0 9.9 HY-U (2) (1) 8.2 7.7 7.9 8.5 AHY-F (64) (7) (2) (4) (1) 8.8 9.3 10.0 8.7 8.4 [0.88] [0.98] 9.6 7.8 7.5 8.7 10.4 9.5 11.8 11.5 AHY-M (2) (84) (25) (3) (8) (4) 9.4 8.7 8.8 9.2 8.6 8.6 8.9 [0.621 [0.45] 8.8 [0.59] 7.6 9.8 7.2 7.7 9.7 8.0 9.8 11.2 9.4 9.5 SY-F (1) (3) 8.1 9.8 8.5 11.3 ASY-F (1) 10.8 ASY-M (2) (17) (1) (1) 8.8 8.8 8.7 9.3 8.5 [0.46] 9.0 7.9 10.0 TY-F (1) 8.9 ATY-F (3) (2) (1) 8.2 8.8 9.8 7.7 8.6 8.9 9.0 ATY-M (13) (1) (2) (1) 8.6 9.1 9.4 8.9 [0.43] 9.1 7.7 9.6 9.4 U-U (1) 8.4 “Brewster's” Warbler (hybrid) Age/Sex Jan. Feb. Mar. HY-U Vennivora chrysoptera x V. pinus Apr. May June July Aug. (1) 7.7 Sept. Oct. Nov. 290 Dec. = 6 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 41 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept AHY-U (4) (I) 8.6 8.8 7.8 9.2 Oct. Nov. “Lawrence's” Warbler (hybrid) Age/Sex Jan. Feb. Mar. HY-M AHY-U TY-M Vennivora vhrysoptera x V. pinii.s r Apr. May June July Aug. Sept. Oct. Nov. (1) 7.8 (I) 9.1 (1) 9.5 (1) 8.1 Blue-winged Warbler Age/Sex Jan. Feb. Mar. HY-M HY-U AHY-F AHY-M U-U Tennessee Warbler Age/Sex Jan. Feb. Mar. HY-F HY-M Apr. Apr. Vennivora pinus May June July Aug. Sept. (1) (4) 7.4 9.2 8.1 9.8 (1) (6) 8.1 8.2 [0.99] 7.3 9.9 (5) (1) (2) 8.3 8.0 8.8 [0.41] 8.5 8.0 9.0 9.0 (8) (2) (5) 8.2 8.0 9.3 [0.84] 7.9 [0.94] 7.2 8.0 8.7 10.0 (1) 11.0 7.8 Venni\ ’ora peregrina May June July Aug. Sept. (9) 8.7 [0.65] 8.1 9.6 (8) 9.4 [0.48] 8.6 10.1 (36) (1588) (117) 8.8 9.0 9.7 [0.62] [0.71] [1.10] 7.7 7.3 7.6 10.3 15.3 14.8 n Nov. n = Nov. Dec. = 4 Dec. = 36 Dec. 2523 Dec. HY-U 42 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. AHY-F AHY-M AHY-U U-U Orange-crowned Warbler Age/Sex Jan. Feb. HY-U AHY-F AHY-M AHY-U U-U Nashville Warbler Age/Sex Jan, Feb. HY-F Mar. Mar. Mar. Apr. Apr. (1) 9.8 Apr. May June July Aug. Sept. Oct. Nov. Dec. (164) (4) 9.8 9.1 [1.09] 8.2 7.8 9.9 13.4 (209) (1) (4) 10.2 9.9 9.2 [1.40] 8.7 7.3 9.8 18.4 (100) (1) (9) (92) (3) 10.2 9.0 9.4 9.6 10.8 [1.33] [0.76] [1.02] 9.3 6.9 8,5 8.0 12.0 13.9 11.0 13.6 (2) (160) (16) 8.4 9.1 10.7 7.9 [0.63] [1.19] 8.8 7.7 8.4 10.9 12.6 Vennivora celatci n = 108 May June July Aug. Sept. Oct. Nov. Dec. (4) (63) (2) 8.4 9.6 10.9 7.9 [0.81] 9,7 8.7 7.9 12.1 11.7 (1) 8.8 (1) 10.5 (12) (20) (1) 9.4 9.2 9.4 [0.47] [0.70] 8.6 8.1 9.9 10.9 (3) 9.4 9.0 10.0 Vennivora ruficapilla n = 920 May June July Aug. Sept. Oct. Nov. Dec (2) (136) (43) 8.0 8.5 8.8 7.9 [0.86] [0.76] 8.0 7.1 7.3 11.7 10.4 (1) (87) (22) (4) 7.1 8.7 9.3 8.2 [0.79] [0.95] 7.2 7.5 7.7 9.0 10.6 12.2 HY-M 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 43 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (2) (35) (6) 8.4 8.6 9.1 8.2 [0.84] [0.96] 8.5 7.3 7.5 10.6 10.1 AHY-F (4) (158) (I) (57) (37) 7.8 8.5 9.5 8.5 8.9 6.8 [0.72] [0.72] [0.72] 8.3 6.7 7.0 7.5 11. 1 10.2 10.2 AHY -M (11) (135) (29) (20) 8.2 9.0 9.0 8.9 [0.74] [0.98] [1.01] [0.89] 7.2 7.2 7.4 7.0 9.6 13.9 12 2 1 1.2 AHY-U (1) (60) (12) (1) 8.7 8.6 9.0 8.0 [0.71] [0.97] 7.1 7.9 10.2 10.5 U-M (1) (2) 8.3 8.4 7.9 8.8 U-U (5) (39) (9) 7.9 8.6 9.1 [0.85] [0.70] [0.97] 7.4 7.4 7.3 9.4 9.9 10.2 Northern Parula Panda americana Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-M (3) 8.3 • 7.2 10.0 HY-U (1) (1) (3) 7.3 7.7 7.4 7.0 7.7 AHY-F (3) (1) (1) 7.7 7.4 7.2 7.1 8.5 AHY-M (8) (1) 8.1 8.4 [0.87] 7.0 9.6 Dec. 23 Dec. U-U (1) 9J 44 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. -S Yellow Warbler Deiulroica petechia n = 376 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) 12.6 HY-U (15) 9.3 [0.80] 8.0 10.7 (11) 10.9 [1.60] 9.2 13.8 (10) 10. 1 [1.02] 8.4 • 11.5 (1) 10.0 AHY-F (3) 9.5 9.1 10.0 (110) 9.1 [0.74] 7.4 12.0 (1) 9.2 (5) 9.4 [0.63] 8.8 10.4 (4) 10.5 10.0 11.2 (1) 16.0 AHY-M (7) 10.2 11.09] 9.3 12.5 (131) 9.8 [0.79] 7.9 12.4 (15) 9.3 [0.60] 8.1 10.2 (6) 10.0 [0.83] 9.0 10.9 (2) 11.8 10.8 12.8 (3) 10.6 10.1 11.0 AHY-U (2) 11.4 10.9 11.8 ASY-F (3) 8.9 8.7 9.0 ASY-M (1) 10.3 (11) 9.5 [0.52] 8.6 10.3 (1) 9.8 ATY-F (1) 9.8 (9) 9.4 [0.75] 8.5 10.7 (1) 9.2 (1) II.O • ATY-M (1) 10.4 (8) 9.9 [0.46] 8.6 12.0 U-U (I) 8.9 (8) 11.0 [1.55] 8.7 13.5 (2) 9.2 9.2 9.2 Magnolia Warbler Dendroica magnolia n = 2561 Age/Sex HY-U Jan. Feb. Mar. Apr. May June July Aug. (85) 8.1 [0.54] 6.0 9.7 Sept. (1214) 8.2 [0.70] 6.6 12.9 Oct. (50) 8.8 [1.06] 6.9 11.5 Nov. Dec 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 45 Age/Sex Jan. AHY-F AHY-M AHY-U SY-M SY-U ASY-M ATY-M U-U Cape May Warbler Age/Sex Jan. HY-F HY-M HY-U Feb. Mar. Feb. Mar. Apr. May June July Aug. Sept. Oct. (1) (238) (1) (2) (165) (23) 7.6 8.5 8.3 8.0 8.4 9.5 [0.81] 7.5 [0.76] [1.30] 6.6 8.4 6.7 7.7 12.6 12.4 11.8 (390) (2) (2) (97) (16) 8.9 9.0 9.0 8.6 9.6 [0.81] 8.6 7.9 [0.62] ]1.06] 7.0 9.4 10. 1 7.2 8.3 12.9 9.9 12.7 (30) (3) (42) (1) 8.2 7.9 8.8 9.4 ]().82] 7.5 [0.83] 6.6 8.2 7.1 10.0 11.2 (ID 8J [0.91] 7.4 10.5 (I) 8J (12) 8.9 [0.91] 7.7 10.6 (1) 9.2 (14) (155) (5) 8.1 8.3 9.1 [0.47] [0.64] [0.87] 7.3 6.9 8.3 8.8 10.9 10.5 Dendroica figrina May June July Aug. Sept. Oct. (2) (280) (13) 9.3 lO.O 10.9 8.9 [0.98] [2.01] 9.7 8.2 8.9 16.5 15.1 (1) (280) (17) 10.4 10. 1 11.3 [0.89] [2.29] 8.5 8.8 16.1 17.3 (1) (51) (7) 9.7 10.0 11.7 [0.92] [2.10] 8.1 8.9 13.1 13.9 (10) (1) (27) (2) I LI 10.5 10.4 12.9 [1.34] [0.85] 10.5 9.7 9.3 15.3 14.0 13.4 Nov. Dec. n = 798 Nov. Dec. AHY-F 46 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. AHY-M (11) (6) (42) (2) 11.2 II.O 11.2 13.2 [1.061 [0.91] [1.79] 11.1 9.9 10.0 9.3 15.2 13.1 12.4 17.3 AHY-U (1) (33) (4) 9.4 10.6 11.0 [1.53] 8.3 8.8 9.6 15.3 U-F (1) 9.6 U-U (5) (1) 10.5 13.5 [0.73] 9.5 11.4 Black-throated Blue Warbler Denciroica caerulescens n = Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. HY-F (3) (27) (7) 8.7 9.2 10.5 8.5 [0.51] [1.39] 8.8 8.2 8.8 10.2 12.7 HY-M (1) (31) (3) 10.0 9.8 10.1 [0.67] 9.5 8.7 10.8 11.8 HY-U (1) 9.3 AHY-F (24) (1) (2) 9.6 9.1 11.0 [0.84] 9.9 8.1 12.1 11.7 AHY-M (9) (1) (4) (2) (1) 9.9 9.4 9.4 10.3 9.9 [0.76] 8.7 10.0 8.8 10.0 10.6 11.0 U-F (10) 9.0 [0.43] 8.5 10.0 U-M (1) (1) 10.5 10.2 Yellow-rumped Warbler (Myrtle Warbler) Denciroica corona ta n = Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. HY-F (15) (323) (7) 11.0 12.0 12.4 [0.80] [1.07] [0.92] 9.7 9.6 11.4 12.4 16.5 13.6 Dec. 129 Dec. 2095 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 47 Age/Sex HY-M Jan. Feh. Mar. Apr. May June July Aug. Sept. (3) II. 6 9.7 12.8 Oct. (193) 12.8 [1.17] 10.6 19.7 Nov. (20) 13.5 [1.12] 11.6 15.7 HY-U (28) II. 4 [0.69] 10.6 13.6 (783) 12.5 [1.25] 9.5 16.6 (24) 13.5 [1.24] 10.9 15.6 AHY-F (5) 12.3 [1.72] 10.3 14.7 (60) II. 8 [0.95] 10.1 14.5 (5) II.5 [0.43] 10.9 12.0 (213) 12.3 [1.07] 9.9 15.3 (5) 12.9 [1.41] 11.5 14.5 AHY-M (18) 13.4 [1.30] 11.8 16.6 (45) 12.4 [0.90] 10.8 15.1 (152) 1 3. 1 [1.20] 10.8 16.7 (1) 14.3 AHY-U (1) II.O (9) 12.6 [1.03] 1 1.0 14.3 (1) 1 2.1 (29) 13.0 [I.IO] 10.8 14.7 (1) 13.4 SY-M (5) 11.3 [0.69] 10,6 12 2 ASY-M (9) 12.2 [0.99] 10.8 13.5 U-F (1) 12.3 (1) 14.5 U-M (1) 12.9 u-u (6) 11.0 [0.87] 10.3 12.6 (121) 12.0 [1.03] 9.7 15.6 (10) 13.3 [0.95] 11.6 14.8 Black-throated Green Warbler Dendroica virens n = Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. HY-F (4) 8.2 7.8 8.9 (18) 9.0 10.51] 8.0 9.9 HY-M 48 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. HY-U AHY-F AHY-M AHY-U SY-F U-F U-M u-u Cerulean Warbler Age/Sex Jan. HY-U AHY-F AHY-M AHY-U Feb. Mar. Apr. (2) 8.9 8.7 9.0 Feb. Mar. Apr. May June July Aug. Sept. Oct. (2) (22) (357) (38) 8.7 8.3 8.6 9.0 8.2 [0.44] [0.64] [0.72] 9.1 7.7 7.1 7.5 9.5 11.2 11.2 (5) (1) (10) (1) 8.8 9.3 8.9 9.9 [0.58] [0.76] 8.1 7.7 9.5 10.1 (11) (1) (2) (25) (2) 10.0 8.6 8.2 9.0 9.8 [1.66] 7.9 10.74] 9.2 8.6 8.5 8,0 10.3 14.4 11.4 (2) (3) 8.1 9.0 7.9 9.0 8.2 9.0 (1) 7.9 (1) 9.2 (6) (1) 8.4 9.0 [0.24] 8.1 8.8 (5) (40) (3) 8.5 8.5 9.1 [0.30] [0.53] 8.6 8.1 7.4 9.0 8.8 9.8 Dendroica ceridea May June July Aug. Sept. Oct. (12) (4) (2) 9.0 8.8 8.7 [0.73] 8.2 8.4 8.0 9.5 9.0 10.2 (1) (1) (2) 9.1 9.2 9.0 8.8 9.1 (2) (3) (4) (1) 9.4 9.5 9.5 9.2 9,1 9.0 9.3 9.7 10.0 9.7 Nov. (1) 9.0 (3) 8.7 8.1 9.2 n Nov. Dec. 36 Dec. U-U 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 49 Blackburnian Warbler Age/Sex Jan. HY-F HY-M HY-U AHY-F AHY-M AHY-U U-U Chestnut-sided Warbler Age/Sex Jan. HY-M HY-U AHY-F Feb. Mar. Apr. Dendroica fiiscii May June July (2) 10.2 10.1 10.2 (3) 10.7 10.0 11.4 (I) 9.3 Aug. (6) 9./ [0.631 8.3 9.8 (4) 9.6 9.2 10.3 (7) 9./ [0.55] 8.3 10.0 (1) 9.6 (3) 8.9 8.6 9.3 Sept. (20) 9.0 [0.65] 8.1 10.5 (30) 9.7 [0.91! 8.0 12.7 (7) 9.2 [0.86] 7.7 10.0 (I) w.o (2) 10.4 9.8 10.9 (1) 9.4 (2) 9.2 9.2 9.3 Oct. n Nov. Dendroica pensylvanica Feb. Mar. Apr. May June July Aug. Sept. Oct. (1) (3) (26) 9.4 9.7 9.5 9.0 [0.74] 11.0 8.0 11.6 (59) (147) (2) 9.1 9.4 9.6 [0.58] [0.73] 8.6 8.2 7.8 10.6 10.4 1 1.6 (26) (2) (1) (2) (14) (1) 9.5 9.7 9.4 9.1 9.3 9.1 [0.72] 9.0 8.7 [0.70] 7.5 10.3 9.5 7.7 10.9 10.5 (45) (1) (1) (2) (17) 9.9 10. 1 lO.O 9.9 9.7 [0.95] 9.8 [0.59] 8.1 10.0 8.9 13.1 11.2 n = Nov. = 90 Dec. 375 Dec. AHY-M 50 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. AHY-U U-U Bay-breasted Warbler Age/Sex Jan. Feb. HY-M HY-U AHY-F AHY-M AHY-U U-U Mar. Mar. Apr. Apr. May June July Aug. Sept. Oct. (3) (6) (4) (1) 9.6 9.7 10.2 9.2 8.5 [0.871 9.0 10.4 8.4 12.1 10.7 (6) (5) 9.1 10.0 [0.53] [0.951 8.7 9.0 10.1 11.6 Dendroica castanea May June July Aug. Sept. Oct. (3) 12.3 II. 1 13.7 (2) (223) (2) 11.5 11.6 II.9 1 1.3 [0.76] 11.5 12.0 9.7 12.3 14.0 (6) (8) 11.9 12.0 [1.021 [0.88] 10.7 10.8 13.6 13.1 (11) (4) (1) U.4 12.0 13.7 [1.171 11.6 12.0 12.3 15.1 (15) 11.9 [0.86] 10.2 13.2 (3) (1) 12.2 14. 1 10.5 12.8 Blackpoll Warbler Dendroica striata Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (192) (83) 11.8 13.5 [0.86] [2.06] 9.8 10.4 14.5 20.8 AHY-F (6) (2) (6) 12.3 14.0 12.3 [0.80] 11.5 [2.38] 11.2 16.4 9.8 13.6 16.8 Dec. 279 Dec. 484 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 51 Age/Sex AHY-M Jan. Feb. Mar. Apr. May (22) 13.0 [1.14] 10.8 15.9 June July Aug. Sept. (2) 13.3 12.3 14.2 Oct. Nov. Dec. AHY-U (53) 12.2 [1.44] 9.7 19.0 (29) 15.5 [2.91] 12.1 21.6 U-U (1) 11.6 (70) 12.2 [1.17] 10.7 18.1 (18) 15.1 [2.96] 12.0 20.9 Pine Warbler Dendroica pinus n = 3 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (1) 11.7 (1) 11.1 (1) 14.2 Kirtland’s Warbler Dendroic a kirllandii n = 3 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-U (2) (I) M.5 15.8 14.0 14.9 (n.b., all 3 weights are from the same individual) Prairie Warbler Age/Sex Jan. Feb. Mar. Apr. HY-F HY-M HY-U AHY-F (2) 6.9 6.8 7.0 (6) 7.2 [0.50] 6.7 7.9 Dendroica discolor May June July Aug. Sept (2) (2) 6.7 7.7 6.5 7.2 6.8 8.2 (1) (3) 7.0 7.6 6.9 8.2 (1) (2) 7.8 6.8 6.6 7.0 (32) (1) (2) (2) (1) 7.0 8.0 7.1 7.2 7.8 [0.48] 7.1 6.9 6.2 7.1 7.5 8.1 (17) (3) (1) (1) 7.2 7.2 8.7 8.7 [0.42] 6.6 6.2 8.0 7.8 n = 83 Oct. Nov. Dec. (1) 8.3 AHY-M 52 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex AHY-U Jan. Feb, Mar. Apr. (2) 7.2 7.0 7.4 May (1) 6.7 June July Aug. Sept. Oct. Nov. Palm Warbler Dendroica palmci rum n = Age/Sex HY-U Jan. Feb. Mar. Apr. May June July Aug. Sept. (233) 10.2 [0.81] 7.9 12.2 Oct. (156) 10.6 [0.97] 8.3 13.2 Nov. (3) 1 1.5 10.2 13.0 AHY-M (2) 10.3 9.8 10.8 (2) 10.9 10.3 11.5 AHY-U (6) 10.2 [0.38] 9.6 10.7 (5) 10.2 [0.70] 9.4 II. 1 (2) 10.5 10.4 10.6 (30) 10.7 [0.92] 9.3 12.9 U-U (91) 10.0 [0.92] 7.0 12.3 (38) 10.5 [0.72] 9.2 11.8 Ovenbird Age/Sex Jan. HY M HY-U AHY-F AHY-M AHY-U Feb. Mar. Apr. Seiurus aurocupillus May June July (9) 19.1 [1.40] 17.3 21. .3 Aug. (2) 20.0 19.6 20.3 Sept. Oct. (48) (193) (16) 18.9 19.4 20.2 [1.60] [1.98] [3.12] 17.0 16.4 17.5 26.3 28.2 28.3 (5) 19.6 [1.53] 18.1 21.7 (3) (9) (1) 18.6 18.8 20.5 18.1 [0.85] 19.3 17.4 20.3 (2) (34) (1) (10) (29) (4) 18.0 18.8 19.4 18.9 19.0 22.4 17.7 [1.87] [1.34] [1.40] 20.3 18.2 15.6 16.5 15.7 26.2 23.4 20.5 22.0 n = Nov. SY-M (I) 18.3 Dec. 568 Dec. 449 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 53 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. ASY-U (2) (1) 19.1 28.8 19.0 19.1 u-u (1) (12) (62) (4) 18.9 18.6 20.0 20.3 [1.42] [2.21] 19.4 16.9 14.0 20.9 21.0 26.5 Northern Waterthrush Seiiirns novehonicensis Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (34) (109) (26) 17.9 18.1 18.9 [1.63] [1.81] [2.50] 15.1 14.3 15.3 22.0 23.9 23.1 AHY-U (16) (169) (32) (26) (4) 16.7 17.6 18.6 18.4 19.0 [1.67] [1.67] [1.31] [1.66] 17.8 13.8 14.2 16.4 15.4 20.2 20.5 24.4 21.6 23.2 U-U (10) (31) (1) 17.9 18. 1 16.2 [0.94] [1.90] 16.5 13.9 19.5 23.3 Louisiana Waterthrush Seiidns motacilla Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-U (2) (12) (9) (1) 20.9 20.0 20.2 17.2 20.6 [2.03] [1.71] 21.1 14.5 18.1 22 1 24.0 AHY-F (5) (2) (2) 21.5 20.2 20.4 [2.83] 20.1 19.4 18.2 20.2 21.4 26.0 AHY-M (15) (1) 19.7 22.7 [1.31] 17.4 22 1 AHY-U (7) (29) (2) 19.0 19.6 22.5 [0.91] [1.23] 21.7 17.3 17.0 23.2 20.1 22.4 ASY-F (I) 21.3 (7) 19.3 [0.86] 18.3 20.6 Nov. n = Nov. n = Nov. Dec. 458 Dec. 103 Dec. ASY-M 54 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex ASY-U ATY-M U-U Jan. Feb. Mar. Apr. (2) 20.7 20.3 21.1 May June July (3) 18.5 18.3 18.7 Aug. Sept. Oct. Nov. (1) 18.7 (2) 19.7 19.3 20.1 Kentucky Warbler Oporornis formosiis Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. HY-F (3) (4) 13.4 13.4 13.3 12.3 13.5 14,3 HY-M (2) (14) (5) 14.4 14.1 14.1 13.8 [1.08] [0.50] 14.9 12.3 13.4 16.7 14.7 HY-U (4) (15) (3) 14.0 14.2 14.0 12.9 [0.86] 12.1 15.2 13.0 15.9 16.7 AHY-F (19) (4) (7) (8) (1) 1 3.1 14.0 13.6 15.0 14.7 [1.18] 12.8 [0.92] [0.65] 11.4 15.8 12.6 14.1 16.5 15.3 16.1 AHY-M (3) (49) (14) (3) (13) (2) 14.2 14.0 14. 1 14.8 14.9 18.5 13.9 [0.67] [0.71] 13.8 [1.36] 16.3 14.5 12.0 12.9 16.3 13.8 20.6 15.5 15.5 18.3 AHY-U (1) (1) (2) 12.4 13.8 14.9 13.3 16.4 ASY-F (1) (1) (1) (1) 13.3 12.7 13.2 14.1 ASY-M (1) (1) (3) 12.2 17.9 14.6 14.3 15.0 ASY-U (1) 14.5 ATY-M (3) (1) (1) 13.7 15.5 15.4 13.2 14.2 Dec. 206 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 55 Age/Sex Jan. U-M u-u Connecticut Warbler Age/Sex Jan. HY-U AHY-F AHY -M AHY-U U-U Mourning Warbler Age/Sex Jan. HY-F HY-M Feb. Mar. Apr. May June July Aug. Sept. (2) 14.0 13.4 14.6 (11) (1) 13.5 14.8 [0.57] 12.7 14.4 Oct. Nov. Oporornis agilis n Feb. Mar. Feb. Mar. Apr. Apr. May June July Aug. Sept. Oct. (4) (107) (12) 13.8 14.0 14.9 13.2 [1.32] [2.11] 14.5 11.5 12.3 19.7 19.7 (4) (1) (17) (3) 13.3 12.7 14.4 13.7 10.4 [2.14] 12.1 19.1 12.4 15.5 21.2 (4) (28) (4) 16.5 14.9 15.9 16.2 [2.34] 12.3 16.8 10.7 22.0 20.2 (8) 14.6 [3.31] 12.1 22.2 (23) (2) 14.3 17.7 [1.93] 13.6 12.5 21.7 21.0 Oporornis Philadelphia May June July Aug. Sept. Oct. (1) (2) 11.3 13.8 12.5 15.0 (3) (2) 1 1.5 15.2 11.1 13.8 11.8 16.5 (21) (50) (7) II.9 12.5 12.9 [1.00] [1.10] [1.20] 10.6 10.4 11.1 14.2 15.0 14.5 Nov. n = Nov. Dec. 217 Dec. 319 Dec. HY-U 56 Age/Sex Jan. AHY-F AHY-M AHY-U U-M Common Yellowthroat Age/Sex Jan. HY-F HY-M HY-U AHY-F AHY-M AHY-U BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Feb. Mar. Apr. May June July (45) (9) (1) II. 9 12.5 12.2 [1.02] [1.00] 1 0.0 11.0 14.7 14.0 (89) (5) 12.9 13.2 [1.02] [0.75] 10.8 12.5 15.6 14.3 (4) 123 11.0 14.0 Aug. Sept. Oct. (5) (24) (5) 12.5 12.0 12.2 [0.56] [1.00] [1.70] 11.8 10.5 10.0 13.3 13.8 14.0 (10) (30) (5) 1 2.6 1 3. 1 13.4 [1.36] [1.54] [1.10] 9.6 10.3 12.2 13.9 17.9 14.6 (I) 13.2 Nov. Dec. Geoflilypis trie has n = 3712 Apr. May June July Aug. Sept. Oct. Nov. 1 (3) (80) (488) (119) (2) 9.2 9.3 9.7 10.5 10.6 8.8 [0.62] [1.03] [1.23] 10.5 9.5 8.0 7.9 8.2 10.7 11.0 14.3 13.7 (11) (112) (695) (156) (1) 9.4 9.9 10.2 II. 0 10.8 [0.89] [0.68] [1.08] [1.33] 8.3 8.7 8.0 8.7 1 1.4 12.5 13.7 14.7 (66) (137) (61) (10) 9.1 9.7 lO.O 10.8 [0.52] [0.75] [1.10] [1..57] 8.1 8.1 7.4 8.6 10.7 12.1 12.7 13.5 (194) (14) (25) (34) (178) (50) (1) 9.5 9.9 9.6 9.6 lO.O tl.O 9.2 [1.02] [0.76] [0.67] [0.70] [1.11] [1.06] 7.6 8.9 8.1 8.1 7.8 8.6 13.0 11.5 10.7 11.7 15.3 12.8 (17) (325) (25) (27) (20) (276) (33) 9.4 10. 1 9.8 9.6 10.3 10.8 1 2. 1 [0.44] [0.99] [0.55] [0.36] [0.70] [1.17] [1.63] 8.6 7.6 8.9 9.1 9.1 8.6 9.5 10.4 13.3 10.6 10.4 11.7 15.2 15.5 (2) (5) 9.9 lO.O 9.0 [0.63] 10.8 9.1 10.6 (2) (4) (2) (2) 8.7 9.2 10.7 10. 1 8.6 8.9 10.2 9.7 8.7 9.9 11.1 10.4 SY-F 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 57 Age/Sex Jan. SY-M ASY-F ASY-M TY-F TY-M ATY-F ATY-M U-F U-M U-U Yellow-breasted Chat Age/Sex Jan. HY-F Feb. Feb. Mar. Mar. Apr. (I) W.7 Apr. May June July Aug. Sept. Oct. (39) (10) (3) (2) (7) 9.7 9.6 9.7 9.4 10. ! [0.641 [0.62] 9.0 8.8 [0.50] 8.4 8.6 10.3 9.9 9.1 11.7 10.4 10.6 (16) (2) (2) (3) (5) (2) 9.4 12.3 10.7 9.5 9.4 9.6 [0.95] 11.6 10.6 9.2 [0.45] 8.5 8.1 12.9 10.7 9.7 8.7 10.6 11.6 9.8 (80) (6) (1) (3) (3) (2) 10.3 9.9 8.8 9.8 10.3 10.7 [1.09] [0.50] 9.2 9.6 10.2 8.6 9.2 10.7 11.0 11.2 13.4 10.4 (4) (2) 8.9 9.5 7.6 9.0 10.7 9.9 (15) (1) (2) (2) (1) 9.6 9.7 9.7 10.7 12.3 [0.57] 9.7 10.5 8.6 9.7 10.8 10.5 (9) (1) (2) (2) (6) (1) 8.8 10.6 8.6 9.0 9.5 10.3 [0.47] 1.1 8.9 [0.93] 8.1 9.5 9.0 8.3 9.6 10.9 (39) (7) (1) (6) (2) 9.4 9.7 9.6 lO.O 10. 1 [0.56] [0.63] [0.33] 9.5 8.5 8.9 9.6 10.7 11.1 10.9 10.4 (4) (71) (5) 8.8 10.3 10.4 8.3 [1.02] [0.98] 9.5 8.0 9.0 12.6 11.7 (1) (7) (I) 9.2 10. 1 12.6 [0.16] 9.8 10.3 (15) (126) (14) 9.4 9.9 10.4 [0.57] [0.97] [1.40] 8.4 8.3 9.1 10.6 12.8 13.7 Icteria virens May June July Aug. Sept. Oct. (7) (6) 28.1 27.3 [1.87] [2.79] 25.6 23.2 30.6 30.1 Nov. (1) lO.I n = Nov. Dec. 564 Dec. 58 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-M (2) (3) 24.2 25.1 23.0 22.9 25.4 26.6 HY-U (12) (43) (36) (5) 24.4 26.0 28.3 29.5 [1.67] [2.62] [2.25] [2.11] 22 2 22.4 24.3 26.7 27.8 34.1 32.8 31.8 AHY-F (1) (!01) (11) (21) (15) (10) (!) 25.5 24.0 26.1 25.6 27.4 29.4 33.8 [2.24] [1.44] [1.57] [2.50] [2.29] 20.2 24.6 23.5 24.0 26.3 32.0 29.5 29.3 33.6 33.1 AHY M (2) (166) (22) (12) (11) (9) 25.7 25.0 25.2 26.0 27.8 28.3 25.6 [1.89] [1.44] [1.24] [2.24] [1.96] 25.7 20.3 22.4 24.5 23.9 25.4 30.4 28.2 28.3 30.0 31.7 AHY-U (1) (9) (5) (7) (1) (1) 22.8 24.1 26.2 27.9 28.8 32.5 [1.23] [2.53] [2.47] 21.5 23.5 24.2 25.8 29.2 31.7 ASY-F (4) (1) 24.7 24.8 23.4 26.1 ASY-M (4) (1) ■ (2) (1) 25.0 23.6 29.6 29.5 24 2 29.5 21. \ 29.6 TY M (1) 25.6 ATY-F (2) 24.2 23.5 24.8 ATY-M (14) (1) (2) 26.1 26.7 27.0 [2.12] 23.2 22.0 30.8 29.3 ATY-U (1) 24.0 U-F (5) (1) 27.0 27.7 [1.12] 25.5 28.5 U-U (2) (1) (1) 27.9 30.4 25.6 27.4 28.3 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 59 Hooded Warbler Wilsoiiia citriiui n = 370 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-F (36) (64) (6) 9.8 9.8 11.6 [0.54] [0.69] [1.85] 8.8 8.1 10.0 11.2 11.6 14.9 HY-M (11) (43) (5) 10.8 10.3 10.9 [0.63] [0.65] [0.60] 9.9 7.4 10.3 12.0 11.8 11.8 AHY-F (36) (5) (2) (2) (7) (1) 9.8 II.O II.O 10.6 10.2 II.2 [0.55] [0.67] 10.5 10.1 [0.43] 8.6 10.5 11.4 II. 1 9.3 11.2 12 2 10.5 AHY-M (18) (3) (3) (4) (2) 10.8 10.7 11.3 lO.O 1 1.2 [0.62] 10.0 10.8 9.7 9.9 9.4 11.3 11.7 10.4 12.4 11.8 ASY-M (I) 1 1.0 ATY-F (1) (1) (1) 9.7 12.1 12.5 ATY-U (1) 10.9 U-F (28) (31) 10.0 10.2 [0.76] [0.69] 8.7 8.2 12.3 11.8 U-M (25) (30) (3) 10.5 II.O II.7 [0.84] [0.82] 10.0 8.1 9.9 13.9 12.3 12.9 Wilson's Warbler Wilsoniu piisilla Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-F (20) (68) (1) 7.5 7.7 8.8 [0.64] [0.78] 6.5 6.0 9.3 11.9 HY-M (17) (74) (3) 7.8 8.0 7.6 [0.66] [0.73] 7.3 6.5 6.7 8.3 9.0 10.9 (7) (I) 7.8 6.8 [0.45] 7.3 8.7 HY-U 60 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex .Ian. AHY-F AHY-M AHY-U U-F U-M u-u Canada Warbler Age/Sex Jan. HY-F HY-M HY-U AHY-F AHY-M Feb. Mar. Apr. May June July Aug. Sept. Oct. (44) (1) (1) (25) 7.4 9.1 7.5 7.9 [0.55[ [0.84] 6.3 6.5 8.8 9.5 (297) (3) (8) (112) (3) 7.6 8.1 7.5 8.0 8.5 [0.54] 7.3 [0.54] [0.72] 7.1 6.4 9.5 6.4 6.7 10.0 9.3 8.1 10.5 (26) (2) (10) 7.3 7.8 8.0 [0.47] 6.9 [1.01] 6.5 8.7 6.6 8.4 9.1 (6) 7.6 [0.69] 7.1 8.9 (1) (I) 7.7 7.6 (1) (14) 7.0 7.4 [0.57] 6.6 8.2 Feb. Mar. Apr. (1) 10.6 WHsonia canadensis May June July Aug. Sept. (5) (1) 10.0 10.3 [0.86] 8.8 11.2 (1) (1) ll.l 9.2 (1) (101) (90) 9.5 10.2 10.3 [0.84] [0.83] 8.7 8.4 12.7 14.7 (226) (4) (1) (15) (16) 10.2 10.9 9.3 10.6 10.2 [0.73] 10.1 [0.63] [0.61] 8.1 11.8 9.5 9.1 12.6 11.6 10.8 (289) (7) (13) 10.6 II.2 10.6 [0.80] [0.50] [0.79] 8.7 10.5 9.5 13.5 11.7 11.9 (5) (-5) (1) 9.5 10.0 10.2 [0.36] [0.70] 9.2 9.2 10.0 11.1 Dec. 825 Dec. AHY-U 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 61 Age/Sex .Ian. Feh. U-U American Redstart Age/Sex Jan. Feh. HY-F HY-M HY-U AHY-F AHY-M AHY-U SY-F SY-M ASY-F ASY-M Mar. Mar. May June July Aug. Sept. Oct. (34) (7) (1) 10. 1 10.2 1 1.0 [0.831 [0.74] 8.6 9.5 11.9 11.5 Setophaga mticilla May June July Aug. Sept. Oct. (5) (42) (203) (1) 8.0 7.8 8.0 7.5 [0.33] [0.371 [0.67] 7.6 7.2 6.5 8.5 9.2 10.3 (5) (26) (138) (6) 8.4 8.1 8.1 9.1 [0.46] [0.641 [0.67] [1.72] 7.8 7.0 6.7 7.5 9.1 9.5 10. 1 12.2 (2) (15) (17) (78) (1) 9.7 8.7 8.2 7.8 7.7 9.5 [0.61] [0.50] [0.66] 9.8 7.3 7.3 6.4 9.6 9.0 10.0 (90) (6) (12) (18) (24) (2) 8.0 8.3 8.2 8.4 8.3 8.3 [0.67] [0.50] [0.46] [0.81] [0.93] 8.2 6.7 7.8 7.5 7.2 7.4 8.3 10. 1 9.2 9.1 10.7 11.2 (15) (2) (2) (9) (27) (1) 8.3 8.5 8.6 9.1 8.2 II. 6 [0.58] 8.1 8.6 [1.26] [0.78] 7.0 8.8 8.7 7.7 7.2 9.1 12.0 11.0 (3) (2) 7.7 9.2 7.3 9.0 8.2 9.4 (1) 7.8 (37) (7) (1) 8.4 8.4 9.1 [0.57] [0.28] 7.2 8.0 9.7 8.7 (5) (2) (1) 8.1 8.6 10.4 [1.20] 7.5 7.2 9.6 10.2 (1) (22) (3) (2) (2) (7) 8.2 8.4 8.7 9.3 8.9 8.2 [0.78] 8.2 8.9 8.6 [0.34] 7.3 9.0 9.7 9.2 7.8 9.9 8.7 (3) 7.5 7.2 8.1 Nov. n = Nov. Dec. 919 Dec. ATY-F 62 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec ATY-M (3) (1) (1) 8.8 9.4 9.6 8.1 9.2 U-F (3) (1) (2) 8.3 8.1 7.3 7.6 7.1 9.2 7.4 U-U (1) (18) (41) (2) 8.0 7.8 7.7 10.1 [0.54] [0.44] 8.6 6.8 6.8 11.6 8.6 8.7 House Sparrow Passer domesticus n = 1541 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) (3) (18) (6) (38) (54) (76) (9) 24.4 23.3 26.2 27.2 27.3 27.9 27.2 27.5 22.2 [1.64] [1.36] [1.70] [1.61] [1.94] [1.23] 24.1 21.9 26.1 23.5 24.6 21.0 25.4 29.1 29.3 30.7 33.5 31.4 29.0 HY-M (5) (5) (19) (2) (19) (63) (102) (6) 22.6 26.6 28.2 25.2 27.7 28.3 28.2 27.8 [2.61] [2.26] [1.84] 24.3 [1.74] [1.58] [1.96] [1.53] 20.3 25.0 25.2 26.0 24.3 24.1 24.3 25.4 26.5 30.5 31.9 30.6 32.7 33.6 29.9 HY-U (11) (44) (46) (1) (3) 26.0 25.8 26.4 24.1 26.9 [1.85] [1.86] [1.98] 26.7 23.1 22.0 21.6 27.0 29.0 29.8 29.5 AHY-F (41) (25) (94) (81) (19) (2) (2) (6) (6) (2) 26.3 27.0 26.9 28.2 28.2 28.9 27.1 28.4 26.4 28.2 [1.52] [1.48] [1.65] [2.21] [2.34] 28.4 23.2 [1.12] [1.95] 27.9 22 2 23.4 20.1 22.7 24.6 29.3 31.0 26.5 23.6 28.4 29.3 29.4 29.7 33.7 32.5 29.3 29.5 AHY-M (44) (37) (109) (64) (14) (6) (3) (1) (1) (3) (14) (3) 27.9 28.0 28.0 28.6 27.6 28.0 28.7 26.8 26.0 27.9 27.8 26.6 [1.83] [1.35] [1.48] [2.00] [2.69] [0.80] 26.3 23.7 [1.72] 25.3 23.1 24.0 23.7 23.2 20.0 27.0 33.3 32.5 24.1 28.3 32.0 30.6 31.2 34.0 30.9 29.1 30.8 AHY-U (1) 27.2 SY-F (7) (6) (19) (7) (9) (4) (2) (1) 26.2 28.0 27.1 28.5 27.8 27.6 23.9 24.4 [1.21] [0.67] [2.07] [2.02] [1.96] 25.5 20.4 24.8 26.9 21.2 26.2 25.3 28.3 27.4 27.7 28.8 30.6 32.5 31.3 SY-M (16) (4) (30) (13) (5) (1) (1) 27.8 26.6 28.1 27.4 28.0 26.5 28.1 [1.14] 24.8 [1.46] [1.44] [0.85] 25.6 29.4 24.6 25.1 27.3 29.4 31.6 30.0 29.4 ASY-F (3) (9) (6) (4) (1) (4) (1) 27.0 26.8 28.3 28.6 27.2 27.2 27.8 26.0 [2.71] [0.76] 25.4 25.1 28.2 22.7 27 2 34.5 28.4 30.8 28.9 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 63 Age/Sex Jan. Feb. Mar. Apr. May June ASY-M (3) (4) (19) (14) (10) (1) 26.3 27.7 27.7 27.6 27.2 27.0 25.3 25.6 [1.28] [1.01] [2.15] 27.6 29.2 25.9 26.3 24.5 30.2 29.7 31.2 TY-F (1) (6) (3) (1) 28.0 27.2 29.6 31. 1 11.42] 27.4 25.6 32.0 29.2 TY-M (1) (1) (7) 28.9 25.4 27.7 [1.43] 25.0 29.4 ATY-F (2) (2) (5) (4) (2) (1) 26.6 27.2 27.6 28.4 26.4 27.5 26.2 25.9 [1.38] 25.9 25.1 26.9 28.4 25.6 31.2 27.6 29.0 ATY-M (2) (2) (11) (3) (5) (3) 26.7 27.2 28.2 27.6 29.7 28.5 26.4 27.0 [1.15] 25.5 [0.96] 27.5 27.0 27.3 26.4 30.2 28.3 29.4 30.0 30.8 July Aug. Sept. Oct. Nov. U-F U-M u-u Bobolink (1) HY-U, Sept.. 34.3. Eastern Meadowlark Age/Sex Jan. AHY-F AHY-U Feb. Mar. (I) 32.0 (2) 29.0 28.2 29.8 Doliclionv.x orvzivoriLS Stiirnella magna Apr. May June (I) (I) 86.1 78.5 (4) (1) 101.7 118.2 83.6 1 19.5 July Aug. Sept. Oct. (4) 29.0 26.8 30.4 Dec. (1) 28.7 (1) (1) (1) 29.1 27.0 27.5 (10) (57) (9) 27.6 27.2 28.3 [1.53] [1.80] [0.89] 25.2 22.6 27.1 30.0 30.3 29.8 (7) (58) (14) 27.7 28.2 28.6 [1.25] [1.70] 11.90] 2S 2 24.6 25.8 28.6 32.3 31.1 n = 7 Nov. Dec. 64 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Red-winged Blackbird Age/Sex Jan. Feb. HY-F HY-M HY-LI AHY-F AHY-M SY-M (1) 67.2 ASY-F ASY-M U-F U-U Orchard Oriole Age/Sex Jan. Feb. HY-F HY-U AgeUiiiis phoeniceiis n — 625 Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. (18) (28) (1) 44.2 45.4 47.3 [2.89] [3.69] 39.0 37.5 48.1 53.0 (6) (21) (6) 66.8 69.5 65./ [9.06] [4.19] [6.01] 55.5 63.1 53.6 80.3 77.7 68.8 (1) (3) (1) 33.0 44.5 47.3 41.1 47.5 (6) (62) (130) (9) (4) (3) (11) 44.2 41.4 40.7 40.8 37.2 42.0 45.4 [2.27] [3.32] [2.49] [2.44] 35.6 40.1 [3.30] 41.4 29.0 36.2 37.6 40.4 45.6 40.0 46.4 55.0 48.3 44.2 50.5 (2) (8) (6) (1) (15) 65.0 63.6 61.8 64.1 71.6 64.2 [6.43] [2.93] [4.03] 65.8 52.9 57.7 62.7 71.8 66.5 78.2 (13) (103) (69) (7) (7) 64.2 61.9 63.1 59.0 60.9 [4.73] [3.11] [3.38] [3.01] [2.53] 57.6 54.1 57.0 54.0 57.1 72.8 69.4 74.9 62.7 64.7 (3) (3) 41.1 41.5 40.2 39.8 42.6 42.4 (11) (32) (12) (3) 64.2 66.7 63.6 76.9 [3.28] [3.50] [2.66] 71.5 59.6 58.2 61.1 81.1 72.1 72.1 70.6 (8) (9) (1) 45.9 43.8 51.0 [4.08] [3.86] 37.1 41.0 48.4 52.4 (1) 44.6 Icterus spuriiis n = 15 Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. (1) 20.0 (3) (1) 22.6 25.6 22.0 23.1 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 65 Age/Sex Jan. AHY-F Feb. Mar. Apr. May (4) 20.2 18.8 21.9 June July (1) 19.5 Aug. Sept. Oct. Nov. Dec. SY-M (4) 22.4 20.6 24.1 (1) 23.5 Northern Oriole (Baltimore Oriole) Icterus galhitla n = 131 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (1) 54.5 HY -M (2) 55.8 34.9 36.7 HY-U (1) 32.0 (2) 29.0 28.4 29.6 (9) 53.0 [3.00] 30.0 38.9 AHY-F (49) 53.2 [2.87] 28.1 41.3 (1) 52.4 (1) 50.8 (1) 38.4 (1) 35.2 AHY-M (1) 55.0 (24) 35.4 [3.60] 22.3 41.5 (2) 57.0 35.1 38.8 (1) 40.0 (3) 57.5 36.6 38.3 SY-M (9) 55.8 [2.53] 31.3 36.2 (2) 54.0 32.4 35.7 (1) 29.2 ASY-F (5) 32.6 [4.70] 28.1 38.7 ASY-M (14) 55.2 [2.51] 31.1 40.0 U-F (1) 54.9 Rusty Blackbird Euphagiis carolinits n = 245 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (11) (9) 56.6 54.8 [4.61] [4.66] 47.3 47.1 64.0 61.3 66 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY Age/Sex HY-M HY-U AHY-F AHY-M AHY-U U-F U-M U-U Jan. Feh. Mar. Apr. May June July Aug. Sept. Oct. Nov. (6) (13) 63.8 65.1 [4.11] [3.06] 58.5 59.3 70.3 69.3 (1) 52.9 (17) (53) (4) (ID (7) 54.0 54.6 56.0 55.7 58.1 [3.31] [5.20] 54.7 [6.42] [2.59] 49.1 47.0 58.0 Al.l 54.1 59.7 76.5 70.4 61.0 (33) (35) (5) (7) 63.1 63.6 70.6 69.8 [4.16] [4.51] [3.59] [6.76] 45.9 52.4 64.9 60.6 69.3 76.8 74.5 80.4 (3) (1) (2) 54.5 61. 1 61.8 53.1 58.5 57.1 65.2 (5) (8) 57.6 57.1 [4.69] [4.24] 51.3 53.5 64.1 66.3 (7) (3) 61.0 68.7 [5.08] 66.5 55.1 71.2 69.1 (3) 57.5 54.3 63.7 Common Crackle Age/Sex Jan. HY-F HY-M HY-U Feb. Mar. Apr. Qitiscaliis quiscida n - May June July Aug. Sept. Oct. Nov. (14) (1) 93.4 86.1 [5.06] 83.2 101.0 (13) 117.3 [10.37] 90.4 136.4 (9) (6) (3) (1) 89.0 94.1 104.2 120.4 [8.39] [12.64] 84.5 75.4 78.2 132.1 102.0 112.4 NO. 5 Dec. 15 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 67 Age/Sex Jan. Feb. Mar. AHY-F (I) 87.8 AHY-M SY-M ASY-M Apr. May June July (8) (18) (11) (2) 101.7 98.1 91.9 97.7 [5.83] [11.53] [4.28] 95.5 95.4 82.1 85.6 99.9 1 13.9 124.7 99.1 (11) (76) (13) (3) 122.4 118.9 120.9 130.6 ;i2.88] [6.92] [9.86] 116.2 96.7 90.9 90.4 144.8 135.4 136.9 (1) 118.1 (4) 121.0 96.5 138.5 131.4 Aug. Sept. Oct. Nov. Dec (2) 96.2 94.9 97.6 (3) 118.2 107.5 126.0 U-F U-M u-u (ID 95.4 [5.061 89.1 103.5 (3) 1 18.6 1 17.1 1 19.5 (I) 94.5 (Note: the extreme variation in weights and the large standard deviations in the grackle sample may be the result of incorrectly aged or sexed birds.) Brown-headed Cowbird Molothrus ater n = 1476 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (1) (1) (1) (3) 32.6 35.5 37.5 40.5 38.1 44.4 HY-U (8) (8) 35.2 35.1 [4.30] [5.45] 31.2 30.4 42.2 45.4 AHY-F (7) (299) (275) (5) (3) (1) 40.4 37.8 39.7 39.0 38.6 34.5 [2.18] [2.39] [3.09] [2.13] 35.5 37.9 32.4 30.5 37.0 40.6 43.9 49.2 51.2 41.8 AHY-M (2) (38) (51) (1) (I) 53.0 48.4 48.8 44.5 47.6 52.1 [2.21] [4.37] 53.8 32.4 36.0 49.2 57.4 AHY-U (1) (1) 37.6 31.8 68 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. SY-M ASY-F ASY-M U-F u-u Scarlet Tanager Age/Sex Jan. HY-F HY-M HY-U AHY-F AHY-M AHY-U Feb. Feb. Mar. Apr . May June (9) (198) (144) (1) 49.3 49.2 48.6 49.4 [2.20] [3.41] [2.781 45.5 34.4 39.8 52.7 57.5 57.4 (39) (58) (3) 38.0 40.5 41.3 [2.78] [2.79] 40.2 32.9 33.2 42.2 47.2 45.5 (4) (169) (133) (6) 51.2 48.8 49.3 49.2 45.8 [3.08] [3.051 [2.73] 54.3 39.7 43.3 46.2 58.0 56.6 52.7 Aug. (I) 31.9 Sept. (1) 33.1 (2) 37.6 36.8 38.4 Oct. Nov. (1) 30.5 Piranga olivacea May June July Aug. Sept. Oct. (3) (-32) (100) (2) 24.5 28.3 28.5 30.7 23.5 [I.-57] [1.78] 27.8 25.1 23.4 24.5 33.5 30.8 33.5 (9) (33) (133) (7) 28.3 28.3 29.5 29.2 [1.53] [1.57] [1.75] [3.-54] 26.2 24.3 26.4 24.2 31.3 31.4 36.0 34.5 (1) (5) (1) 26.9 28.5 29.0 [2.18] 26.1 31.6 (39) (8) (4) (6) (9) (1) 27.9 30.3 28.4 28.8 27.2 31. 1 [2.57] [2.83] 25.8 [1.09] [2.74] 21.8 25.8 29.8 27.2 21.5 35.1 35.2 30.0 30.5 (10) (2) (15) 28.5 27.1 30.3 [1.24] 26.3 [2.32] 26.6 27.9 27.6 30.6 (1) 33.5 25.8 (2) 27.0 26.3 27.7 Dec. 544 Dec. SY-F 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 69 Age/Sex SY -M ASY-M TY-M ATY-M U-F U-M U-U Jan. Feb. Mar. Apr. May June July Aug. (23) (3) (1) (2) 28.2 28.3 28.7 29.1 [1.56] 28.0 27 2 26.2 31.9 28.6 30.9 (18) (2) (1) (3) 28.3 27.3 28.7 31.6 [2.41] 27.0 29.7 24.4 34.3 27.5 34.6 (I) 29.3 (2) 29.5 27.6 31.4 (9) 28.1 [1.181 26.4 30.4 (4) 28.0 27.1 28.8 Sept. Oct. Nov. Dec. (27) 28.9 [1.78] 25.0 33.6 (24) 29.3 [3.261 17.5 33.1 (1) 30.3 Summer Tanager Piranga rahra Age/Sex Jan. Feb. Mar. Apr. May June SY-M (1) (3) 25.5 29.2 27.2 31.2 Aug. Sept. Oct. n = 4 Nov. Dec. Cardinal Age/Sex Jan. Feb. Mar. HY-F HY-M Cardinalis cardinal is Apr. May June July (8) 38.4 [3.22] 34.7 42.5 (6) 42.4 [3.25] 38.6 47.7 (2) 37.5 35.4 39.5 n = 1471 Aug. Sept. Oct. Nov. Dec. (39) (55) (59) (38) (3) 39.1 40.5 40.5 40.6 49.7 [2.89] [2.34] [2.48] [2.20] 43.3 31.7 34.4 30.7 35.4 58.2 43.9 44.0 46.2 45.0 (17) (35) (48) (29) (4) 41.5 42.1 42.4 43.8 46.0 [2.84] [3.04] [3.24] [2.80] 44.8 35.9 34.9 32.1 38.8 47.1 46.1 51.8 48.7 48.4 (6) (12) (2) 42.7 41.5 42.2 [3.47] [2.86] 42.0 38.3 35.6 42.4 48.5 46.2 HY-U 70 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. AHY-F (29) (21) (112) (63) (18) (5) (3) (4) (13) (12) 49.3 47.8 44.6 41.1 41.4 43.1 41.4 38.5 42.2 42.3 [4.68] [3.88] [3.62] [2.61] [2.28] [1.-58] 39.1 36.6 [1.88] [2.44] 40.1 39.3 36.9 35.4 37.0 42.0 44.0 39.9 38.6 38.7 64.9 54.3 52.7 47.6 45.8 45.7 45.3 47.0 AHY-M (27) (15) (100) (51) (24) (2) (1) (4) (17) (13) 51.6 48.3 46.3 43.2 41.6 43.6 37.7 44.2 43.4 45.1 [4.11] [4.25] [3.48] [3.29] [2.72] 42.3 42.2 [2.49] [3.48] 45.0 43.0 38.3 34.0 34.4 44.9 48.8 40.3 39.1 63.2 57.2 57.5 51.0 46.2 48.1 50.2 SY-F (22) (5) (43) (18) (9) (1) (2) (1) (2) 46.1 44.4 43.7 41.3 40.5 39.5 41.2 39.9 42.2 [3.93] [0.94] [3.34] [4.35] [3.48] 36.0 41.3 40.3 43.5 36.0 36.0 36.3 46.4 43.1 57.3 45.7 52.1 47.2 46.0 SY-M (29) (11) (61) (23) (8) (1) (1) (2) (1) (5) 47.5 46.2 46.2 41.9 40.4 41.6 40.9 43.1 47.6 45.2 [4.19] [3.34] [3.75] [2.52] [1.73] 40.7 [1.17] 41.4 40.8 37.7 37.0 38.5 45.5 43.6 56.3 50.6 53.3 47.2 43.0 46.1 ASY-F (4) (5) (13) (4) (1) (1) (1) 44.6 47.4 44.4 43.5 42.2 44.2 44.8 41.4 [3.19] [2.46] 41.4 47.8 42.9 40.9 49.4 50.3 48.1 ASY-M (9) (7) (14) (12) (3) (1) (1) (2) 49.0 49.5 46.2 40.6 42.0 40.8 38.8 47.1 [2.91] [3.32] [2.10] [1.55] 40.1 46.2 42.9 45.2 43.1 37.8 44.3 48.0 52.5 53.6 49.4 43.1 TY-F (2) (2) (2) (2) (1) (1) (1) 46.6 49.9 44.1 42.8 50.1 45.0 42.6 45.9 48.5 43.7 42.2 47.2 51.2 44.5 43.3 TY-M (1) (5) (3) (4) (3) 52.8 44.7 42.8 44.9 47.8 [1.35] 42.3 42.4 42.9 43.0 43.2 50.2 50.4 46.4 ATY-F (7) (4) (19) (7) (2) (1) (4) (4) 46.0 44.9 44.0 41.9 43.0 42.2 42.5 41.7 [2.86] 42.1 [2.63] [4.86] 42.7 40.3 41.0 42.4 47.8 41.0 33.6 43.3 43.6 43.2 50.1 50.9 49.2 ATY-M (5) (9) (33) (17) (9) (3) (1) (4) (1) (4) 46.9 48.7 45.6 43.0 43.3 43.7 38.9 41.8 41.8 45.1 [1.16] [5.01] [2.73] [2.97] [2.34] 43.3 38.8 42,1 45.5 42.0 40.8 38.0 39.5 44.1 44.5 47.1 48.2 55.0 51.4 50.7 47.9 U-F (1) (17) (5) 45.2 40.8 42.2 [2.69] [3.05] 34.6 39.5 45.3 47.3 Dec. (10) 49.9 [4.91] 43.6 60.5 (21) 5/.2 [4.34] 42.6 57.5 (1) 41.0 (1) 43.5 (3) 46.3 45.3 47.0 (1) 46.3 (9) 45.5 [3.63] 41.3 51.2 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 71 Age/Sex Jan. Feb. U-M Mar. Apr. May June July Aug. Sept. (1) 42.6 Oct. (10) 41.6 [3.16] 33.7 45.2 Nov. (9) 43.4 [3.46] 36.6 48.1 Dec (1) 53.7 Rose-breasted Grosbeak Plieuc ficus ludovicianiis n = 494 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (4) 41.4 37.7 45.0 (14) 41.7 [1.77] 39.0 45.3 (70) 45.4 [3.68] 33.8 56.0 (7) 47.9 [3.14] 43.3 52.2 HY-M (4) 39.0 36.7 41.3 (21) 40.8 [3.30] 32.8 48.6 (88) 45.0 [3.63] 37.5 53.4 (7) 54.5 [6.76] 46.3 63.0 HY-U (1) 40.6 AHY-F (67) 46.2 [4.13] 38.3 56.1 (4) 41.8 40.7 43.3 (7) 41.6 [3.51] 35.4 46.8 (21) 45.4 [4.37] 38.7 52.6 (1) 62.4 AHY-M (2) 41.0 39.9 42.1 (6) 41.1 [4.44] 36.9 49.2 (20) 48.3 [3.15] 43.3 55.4 (2) 49.0 47.6 50.5 SY-M (2) 43.7 42.2 45.3 (28) 44.4 [4.24] 37.3 52.5 (1) 40.5 (8) 43.7 [1.83] 39.7 45.7 (2) 46.7 46.6 46.8 ASY-M (2) 44.9 44.1 45.7 (59) 46.3 [4.80] 38.6 57.5 (5) 40.3 [2.40] 36.3 42.5 TY-M (1) 41.9 (1) 43.0 ATY-M (2) 42.7 42.3 43.1 (1) 40.5 U-F (1) 40.2 (29) 45.9 [4.00] 38.5 56.3 (2) 58.5 52.1 65.0 U-M (3) 46.6 42.2 49.5 U-U (1) 43.2 72 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Indigo Bunting Passerina cyanea Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. HY-F (2) (32) (101) (16) 13.7 13.8 14.2 15.9 13.1 [0.98] [1.26] [2.30] 14.2 12 2 11.7 12.9 16.7 20.0 19.9 HY-M (10) (22) (65) (8) 14.2 14.5 14.6 15.2 [0.87] [1.14] [1.05] [1.48] 13.0 12.0 12.5 13.2 16.1 16.5 19.0 17.8 HY-U (29) (71) (76) (16) 14. 1 14.7 14.4 14.6 [0.96] [1.26] [1.00] [0.88] 12.5 12 2 11.6 13.5 16.2 17.7 16.9 17.1 AHY-F (137) (29) (35) (25) (33) (12) 13.7 14.2 14.1 13.8 14.8 15.9 [0.97] [1.16] [1.24] [1.38] [1.20] [1.64] 11.4 12.1 12.4 11.5 12.7 13.8 18.0 17.2 17.5 18.0 18.3 18.5 AHY-M (1) (50) (6) (9) (10) (52) (14) 16.7 14.7 14. 1 14.4 15.0 16.1 16.5 [0.88] [0.80] [0.92] [0.73] [1.77] [1.38] 13.0 13.2 13.1 14.1 13.4 14.4 16.7 15.3 15.9 16.4 21.4 18.5 SY-F (2) (2) (2) (3) 15.2 14.4 12.8 12.2 13.6 13.7 12.3 11.4 16.7 15.1 13.2 13.6 SY-M (101) (23) (19) (4) (7) 14.5 14.5 14.5 14.5 15.3 [0.92] [0.96] [0.70] 13.0 [1.81] 12.7 12.8 12.8 15.4 13.3 16.7 17.3 15.6 18.8 ASY-F (6) (2) (5) (5) (1) 13.7 14.4 13.0 13.8 16.6 [0.74] 14.4 [1.83] [0.83] 12.5 14.4 11.2 12.7 14.5 15.8 14.7 ASY-M (2) (40) (5) (1) (10) (3) 15.0 14.7 14.8 15.6 15.4 16.1 14.2 [0.94] [0.64] [0.98] 15.2 15.8 13.1 14.2 14.0 16.7 17.2 15.9 16.8 TY-F (3) (2) (1) (3) 13.7 14. 1 13.6 13.2 12.7 13.9 12.7 15.4 14.2 13.8 TY-M (13) (5) (4) (5) (3) 14.7 14.5 14.8 14.4 14.8 [0.55] [0.70] 14.4 [0.59] 12.6 14.0 13.7 15.2 13.4 16.2 15.6 15.2 14.9 1288 Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 73 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. ATY-F (14) (1) (4) (6) (5) 14.4 18.6 14.0 15.0 14.5 [1.69] 13.4 [0.93] [1.02] 12.6 14.6 13.6 13.5 18.5 16.3 16.1 ATY-M (45) (3) (7) (7) (12) 14.5 13.9 14.7 15.4 16.2 [1.30] 13.8 [0.72] [0.76] [1.87] 12.3 14.1 13.7 13.8 14.3 20.7 15.9 16.0 20.1 U-M (2) (1) 16.4 14.5 15.0 17.7 U-U (8) (27) 14.2 [0.90] 13.1 15.9 14.9 [0.96] 13.2 16.8 Oct. (I) 13.7 Nov. (2) 16.0 14.5 17.4 Evening Grosbeak Age/Sex Jan. Feb. HY-F Mar. Hesperiphomi vespertina Apr. May June July Aug. Sept. Oct. Nov. (3) 56.4 54.7 57.3 HY-M (2) 61. 3 57.0 65.6 AHY-F (77) (111) (227) (138) (21) (1) 60.5 60.7 59.5 58.2 63.3 56.3 [3.77] [4.17] [4.00] [4.61] [5.01] 51.8 52.6 47.0 43.2 57.3 72.5 70.1 69.8 73.0 73.2 AHY-M (59) (100) (151) (87) (20) (3) 61.5 62.0 60.1 59.3 66.8 60.9 [3.51] [4.16] [3.88] [5.61] [5.96] 55.3 53.7 51.2 49.7 45.4 54.9 65.2 70.3 72.1 70.2 86.1 77.7 SY-F (8) (4) (92) (124) (4) 58.9 64.1 56.5 56.3 62.4 [1.82] 58.7 [7.25] [4.08] 53.2 56.0 68.8 46.5 45.5 70.3 62.0 68.6 68.3 SY-M (1) (37) (97) (5) 60.3 56.9 57.9 59.7 [2.90] [3.26] [4.01] 51.4 51.0 55.1 65.0 69.3 63.3 ASY-F (5) (8) (110) (118) (7) 58.3 61. 1 57.9 57.6 62.2 [1.52] [3.47] [3.83] [4.35] [7.11] 55.8 57.5 51.1 49.0 52.1 59.7 68.6 70.8 73.5 69.6 Dec. 2047 Dec. (22) 61.2 [4.21] 55.8 70.6 (11) 59.1 [4.23] 51.6 64.3 (33) 59.3 [3.33] 51.5 65.6 (75) 62.0 [3.93] 54.8 73.4 74 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May ASY-M (15) (6) (76) (Ill) (4) 63.1 62.4 58.9 58.7 67.4 [3.76] [3.04] [4.10] [4.60] 63.5 58.0 57.9 49.6 38.7 70.5 70.9 65.5 67.4 73.3 U-F U-M Purple Finch Age/Sex Jan. Feb. Mar. HY-M HY-U Carpodacus piirpiireiis Apr. May June July (I) (I) 21.5 22.9 AHY-F (1) (I) 24.0 22.0 AHY-M (1) (13) (2) 25.7 25.6 27.2 [1.26] 25.1 23.5 29.3 27.6 AHY-U (1) (6) (72) (30) 25.9 24.7 24.6 26.1 [1.72] [2.36] [2.95] 23.1 19.8 22 1 28.0 35.3 33.1 SY-M (1) 22.1 ASY-M (4) (54) (2) 27.4 24.9 27.1 24.5 [2.55] 23.7 28.7 18.6 30.4 32.0 u-u Common Redpoll Age/Sex Jan. AHY-F Feb. Mar. Apr. Cardiielis fhimmea May June July Sept. Oct. Nov. Dec. (2) (4) (63) 67.7 57.8 59.8 67.2 55.1 [3.11] 68.2 60.5 53.5 66.6 (1) (2) (2) 65.4 60.0 63.0 56.3 62.7 63.8 63.3 n = 525 Sept. Oct. Nov. Dec. (1) 25.1 (43) (142) (17) (4) 22.7 24.2 25.0 25.6 [1.58] [1.80] [3.30] 24.7 18.3 20.5 19.1 26.6 26.0 29.9 33.3 (1) (24) 21.7 24.1 [1.99] 18.1 28.0 (1) (36) (5) (2) 23.5 25.0 25.6 23.4 [1.61] [0.50] 23.4 21.4 25.0 23.4 28.7 26.3 (21) (2) 23.8 26.8 [1.78] 23.6 21.0 30.0 26.7 (1) (1) 21.6 (1) 27.1 22.8 (3) (24) (5) (1) 22.5 24.1 24.2 22.8 20.1 [2.05] [1.57] 23.9 19.7 22.0 28.0 26.0 n = 13 Sept. Oct. Nov. Dec. (1) 13.2 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 75 Age/Sex AHY-U ASY-M Jan. Feb. Mar. (2) 11.7 11.5 11.9 (i) 14.6 Apr. May June (3) 15.0 13.1 16.6 (6) 15.0 [1.03] 13.6 16.5 July Aug. Sept. Oct. Nov. Dec. Pine Siskin Carduelis pinus n = 356 Age/Sex HY-U AHY-U U-U Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. (25) 13.5 [1.24] 11.7 16.8 Nov. (3) 12.3 12 2 12.6 (7) (11) (132) (111) (21) (1) (13) (4) 16.2 17.6 15.1 14.1 14.5 12.3 13.4 13.6 [1.17] [1.20] [1.31] [1.18] [1.75] [1.38] 11.6 15.0 16.0 12.1 11.8 12.1 10.8 15.3 18.6 20.1 19.0 19.0 18.7 15.6 (16) (12) 12.9 13.5 [0.79] [0.60] 11.3 12.5 14.1 14.7 Dec. American Goldfinch Carduelis Iristis n = 5180 Age/Sex HY-F HY-M HY-U AHY-F Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. (2) (110) (269) (79) (2) 11.7 11.4 12.1 12.7 15.0 11.2 [0.92] [0.81] [0.87] 14.1 12 2 8.6 9.7 10.9 15.8 13.7 14.5 14.9 (6) (116) (226) (68) (1) 12.1 12.0 12.4 13.1 13.8 [0.86] [1.08] [0.80] [1.03] 11.0 9.4 9.6 10.5 13.4 15.3 14.8 16.3 (13) (223) (101) (135) 11.7 11.8 12.2 12.8 [1.12] [0.97] [0.80] [0.88] 9.8 8.4 10.1 10.5 13.7 14.8 13.9 15.1 (7) (6) (74) (U58) (520) (24) (15) (122) (281) (152) (133) 14.2 14.1 13.6 13.2 12.4 12.1 12.5 12.5 12.4 12.5 12.9 [1.20] [1.12] [1.00] [0.81] [0.90] [0.72] [0.60] [0.90] [0.85] [0.76] [0.97] 13.4 12.7 11.8 10.5 10.3 10.6 1 1.6 10.7 10.0 10.4 10.4 16.8 15.6 17.1 15.7 16.5 13.5 13.4 15.3 15.2 15.0 16.1 (9) (14) (142) (257) (327) (17) (11) (129) (264) (158) (189) 14.5 15.4 14.2 13.9 12.6 11.9 11.9 12.5 13.0 13.3 13.4 [0.80] [1.35] [1.05] [0.83] [1.03] [0.63] [0.47] [0.83] [0.86] [1.00] [0.88] 13.0 13.6 11.5 11.6 10.6 10.5 11.1 10.5 10.0 10.5 11.5 15.8 18.8 17.1 16.6 17.1 13.0 13.0 14.6 15.4 19.4 19.4 AHY-M 76 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. AHY-U (7) (3) (4) (3) (4) (6) 14.2 13.6 12.3 10.9 12.6 12.9 [1.02] 13.1 11.6 9.0 12.2 [0.46] 12.9 14.5 13.3 12.0 13.2 12.4 15.5 13.7 SY-F (1) (1) 11.2 11.2 SY-M (i) (46) (83) (159) (5) (2) (10) (22) (3) 14.0 14.3 13.6 12.7 11.6 11.3 12.7 12.7 13.4 [0.93] [0.90] [1.00] [0.63] 10.9 [0.78] [0.85] 13.0 12.6 12.1 8.6 10.8 11.6 11.2 11.1 14.0 17.1 16.3 16.2 12.4 13.9 14.2 ASY-F (11) (10) (8) (1) (2) 11.9 12.2 12.2 10.7 13.1 [0.74] [0.48] [0.67] 12.6 10.6 11.4 11.0 13.5 13.1 12.9 13.0 ASY-M (1) (1) (30) (92) (107) (2) (3) (10) (19) (1) (1) 13.4 20.7 14.3 13.9 12.8 12.6 12.3 12.6 12.8 13.0 9.9 [1.02] [0.91] [1.13] 11.9 12.0 [0.48] [0.83] 12.6 11.6 9.2 13.3 12.5 11.7 11.5 16.4 17.6 15.6 13.3 14.3 TY-F (1) (1) 13.3 13.3 ATY-F (2) (1) (4) (2) (1) (3) (4) 12.8 12.8 12.1 12.1 12.8 12.2 11.9 12.6 11.2 11.7 11.5 11.4 12.9 12.7 12.4 13.4 12.3 ATY-M (1) (3) (5) (17) (3) (1) (10) (8) (4) (3) 15.2 13.3 14.3 11.9 12.3 12.7 12.7 12.7 12.8 13.5 12.5 [1.30] [0.67] 11.7 [0.73] [0.45] 11.9 11.8 13.9 12.4 11.0 12.7 11.7 11.8 13.9 14.7 15.4 13.3 14.2 13.4 U-F (1) (1) 11.6 11.5 U-M (1) (6) (1) 11.6 13.0 13.7 [0.92] 12.2 14.1 U-U (1) (8) (48) (20) 11.7 12.0 12.4 12.9 [0.91] [1.14] [0.68] 10.5 10.2 10.9 13.2 15.5 13.8 Rufous-sided Towhee Pipilo erythrophthalmus n = 711 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec HY-F (5) (16) (42) (98) (4) 36.3 36.8 39.7 39.1 39.7 [2.08] [1.82] [2.10] [2.64] 37.2 33.9 34.5 35.7 30.8 43.1 39.1 41.2 44.5 47.6 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 77 Age/Sex Jan. HY-M AHY-F AHY-M SY-M ASY-F ASY-M TY-M ATY-F ATY-M U-F U-M Savannah Sparrow Age/Sex Jan. HY-U Feb. Mar. Apr. May (1) 40.4 (21) (31) 38.0 38.6 [2.98] [4.16] 33.7 32.1 46.8 52.3 (6) (15) (15) 42.9 40.7 40.2 [2.31] [1.95] [2.37] 39.4 38.4 36.1 45.8 45.6 46.5 (4) (35) (21) 42.8 40.9 40.6 42.3 [2.26] [3.26] 43.5 36.9 32.1 44.5 45.2 (2) 36.2 34.0 38.4 (2) (19) (10) 40.8 41.8 41.4 37.8 [2.03] [1.63] 43.8 37.9 39.0 45.0 43.9 (3) (3) 40.3 43.1 38.6 42.2 43.6 43.9 (4) (2) 39.5 40.2 33.6 39.5 43.1 40.9 (1) (1) (2) 48.2 39.5 40.6 39.5 41.7 June July Aug. (1) (8) (36) 41.3 38.5 39.0 [2.66] [2.12] 35.5 34.8 42.3 44.5 (2) (2) (6) 40.7 36.5 39.5 40.0 34.3 [1.86] 41.4 38.7 37.9 43.1 (1) (3) (3) 40.4 45.5 41.3 43.4 39.6 47.4 42.2 (3) (3) (2) 41.3 42.2 41.4 40.4 38.8 39.7 42.3 44.7 43.4 (1) 40.5 (1) (1) (2.5 40.4 Sept. Oct. Nov. (62) (102) (3) 42.7 42.8 48.4 [3.45] [2.59] 42.5 29.3 35.7 51.5 51.5 49.0 (14) (37) 39.6 39.9 [2.26] [2.88] 36.2 33.9 43.6 44.7 (8) (33) 42.8 43.0 [2.31] [2.99] 39.2 37.8 45.9 50.0 (2) (2) 44.7 42.0 44.7 41.4 44.7 42.6 (1) 42.7 (1) 41.3 (I) 43.2 (2) 47.8 47.1 48.5 (1) (3) 45.4 39.5 37.4 40.9 (2) 45.1 44.1 46.1 Dec. (1) 50.4 Passerciilus sandwichensis Mar. Apr. May June July n = 60 Aug. Sept. Oct. (1) (5) (5) 15.6 18.1 19.3 [0.42] [2.09] 17.4 16.9 18.5 22.3 Feb. Nov. Dec. 78 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. AHY-U U-U Grasshopper Sparrow Age/Sex Jan. Feb. HY-U AHY-M AHY-U U-U Mar. Apr. May (19) (10) (2) 18.3 18.6 17.3 [1.27] [1.85] 16.2 15.4 16.1 18.4 20.3 23.0 June July Mar. Ammodramiis savannarum Apr. May June July (1) 16.6 (6) 16. 1 [0.90] 15.1 17.3 Aug. Sept. Oct. (1) (1) 16.8 17.2 (4) (ID (1) 16.0 16.7 21.9 14.9 [1.93] 17.4 12.4 20.5 Aug. Sept. Oct. (4) (2) (2) 16.6 16.1 19.0 15.4 15.7 17.0 17.8 16.5 21.0 (4) 16.3 15.3 17.4 (2) 17.2 16.7 17.6 Nov. n = Nov. Hensiow’s Sparrow Age/Sex Jan. Feb. Mar. HY-U Ammodramiis henslowii Apr. May June July U-U Aug. Sept. Oct. (2) (1) 13.8 12.3 11.7 15.9 (1) 12.2 Vesper Sparrow Age/Sex Jan. Feb. Mar. HY-U AHY-F AHY-M (2) 25.8 25.4 26.2 Pooecetes gramineiis Apr. May June July (1) 22.2 (2) 25.0 24.7 25.2 (5) (3) 23.6 24.5 [1.36] 21.8 22.0 26.2 25.4 Aug. Sept. Oct. (1) (2) (1) 26.3 22.6 21.5 22.0 23.1 (1) 24.3 n = Nov. Dec. 21 Dec. 4 Dec. 21 Dec. AHY-U 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 79 Age/Sex U-U Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. (3) 25.4 24.6 26.1 Nov. Dec. Dark-eyed Junco (Slate-colored Junco) Junco hyenuiHs n = 7715 Age/Sex HY-F HY-M HY-U AHY-F AHY-M AHY-U SY-F SY-M SY-U ASY-F Jan. Feb. Mar. Apr. (24) (6) (213) (776) 21.2 20.5 19.1 18.7 [1.64] [0.79] [1.41] [1.51] 18.2 19.4 14.3 14.6 24.9 21.5 23.4 23.6 (28) (13) (1016) (982) 22.7 22.0 20.6 20.4 [1.52] [1.26] [1.53] [1.61] 20.0 19.9 16.4 16.3 25.6 24.0 25.4 26.7 (23) (ID (790) (951) 22.3 22.1 20.0 19.5 [1.15] [2.30] [1.54] [1.53] 20.4 18.1 13.5 14.2 26.0 26.2 25.3 25.1 (1) (2) (1) 23.0 19.4 17.7 18.6 20.1 (1) 22.6 (3) (6) (3) 22.2 19.4 20.8 21.8 [1.28] 19.6 22 5 17.0 21.5 20.3 (6) (3) 19.3 21.0 [0.28] 20.3 19.0 21.6 19.7 (1) (16) (9) 23.4 20.7 20.7 [1.32] [1.54] 18.7 17.4 23.0 22.6 May June July Aug. (4) 20.! 18.1 22 2 (6) 19.7 [1.15] 17.9 21.0 Sept. Oct. Nov. Dec. (Ill) (136) (7) 17.9 18.0 19.9 [1.08] [1.30] [1.39] 15.0 15.5 17.9 21.3 22. 1 21.9 (67) (132) (2) 19.5 20.0 23.3 [1.10] [1.48] 21.9 17.2 15.6 24.6 23.4 23.6 (5) (268) (269) (14) 18.0 18.6 19.5 21.1 [0.90] [1.31] [1.39] [1.33] 16.9 15.6 16.3 18.2 19.4 22.8 25.8 23 2 (1) (169) (94) (4) 17.4 18.2 18.9 21.0 [1.22] [1.13] 18.5 14.4 16.1 25.1 21.2 21.3 (1) (418) (285) (13) 18.6 19.5 20.2 22.2 [1.21] [1.42] [1.20] 14.3 15.5 20.5 23.1 25.1 24 2 (3) (313) (171) (4) 19.4 18.6 19.7 21.4 18.8 [1.29] [1.18] 19.6 20.2 13.6 17.0 23.4 22.6 23.1 (1) 20.1 (3) (5) 18.8 19.8 17.4 [1.24] 19.9 17.7 20.7 ASY-M 80 Age/Sex ASY-U ATY-M ATY-U BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY May June July Aug. Sept. Jan. Feb. Mar. Apr. (3) (5) (6) 20.9 19.6 19.4 20.2 [0.98] [1.50] 21.5 18.0 17.7 20.6 21.3 (4) (5) 20.8 22.0 19.3 [0.67] 22.0 21.5 23.1 (1) (7) 22.2 21.0 U-F U-M u-u Dark-eyed Junco hybrids (Slate-colored x Oregon) Age/Sex HY-F HY-U AHY-F AHY-U Jan. Feb. [0.61] 20.2 22.0 Mar. (1) 18.7 (5) 19.7 [0.56] 19.1 20.6 Dark-eyed Jiinco (Oregon Junco) Age/Sex AHY-F AHY-M AHY-U Jan. (1) 21.1 Feb. Mar. (1) 22.5 Oct. (1) i8.7 Junco li. hyemalis x J. h. montanus Apr. May June July Aug. Sept. (I) 20.1 (34) 18.8 [1.41] 16.4 21.7 Junco hyemalis montanus Apr. May June July Aug. Sept. Oct. (1) 16.5 Nov. NO. 5 Dec. (I) 19.1 (1) 18.7 (9) (3) 19.6 21.5 [1.27] 20.7 18.5 22 1 22 3 (13) (3) 20.4 21.2 [1.47] 20.1 18.1 22.9 23.3 (138) (115) (9) 18.5 19.2 20.6 [1.47] [1.29] [1.90] 12.9 13.9 18.0 22.7 22.0 23.0 n = 44 Oct. Nov. Dec. (1) 18.0 (2) 17.5 17.2 17.8 = 3 Nov. Dec. 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 81 Tree Sparrow Age/Sex Jan. HY-U Feb. Mar. Apr. Spizella arhorea May June July Aug. Sept. Oct. (2) 15.8 15.3 16.2 n = 1585 Nov. (78) 16.9 [1.16! 13.8 19.7 [1.49] 14.0 22 7 Dec. AHY-U (169) ( 105) (590) (153) (2) (4) (160) (11) 18.9 19.4 18.5 18.4 19.7 17.2 17.3 17.8 [1.56] [1.33] [1.51] [1.73] 18.1 15.3 [1.25] [1.64] 15.7 16.5 14.2 14.0 21.3 19.1 14.5 15.4 25.6 22.5 23.7 25.7 20.5 21.2 SY-U (3) (4) (6) (5) 18.6 19.0 18.8 17.9 17.3 17.7 [1.29] [1.05] 20.0 21.7 16.7 16.8 20.4 19.5 ASY-U (6) (8) (37) (12) (7) (1) 18.5 18.8 18.7 18.5 17.6 16.9 [0.87] [1.14] [1.28] [2.01] [0.99] 17.0 17.6 16.1 15.2 16.0 19.5 20.2 21.2 22.2 18.9 TY-U (2) (4) 18.9 18.7 18.6 17.6 19.1 21.1 ATY-U (5) (6) (40) (9) (5) 19.2 19.8 18.6 19.2 17.9 [0.90] [0.70] [1.45] [2.23] [1.31] 18.2 18.5 15.1 16.5 16.0 20.6 20.4 21.3 22.6 19.5 U-U (1) (98) (52) 16.4 17.8 18.9 [1.75] 15.6 23.9 Chipping Sparrow Age/Sex Jan. HY-U AHY-F Spizella passerina 1143 Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. (5) (2) (9) (75) (115) (3) 12.3 12.4 11.5 12.1 12.4 13.2 [0.17] 11.7 [1.12] [0.84] [0.92] 12.5 12.2 13.0 10.1 10.1 10.5 13.8 12.6 13.7 14.7 15.4 (2) (14) (6) (1) (1) 13.6 12.4 11.8 12.7 11.4 13.5 [1.17] [1.29] 13.7 10.6 10.5 14.6 14.0 (15) (53) (5) (2) 12.3 11.9 12.5 12.2 [0.57] [0.81] [0.57] 11.1 11.5 9.9 11.5 13.2 13.4 13.9 12.9 Dec. AHY-M 82 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. AHY-U (2) (361) (316) (1) (1) (1) (17) (46) (2) 13.8 12.3 11.7 12.5 12.9 11.9 12.2 12.7 12.3 13.3 [1.07] [0.87] [0.66] [0.83] 12.0 14.2 10.1 9.8 11.4 11.1 12.6 18.8 15.1 14.0 14.3 SY-U (6) (4) 12.2 12.4 [0.42] 11.9 11.9 12.8 12.9 ASY-F (1) (1) (1) 11.7 13.7 12.7 ASY-M (8) (13) (1) 12.4 12.4 12.5 [0.68] [0.62] 11.7 11.1 13.6 13.1 ASY-U (28) (15) (1) (4) 12.5 12.5 11.9 12.8 [0.67] [0.67] 11.7 11.2 11.5 14.5 14.3 13.9 u-u (3) (1) (1) 12.4 12.6 12.4 11.7 12.8 Field Sparrow Spizella pusilla n = 4778 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. HY-F (1) (1) (2) 10.2 12.8 12.5 12.4 12.6 HY-M (1) (2) (1) (3) (4) 12.8 11.7 12.3 12.1 12.4 11.5 11.2 11.3 11.8 13.5 13.3 HY-U (3) (128) (222) (189) (1057) (103) 11.6 12.0 12.0 12.6 12.9 13.2 9.2 [0.72] [0.94] [0.95] [0.92] [1.14] 12.8 10.2 10.0 10.3 9.0 10.9 14.3 15.7 15.7 17.2 16.5 AHY-F (I) (4) (44) (17) (22) (32) (12) (8) 13.4 12.1 12.6 12.9 12.1 12.1 12.6 11.9 11.4 [1.15] [1.19] [1.34] [0.76] [0.93] [0.94] 12.9 11.0 10.3 10.4 11.2 11.1 10.4 15.9 14.7 16.5 14.0 14.4 13.3 AHY-M (1) (43) (99) (35) (18) (23) (18) (34) (1) 12.0 12.6 12.3 12.3 12.3 12.6 13.0 13.5 16.2 [0.80] [0.73] [0.76] [0.72] [0.69] [0.93] [1.34] 10.9 10.2 10.3 10.9 11.0 10.5 11.8 14.7 13.8 14.1 13.6 13.7 14.5 15.3 AHY-U (6) (2) (131) (876) (289) (2) (3) (38) (50) (525) (57) 14.2 13.2 13.3 12.5 11.9 12.8 12.4 12.3 12.6 13.2 13.8 [2.00] 12.7 [1.04] [1.00] [0.98] 12.5 11.8 [0.81] [0.95] [1.06] [1.12] 11.0 13.6 10.8 9.7 9.4 13.1 13.2 10.8 10.8 10.1 10.1 16.6 17.2 16.9 15.5 14.0 14.8 16.7 16.0 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 83 Age/Sex SY-F Jan. Feb. Mar. Apr. (1) 10.5 SY-M (I) 13.5 (1) 12. 1 SY-U (2) 14.4 14.0 14.8 (1) 10.5 (8) 12.5 [0.52] 11.9 13.5 ASY-F (4) 12.8 12.2 13.4 ASY-M (1) 13.7 (1) 12.3 (32) 12.2 [0.63] 10.7 13.6 ASY-U (8) 12.8 [0.74] 11.6 14.0 TY-F TY-M (6) 12.8 [0.67] 12.1 13.7 May June July Aug. Sept. Oct. (2) (2) (4) 11.8 11.6 13.8 11.5 10.8 12.8 12.0 12.3 15.3 (7) (1) (1) (4) 12.6 12.7 11.9 12.6 [0.74] 12.0 11.6 14.0 13.6 (8) (1) (1) (5) 11.9 12.4 12.6 12.8 [0.24] [0.83] 11.5 11.8 12.3 13.6 (5) (5) (3) (2) 12.5 11.6 12.1 12.6 [1.48] [0.81] 11.8 12.5 11.5 10.9 12.2 12.7 15.1 13.0 (31) (4) (2) (4) (6) (10) 12.4 12.2 12.5 13.2 12.9 13.1 [0.53] 11.8 11.8 12.7 [1.26] [0.69] 11.3 12.6 13.1 13.5 10.7 11.8 13.5 14.4 14.2 (8) (5) (8) 12.1 13.5 12.4 [1.01] [1.25] [0.54] 10.4 11.9 11.8 13.5 15.0 13.6 (1) (1) 12.3 11.5 (3) (2) 12.0 12.7 11.6 12.5 12.2 12.8 Nov. (U 12.7 (I) 15.0 TY-U ATY-F ATY-M (7) 12.3 [0.64] 11.6 13.5 (1) 13.0 (1) (6) (1) (3) (4) (2) 13.7 13.4 13.8 12.0 11.7 13.7 [1.55] 11.5 10.2 13.3 11.5 12.5 13.3 14.0 15.6 (1) (22) (20) (1) (3) (4) (2) (5) (1) 13.0 12.6 12.7 11.9 12.4 12.5 13.1 12.6 15.3 [0.92] [0.73] 12.3 11.3 12.9 [0.40] 11.4 11.5 12.6 13.6 13.3 12.2 14.8 14.6 13.2 (2) (6) (8) (3) (1) (1) (8) (1) 15.3 13.4 12.5 12.0 10.5 13.7 13.5 12.7 14.6 [0.81] [0.55] 11.5 [1.28] 15.9 12.5 11.3 12.6 12.2 14.8 13.0 15.6 Dec. (1) 14.0 ATY-U 84 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 5 Age/Sex Jan. Feb. Mar. Apr. May June July U-U White-crowned Sparrow Age/Sex Jan. Feb. Mar. HY-U Zonotrichia leiicophrys Apr. May June July AHY-U (H8) 30.0 [3.22] 24.6 38.5 White-throated Sparrow Age/Sex Jan. Feb. Mar. HY-U Zonotrichia alhicollis Apr. May June July AHY-U (10) (2) (20) (232) (163) 28.2 24.7 26.8 27.7 25.8 [3.78] 24.1 [3.55] [2.39] [2.45] 20.7 25.3 ■>2 2 22.0 21.2 34.3 34.1 35.4 33.7 SY-U (2) (1) (7) (2) (1) 30.1 30.0 25.8 27.2 23.6 29.4 [2.46] 26.5 30.8 22.3 27.8 28.6 ASY-U (1) (3) (1) 32.1 29.1 28.3 28.2 29.6 U-U (1) 29.2 Fox Sparrow Age/Sex Jan. HY-U Feb. Mar. Apr. Passerella iliaca May June July Aug. Sept. Oct. Nov. Dec. (3) (25) (261) (50) (12) 11.2 12.7 13.0 13.6 13.4 10.6 [0.92] [0.96] [0.81] [1.37] 12.0 10.8 10.0 11.9 10.2 14.7 17.1 15.2 15.0 n = = 231 Aug. Sept. Oct. Nov. Dec. (1) (64) (4) 27.0 27.4 30.2 [2.64] 23.1 18.9 33.3 33.1 (3) (41) 25.3 28.0 23.1 [2.56] 26.5 21.6 33.3 n = 3501 Aug. Sept. Oct. Nov. Dec. (88) (1220) (306) (3) 24.2 25.5 26.8 27.8 [1.65] [2.06] [2.74] 26.3 19.5 19.3 20.3 29.2 28.2 32.6 37.1 (29) (1061) (59) (1) 24.2 25.7 27.1 28.5 [1.36] [2.10] [2.35] 22.0 19.9 19.0 27.9 33.1 32.2 (27) (232) (23) (6) 24.4 25.0 26.6 28.8 [1.60] [2.02] [1.99] [2.42] 22.1 19.6 23.5 25.4 29.0 30.8 31.1 31.8 n = 446 Sept. Oct. Nov. Dec. (56) (50) 35.2 37.4 [2.40] [2.78] 29.6 32.6 40.4 43.9 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS 85 Age/Sex Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. AHY-U (14!) (69) (2) (37) (47) 36.7 37.4 35.0 35.3 38.4 [2.71] [2.70] 34.0 [2.32] [3.51] 29.4 32.3 36.0 30.. 3 27.2 43.1 47.0 40.0 49.0 u-u (17) (26) JJi.6 40.3 [2.97] (3.48] 32.0 33.3 43.6 46.2 Dec. (1) 44.4 Lincoln's Sparrow Melospiz.li lincohiii n = 845 Age/Sex Jan. HY-U AHY-U U-U Swamp Sparrow Age/Sex Jan. HY-U AHY-F AHY-M AHY-U (I) 19.6 SY-M Feb. Feb. (I) 18.3 (1) 18.9 Mar. Apr. May June July Aug. Sept. Oct. Nov. (2) (212) (268) (3) 17.7 16.6 17.2 20.3 17.3 [1.61] [1.51] 16.3 18.0 1 1.2 1 1.5 24.6 20.6 22.5 (76) (2) (62) (63) (1) 18.6 17.7 17.2 17.4 17.5 [1.82] 17.1 [i.-‘;o] [1.78] 14.8 18.3 13.7 1 1.5 24.0 20. 1 21.6 (101) (54) 16.7 17.2 [1.52] [1.23] 10.4 14.8 20.3 20.2 Melospiz a georpiana n = 2178 Mar. Apr. May June July Aug. Sept. Oct. Nov. 1 (3) (105) (964) (115) 14.8 16.0 16.8 17.4 14.2 [1.26] [1.44] [1.85] 15.4 13.8 11.1 14.0 20.0 22.5 22.1 (1) (1) 19.3 15.8 (2) (3) 16.8 17.6 16.7 16.1 16.8 19.0 (8) (191) (154) (12) (432) (27) 17.8 17. 1 16.7 15.9 17.2 17.9 [2.03] [1.58] [1.64] [1.00] [1.51] [1.13] 15.1 13.3 12.8 14.4 10.9 16.1 21.1 22 20.5 17.5 22.0 20.5 (1) (1) 16.5 17.5 (2) 17.3 16.6 17.9 SY-U 86 Age/Sex U-U Song Sparrow Age/Sex HY-F HY-M HY-U AHY-F AHY-M AHY-U SY-F SY-M SY-U ASY-F BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 5 Jan. Feh. Mar. Apr. May June Jan. Feb. Mar. Apr. (2) 20.1 19.7 20.4 Melospizd melodia May June (2) 20.3 19.8 20.7 (14) /9.5 [1.101 17.5 21.7 (4) (30) (25) (11) 19.5 20.9 21.4 20.6 17.1 [1.78] [2.14] [1.97] 22.6 18.1 17.2 17.2 25.6 26.1 25.2 (1) (1) (19) (43) (60) (8) 22.8 21.5 20.6 20.9 20.6 20.8 [1.19] [1.56] [1.31] [1.31] 18.3 19.0 18.2 18.5 22.3 25.4 23.5 23.2 (6) (5) (1090) (646) (27) 25.0 23.9 21.6 20.8 20.1 [1.43] [1.96] [2.05] [2.03] [1.39] 22 9 21.0 15.1 11.4 17.4 27.3 25.9 27.9 29.1 21 2 (7) (5) (1) 20.4 21.2 22.2 [0.71] [0.67] 19.3 20.4 21.1 22.1 (1) (4) (15) (8) 26.1 21.3 22.3 20.9 20.2 [2.37] [0.67] 23.1 19.7 20.3 29.8 21.5 (1) (8) (3) (3) 24.3 21.6 19.2 20.3 [0.60] 18.5 19.9 20.7 19.8 20.6 22 5 July Aug. Sept. Oct. Nov. (9) (140) (3) 16.2 16.5 18.0 [1.17] [1.47] 17.7 14.3 12.4 18.6 17.6 20.8 a n = : July Aug. Sept. Oct. Nov. (3) (1) (3) (3) (1) 18.6 18.6 19.6 19.9 21.5 18.3 19.6 18.3 19.0 19.7 21.1 (3) (3) (4) (4) (7) 20.3 20.3 20.6 20.8 20.6 20.0 19.3 19.7 19.5 [0.97] 20.7 21.2 21.6 22.4 19.6 22.5 (141) (163) (261) (1202) (208) 19.2 18.9 19.6 20.4 21.1 [1.33] [1.33] [1.31] [1.53] [1.60] 16.3 12.8 16.2 14.7 14.8 22.7 23.0 24.1 26.6 25.0 (15) (32) (14) (9) 19.6 19.3 20.9 20.4 [1.17] [1.83] [1.64] [0.78] 16.9 11.9 18.1 19.2 22.3 22.4 23.6 21.3 (8) (8) (5) (6) (1) 20.7 20.1 22.3 21.0 21.4 [0.65] [0.94] [1.95] [1.01] 19.4 18.8 20.1 19.2 21.4 21.4 25.3 21.9 (4) (14) (18) (618) (102) 19.9 20.1 19.8 20.8 21.8 17.6 [1.71] [1.65] [1.53] [1.48] 23.4 18.1 14.3 15.6 18.3 23.0 21.7 28.0 24.9 (1) (1) (1) (3) 22.5 19.1 23.5 20.3 19.2 21.6 (1) 23.1 (1) (2) (5) (2) 20.8 23.9 20.9 21.5 21.7 [1.01] 21.3 26.0 20.1 21.6 22.4 (1) (2) 19.7 18.1 Dec. (1) 15.8 Dec. (3) 23.4 22.1 25.0 17.6 18.6 1978 CLENCH AND LEBERMAN— PENNSYLVANIA BIRD WEIGHTS Age/Sex ASY-M ASY-U TY-F TY-M TY-U ATY-F ATY-M U-F U-M U-U 87 Jan. Feb. Mar. Apr. May June July Aug. Sepl. Oct. Nov (6) (5) (1) (1) (2) 21.8 21.0 20.5 21.4 21.8 [1.01] [0.80] 20.2 20.4 20.0 23.3 23.5 22.0 (1) (1) (2) (2) (2) (1) 28.1 20.2 20.2 20.2 20.0 24.1 20.1 19.3 20.0 20.3 21.1 20.0 (I) (4) 25.0 21. 1 20.2 22.2 (3) (I) 22.4 22.0 21.4 24.2 (1) (I) 20.0 23.3 (2) (1) (1) 20.7 21.2 22.5 20.0 21.3 (3) (12) (2) (1) (3) (3) (1) 22.1 21.3 20.8 23.2 21.0 20.3 22.7 21.0 [1.20] 20.7 19.9 19.3 23.0 18.7 20.8 21.6 22 2 22.9 (I) 22.1 (I) 20.5 (I) (3) 20.0 21.0 20.0 21.8 (1) (18) (73) (430) (54) (16) 18.2 20.5 20.2 20.7 21.2 23.1 [1.46] [1.31] [1.59] [2.14] [1.58] 18.5 16.7 11.7 11.9 20.2 22.8 23.0 26.6 24.6 26.2 4 (V hs* ) ■1^ ■ ■? ! I I BULLETIN of CARNEGIE MUSEUM OF NATURAL HISTORY ECOLOGY AND TAXONOMY OF AFRICAN SMALL MAMMALS Edited by DUANE A. SCHLITTER Associate Curator, Section of Mammals NUMBER 6 PITTSBURGH, 1978 BULLETIN OF CARNEGIE MUSEUM OF NATURAL HISTORY Number 6, pages 1-214, 48 figures, 58 tables, 1 appendix Issued 7 November 1978 Price: $15.00 a copy Craig C. Black, Director Editorial Staff: Hugh H. Genoways, Editor-, Duane A. Schlitter, Associate Editor-, Stephen L. Williams, Associate Editor-, Teresa M. Bona, Technical Assistant. ® 1978 by the Trustees of Carnegie Institute, all rights reserved. CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE PITTSBURGH, PENNSYLVANIA 15213 PARTICIPANTS Karl F. Koopman Honorary Chairman American Museum of Natural History Central Park West at 79th Street New York, New York 10024 U.S.A. I. R. Aggundey National Museum of Kenya P.O. Box 40658 Nairobi, Kenya Wim Bergmans Instituut voor Taxonomische Zoologie Universiteit van Amsterdam Plantage Kerklaan 36 NL 1018 CZ Amsterdam, The Netherlands Craig C. Black Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, Pennsylvania 15213 U.S.A. Donald P. Christian The Museum Michigan State University East Lansing, Michigan 48824 U.S.A. Gordon B. Corbet British Museum (Natural History) Cromwell Road London SW7 5BD, England Fritz Dieterlen Staatliches Museum fiir Naturkunde Schloss Rosenstein 7000 Stuttgart 1, West Germany M. Brock Fenton Department of Biology Carleton University Ottawa, Ontario KIS 5B6 Canada David L. Harrison Harrison Zoological Museum Bowerwood House St. Botolph’s Road Sevenoaks, Kent, England Robert J. Baker The Museum Texas Tech University Lubbock, Texas 79409 U.S.A. Ian Bishop British Museum (Natural History) Cromwell Road London SW7 5BD, England Pierre Charles-Dominique Museum National d’Histoire Naturelle Laboratoire d’Ecologie Generale 4 Av. Petit Chateau 91800 Brunoy, France C. G. Coetzee State Museum P.O. Box 1203 Windhoek 9100, South West Africa Michael J. Delany School of Environmental Science University of Bradford Bradford BD7 IDP, England Mohammed El Rayah Sudan Natural History Museum P.O. Box 321 University of Khartoum Khartoum, Sudan Peter Grubb Department of Zoology University of Ghana P.O. Box 67 Legon, Accra, Ghana Hendrik N. Hoeck Max-Planck-lnstitut fiir Verhaltens- physiologie D-8131 Seewiesen, West Germany Bernard Hubert Laboratoire de Zoologie Appliquee O. R. S. T. O. M. B. P. 1386 Dakar, Senegal Clyde Jones National Fish and Wildlife Laboratory National Museum of Natural History Washington, D.C. 20560 U.S.A. Lan A. Lester Natural History Museum of Los Angeles County 900 Exposition Boulevard Los Angeles, California 90007 U.S.A. Xavier Misonne Institute Royal des Sciences Naturelles de Belgique La Rue Vautier, 31 Bruxelles, 1040, Belgium J. A. J. Nel Department of Zoology University of Pretoria Pretoria 0002, Republic of South Africa Nathan O. Okia Department of Zoology Makerere University P. O. Box 7062 Kampala, Uganda Randolph L. Peterson Royal Ontario Museum 100 Queen’s Park Toronto, Ontario MSS 2C6 Canada Galen Rathbun Office of Zoological Research National Zoological Park Smithsonian Institution Washington, D.C. 20560 U.S.A. Jose M. Rey Instituto “Jose de Acosta’’ Museo de Ciencias Naturales Castellana 84 Madrid, Spain Jennifer U. M. Jarvis Zoology Department University of Cape Town Rondebosch 7700, Republic of South Africa Dieter Kock Natur-Museum und Forschungs-Institut Senckenberg Senckenberg Anlage 25 6 Frankfurt-Main 1, West Germany J. A. J. Meester Department of Zoology University of Natal P.O. Box 375 Pietermaritzburg 3200, Republic of South Africa lyad A. Nader College of Education at Abha P.O. Box 157 Abha, Saudi Arabia Francis Fetter Museum National d’Histoire Naturelle 55 rue de Buffon 75005 Paris, France Eyo E. Okon Department of Biology University of Ife Ife-Ife, Nigeria Urs Rahm Naturhistorisches Museum Augustinergasse 2 CH-405I Basel, Switzerland I. L. Rautenbach Transvaal Museum P.O. Box 413 Pretoria 001, Republic of South Africa C. Brian Robbins Division of Mammals National Museum of Natural History Washington, D.C. 20560 U.S.A. Lynn W. Robbins Museum of the High Plains Fort Hays State University Hays, Kansas 67601 U.S.A. Pierre Swanepoel Kaffrarian Museum King William’s Town, 5600 Republic of South Africa Erik van der Straeten Laboratorium voor Algemene Dierkunde Rijksuniversitair Centrum van Antwerpen Groenenborgerlaan, 171 B2020 Antwerpen, Belgium Duane A. Schlitter Organizing Chairman Carnegie Museum of Natural History 4400 Forbes Avenue Pittsburgh, Pennsylvania 15213 U.S.A. Thomas T. Struhsaker New York Zoological Society Bronx, New York 10460 U.S.A. Kamal M. Wassif Department of Zoology Faculty of Science Ain Shams University Abbasia, Cairo, Egypt 1^ k. If. i 4 i •'»1 N RUMP REGION EVOLUTION 17 Fig. 6. — Location of 13 dermal scars (five female and eight male) from 43 female and 41 male Rhynchocyon chrysopygiis study skins. a predator, which is signalled by the rump display. The dispersed nature of the elephant-shrews, to- gether with their flight in different directions when disturbed, does not support the antipredator cohe- sive signal function for the golden rump of R. chry- sopygiis . Smythe (1970«, 1977) suggests that rump patches may serve as predator "pursuit invitation signals." By stimulating a predator to attack prematurely, a healthy individual is able to flee successfully and the location and intent of the predator are revealed, thereby reducing the danger of ambush. Smythe (1970«) suggests that stotting-like gaits and rump patch displays are used to invite pursuit. R. chry- sopygiis is probably frequently preyed upon by the harrier hawk, which hunts visually by perching 2 to 4 m above the relatively open forest floor waiting for an opportunity to make a quick, concealed am- bush (Rathbun, in press). This is precisely the type of predation needed for visual invitation signals to evolve. The structure of the Kenyan coastal forest results in direct-line visibility being restricted to about 10 m, which is conducive to the evolution of auditory invitation signals. I believe the golden- rumped elephant-shrew's rump patch and behav- ioral complex of stotting, rear leg hammering, tail slapping, and short, obvious flight followed by a pause, are directed toward diurnal bird and mam- mal predators in an attempt to attract their attention and induce their exposure and/or premature attack, as proposed by Smythe (1970u, 1977). The three allopatric species of Rhynchocyon oc- cupy similar habitats, and except for distinctly dif- ferent body coloration, their morphology is very similar (Corbet and Hanks, 1968). Kingdon (1974) has suggested several explanations for the variation in color. He thinks the checkered back and rump of R. cirnei may be an adaptation for camouflage. 18 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 whereas the difference between Rhynchocyon pe- tersi, with its pure black back and rump, and R. chtysopygus may be related to differences in sexual behavior which evolved as an "ethological bar- rier.” More information is needed on the ecology and behavior of the Rhynchocyon species in order to verify these ideas. I have discussed briefly some hypotheses which might conceivably explain the evolution of the dermal shield and the rump patch in R. chry- sopygus as though the two structures were unrelat- ed in this animal. The dermal shield lies directly beneath the rump patch (Fig. 4) and 1 feel that there is probably an integrated function in the two struc- tures. The rump patch may serve as a conspecific target organ, attracting the blows of a pursuing el- ephant-shrew to a region of the body, which is not only relatively immune to serious injury, but also morphologically adapted to receive the blows. This is a similar mechanism to the “deflective marks” extensively described by Cott (1940), where a pred- ator is attracted to a less vital part of the body by a very distinctive coloration or marking. In the case of the elephant-shrew, the predator is a conspecific individual defending its territory through aggression and the distinctive mark is the golden rump. CONCLUSIONS I think the data for R. chrysopygiis support a multifactor explanation for the evolution of its rump patch. The location of the dermal shield, the scar- ring pattern, and the aggressive behavior indicate that the rump patch, with its very distinctive col- oration, may serve as a conspecific target organ. On the other hand, because there is no significant sexual dimorphism in the rump patch corresponding with the rump skin, this suggests that the patch may have an additional, even distinct, function. The el- ephant-shrew’s attention-gathering behavior while fleeing a potential predator, and its distinct rump patch, support Smythe’s predator pursuit invitation hypothesis (1970u, 1977). Predator selection pres- sure probably acts equally on both sexes, resulting in the male and female rump patches being equal in size. Thus two distinct selection pressures, conspe- cific territorial aggression and diurnal predation, are probably involved in the evolution of the rump patch in R. chrysopygiis . It is unlikely that the dual rump patch function proposed for the elephant-shrew can be used as a general model for mammalian rump patch evolu- tion. It is inappropriate to try to apply models de- veloped for relatively social animals, such as many of the larger ungulates (Hirth and McCullough, 1977), to less social species. It would, however, be interesting to compare mammals that have similar life histories, such asR. chrysopygiis and an agouti, Dasyprocta . The latter inhabit South American for- ests, and, like R. chrysopygiis , are diurnal, curso- rial, territorial, probably monogamous, and some bear a distinct rump patch (Morris, 1962; Smythe, 1970^). The yellow-back duiker, Cephalophiis sil- viciiltor, may also be comparable to R. chrysopy- giis. If the rump patch functions proposed above for the golden-rumped elephant-shrew are valid, and Geist (1971), Alvarez et al. (1976), and Hirth and McCullough (1977) are correct in their interpreta- tion of rump patch function in the species that they consider, it would not only seem fairly certain that rump patches have evolved independently several times within the Mammalia (Guthrie, 1971), but probably also due to several different selection pressures. ACKNOWLEDGMENTS I am grateful for the support of the National Geographic So- ciety, East African Wildlife Society, and the National Museums of Kenya. Mr. Kashuru Mumha and his family greatly assisted in the collection of specimens. Drs. H. Croze, P. Jarman, H. Kruuk, and A. Walker offered stimulating ideas and direction to the rump patch region aspects of the elephant-shrew project. which was undertaken as part of a Doctor of Philosophy program in the Department of Zoology, University of Nairobi, Nairobi, Kenya. The participants of the colloquium as well as Dr. Devra Kleiman and Ms. Carolyn Dorsey generously offered helpful comments on drafts of this paper. The National Zoological Park kindly supported my participation in the colloquium. 1978 RATHBUN— 7?//F7VC//6>CFON RUMP REGION EVOLUTION 19 LITERATURE CITED Ansell, W. F. H., AND P. D. B. Ansell. 1973. Mammals of the northeastern montane areas of Zambia. The Puku, 7:21-69. Alvarez, F., F. Braza, and A. Narzagaray. 1976. The use of the rump patch in the fallow deer (Duma damn). Behav- iour, 56:298-308. Corbet, G. B., AND J. FIanks. 1968. A revision of the elephant- shrews, family Macroscelididae. Bull. British Mus. Nat. Flist. (ZooF), 16:47-113. CoTT, FI. B. 1940. Adaptive coloration in animals. Methuen, London, 508 pp. Dubost, G., and R. Terrade. 1970. La transformation de la peau des TniguUdae en bouclier protecteur. Mammalia, 34:505-513. Eisenberg, J. F., AND E. Gould. 1970. The tenrecs: a study in mammalian behavior and evolution. Smithsonian Contr. ZooF, 27:1-138. Eisenberg, J. E., and M. LocKitARi . 1972. An ecological re- connaissance of Wilpattu National Park, Ceylon. Smithson- ian Contrib. ZooF, 101:1-118. Geis i , V. 1967. On fighting injuries and dermal shields of moun- tain goats. J. Wildlife Mgmt., 31:192-194. . 1971. Mountain sheep: a study in behavior and evolu- tion. Univ. Chicago Press, Chicago, xv + 383 pp. Gui hrie, R. D. 1971. A new theory of mammalian rump patch evolution. Behaviour, 38:132-145. Hirth, D. FI., AND D. R. McCullough. 1977. Evolution of alarm signals in ungulates with special reference to white- tailed deer. Amer. Nat., 111:31-42. Jarman, P. J. 1972. The development of a dermal shield in im- pala. J. ZooF, London, 166:349-356. Kingdon,J. 1974. East African mammals: an atlas of evolution in Africa. Academic Press, London, 2A: xi + 1-392. Kleiman.D.G. 1974. Patterns of behaviour in hystrichomorph rodents. Symp. ZooF Soc. London, 34:171-209. Morris, D. J. 1962. The behaviour of the green acouchi (My- oproctd pnitti) with special reference to scatter hoarding. Proc. ZooF Soc. London, 139:701-732. Raihbun, G. B. 1976. The ecology and social structure of the elephant-shrews Rhynchocyon clirysopygiis Gunther and Elephantulus rufescens Peters. Unpublished Ph.D. thesis, Univ. Nairobi, Kenya, 236 pp. . In press. The ecology and social structure of elephant- shrews. Z. TierpsychoF, Advances in Ethology Supplement. Smmhe.N. I970u. On the existence of “pursuit invitation" sig- nals in mammals. Amer. Nat., 104:491-494. . 19706. Ecology and behavior of the agouti (/)u.vy/)/-()c7« pinwtalu ) and related species on Barro Colorado Island, Pan- ama. Unpublished Ph.D. thesis, Univ. Maryland, College Park, 123 pp. . 1977. The function of mammalian alarm advertising: so- cial signals or pursuit invitation .’ Amer. Nat., 111:191- 194. REVIEW OF DRINKING BEHAVIOR OF AFRICAN FRUIT BATS (MAMMALIA: MEGACHIROPTERA) WIM BERGMANS Instituut voor Taxonomische Zodlogie (Zoologisch Museum), Plantage Kerklaan 36, NL 1018 CZ, Amsterdam, The Netherlands ABSTRACT Certain captive African Megachiroptera drink regularly from a hanging posture. There are very few reports of these animals drinking in the wild, nearly all claiming that they would drink by skimming over open water surfaces and scooping water or wet- ting their fur and licking water from this afterward. This paper summarizes and critically deals with the literature on the subject, and presents some new evidence suggesting that the skimming maneuvers of some fruit bats might serve a purpose other than drinking. The discussion includes some related observations on Australasian Megachiroptera. INTRODUCTION Little has been written about possible drinking habits of African Megachiroptera in the wild. About fruit bats in general Novick (in Novick and Leen, 1969) wrote; “Fruit bats like the flying foxes and their smaller relatives rarely drink water, apparent- ly they derive all the water they need from their moist food." Rosevear (1965) stated about West African fruit bats that "the demand of the body for water must be satisfied ... to a consid- erable extent by the intake of liquified fruit pulp or floral nectar." Indeed one might expect that these animals, whose diet probably consists mainly of juicy fruits and equally water-rich nectar, would not need to drink water too. There are reports however which suggest that at least some wild fruit bats do drink. Two ways of drinking have been described, the first for captive animals and the second for an- imals in the wild. Quite a few species of Megachi- roptera have been kept in captivity, but hardly any author paid attention to whether or not his animals drank. Notwithstanding discussions of their diet, it may not even be stated if water was offered (for example, Blackwell, 1966, 1967; Brosset, 1966r/; Coe, 1976; Jones, 1972; Lawrence and Novick, 1963). DRINKING RECORDS Kulzer ( 1958, 1969) related how his captive Rou- settiis aegyptiacits (Geoffroy, 1810) and Eidolon helvnm (Kerr, 1792) drank water from bowls, rap- idly with their tongues, in a dog-like manner. Ron- settns came down to the cage floor to drink, but Eidolon preferred to remain hanging. This is most probably related to the ability of Rouse tins aegyp- liacus to take flight from an even surface without effort, as I have observed in the zoo in Amsterdam where a small number of unknown provenance is kept (1977), and the apparent lack of this ability in Eidolon (Kulzer, 1968). At Windhoek, Namibia, Dr. C. G. Coetzee kept a crippled specimen of Eidolon helvnm, which drank regularly from a bowl (per- sonal communication, 20 September 1977.) Brosset (19666) observed wild Eidolon helvnm, probably in northeast Gabon, drinking from the sap of palm trees, which had been collected by Africans for the production of palm wine. As far as I know this is the only record of wild African Megachiroptera drink- ing in a manner more or less comparable to that observed by Kulzer ( 1958, 1969) in his captive spec- imens. The other manner of drinking described for some Megachiroptera is to skim close to the surface of water and, in some way or other, take in water. In 1841, Prof. J. A. Wahiberg observed unidentified fruit bats near Port Natal ( = Durban), South Africa, drinking in this way (Petermann, 1858). Rousselot ( 1950) related how he saw Eidolon helvnm and Epomophorns gamhianns (Ogilby, 1835) flying so close to the surface of a river that they could wet 20 1978 BERGMANS— MEGACHIROPTERAN DRINKING BEHAVIOR 21 their ventral fur from which they would lick the water afterward. He observed Eidolon doing so at Segou in Mali, and observed Epomophorns gam- hianns both at Segou and at Maradi in Niger but does not state where he saw this species perform in this manner. Rosevear (1965) suggested that in- sectivorous bats would exhibit this behavior much more often than fruit bats would, but does not seem to have observed fruit bats drinking in this way him- self. Novick (in Novick and Leen, 1969) mentioned the habit for insectivorous bats but, as appears from the above quotation, remained equally vague about fruit bats on this point. The ornithologist Mr. Roy Parker, at the time Curator of the Zoology Museum of the University of Ibadan, Nigeria, told me that he had often seen how Eidolon helviun came down at dusk to the water of the swimming pool of the Staff Members Club on the campus of the univer- sity. He believed the animals to be drinking (per- sonal communication, August 1976). Dr. M. El Ray- ah saw Eidolon helviun skimming over the water of the Nile at Khartoum in Sudan, but did not observe further details (personal communication, 20 Sep- tember 1977). At Makokou, northeast Gabon, Mr. J. Bradbury kept Hypsignathns monstrosns Allen, 1861, in captivity in a large enclosure with a large bowl of water on the floor. This bat regularly drank by scooping water with its mouth while flying low over the bowl (Dr. P. Charles-Dominique, personal communication, 20 September 1977). PADDLING BATS When traveling through Nigeria in the summer of 1976 I stayed for 2 weeks in Pandam Wildlife Park, a small game reserve just north of the village of Pandam (9°15'N, 7°50'E) that includes the 2 square mile Lake Pandam. There I enjoyed the hos- pitality of the zoologists Chris Smeenk and his wife Nellie, who told me of their observations of Eidolon helviun skimming over the lake. According to them. Eidolon did not drink when close to the water sur- face but touched it with its legs. On 7 July 1976, the three of us made a canoe trip on the lake for a more detailed study of this phenomenon. At twilight, from approximately 1840 to 1920 h, hundreds of Eidolon approached the lake, all coming from the southwest and eventually continuing roughly to the northeast. Only a few bats continued their flight without interruption. Most of them, when they had reached open water (much of the lake was covered with water lilies), came down to it in one circular movement (infrequently two or three circles were described). Then they flew parallel with and very close to the water surface for a short time and at a certain moment dipped their hind legs into the water. This dipping lasted much less than a second. Thereafter they flew up and continued on their way. A striking detail was the apparent hesitation, which most bats showed in their movements before dip- ping. We observed several individuals, who at the last moment refrained from dipping and continued on their way without having dipped. However, we also saw a small number of bats who dipped more than one time. These continued flying low over the water after their first dip, and dipped up to four times. On one occasion I counted five dips. These repeated dips followed shortly upon one another, reminding one of a well flinched drake-stone. After that these bats also resumed their flight, undoubt- edly towards their feeding grounds. The bats flew in small groups of a few dozens or less, arriving at different times during about 40 min, and all these groups displayed the described behav- ior. This enabled us to focus on new individuals and observe this behavior many times. Nevertheless, it was very hard to establish what was actually hap- pening. The bats kept considerable distance, ap- parently avoiding our presence. It proved moreover very difficult to observe correctly what was hap- pening at the division of air and water, and of course our sight was hindered by the gradual fading of light. Later 1 heard of a possibly similar observation. The zoologist Rob Robelus told me that he had watched large bats skimming over the water of a swimming pool at Bebedjia near Moundou, in southwest Tchad, in September 1969. These bats touched the water with some part of their ventral side, but were not believed to be drinking (Dr. R. A. Robelus, personal communication, July 1977). DISCUSSION AND CONCLUSIONS The more conventional and seemingly more like- Megachiroptera, that is, lapping while in a hanging ly way of drinking as described for certain African or clinging position, has been observed in the wild BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 only once (Brosset, 1966/0. Because in this case, the bats (Eidolon lielvnm) drank no water but the probably tasty sap of a palm tree, the question re- mains whether they were hunger- or thirst-induced. The habit shown by captive Rouseltiis aegyptiacus and Eidolon helvinn of drinking water regularly does not necessarily reflect natural behavior. Their de- mand for water may well have been brought about by a shortage of liquid in the food offered. Kulzer (1958, 1969) gave his Roiisettns bananas, apples, and oatmeal porridge, and his Eidolon bananas and figs. In this respect, it is possibly of significance that his Ronsettns always drank immediately upon feeding (Kulzer, 1958). On the other hand, Ronset- tns aegypticns in the Amsterdam zoo, which are given a much more varied diet with juicier fruits, also drink regularly (Mr. F. Gangel, personal com- munication, August 1977). The fact that wild Afri- can Megachiroptera drink regularly, and in a similar manner, is still to be proven. The other manner of drinking water, by scooping or lapping it directly while skimming close to the water surface or indirectly by wetting the fur while skimming over the water and licking water from the fur afterward, has so far been ascribed to three Af- rican species-Eponiopliorns gainhianns , Eidolon helvnin, and Hypsignathns monstrosns (Rousselot, 1950; Parker, personal communication, 1976; Charles-Dominique, personal communication, 1977). The first observation, of unidentified species, was made by Prof. J. A. Wahlberg in 1841 near Port Natal, South Africa. The first part of his account reads (in translation): "Port Natal, October 5th, 1841 .... Some time ago I shot two species of Pteropns, one of which I think may be new (pos- sibly a new genus); it has four upper and five lower cheek teeth." Thereafter he relates how he shot them: in the evening over a spring where they came to drink while flying low over the water (in Peter- mann, 1858). Unfortunately he does not mention the name of the species he knew, which would have given a clue to the identity of the other, supposedly new species. Of the fruit bat species now known to occur there, only Eidolon lielvnm and Ronsettns aegyptiacus had been described at the time of Wahlberg’s writing. Lock’s assumption (1969) that an Epoinoplwrns species was involved here is prob- ably based on the fact that the holotype of Epo- mopliorns wahlhergi (Sundevall, 1846) had been col- lected by Wahlberg near Port Natal. However, this happened 2 years after Wahlberg's observation and shooting of skimming bats, namely on 27 No- vember 1843 (Andersen, 1912). Moreover, it is clear from Wahlberg’s report that his observation applied to both "species” shot by him, one of which he knew (thus probably Eidolon or Ronsettns), and one of which was new to him and which had four upper cheek teeth instead of three as in Eponwpho- rns . To my knowledge, cases of additional upper cheek teeth have not been described for this genus (compare the footnote in Andersen, 1912:516). I think, therefore, that Wahlberg had before him a sub- adult specimen of Eidolon helvnni or of Ronsettns aegyptiacus with unerupted M“ and Mg. Wahlberg’s bat collection, now in the Naturhistoriska Riks- museet in Stockholm, did not contain other Mega- chiroptera than the just mentioned holotype of Eponwphorns wahlbergi (Dr. C. Edelstam, letters of 1 April 1976 and 12 September 1977), although it is not impossible that part of his material was lost, as were many of his original notes (Petermann, 1858). Precise accounts of what African Megachiroptera actually do when skimming over water are very few. In three cases, they are reported to drink somehow directly (Wahlberg, in Petermann, 1858; Parker, this paper; Bradbury, this paper). On two occasions, they were seen to dip with some ventral part of their bodies and definitely not with their snouts (Rousselot, 1950; Robelus, this paper), while one of these observers reported that they lick water from their fur afterward (Rousselot, 1950). On yet another occasion three observers saw Eidolon liel- vnin dipping with its legs (this paper). Wahlberg went to the spring where he saw the skimming bats to shoot them in the first place. Moreover he had fires lit at the sides of the spring to enable him to shoot bats coming near the fire, because otherwise he could not hit them. When looking past these fires at the skimming bats, his chances to observe accurately can hardly have been great, to say the least. I know that Parker made his observations without binoculars, while the bats were skimming the water of a scarcely illuminated swimming pool, and he himself was seated at some distance on a well-illuminated terrace. Again, the circumstances were rather poor. Rousselot’s para- graph (1950) dealing with skimming Eidolon at Se- gou is worth quoting (in translation): "... they fly very low over the water and only wet the fur of their breast and belly. Having risen immediately thereafter for two or three meters, they lick this wet fur, simultaneously interrupting their flying move- ments. Then they rise to make up for the lost height 1978 BERGMANS— MEGACHIROPTERAN DRINKING BEHAVIOR 23 and lick again. This happens three or four times, after which they repeat the whole maneuver, until they are satisfied." He wrote that Epomophorus gamhianns acted in the same way, but where he observed this is not stated. With regard to Ei- dolon, it looks very much like Rousselot may have seen what my companions and I saw at Lake Pan- dam in Nigeria, in which case the aberrant details in his story could be due to careless observation and interpretation. We observed a flying height of at most 30 to 50 cm. Most of the bats we followed dipped (their legs) only one time, some up to four or five times, and they touched the water each time they were near it (we actually saw the splashes). Their movements down to the water were active and deliberate. We observed no interruptions of wing movements, but of course, the bats had to brake their flight in order to descend. We did not see individuals flying more than one stretch at the reported height. All in all, the evidence that certain wild African Megachiroptera skim over open water surfaces in order to drink is meager and, in part, doubtful. Of course, it is easily conceivable that at least some spe- cies do drink water under certain conditions. Forest species might lap water from small reservoirs of different types, such as holes in trees, but popula- tions inhabiting the dryer savanna regions may have to search for larger bodies of water, such as lakes and streams. This may be true for Eidolon helvitm in particular, because this species roosts exposed to the sun, unlike other African Megachiroptera, and may need extra water to make up for its po- tentially stronger transpiration. However, why should they drink by means of the seemingly haz- ardous skimming maneuvers instead of getting at the water through climbing down to it along tree branches? This latter behavior has been reported for wild specimens of an unidentified Pteropns spe- cies from Sumatra (van Balen, 1914). Ratcliffe (1961) saw flying foxes (most probably species of Pteropns) drinking by the skim-and-scoop method in New South Wales on the Nambucca River in 1930, and offshore on Bougainville, Solomon Is- lands, in 1945. He tried to explain this by assuming that these bats were drinking salt (or salty) water for the sake of mineral salts, which their normal food would not supply in adequate quantities. His observations on the Bougainville bats reminded me of our own of Eidolon at Lake Pandam. He also emphasized the importance of focusing on individ- ual bats to obtain an idea of what they were doing. He observed large groups of bats flying to the sea just after sunset. Each bat would cruise up and down for a while, and then fly right down to the sea once or twice (some three times) and then leave the flying mob and head inland, presumably to set out on its nightly food search. If the water was calm enough, he and his co-observer did see the actual lap-splashes. From the latter remark in his account, it is clear that it must have been very difficult to see (after sunset) whether the bats were scooping water or wetting their fur or dipping their legs. Another interesting report is that of Mr. J. V. de Bruijn (in Ripley, 1960), who at the north coast of New Guin- ea observed large flying foxes trying to pick up fruits, which were floating on the sea surface just offshore. Ratcliffe (1961) thought that these bats may also have been drinking salt water instead of picking up fruits. I do not know of other reports on Megachiroptera outside Africa skimming over water in order to drink. There must be more, be- cause in a popular account on Indonesian flying fox- es van Bemmel ( 1974) wrote (in translation): "Their habit of scooping water while flying low over its surface is well-known." The salt-water story does not hold for the fresh-water skimming African bats, and in general I doubt whether Ratcliffe’s conclu- sion is correct. Many fruit bats will never have an opportunity to drink sea water, and quite a few, apparently, skim over fresh water. Moreover, the wild Sumatran Pteropns mentioned by van Balen (1914) drank fresh water, albeit from a hanging pos- ture, and captive Pteropns gigantens (Briinnich, 1782) are reported to "drink often, especially during the summer" (Sanyal, 1892). The question remains why Eidolon, at least on one well-documented occasion, dipped its legs. One can think of several explanations, none of which seems good enough. Ronsettns aegyptiacns in the Amsterdam zoo not only drinks from its water bowl, but also uses it to take baths. A freshly bathed bat licks itself and is licked by the others (Mr. F. Gangel, personal communication, August 1977). This could hardly be a way of drinking, as the animals also drink "normally." Could it be a means to clean the body, maybe even of certain parasites, or could it have some thermoregulatory function? If one of these questions would even- tually produce a satisfying explanation with re- gard io Ronsettns , would this explanation also sat- isfy us with regard to the wild, leg-dipping Efi/o/o/;? I do not think so. In this regard. Eidolon helvnin is easy to study, as it roosts in large numbers and fully 24 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 exposed near many human settlements. It is known that Eidolon passes its day, like so many other mammals, partly with grooming its body and fur. It is also sufficiently known that common external parasites of Eidolon , such as the nycteribiid fly. Cyclopodia greeffi Karsch, and the spintumicid mite, Meristaspis sp. (Adeosun, 1974), have no prefer- ence whatsoever for the legs or lower abdomen of their hosts. Moreover these parasites attach them- selves so firmly to the skin of the bat that a little water will not harm them in the least. Likewise, a thermoregulatory purpose of the wetting of so small a part of the body is hard to understand. Another possibility is that the water itself does not play the important role, which we have tried to attach to it. Could Eidolon be picking something out of the water (as there were, on Lake Pandam, hardly any fruits, these can be disregarded)? Could Eidolon possibly be fishing? Carnivorous behavior has recently been reported of two other African Megachiroptera. Van Deusen (1968) cited someone, who observed wild Hypsignatliiis monstrosns pick up and eat the skinned bodies of birds, which had been tossed outside a house, and also how it attacked live chickens near the same house, at M’Bigou in Gabon. Coe ( 1976) kept Lissonycteris angolensis (Bocage, 1898) in a cage together with Myonycteris tor quota (Dobson, 1878), at Mount Nimba, Liberia, and saw how specimens of the former species actually attacked and devoured specimens of the latter. It is of interest to quote here (in translation) the report on a Sumatran Pter- opiis by van Balen to which I referred before in relation to its drinking behavior. “My son Mr. D. J. van Balen, living at Serbadjadi (Galang), Deli, observed that kalongs (=flying foxes) do drink. He used to spend the night now and then on a platform in a forest tree to watch animals. One night he saw a number of flying foxes alight in a dead tree, climb down along branches which hang over the water of a deep pool, and drink while hanging from these branches. After drinking they let themselves fall onto the surface of the water, where they stayed for some moments, constantly flapping their wings, and then flew up in the limited space in circular movements until they could take flight in the open air above the trees." This observation was repeated on request, and I think it should be taken seriously. Later, van Balen’s son was told by a native that the flying foxes did not descend to the pool to drink, but to catch the little fishes which abounded in such pools. Dr. F. A. Jentink, the well-known Dutch mammalogist, wrote to van Balen that he did not believe that kalongs drank, and he warned him against native stories. But why then, asked van Bal- en, would kalongs perform such breakneck stunts? (van Balen, 1914). Apparently this belief in fishing kalongs is more widespread (though not so much in the literature), as van Bemmel ( 1974) also mentioned it (and re- jected it). Apart from the hint at possible piscivo- rous behavior of Pteropiis van Balen’s account ( 1914) deals with another aspect of the relation be- tween Megachiroptera and water. It confirms that they, or at least certain Pteropus species, can swim. Ryberg (1947) cited Trouessart’s story on how, dur- ing James Cook’s second journey around the world (1772-1775), swimming Megachiroptera were ob- served in the Tonga Archipelago. These were prob- ably Pteropus tonganns Quoy & Gaimard, 1830. Some of the fruit bats seen picking up fruits from the sea surface in North New Guinea (Ripley, 1960) “dipped too far into the water, became caught and splashed in. The instant this happened, the flying foxes came to rest quietly on the surface with wings well spread. Thence they gradually rode in on the waves of the rising tide. The sea was not too rough, and gradually they tumbled into the light surf and quickly proceeded to drag themselves out of the wave area. Without exception, the bats then crawled up the beaches, occasionally shaking them- selves, and made for the low dunes above high water mark." Beyond these dunes the bats climbed up small trees and launched themselves into the air again. Novick and Leen ( 1969) gave a splendid se- ries of photographs of a swimming Pteropus gigan- teiis (Briinnich), but unfortunately it is not stated where they were taken. Van Bemmel ( 1974) knew of a Pteropus kept as a pet on a houseboat on Su- matra. This animal once fell into the water. There- after, it made a daily habit of plunging into the water and swimming about for a while (until, 2 years later, it was seized by a crocodile). Summarizing, it has not yet been sufficiently doc- umented that, in the wild, African Megachiroptera need to drink water. In captivity Rousettus aegyp- tiacus and Eidolon lielvum may drink from bowls, and Hypsignathus monstrosns is reported to drink through the skim-and-scoop method. In the wild Eidolon helvum and Epomophorus gambianus have been seen to skim over open water, but none of the reports in which they would do so in order to drink seem to stem from really careful observation. The behavior of captive Hypsignathus monstrosns and 1978 25 BERGMANS— MEGACHIROPTERAN DRINKING BEHAVIOR of many wild Microchiroptera supports the idea that physically Megachiroptera may be able to drink while skimming. Eidolon helviim has been seen to skim over open water not to drink but to dip its legs for an unknown purpose. Of skimming and swim- ming Australasian species of the genus Pteropns (which in its ecology and ethology shows a great resemblance to Eidolon) it has been suggested that they might be fishing, and as carnivorous behavior has been reported for some African fruit bats, I do not exclude fishing as a possible explanation for the leg-dipping of Eidolon. Against this would be that Eidolon is also attracted by swimming pools, but I do not suggest that drinking as one of the possible aims of skimming Megachiroptera should be ne- glected. More critical observation in the wild, and possibly some experimenting with captive fruit bats (large enclosures with large, fish-inhabited water containers on the floor; small, whole fishes as part of their diet) should provide the final answers. LITERATURE CITED Adeosun, D. a. 1974. Nycteribiid and Spintumicid ectopara- sites of bat. Eidolon helvum, Kerr, roosting on campus of University of Ife, lle-lfe, Nigeria. Unpublished thesis, Univ. Ife, Ife, Nigeria, 48 pp. Andersen, K. 1912. Catalogue of the Chiroptera in the collec- tion of the British Museum. 1. Megachiroptera. British Mu- seum (Nat. Hist.) London, ci -t- 854 pp. Balen, J. H. VAN. 1914. De dierenwereld van Insulinde. Thieme & Cie, Zutphen, 6 -I- xi -f 505 pp. Bemmel, a. C. V. VAN. 1974. Kalongs. Artis, 20:75-79. Blackwell, K. 1966. A note on fruit bats in free-flight aviaries. Internat. Zoo Yearbook, 6:247. . 1967. Breeding and hand-rearing of fruit bats Epomops friiiupieti and Eidolon helvum at Ife University. Internat. Zoo Yearbook, 7:79-80. Brosset, a. 1966(1. Les Chiropteres du Haut-Ivindo (Gabon). Biol. Gabon., 2:47-86. . 19666. La biologie des chiropteres. Mason, Paris, viii -I- 240 pp. Coe, M. J. 1976. Mammalian ecological studies on Mount Nim- ba, Liberia. Mammalia, 39:523-587. Jones, C. 1972. Comparative ecology of three pteropid bats in Rio Muni, West Africa. J. Zook, London, 167:353-370. Kock, D. 1969. Die Fledermaus-Fauna des Sudan. Abh. Senck- enb. Naturforsch. Ges., 521:1-238. Kolb, A. 1975. Schwimmen, Schwimmtechnik und Auffliegen vom Wasser bei einheimischen Fledermausen. Ber. Naturf. Ges. Bamberg, 50:75-88. Kulzer, E. 1958. Unlersuchungen iiber die Biologie von Flug- hunden der Gattung Gray. Z. Morph. Okol. Tiere, 47: 374-402 . . 1968. Der Plug des afrikanischen Flughundes Eidolon helvum. Nat. Mus., Frankfort, 98:181-194. . 1969. Das Verhalten von Eidolon helvum (Kerr) in Ge- fangenschaft. Z. Saugetierk., 34:129-148. Lawrence, B., AND A. NoviCK. 1963. Behavior as a taxonomic clue: relationships of Lissonycteris (Chiroptera). Breviora, 184:1-16. Novick, A., AND N. Leen. 1969. The world of bats. Holt, Rine- hart, and Winston, New York, 172 pp. Petermann, a. 1858. Aus den Briefen des Siid-Afrikanischen Reisenden Prof. Wahlberg. Mitt. Perthes' Geogr. Anst., pp. 414-418. Ra i ci iffe,F. 1961. Flying foxes drinking sea water. J. Mamm., 42-252-253 Ripley, S. D, 1960. Behavior of flying foxes in coastal New Guinea. J. Mamm., 41:264-265. Rosevear, D. R. 1965. The Bats of West Africa. British Museum (Nat. Hist.), London, xvii + 418 pp. Rousseloi,R. 1950. Les rousettes du Soudan (Animaux nuisi- bles). C. r. First Conf. Int. Afr. Guest, 1:233-238. Sanyal, R. B. 1892. A hand-book of the management of ani- mals in captivity in Lower Bengal. Bengal Secretariat Press, Calcutta, 351 pp. Van Deusen.H. M. 1968. Carnivorous habits of //v/w(To;«t/»/.s monsirosu.s . J. Mamm.. 49:335-336. ECOLOGICAL POSITION OP THE PAMILY LORISIDAE COMPARED TO OTHER MAMMALIAN PAMILIES PIERRE CHARLES-DOMINIQUE Laboratoire de Primatologie et d'Ecologie Equatoriale, C.N.R.S. and Museum National d’Histoire Naturelle, 4, avenue du Petit Chateau 91800 Brunoy, Prance ABSTRACT African primates of the family Lorisidae are compared to other families of African mammals from the perspective of ecological niche occupancy. It is shown that lorisids, along with other ro- dents and carnivores, occupy an ecological niche characterized by arboreal-nocturnal-climbers and/or leapers-animalivores-fru- givores-gumivores habits (A.N.C.L.A.F.G.). Generally, mam- mals (especially primates and carnivores) occupying this type of niche exhibit many primitive evolutionary characters. INTRODUCTION Among primates, Strepsirhini (Afro-Asian and Madagascan lemurs) represent 18% of the living species. In terms of their anatomy, physiology, and behavior, they are considered more primitive than the “higher primates” (Catarhini and Platirrhinii). Although five Malagasy families (of which two are subfossils) are the result of a radiative evolution, the Cheirogaleidae and the Lorisidae (in Africa and South Asia) have retained numerous primitive fea- tures (Charles-Dominique and Martin, 1970). The taxonomic position of these two groups have been discussed by different authors (Cartmill, 1975; Charles-Dominique and Martin, 1970; Charles- Dominique, 1977; Goodman, 1975; Hoffstetter, 1974, 1977; Martin, 1972; Pocock, 1918; Simpson, 1945; Szalay and Katz, 1973; Szalay, 1975; Tatter- sall and Schwartz, 1975; Van Kampen, 1905; Van Valen, 1969; Weber, 1928). According to the inter- pretation of early primate history by each author, the Lorisidae and Cheirogaleidae have been clas- sified either in two different suborders, two differ- ent families, or two different subfamilies. Never- theless, even if the Cheirogaleidae and the Lorisidae have been geographically separated since the early Tertiary, they share numerous characters inherited from a common ancestor, and presently occupy similar ecological niches. In this paper, the ecological position of one of these two primitive families (Lorisidae) and its re- lationship to other mammalian groups living in the same forest ecosystem are discussed. Different eco- logical parameters shall be analyzed including diet, activity rhythms (nocturnality or diurnality), habitat preference, and locomotor patterns. EOOD PROVIDING ENERGY AND FOOD CONCERNED WITH THE RESTORATION OF TISSUES It is necessary to make a distinction between food which is used for energy (glucids) and food which balances the loss due to tissue turnover (for ex- ample, loss of proteins, see Hladik, 1977). Food providing energy. — Every molecule catab- olized by an organism can be used for energy. For example, strict carnivores and strict insectivores ( = “animalivores”) use glucids, lipids, and proteins for energy. Proteins are difficult to find in large quantities and many animals eat small quantities of proteins necessary for their balance, but use lipids and especially glucids as energetic food. Soluble glucids (C5, C6, and C12 sugars) are directly assim- ilable by a nonspecialized gut. They are found in fruits like berries, drupes, and so on (especially in fruits with a fleshy pericap around the stone). Non- soluble glucids (celluloses, hemicelluloses, and oth- ers) are made up of soluble sugars polymerized in complex chains. Mammalian enzymes cannot break these long chains, but they are hydrolized by bac- teria in specialized digestive ducts (stomach of ru- minants, saculated stomach, caecum). Celluloses and hemicelluloses are the principal constituents of leaves and wood. Gums are produced by some trees and lianas after different actions (insect bites, xy- lophages, other injuries). Composed of C5 and C6 sugars polymerized in long chains, they are biode- graded in the caecum by bacteria (for Prosimians). 26 1978 CHARLES-DOMINIQUE— ECOLOGICAL POSITION OE LORISIDAE 27 Food concerned with the restoration of tissue. — In addition to vitamins and mineral elements, which are generally present in sufficient quantities in nat- ural food, amino acids (associated in proteins) play a major role in the constitution of the tissues (turn- over, growth, gestation, and lactation). In the for- est, proteins can be found in prey, in kernels, and in the green part of plants (especially young shoots). These three sources are correlated with different specializations; prey capture is correlated with mor- phological and behavioral specializations of preda- tors; kernel and seed intake is correlated with den- tal specializations allowing opening of hard stones, and gut specializations for detoxification of second- ary compounds often present in seeds; leaf intake is correlated with the specialization of the digestive duct. This last solution, generally used by animals which can digest the cellulose, requires large quan- tities of bulky food relatively poor in assimilable proteins. As “energetic food,” proteins can be a limiting factor for animal populations. NOCTURNALITY AND DIURNALITY Theoretically, food resources are available around the clock but consumers are generally spe- cialized to feed during either the day or night (Charles-Dominique, 1975). It seems that among higher vertebrates, birds are specialized to the diur- nal way of life and mammals to the nocturnal. Of course, secondary specializations allow some birds to feed during the night and some mammals to feed during the day, but they are limited. Eor example, in the Gabon forest, 70% of the mammals are noc- turnal, 10% are nocturnal and diurnal, and 20% are strictly diurnal (among 120 species of mammals); in the same ecosystem, 96% of the 216 bird species are diurnal. The same picture exists in the tropical forest of Panama where 86% of the mammals are nocturnal, 6% nocturnal and diurnal, and 8% strict- ly diurnal (Eisenberg and Thorington, 1973). The diurnal mammals compete successfully in four dif- ferent ways, which can be listed as follows: in- crease in body size (monkeys, apes, antelopes, and big carnivores); continually growing incisors (a sys- tem which enables squirrels to open tough fruits and nuts); claws to dig out hidden prey (some squir- rels, Callitrichidae monkeys, and some anteaters); development of intelligence (especially for mon- keys). In terms of competition for food, it is necessary to separate the “nocturnal world” from the “diur- nal world.” ECOLOGICAL POSITION OE THE EAMILY LORISIDAE Lorisids (10 species of which there are eight in Africa) are all arboreal, nocturnal animals. They find proteins by hunting prey; the energetic food is found mainly in soluble sugars (soft fruits) and/or in gums which are biodegraded in the caecum. Sec- ondary specializations allow several species to live sympatrically, avoiding competition for food. Eor example, five species of Lorisidae live in the same ecosystem of the rain forest in Gabon (two lorisi- naes = slow-moving climber animals and three gal- aginaes = fast-running and leaping animals). The two Lorisinae (Arctocebns calabarensis and Per- odicticns potto) are specialized in the capture of irritant and pungent-smelling prey ignored by the three Galaginae species (Galago elegantidus , G. alleni, and G. demidovii). A. calabarensis live in the understory, in recent tree-fall zones and P. pot- to in the canopy. An equivalent separation exists among the three galaginaes — G. elegantidiis live in the canopy where they feed on insects and gums; G. alleni live in the undergrowth and feed on insects and fruits; G. demidovii, the dwarf bushbaby, hunt prey (70% of their diet) in thick vegetation com- posed of a mixture of lianas and tree foliage (Charles-Dominique, 1971, 1977). A. calabarensis and G. demidovii are the two smallest species of their respective subfamilies. They succeed in feed- ing almost entirely on prey (85% and 70%). The three other much larger species capture the same absolute quantities of prey but they must comple- ment this food by fruits and/or gums (food providing energy). The ecological niches of the Lorisidae can be defined as Arboreal-Nocturnal-Climbers and/or Leapers- Animal ivores-Erugi vores-Gu mi vores ( = A.N.C.L.A.E.G.). 28 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 DIRECT COMPETITORS OE THE LORISIDAE The Lorisidae are not the only mammals to oc- cupy such ecological niches (A.N.C.L.A.E.G.); in Gabon, the Paradoxurinae Nandinia hinotato (palm civet) and several rodents can be found in them: Graphiurus miiriniis (Gliridae), and Thamnomys ni- tilans, Praomys tulhergi, Praomys liikolelae, Sto- chomys longicaiidatus, Hylomysciis Stella, Hylo- myscns fumosiis, Hylomysciis aeta, ?ind Hylomysciis parvus (Muridae). With the exception of Graphiurus murimis and Thamnomys rutilans, which are true arboreal ani- mals, these species generally live at a lower level of the forest but can interfere with lorisids in the undergrowth. Because of a poor knowledge of their ecology, an estimate of the total number of A.N.C.L.A.E.G. mammals is between eight and 15 species. In addition, other categories of sympatric mammals whose diet can interfere with those of A.N.C.L.A.E.G. mammals must be considered (for example, some bats which hunt prey hidden in the foliage). The more complex the forest, the more different biotopes and, thus, different ecological niches it presents. In Gabon, where the forest is one of the most complex in the world (1,000 to 2,000 species of trees and lianas) eight to 15 mammal species oc- cupy these A.N.C.L.A.E.G. ecological niches (among 120 sympatric mammalian species). In com- parison, the dry forest of the west coast of Mada- gascar (Morondava region) is composed of about 150 species of trees and lianas; only five species of mammals (four primates — Microcehus murimis, Microcehiis coquereii, Cheirogaleiis mediiis, and Phaner furcifer — one rodent — Eiiurus sp.) occupy the A.N.C.L.A.E.G. ecological niches among 19 sympatric mammalian species. In South Africa and East Africa, gallery forest and woodland-savanas are poor in species of trees and lianas (about 100 species per ecosystem); only one species of Loris- idae, or rarely two {Galago senegalensis and/or Galago crassicaudatus), can be found in these eco- systems (Bearder, 1974; Bearder and Doyle, 1974). Population densities change as a function of the abundance of arboreal mammalian species in rela- tion to the complexity of the forest. In Gabon, if we consider the five lorisid species together, they form an overall population of about 160 individuals/ square km. In South Africa and East Africa, the population of Galago senegalensis and/or Galago crassicaudatus is generally more abundant — 110, 112, 125, 175, 183, 200, 175, 500 individuals/square km according to the area (Bearder, 1974). In the dry forest of western Madagascar, the four primate species, ecologically equivalent to the Lorisidae (Cheirogaleidae) form an overall population of about 500 individuals/square km (Charles-Domi- nique and Fetter, 1978). At first sight, in spite of the scarcity of plant and animal species, the animal populations of dry for- ests are highly concentrated (Table 1). In fact, for an ecosystem we must consider all the species oc- cupying the A.N.C.L.A.E.G. ecological niches. In Gabon, the Lorisidae constitute only five of the eight to 15 mammalian species adapted to these niches; the bushbabies in South Africa and the Cheirogaleidae in western Madagascar are practi- cally the only species to occupy homologous eco- logical niches in their respective ecosystems. At the present time, it is impossible to make an evaluation in Gabon, but it is probable that the overall density of the eight to 15 Gabonese species living in A.N.C.L.A.E.G. niches (or at least their biomass) is roughly equivalent to those calculated for their relatives in dry forests. In the Gabon rain forest ecosystem, the different families which occupy similar niches of the A.N.C.L.A.E.G. categories avoid food competition by developing some morphological, ethological, and physiological specializations, in the same man- ner as the different lorisid species avoid food com- petition. These specializations are differences in body weight (rodents — 10 to 50 g; lorisids — 60 to 1 ,000 g; palm civet — 2,000 to 4,000 g, and prey cap- ture techniques (prey size is generally related to predator size). Bushbabies detect prey by vision and hearing. Capture is achieved by a rapid stereotyped move- ment of the hand, often when the insect is flying or ready to escape (Charles-Dominique, 1971, 1977). This elaborate system provides the bushbabies with exclusive access to some categories of mobile prey living in the foliage. CONCLUSION— DISCUSSION If animals which are phyletically as different as the primates, rodents, and carnivores discussed in this paper can occupy ecological niches of the A.N.C.L.A.E.G. types (Arboreal, Nocturnal, 1978 CHARLES-DOMINIQUE— ECOLOGICAL POSITION OF LORISIDAE 29 Table I. — Comparative situations of A.N.C.L.A.F.G. (=Arhoreal, Nocturnal, ClimherlLeaper, Animalivore-Fru^ivore-Gumivore) le- murs ( Lorisidae in A frica. Cheirogaleidae in Madagascar} and sympatric mammals in three different ecosystems — Gabon, rain forest of Makokou area: South Africa, different gallery forest and wooden savannas of Tran.svaal, Zululand, and eastern Rhodesia: Mada- gascar, West Coast, Morondava area. Even with a higher number of species, the lemurs of the rain forest occur in lower population densities compared to dry forest species. This low value. In rain forests, can he related to the presence of numerous other mammalian competitors living in sympatry (A.N.C.L.A.F.G. mammals). Gabon South Africa Madagascar Parameters rain forest dry forest dry forest Number of plant species (trees — shrubs — lianas) 1 ,000-2,000 #100 #150 Number of sympatric mammals 120 20 Number of sympatric arboreal mammals 38 10-15 8 Number of sympatric A.N.C.L.A.F.G. mammals 8-15 #5 #5 Number of sympatric A.N.C.L.A.F.G. lemurs Total density of sympatric A.N.C.L.A.F.G.* 5 1-2 4 lemurs (square km) 160 110-500 600 Climber/Leaper, Animalivore-Frugivore-Gumi- vore), it is probably because these modes of life do not require such complex specializations as animals exploiting areas or food of “difficult" access. In South America, the didelphid marsupials of the gen- era Mannosa, Philander, and Caluromys, which present numerous primitive characters, also occupy similar ecological niches. The Jurassic panthother- ian ( 140 million years old) found in Portugal is con- sidered to have been insectivorous and arboreal in habits (Buffetant, 1977) probably close to those ob- served for the above species. One can consider that these modes of life were adopted very early by different primitive mammals which, later, could give rise (or not) to taxa ecolog- ically adapted to other types of life. Among strepsirhines (African and Malagasy le- murs) the two families presenting the greatest number of primitive characters — Lorisidae and Cheirogaleidae — occupy ecological niches of the A.N.C.L.A.F.G. type. The situation is identical to that of Paradoxurinae (palm civets) as compared to other subfamilies of Viverridae. These A.N.C.L.A.F.G. ecological niches can be consid- ered as "conservative"; they are not related to ma- jor modifications of the primitive mammalian mod- el. Among Muridae, most of which are terrestrial or semiterrestrial animals, some species have been able to colonize these ecological niches without de- veloping particular adaptations. LITERATURE CITED Bearder, S. K. 1974. Aspects of the behaviour and ecology of the thick-tailed bushbaby Galago crassicaudatus . Ph.D. dis- sertation, Univ. of Witwatersrand, South Africa. Bearder, S. K., and G. A. Doyle. 1974. Ecology of bushba- bies. Galago senegalensis and Galago crassicaudatus . with some notes on their behaviour in the field. Pp. 109-130, in Prosimian biology (R. D. Martin, G. A. Doyle, and A. C. Walker, eds.), Duckworth, London, xiv -i- 983 pp. Buffetant, E. 1977. Le premier squelette d un mammifere Jurassique. La Recherche, 82:894—895. Carimill, M. 1975. Strepsirhine basicranial morphology and the affinities of the Loiisiformes. Pp. 313-351, in Phylogeny of the primates: a multidisciplinary approach ( W. P. Luckett, and F. S. Szalay, eds.). Plenum Press, New York, xiv -i- 483 pp. Charles-Dominique, P. 1971. Eco-ethologie des Prosimiens du Gabon. Biol. Gabon, 7:121-228. . 1975. Nocturnality and diurnality: an ecological inter- pretation of these two modes of life by an analysis of the higher vertebrate fauna in tropical forest ecosystems. Pp. 69- 87, in Phylogeny of the primates: a multidisciplinary ap- proach (W. P. Luckett, and F. S. Szalay, eds.). Plenum Press, New York, xiv -f 483 pp. Charles-Dominique, P., AND R. D. Martin. 1970. Evolution of lorises and lemurs. Nature, 227:257-260, Charles-Dominique, P., and J. J. Pe I I ER. 1978. Ecology and social life of Phaner furcifer . In Ecology, physiology and be- haviour of five nocturnal Lemur of the west coast of Mada- gascar, Contribution to primatology (P. Charles-Dominique, ed.), Karger, Basel, in press. Eisenberg, J. F., AND J. R. Thoringion. 1973. A preliminary analysis of a Neotropical mammal fauna. Biotropica, 5: 150- 161. Goodman, M. 1975. Protein sequence and immunological spec- ificity in the phylogenetic study of primates. Pp. 219-245, in Phylogeny of the primates: a multidisciplinary approach (W. P. Luckett, and F. S. Szalay, eds.). Plenum Press, New York, xiv -I- 483 pp. 30 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Hladik, C. M. 1977. Chimpanzees of Gabon and chimpanzees of Gombe. Some comparative data on the diet. Pp. 481-501, ill Primate ecology ranging and feeding behavior (T. H. Clut- ton-Brock, ed.). Academic Press, 631 pp. Hoffstetter, R. 1974. Phylogeny and geographical deploy- ment of the primates. J. Hum. Evol,. 3:327-350. . 1977. Phylogenie des primates, confrontation des resul- tats obtenus par les diverses voies d’approche du probleme. Bull. Mem. Soc. d'Anthrop. Paris, 4, ser. 8:327-346. Martin, R.D. 1972. Adaptative radiation and behaviour of the Malagasy lemurs. Phil. Trans. Roy. Soc. London (B), 264:295-352. PococK, R. I. 1918. On the external characters of the lemurs and of Tarsins. Proc. Zool. Soc. London, 1918:19-53. Simpson, G.G. 1945. The principles of classification and a clas- sification of mammals. Bull. Amer. Mus. Nat. Hist., 85:1- 350. SzALAY, L. S. 1975. Phylogeny of primate higher taxa: the basi- cranial evidence. Pp. 91-124, in Phylogeny of the primates: a multidisciplinary approach (W. P. Luckett, and L. S. Sza- lay, eds.). Plenum Press, New York, xiv -I- 483 pp. Tattersall, T, and J. H. Schwartz. 1975. Relationships among the Malagasy lemurs: the craniodental evidence. Pp. 299-311, in Phylogeny of the primates: a multidisciplinary approach (W. P. Luckett, and L. S. Szalay, eds.). Plenum Press, New York, xiv + 483 pp. Van Kampen, P. N. 1905. Die Tympanalgegend des Saugetier- schadels. Physiol., 54:26-57. Van Valen, L. 1969. A classification of primates. Amer. J. Phys. Anthrop., 30:295. Weber, M. 1928. Die Saugetiere, Vol. 2: Systematischer Teil. Lischer Verlag, Stuttgart. THE IDENTITY OE GERBILLUS BOTTAI LATASTE, 1882 (MAMMALIA: RODENTIA), EROM SENNAR, SUDAN DIETER KOCK Forschungsinstitut Senckenberg, Senckenberg-Anlage 25, D-6000 Frankfurt a.M., West Germany ABSTRACT Two recently collected gerbils from the Gezira, Sudan, are identified as GerhiUus holiai Lataste 1882. They are compared with other species of the subgenus Hendecapleura (campesths , c. wassifi, nanus, n. garamantis , pusillus, harwoodi , waters!, and henleyi Jordan!) and discussed with further described taxa (st!gmonyx , litteolus, ve mist us). The proposal is made to con- sider G. st!gmony.\ as 'a. nomen duh!um, and to unite G. luteolus and G. harwood! under the older name G. hotta! . The use of the geographical name Sennar is briefly explained for a more careful consideration in connection with restrictions of type localities and animal distribution. INTRODUCTION During an investigation of harmful rodents in the Sennar region of the Sudan, Mr. Arno H. Hoppe of the Deutsche Gesellschaft Fiir Technische Zusam- menarbeit (GTZ) collected several specimens of small mammals, which were sent for identification to the Forschungsinstitut Senckenberg, Frankfurt a.M. Among the specimens received were Cro- cidura s. sericea (Sundevall 1843) from Fadasi Hal- imab (Southern Gezira Dist.), Crociditra flavescens hedenhorgiana (Sundevall 1843) from Ma’tuq (Dueim Dist.), Arvicanthis ahyssinicits (Ruppell 1942) from Fadasi Halimab and Wad Medani, Mas- toinys natalensis macrolepis (Sundevall 1943) from Wad Medani and Barakat (Southern Gezira Dist.), Taterilliis gyas Thomas 1918 and Acotnys dimidia- tiis (Cretzschmar 1826) from Gadambalia (Gedaref Dist.), and two gerbils discussed below. SYSTEMATIC DISCUSSION Gerbillus (Hendecapleura) bottai Lataste, 1882 Specimens e.xanuned. — Bageir = El Baqeir, Northern Gezira Dist., Blue Nile Prov., l5°2i'N, 32°45'E, 19 (skin, skull) SMF 50158; Managil, Southern Gezira Dist., Blue Nile Prov., 14°15'N, 32°58'E, 19 (skin, skull) SMF 50159. Comparath'e materkil of the suhgenus Hendecapleura 1882. — G. campestris LeVaillant, 1857: S. W. Matmata, S. Tu- nisia, 29 (2 skins, 2 skulls) SMF 52328, 52329; El Golea, Algeria, 1 sex unknown (skin, skull) SMF 51629; Imlaoulaouen, Hoggar, Algeria, Id (skin, skull) SMF 51630; Guelta Afilale, Hoggar, Algeria, 1 d , 4 sex unknown (5 skins, 5 skulls) SMF 51631-51635. G. campestris wassifi Setzer, 1958: Bahig, Western Desert, Egypt, 2d , 19 (3 skins, 3 skulls) SMF 26804-26806. G. n. nanus Blanford, 1875, Rafsanjan, Kerman Prov., Iran, 19,3d (4 skins, 4 skulls) SMF 46370-46372, 46374. G. n. garamantis Lataste, 1881 : Oued Noun, S. E. Goulimine, S. Morocco, 16 ,29 (7 skins, 2 ale., 9 skulls) SMF 47676-47684. G. stigmonyx Heuglin, 1887: Jebelein, E. bank White Nile, Blue Nile Prov., Sudan, 19 (skin, skull broken) AMNH 82219. G. pusillus Peters, 1878: Iraka, Tan- zania, 29 juvenile (2 ale., 1 skull) SMF 11494, 11495. G. harwoodi Thomas, 1901: Mt. Suswa, Rift Valley. Kenya, 10 sex unknown (broken skulls from owl pellets) SMF 415 13-415 19, 41527, 41530-41531; S. shore L. Naivasha, Kenya, 19 juvenile (ale., skull) SMF 42297 [This species averages larger (most ob- vious in M'-M'*) than the sympatric G. pusillus (Petter, 1975; Roche 1976)]. G. waters! DeWinton, 1901: Shendi, Northern Prov., Sudan, 2d (2 skins, 2 skulls) B.M. 1.5.5.55, 1 .5.5.56 (para- types). [Listed by Petter ( 1975) as a valid species and as a sub- species of G. nanus, said to occur in Upper Nile, Sudan, which places it in the southern Sudan. The occurrence of a subspecies of nanus in the horn of Africa besides waters! (Petter, 1975; Roche, 1976) hints to the specific validity of waters!. \ G. henleyi Jordan! Thomas, 1918: Gabes, Tunisia, Id, 19 (2 skins, 2 skulls) SMF 19543, 19544; Jebel Mrhila. Tunisia, 29 (2 ale., 1 skull) SMF 28598. 29599. Lataste (1882c/) recognized the new species G. bottai (two skulls, one skin from Sennar, collected by Botta in 1834) among misidentified gerbils and had already used the new name in a session of the Societe Zoologique de France on 13 December 1881 in explaining the synonymy of GerhiUus gerbillus Olivier, 1801 (Lataste, 18826), but the original de- scription was not published until 1 March 1882 (La- taste, 1882c'). This description is based on two skulls and one skin with incomplete tail; the latter nearly unicol- ored, dorsal coloration not clearly set off from ven- tral side; hind feet naked, shorter than in G. cjuad- rimaculatus Lataste, 1882, from Nubia, a supposed synonym of G. campestris LeVaillant, 1857. The two skulls were both 25 mm long (which is probably 31 Table 1. — External and cranial nieasurements of selected specimens of the genus Gerbillus. BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 32 ^ o ^ pH S s on o C5 S S o SI ^ H c:) K ON ,00 ^ s u - i' < °° e; ~ o u o s O S 'O X " t: U. S c/5 Ji; 00 — S > 1/1 m Ofi Tin ^ o oa C >. 1) VD N iC w ON (U — 00 ' ■- iJ Cj .2 n ■ 5 <" <“ ) o g: U 5; 2 °° K - ^ ^ j 22 ( — r<^ Tf ON ON z z z z z oo Tt lO Tt NO II II II II II s s s s s 00 rl 6s 'ir 00 O 00 ITl Tj- tJ- ri m o NO I On 00 — rj ^ ri I — NO ri ON rn r i On rn tt Tt 00 Tf Tf ON On ON ON Tt ON ON ^ I rj tt m m ON r- — 00 ri Tf O ON 0^4 n o r4 Tt »TN Ti* r^ N© r- NO ri ON JO ■5 -2 ^ j= -ii 2 ob Sb y X) t= o cj (1> z z O ri NO ON m rr II II rn m ■^«r-ir--O>r-i0NN000N0 iri Tt r^' Tt ON On ri ro oor-w-iONr--oorrNor-oooo — r D 'O 00 C c o r- Tf NO rn rr rr rr cr ’— NO rr rr rr — i ° o c § O E o V-* c3 60 .g|x: xl|^« Fig. I. — Karyotype of a male Lemniscomys s. striatus (coll. no. 846) and sex chromosomes of a female L. s. striatus (coll. no. 807). The chromosome slides with the air-dried preparations were made in the Laboratoire d’Ecologie animale of the O.R.S.T.O.M. in Adiopodoume, Ivory Coast, during February and March 1972. The eight specimens of Lemniscomys hellieri (six males and two females) were captured in Lamto and produced 74 (57 male and 17 female) countable mitoses spreads from which 12 (seven male and five female) were selected for measuring. All of the karyol- ogical results for Lemniscomys striatus striatus were obtained with five specimens (two males and three females) from Dabou and 10 (four males and six females) from Adiopodoume. To- gether they produced 198 (62 males and 136 female) countable mitoses spreads, from which 55 (21 male and 34 female) were measurable. The following specimens in the Koninklijk Museum voor Midden- Afrika (collector numbers in parentheses) were those from which air-dried karyological preparations were made: Lem- niscomys hellieri, Lamto (males — 1275, 1276, 1282, 1331, 1334, 1339; females — 1266, 1279); Lemniscomys s. striatus, Adiopo- doume (males — 29, 569, 1010, 1104; females — 119, 568, 1145, 1176, 1312, 1386); L. s. striatus, Dabou (males — 830, 846; fe- males—807, 831, 848). Skull measurements and statistical methods are the same as used in former studies (Van der Straeten and Van der Straeten- Harrie, 1977; Van der Straeten and Verheyen, 1977). The 18 measurements which were used are enumerated in Table 3. The external measurements were copied from the labels. The hind foot length was measured with the nail. The air-dried preparations were made following the slightly modified method of Hsu and Patton (1969). The chromosomal measurements were taken with a curvimeter; each measurement was taken 10 times, after which the mean was calculated. These data were used to pair the chromosomes. This pairing procedure was executed with an IBM 1130 computer and a FORTRAN pro- gram. RESULTS Description of Karyotypes Lemniscomys striatus striatus. — Of the 198 counted mitoses spreads, 78.8% have 2N = 44 and NE = 72. The karyotype is composed of eight pairs of telocentric chromosomes, seven pairs of subtel- ocentric chromosomes, six pairs of metacentric chromosomes, and 1 pair of sex chromosomes XX' or XY (see Eigs. 1 and 2). The X-chromosome is the largest metacentric chromosome; the X' the only submetacentric one. The Y-chromosome could not be identified with absolute certainty but we sup- pose that it is the smallest metacentric chromo- some. It is remarkable that in the females the sex chro- mosomes form a heterogenic pair. The difference between the X and X' chromosome is due to a dif- ference in length of the short arms. In some of the telocentric chromosomes a very short additional arm can be observed; it is possible that a number of them should be considered to be acrocentric. However the quality of the preparation 1978 VAN DER STRAETEN AND VERHEYEN— LEM/V/5COMF5 SYSTEM ATICS 43 0 c- - Fig. 2. — Graph showing karyotype of Lernniscomys s. striutus. Mean, mean ± standard deviation, and range are indicated for each chromosome. and the photographs did not allow further investi- gation. With the exception of the sex chromosomes we could not find a difference between the male and female karyotype. Lernniscomys hellieri . — Of the 74 studied mitoses spreads, 55.4% show a diploid number of 56 (= 2N) and a fundamental number of 78 ( = NE). The karyo- type is composed of 17 pairs of telocentric chro- mosomes, seven pairs subtelocentric chromosomes, three pairs metacentric chromosomes, and one pair of sex chromosomes XX' or XY (see Figs. 3 and 4). In this species the X-chromosome is also the larg- est metacentric of the karyotype and the X' chromo- some is submetacentric. The Y-chromosome is the Table 1 . — Eigenvalues of the canonical transformation with test of significance. No. Eigen- value Relative impor- tance (%) Chi- square Degrees of freedom Proba- bility 1 373.792 81.7 872.046 76 1.000 2 61.164 13.4 269.359 54 1.000 3 14.264 3.1 83.239 34 0.999 4 8.260 1.8 31.069 16 0.987 second longest metacentric or is perhaps submeta- centric. Here also the sex chromosomes of the fe- male form a heterogenic pair XX'. There is a certain similarity between the karyo- types of L. striatiis striatus and L. hellieri when we Table 2.- -Eigenvectors of 19 variables for the first three canon- ical variates. Variable code 1 2 3 HL 0.5831 0.1932 -0.3783 GRLE -0.9375 -0.4675 0.3525 PRCO 0.4559 3.8381 -0.2940 HEBA 0.3818 -2.6529 0.4863 HEP A 0.7462 -0.4103 0.9431 PAF -0.0886 1.1638 0.5596 DIAl -0.3928 -3.4485 -1.6678 DIA2 -0.6140 0.5533 -0.6217 INT 2.2791 -0.6067 1.6093 ZYG 0.2112 0.4286 0.9798 UPTE 0.0120 -1.9578 -0.2864 UPDE -0.7466 1.6011 1.4716 M' 4.3813 -1.5595 -13.0248 ZYPL -0.1754 0.2801 -0.8379 BNAS -0.5152 -1.2736 -2.1432 LNAS 1 .0952 0.2441 -0.5363 LOTE 0.3778 0.7545 -0.1532 BUL -4.1282 -0.0960 -1.4002 BRCA 0.4681 -0.8785 -0.0589 44 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 A/I A<^^***^ A* 4A *<• 4*k#v 4V*% l» M II • * • J|« «||v Fig. 3. — Karyotype of a male Lemniscomys bellieri (coll. no. 1275) and sex chromosomes of a female L. hellieri (coll. no. 1279). consider the X-chromosome, the X'-chromosome, and the seven pair of subtelocentric chromosomes. Canonical Analysis For the canonical analysis, we used 397 speci- mens collected in four localities and divided as fol- lows in five groups: I, Lemniscomys s. striatns, Adiopodoume (25 specimens); 2, Lemniscomys s. striatns , Dabou (74); 3, Lemniscomys s. striatns , Mopoyem (88); 4, Lemniscomys s. striatns, Lamto (43); 5, Lemniscomys hellieri, Lamto (167). We based this analysis on 18 skull measurements (see Table 2) and one external measurement (the hind foot length). The four canonical variates differ significantly from O. The first contains 81.7% of the total variation; the first and second together repre- sent 95. 1% of the total variation. In Table 1 we give the eigenvalues of the canonical transformation and in Table 2 the eigenvectors of the 19 variables for the first three canonical variates. Starting from these eigenvectors a graphical representation is made. For each group the center and the most ex- treme values are indicated by a polygonal delimi- tation. In Fig. 5 we give the graphical representa- tion of the first and second canonical variate. The first canonical variate shows a very clear di- vision between two groups — Lemniscomys hellieri and L. s. striatns. The overlap is very small. The second canonical variate separates Lemniscomys s. striatns in two subgroups depending upon their geo- graphic origin; one subgroup includes the speci- mens from Adiopodoume-Dabou and the other in- cludes animals from Lamto-Mopoyem. The overlap, however, is extensive. The third canonical variate gives no further information. When we consider the same species but now with the specimens grouped in age-classes we obtain similar results. Indeed, the first canonical variate separates L. hellieri fromL. s. striatns, whereas the second variate separates the specimens on the basis of age. Finally, if we consider the same species but now grouped following age and collecting locality similar results are obtained. The first variate separates hei- iieri from striatns and the second one separates the specimens according to age. In the second variate the differences between the subgroups Adiopo- doume-Dabou and Lamto-Mopoyem remain when one compares animals of comparable age groups. Taxonomicai Characters and Discrimination Analysis Lemniscomys hellieri is in its overall measure- ments clearly smaller than L. s. striatns (Table 3). This is especially true for the hind foot length of the adults — L. hellieri 25.4 mm (23.0-27.0); L. s. stria- tns 26.0 mm (24.0-31.0). Both species differ also strikingly in their dorsal 1978 VAN DER STRAETEN AND VERHEYEN— LEMN/5COME5 SYSTEMATICS 45 Fig. 4. — Graph showing karyotpye of Lemniscomys hellieri. Mean, mean ± standard deviation, and range are indicated for each chromosome. pelage pattern. L. s. striatus has on both sides of its black mediodorsal stripe, eight light longitudinal stripes, each of which on closer examination is composed of a row of individualized spots (more to the eastern part of Africa-Zaire and Ethiopia — we find specimens of L. s. striatus where the individual spots show the tendency to blend into a set of con- tinuous longitudinal light stripes). Lemniscomys bellieri also shows eight light longitudinal stripes on each flank and a sharply defined black mediodorsal line. It must be noted, however, that only the third, fourth, and fifth stripes are clearly outlined. The in- dividual spots are not neatly defined and melt more or less away in the darker background, resulting in an even more fuzzy dorsal pattern. Lemniscomys s. striatus is craniometrically larg- er thanL. bellieri and Mi ofL. 5. striatus has mostly an Sm, which is generally absent in bellieri. Notwithstanding all these morphological and metrical characters, a correct determination can give some difficulties when one examines only a skull. Eor this reason, we calculated two discrimi- nant functions. Because these functions are based on museum material exclusively from Ivory Coast, it is obvious that the proposed functions can only be used on Ivory Coast specimens. Eor the first discriminant function we used all of the 18 cranial measurements as well as the hind foot length (including nail). The data for 229 specimens of L. s. striatus were included as well as the data for 167 specimens of L. bellieri . The first discrimi- nant function is calculated as follows: K = 2.416 X HE + 7.517 X INT + 16.284 x M> + 2.553 x ENAS - 16.525 x BUL - 70.229. If K > 0, then we are dealing with L. s. striatus, if K < 0, then it is L. bellieri. The chance of misidentification is 3.5%. The second discriminant function concerns only the cranial measurements and was computed with the same data as the first. It is calculated as follows: K = 4.320 X HEPA - 3.985 x DIA 2 + 8.427 x INT + 2.772 X ENAS - 13.470 x BUL - 23.793. If K > 0 then the specimen is a L. v. striatus; if K < 0, it is a L. bellieri. The chance of misidentifi- cation here is 6.2% DISCUSSION AND CONCLUSIONS There can be no doubt that two species of the Coast — the smaller L. bellieri and the larger L. Lemniscomys striatus-complex exist in the Ivory striatus striatus. It was possible to describe L. bel- 46 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Fig. 5. — Canonical analysis: canonical means (solid circles) and extreme limit of each cloud of points; canonical variates I (abscissa) and II (ordinate); 1 (A) Lemniscomys s. sthatus from Adiopodoume, 2 (B) L. s. striatiis from Dabou, 3 (C) L. s. striatus from Mopoyem, 4 (D) L. striatus from Lamto and 5 (E) L. hellieri from Lamto. lieri as a new species because we are in the pos- session of a very important series of skins and skulls, which allowed a biostatistical approach, and we made air-dried preparations for karyological study. In all of the older museum collections we were able to examine, we found only two specimens of L. hellieri (British Museum Natural History, 23.2.3.42 and 23.2.3.43, collected in 1922 by Lowe and Hardy in Beoumi, Ivory Coast). Both animals were labeled as belonging to the species L. striatus . Rosevear ( 1969), who studied these specimens, not- ed that they are smaller and show an aberrant dorsal pattern. He explained both characteristics by stat- ing that the specimens are juveniles, which is clear- ly an error because the teeth and the skull sutures show that we are dealing with adults. The dorsal stripe pattern of L. hellieri shows a certain similarity with that of L. striatus venustus. However, in all other respects both are very differ- ent. Indeed, the latter is the largest form of the Lemniscomys striatus -complex, whereas hellieri is the smallest. The karyotypes of L. s. striatus and L. hellieri show certain similarities in the X and X' chromo- some as well as in the seven pairs of subtelocentric chromosomes. Matthey (1959) first described the chromosomes ofL. striatus. The results of his study were 2N = 48, four metacentrical chromosomes, and NE difficult to establish but probably around 60. It is probable that this specimen belongs to an East African form of L. striatus. Regrettably, skin and skull were not prepared so that an exact deter- mination of his specimen will remain impossible. 1978 VAN DER STRAETEN AND VERHEYEN— LE/WN/5CO/WK5 SYSTEMATICS 47 Table 3. — Measurements in mm of adult Lemniscomys. Number of specimens, mean, and range are given from left to right for each species. Variable code Description L. s. striatus L. hellieri HB Head and body length 331; 118.1 (61.0-170.0) 220 109.8 (91.0-127.0) TL Length of tail 300; 114.1 (30.0-143.0) 186 113.3 (70.0-134.0) HL Length of hind foot + nail 340; 26.0 (24.0-31.0) 211 25.4 (23.0-27.0) EL Length of ear 317; 15.2 (10.0-18.0) 203 15.9 (13.0-19.0) GRLE Greatest length of skull 317; 29.63 ( 23.30-33.35) 223 28.25 (24.35-31.30) PRCO Prosthion — condylion 346; 26.66 (23.10-30.10) 240 25.31 (22.35-27.75) HEBA Henselion — basion 346; 22.83 (19.10-25.70) 230 21.47 (18.75-23.70) HEPA Henselion — palation 349; 12.27 (10.40-13.70) 244 11.50 (10.20-12.85) PAF Length of palatal foramina 349; 5.83 (4.85-6.85) 240 5.53 (4.80-6.30) DlAl Length of diastema 349; 7.17 (5.75-8.50) 242 6.72 (5.80-7.75) DIA2 Distance between the anterior border of the alveole of M* and the edge of upper incisor 331; 7.70 (6.20-9.20) 225 7.21 (6.15-8.20) I NT Interorbital breadth 349; 4.87 (4.30-5.55) 241 4.42 (3.85-5.05) ZYG Zygomatic breadth on the zygomatic process of the squamosum 328; 13.54 ( 11.95-15.30) 229; 12.82 (11.20-14.40) UPTE Length of upper cheekteeth 353; 5.35 (4.60-5.90) 250 5.11 (4.60-5.65) UPDE Breadth of upper dental arch 330; 6.06 (5.20-6.95) 233 5.82 (5.15-6.50) M‘ Breadth of M‘ 355; 1.66 (1.40-1.85) 248 1.61 (1.45-1.75) ZYPL Breadth of zygomatic plate 350; 3.56 (2.95-4.20) 242 3.35 (2.65-4.10) BNAS Greatest breadth of nasals 329; 3.34 (2.70-4.30) 230 3.16 (2.70-3.65) LNAS Greatest length of nasals 317; 11.59 (9.30-13.60) 223 10.64 (8.55-12.20) LOTE Length of lower cheekteeth 350; 5.09 (4.50-5.55) 251 4.88 (4.45_5.45) BUL Length of auditory bulla 339; 4.94 (4.25-5.60) 239 5.08 (4.55-5.70) BRCA Braincase breadth 345; 12.15 11.05-13.00) 240 11.69 (10.50-12.90) ACKNOWLEDGMENTS We wish to thank Dr. L. Bellier for the hospitality and help he provided us during our stay in Adiopodoume (1972). More- over this study had been made possible thanks to the important collections he made in Ivory Coast and which were kindly put at our disposal for further study by the Authorities of the O.R.S.T.O.M. Our gratitude goes also to our colleagues Dr. G. Corbet and Mr. I. Bishop, who allowed us to examine the murid collections of the British Museum of Natural History, to Mr. M. Michiels, who made the air-dried preparations and worked out the karyo- types, to Dr. F. Hebrant and Dr. Ir. J. Desitter, who assisted us with the computer programs and to Mrs. B. De Vry-Vanlinden for technical assistance. A large part of this research was made possible by the research grant F.K.F.O. no. 955. LITERATURE CITED Hsu, T. C., AND J. L. Patton. 1969. Bone marrow preparations for chromosome studies. Pp. 454-460,/;; Comparative Mam- malian Cytogenetics (K. Benirschke, ed.), Springer-Verlag, Berlin. Linnaeus, C. 1758. Sy sterna naturae per regna tria naturae, se- cundum classes, ordines, genera, species, cum characteri- bus, differentiis, synonymis, locis. Impensis Direct. Lauren- tii Salvii, Halmiae, 10th ed. 1:1-77. Matthey, R. 1959. Formules chromosomiques de Muridae et de Spalacidae. La question du polymorphisme chromoso- mique chez les Mammiferes. Revue Suisse Zook, 66:175- 209. Rosevear, D. R. 1969. The rodents of West-Africa. Trust. Brit. Mus., London, 604 pp. Thomas, O. 1911. The mammals of the tenth edition of Lin- naeus; an attempt to fix the types of the genera and the exact bases and localities of the species. Proc. Zool. Soc. London, pp. 120-158. Van DER Straeten, E. 1975«. Systematische en biostatistische studie van het Afrikaanse Lemniscomys striatus species- complex (Mammalia, Muridae). Doctoraatsthesis weten- schappen, Universiteit Antwerpen, 2 delen. . 1975/). Lc;;»;/5co;;;y.y /)e///c;7, a new species of Muridae from the Ivory Coast (Mammalia, Muridae). Rev. Zool. Afr., 89:906-908. Van DER Straeten, E., and B. Van der Si raeten-Harrie. 1977. Etude de la biometrie cranienne et de la repartition d'Apodemus sylvaticus (Linnaeus, 1758) et d'Apodemus flavocoUis (Melchoir, 1834) en Belgique. Acta Zool. Pathol. Antverp., 69:169-182. Van der Straeten, E., and W. N. Verheyen. 1978. Taxo- nomical notes on the West- African Myomys with the descrip- tion of Myomys derooi (Mammalia — Muridae). Z. Sauge- tierk., 43:31-41. REVISION OF THE GENUS SACCOSTOMUS (RODENTIA, CRICETOMYINAE), WITH NEW MORPHOLOGICAL AND CHROMOSOMAL DATA FROM SPECIMENS FROM THE LOWER OMO VALLEY, ETHIOPIA BERNARD HUBERT Laboratoire de Zoologie Appliquee, O.R.S.T.O.M., BP 1386, Dakar, Senegal ABSTRACT Based upon new records of Saccostomus from southwestern ognized — 5. campestris from southern Africa and S. mearnsi Ethiopia, a revision of the genus is considered. Morphometric from eastern Africa, and karyological evidence indicate two species should be rec- INTRODUCTION Seven rodents of the genus Saccostomus were collected during the summer of 1973 in the Lower Omo Valley, Ethiopia, during a paleontological ex- pedition directed by Y. Coppens. These individuals are noteworthy because of their shaggy fur, uni- formly brownish-gray in color, and consisting of long, soft, and silky hairs. The belly of these spec- imens is paler than the back, but still definitely gray- ish, unlike the forms from southern Africa with their pure white belly and their sleek but less shaggy fur. RESULTS AND DISCUSSION Table 1 presents the main external and cranial measurements for the specimens collected in the Lower Omo Valley, Ethiopia; three specimens from the British Museum, one from Ethiopia (reported by Yalden et al., 1976) and two from Uganda; three specimens from the Museum National d’Histoire Naturelle (MNHN) de Paris with southern African origin, one of which is the specimen studied by Matthey (1958). The measurements published by Roche (1976) for a specimen from Somalia are also reported here. External measurements are given to the nearest millimeter; cranial measurements were taken with a dial caliper and reported to the nearest tenth of a millimeter. The specimens from northeast Africa (Uganda, Ethiopia, and Somalia) have a gray belly, a long ear (more than 20 mm), a long tail (more than 50 mm), and a long maxillary toothrow (more than 5.0 mm). The three southern specimens from the MNHN have a white belly, a short ear (less than 20 mm), a short tail (40 mm), and a short maxillary toothrow (less than 5.0 mm). The same characters are present on all the specimens from southern Africa that I studied in the British Museum, particularly the co- types of Saccostomus campestris Peters. The karyotypes were observed for two individu- als in 1973 by the “squash” method as described by Matthey (1958), and for two others in 1976-1977 by the “air-drying” method. The slides are not very easily interpreted, but the chromosome number was always between 40 and 42, never exceeding this last number, in more than 20 metaphases observed (Pig. 1). Lord and Hamerton (1956) published a ka- ryotype of 2N = 44 for an individual of unknown origin, but Matthey (1958) considers that it is diffi- cult material to interpret and that the number of chromosomes for Saccostomus campestris is 2N = 46, as he personally observed on a specimen col- lected in the Cape Province, South Africa. Ellerman (1941) thought that there was only one species of Saccostomus — S. campestris Peters. Misonne (1968) agreed with him, but Roche (1976) thought that his specimen from Somalia was closely related to the other specimens from the “north,” and different from the southern African Saccosto- mus campestris . I think that Ethiopian specimens, which are different by their morphological charac- ters, by their measurements, and by their karyo- type, represent a distinct species together with all the Saccostomus from the north of this area in Af- rica (Ethiopia, Uganda, Somalia, and Kenya). Prior to the revision of Ellerman (1941), different species had been described from these countries as follows: Saccostomus umbriventer. Miller, 1910; Saccosto- 48 1978 HUBERT— REVISION 0¥ SACCOSTOMUS 49 5 ^3 .15 o c C/5 5 o ^ CJ ^ U 5 c/5 < 7 -D (?3 I .ti c X> O aiS 5= a o o P E ^ b •— -n N E i 2 15 - O .-S C/5 Of) O a c •“CO/ “O *- c o o ^ ^ "o >>>>>.>. >^ -D 55>n >-, >^ -D X) X) X) -D -D -D >> .ti ^ X r“ r- r- uuuuuuuuuuuu.r:T^:r" 000000000000:>^> r4 r^i o ON O ON ri mi ON n 1 1 o o rr X X X X X mi 1 1 od X X X X ON ON o m n ON 00 X 1 r4 mi X rr in in n m, X X mi mi mi 1 X rr 'O' rt X n ON in mi Tj- <*r 1 , r- rr, in Tf Tj* Tf 1 rj- rr rr rr rj _ 00 X mi in O o r^i 00 r^ ON 00 o^ On ON ON r- 00 1 r--’ 1 1 o o ri ON n n m, rj rr, r- 00 rr rj n X X X ri 1 mi mi o ri f^< rn r^i r^i r^i rn m rr r^i r^t r«-i r^, rr, rr, n O rn mi ri O r^i tT o 1 o — O r- X n ri r 1 ri ri n rj rj 1 r4 ri rji Tf ri mi r^> ri rn Tf ri 1 ON rj r^i r I On ON n n ri ri ri r-4 r j ri 1 r\ rj ri o X O 00 00 r^ o 1 ON Tf ri o O o cr 6 6 o o d d d 03 03 03 . __*■ X c ~o s E E E E E E 03 03 03 d' o p3 X X o O O O O O O ■jt at; s o H QC Qi 0+ 0+ 0+ 0+ 0+ O ’TO o o CH — ri rn »A1 <5rixr- r\ ^ Tf •T', X X O On On On Table 2.— Means, standard errors, and standard deviation of the lengths of the tail, the ear, and the upper molar row from different populations from 50 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 nJ S § Ci. O cij O "5 2 it o 2 cx ^ 2. c. ^ 2 ^ 5 2 "5 oo CO 2 a « K, O Si ^ a > w o> CO "O ^ s C u Cd •*-« QJ CO “O -S o C 1- a +- 4> CO "O E o Z _D .« c/5 a c X) o ^ E t/D ^ ON O 3 C o o 0 S' 1 ig 5 S --si II A S ^ 3 C V o o " rj^ooio r-TtNOoooor'-^r'4 0\ONON Or-i— io— ^ — ^ooo — O C'l -O 00 rn n -tT 0 — 000000000 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 +1 r'4 0<^i^ov^r''0s<^o\t'^ oo'sdu-''dN6oor''’vor'-NOr'- i/“j«/Yr0‘/^i00N000'O’^rnrN r-'«noo — rn — r4 — Tf0\0\ O r4 O — r-4 ri — — O — +1 +1 +1 +1 +1 +1 +1 -H +f +1 +1 co-^00ror«N — r4 — o — O Os 00 o NO C/^l NO ® s ® — o — o r-' o o o +1 -fi +1 O 00 ON O U- ro ^ - 00 os — — +1 +1 +1 +1 ■'d- ON vY r- z s X 'Nl § U E S Z • ' ^ o .V 3 c 2 OK) O 0> fc D ^ O 13 cd 3 >> Tanganyika (AMNH) 29 48.8 ± 1.2 7.2 — — 5.38 ± 0.04 0.21 significant difference 1978 HUBERT— REVISION 0¥ SACCOSTOMUS 51 X X Fig. 1. — Karyotype of Saccostonuis mearnsi from Omo Valley, Ethiopia. mus mearnsi, Heller, 1910; Saccostonuis isiolae, Heller, 1912; Saccostonuis cricetulus, Allen and Lawrence, 1936. The holotypes of these species are deposited in the Smithsonian Institution, Washing- ton, D.C., and at the Museum of Comparative Zo- ology, Harvard University. The four species are closely related. The first de- scribed is S. mearnsi, the holotype of which is an adult male in good condition (better condition than S. isiolae, but younger than S. iimhriventer, which is a very old female). S. cricetulus is not very dif- ferent and when described by G. M. Allen and B. Lawrence they thought that the four species might be the same (Allen and Lawrence, 1936). I think that Saccostonuis mearnsi is the available name for the northeastern species of Saccostonuis character- ized by a gray belly, a long tail (more than 49.0 mm), a large ear (more than 19.0 mm), a long max- illary toothrow (more than 5.0 mm), and a karyo- type of 2N = 40-42 chromosomes. Table 2 shows the means of the length of the tail, the ear, and the upper molar row for a group of specimens from the American Museum of Natural History, New York, from the United States Na- tional Museum, Washington, D. C., and from the Museum of Comparative Zoology, Harvard, col- lected in different countries of eastern and southern Africa. It shows the difference between the speci- mens from Zambia (Northern Rhodesia), Angola, Botswana (Bechuanaland), Malawi (Nyasaland), Zimbabwe (Southern Rhodesia), Transvaal, and Mozambique on one hand and from Tanzania, Uganda, Kenya, and Ethiopia on the other hand. The specimens from Kijungu, Tanzania, are re- ferred to S. mearnsi, hut they are a bit smaller than the individuals from the other northeastern coun- tries, and statistically significantly different. Their karyotype is unknown. An hypothesis to test is whether the division be- tween the two species is the Rift Valley. Lurther- more, it would be very interesting to collect some specimens in southern Tanzania or in northern Mo- zambique to test the possibility of speciation ac- cording to the latitude in this region. SUMMARY Two different species of Saccostonuis are rec- acterized by a white belly, a short ear (less than 19 ognized — S. campestris from southern Africa, char- mm), a small tail (less than 49 mm), a short upper 52 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 molar row (less than 5.0 mm) and a karyotype of a long tail, a long upper molar row and a karyotype 2N = 46 chromosomes; 5. mearnsi from eastern of 2N = 40-42 chromosomes. Africa, characterized by a gray belly, a large ear, LITERATURE CITED Allen. G. M., and B. Lawrence. 1936. Reports on the sci- entific results of an expedition to rain forest regions in East- ern Africa, III, Mammals. Bull. Mus. Comp. Zool., 79:100- 102. Ellerman, J. R. 1941. The families and genera of living rodents. Vol. II. Eamily Muridae. Trustees of the British Museum, London, 690 pp. Ford, C. E., AND J. L. Hamerton. 1956. Chromosomes of five rodent species. Nature, 4499:140-141. Malthey, R. 1958. Les chromosomes et la position systema- tique de quelques Murinae africains (Mammalia, Rodentia). Acta Tropica, 15:97-117. Misonne, X. 1968. Rodentia: main text (19). Pp. 1-55, in Pre- liminary identification manual for African Mammals (J. Meester and H. W. Setzer, eds.), Smithsonian Inst., Wash- ington, D.C. Roche, J. 1976. Capture de rongeurs appartenant aux genres Lemnisconiys, Saccostomus, Steatomys et Lophioinys en Republique Somalie. Monit. Zool. Ital., n.s., 8(6):195-201. Yalden, D. W., M. J. Largen, AND D. Kock. 1976. Catalogue of the Mammals of Ethiopia. 2 — Insectivora and Rodentia. Monit. Zool. Ital., n.s., 8(1):1-118. TAXONOMIC REVIEW OF THE FAT MICE (GENUS STEATOMYS) OF WEST AFRICA (MAMMALIA: RODENTIA) PIERRE SWANEPOEL Resident Museum Specialist, Carnegie Museum of Natural History, 4400 Eorbes Avenue, Pittsburgh, Pennsylvania 15213 U.S.A., and Kaffrarian Museum, King William’s Town, Republic of South Africa DUANE A. SCHLITTER Section of Mammals, Carnegie Museum of Natural History, 4400 Eorbes Avenue, Pittsburgh, Pennsylvania 15213 U.S.A. ABSTRACT Data from specimens of West African taxa of Steatomys were subjected to univariate and multivariate statistical analyses. Geographic and nongeographic variation were analyzed based upon an examination of 385 conventional museum specimens. From the results of the analyses, three species of Steatomys are recognized in West Africa. Steatomys ciippedius, a small mono- typic species, occurs in the Sahel and Sudan Savanna zones of West Africa. S. caurinus exhibits geographic variation; two sub- species are recognized, one of which is described as new. S. jacksoni is retained as a unique species known only by the ho- lotype from Ghana. Both S. caurinus and S. cuppedius exhibit a high degree of variation with age. Geographic distributions of the West African taxa are plotted based upon specimens exam- ined and literature records. INTRODUCTION Fat mice of the genus Steatomys are members of the rodent family Muridae and together with the genera Dendromus, Prionomys, Dendroprionomys , Leimacomys, Deo my s, and Malacothrix are includ- ed in the subfamily Dendromurinae. The phyloge- netic affinities of the genera of this subfamily are little understood and the taxonomic relationships of the many described taxa are poorly known. Geo- graphically, genera of this subfamily are distributed throughout subsaharan Africa in habitats ranging from subdesert to forest. Species of the genus Steatomys occur throughout subsaharan Africa in habitats ranging from subdes- ert to degraded forest or forest edge. Generally, however, they are found in savannah. In spite of their widespread occurrence, specimens of the ge- nus are not often taken because they do not readily enter any type of traps. In areas where individuals can be captured by hand at night or dug from their burrows, significant samples are being accumulat- ed. Initially, specimens were few and wide hiatuses existed between samples. During the last half of the nineteenth and early part of the twentieth centuries, individuals from each locality were usually de- scribed as a new taxon. There have been two major listings of the taxa in the genus. Allen ( 1939) listed 18 species with 26 subspecies, whereas Ellerman (1941) gave 18 species with 24 subspecies, some of which differed from Allen’s list. Three species of Steatomys have been described from West Africa — S. caurinus Thomas, 1912, from Panyam, Nigeria; S. cuppedius Thomas and Hin- ton, 1920, from Farniso [ = Panisau], Nigeria; 5. jacksoni Hayman, 1936, from Wenchi, Ghana, Al- len ( 1939), Ellerman (1941), and Rosevear ( 1969) considered these three taxa to be distinct species, based generally on the original descriptions. In a recent provisional checklist of the genus, Coetzee (1977) listed both S. caurinus and 5. jacksoni as subspecies of S. pratensis Peters, 1846, and S. cup- pedius as a subspecies of S. parvus Rhoads, 1896. Both S. pratensis and S. parvus were considered by Coetzee (1977) to be widespread savannah-oc- curring Panafrican species. However, he has ad- mitted to perhaps oversynonymizing the nominal taxa in his provisional checklist, wherein he rec- ognized three species in the genus. Relationships of the three West African forms to other species within the genus in the remainder of Africa is beyond the scope of the present study. A short review of described forms, which might occur sympatrically or parapatrically with these three taxa follows. Steatomys opimus Pousargues, 1894, originally described from material from the region now included in the Central African Empire, is 53 54 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 found in degraded forest and woodland savannah in central Africa (Coetzee, 1977). Because it is known to occur in the Cameroons, it might therefore also occur in the southern parts of West Africa, at least in southern Nigeria. A large species (greatest length of skull above 30), Steatomys opimits, is considered by Coetzee ( 1977) to be only subspecifically distinct from S. pratensis, a more geographically wide- spread species described originally from southern Africa. Three additional species occur in the Sudan (Setzer, 1956) — S. aqiiilo, S. gazellae, and S. tho- masi. Coetzee (1977) also synonymized S. (icpiilo and S. gazellae with S. parvus, and S. thomasi with S. prate ns is. The purpose of this study was to examine the systematic relationships of the nominal taxa of the genus Steatomys in West Africa, based on sizeable unreported collections in the Smithsonian Institu- tion, Washington, D.C. Here West Africa is con- sidered to be that portion of Africa west of the Ni- gerian border and south of the Sahara. This geographic limitation is not the same as that used by Rosevear (1965, 1969) who included those parts of Cameroon west of the Sanaga River. MATERIALS AND METHODS Conventional museum skins and skulls of 385 specimens of the genus Steatomys from West Africa were examined. All of the holotypes of western African taxa were examined by both authors. Five external and 14 cranial measurements were re- corded. in millimeters, from nearly all specimens examined. Ex- ternal measurements were listed from the specimen labels; cra- nial measurements were taken with dial calipers. Definitions of these cranial measurements are given below; Greatest length of skull. — Greatest distance from the anterior edge of the nasal bones to the posteriormost edge of the occipital bone. Condylohasal length. — Greatest distance from the anterior- most projection of the premaxilla between the incisors to the posterior edge of the occipital condyles. Zygomatic breadth. — Greatest distance across zygomatic arches at right angles to longitudinal axis of cranium. Interorhital breadth. — Least distance across interorbital con- striction. Rostral breadth. — Least distance across the rostrum imme- diately anterior to the zygomatic plate. Oblique length of bulla. — Greatest oblique length of auditory bulla, taken from a point adjacent to the paraoccipital process to the anteriormost edge of bulla. Greatest length of bulla. — Greatest oblique length of bulla tak- en from the posterior edge of mastoidal bulla to the anteriormost edge of auditory bulla. Length of maxillary toothrow. — Least distance from the an- terior edge of alveolus of M* to posterior edge of the alveolus of ML Breadth across upper molars. — Least distance, measured at right angles to the longitudinal axis of the skull, from the widest point on the labial edge of the crowns of M‘ in each maxillary toothrow. Length of anterior palatal foramen. — Greatest distance from anterior edge to the posterior edge of the anterior palatal fora- men. Length of posterior palatal foramen. — Greatest distance from the anterior edge to the posterior edge of the posterior palatal foramen. Length of diastema. — Greatest distance from the posterior edge of the alveolus of the incisors to the anterior edge of the alveolus of M'. Height of skull. — Greatest height of skull (taken perpendicular to the horizontal plane of skull when placed on a microscope slide). Length of mandibular toothrow. — Least distance from the an- terior edge of the alveolus of M, to posterior edge of the alveolus of M;j. Univariate statistical analyses were performed on an IBM 360 computer at Carnegie-Mellon University, Pittsburgh, and mul- tivariate statistical analysis on an IBM 370 computer at Texas Tech University, Lubbock. Univariate analyses of secondary sexual, age, and individual variation were performed using the UNIVAR program developed and published by Power (1970). Standard statistics (mean, range, standard deviation, standard error, variance, and coefficient of variation, among others) are generated by this program. In the event of two or more groups being compared, a single-classification analysis of variance (AN- OVA) to test for significant differences between or among means is employed. When means were found to be significantly differ- ent, the Sums of Squares Simultaneous Test Procedure (SS-STP) was used to determine maximally nonsignificant subsets. Multivariate statistical analyses were performed using the Nu- merical Taxonomy Systems (NT-SYS) package developed by F. J. Rohlf, R. Bartcher, and J. Kishpaugh at the University of Kansas. In these analyses the OTUs were grouped localities (geographical samples) and the values used for each measure- ment were arithmetic means for the measurement. Matrices of Pearson's product-moment correlation and taxonomic distance coefficients were computed. Cluster analyses were performed using the unweighted pair-group method using arithmetic aver- ages (UPGMA) on correlation and distance matrices, and phe- nograms were generated for both. Only the distance phenogram is given as it gave larger coefficients of cophenetic correlation and the results of the distance phenogram also agree more closely with other analyses. Three principal components were extracted from the matrix of correlations, and a three-dimen- sional projection of the samples onto the three principal com- ponents was made. For the theory and use of these tests, see Sokal and Sneath (1963), Schnell (1970), Atchley (1970), Geno- ways (1973), and Sneath and Sokal (1973). 1978 SWANEPOEL AND SCH LITTER— 57EA 7 OMK5 SYSTEMATICS 55 Fig. 1. — Diagrammatic representations (drawn from camera lucida) of left maxillary toothrows of Steatomys cuppedius illustrating wear pattern for age categories II, III, and IV individuals. Age category I, not shown, is the same as shown for category II, except that is not fully erupted. NONGEOGRAPHIC VARIATION The kinds of nongeographic variation considered in the study include secondary sexual, age, and in- dividual variation. Individuals were first sorted by sex and age and tested for nongeographic variation with samples maintained in separate age and sex categories (Straney, 1978). This method should keep separate the effect of age and sex upon non- geographic variation. Then, in the absence of sig- nificant secondary sexual variation in each age cat- egory, the sexes were combined in the geographic samples used for the analysis of age variation. Secondary Sexual Variation Specimens from selected populations of Steato- mys cuppedius and S. caurinus were assigned to one of four age categories (see section on Age Vari- ation for details of ageing method and criteria). Steatomys cuppedius males and females of age cat- egories II and III from Panisau, Nigeria, and age category II from Senegal were tested for significant differences in size. Similar tests between various ages of samples of male and female S. caurinus from Pirisi, Ghana (age II and III); Gudi, Nigeria (age III); and Celia, Upper Volta (age II) were per- formed. These samples were tested as only those populations and age categories from which suffi- cient individuals of both sexes were available could be analyzed. The results of these tests on age category II in- dividuals of 5. cuppedius from Panisau, Nigeria, and of S. caurinus from Pirisi, Ghana, are shown in Table 1, whereas the results for age category III of the same geographic samples are listed in Table 2. The data for the other comparisons given above are not presented but are on file in the Section of Mammals, Carnegie Museum of Natural History. Steatomys caurinus. — No significant differences in measurements between 16 males and 20 females from Gudi, Nigeria (age III), 13 males and 14 fe- males from Celia, Upper Volta (age II), and 14 males and nine females from Pirisi, Ghana (age II) were found. The results of the latter comparison are 56 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 1. — Secondary sexual variation in external and cranial measurements of age category II specimens of Steatomys caur- inus from Pirisi, Ghana, and S. cuppedius from Panisau, Ni- geria. Means for males and females that are significantly dif- ferent at the 5% level are marked with an asterisk. Measurements and sex N Mean ± 2 SE Range CV Steatomy caiirinus Total length Male 11 139.4 5.38 130-156 6.4 Female 9 133.4 4.10 128-147 4.6 Length of tail Male II 38.8 ± 1.54 34-41 6.6 Female 9 39.4 2.16 35-46 8.2 Length of hindfoot Male 11 19.0 0.38 18-20 3.3 Female 9 18.4 -4- 0.59 17-20 4.8 Length of ear Male 11 16.3 ±_ 0.55 15-18 5.6 Female 9 15.8 ± 0.44 15-17 4.2 Weight Male 11 34.4 ± 7.40 19-60 35.7 Female 9 32.2 4.73 22-47 22.0 Greatest length of skull Male 8 24.4 0.54 23.4-25.7 3.1 Female 4 23.7 0.46 23.2-24.2 1.9 Condylobasal length Male 7 23.2 ± 0.46 22.5-24.3 2.6 Female 6 22.7 ± 0.53 22.0-23.4 2.8 Zygomatic breadth Male 8 12.1 ± 0.28 11.4-12.6 3.3 Female 3 11.8 0.35 11.5-12.1 2.5 Interorbital breadth Male II 3.8 ± 0.09 3.7-4. 1 4.1 Female 8 3.8 0.14 3.6-4. 1 5.1 Rostral breadth Male 11 4.5 0.10 4.3-4. 8 3.5 Female 9 4.3 ± 0.15 4. 0-4. 7 5.1 Oblique length of bulla Male 7 5.4 ± 0.11 5. 1-5.5 2.8 Female 4 5.3 0.19 5. 1-5.5 3.6 Greatest length of bulla Male 7 8.0 ± 0.18 7. 8-8. 4 3.0 Female 3 Length of maxillary toothrow 8.2 0.07 8. 1-8.2 0.7 Male 10 4.1 0.13 3. 8-4. 4 5.0 Female 9 Breadth across upper molars 3.9 — 0.09 3.6-4. 1 3.6 Male 11 5.6 0.10 5.4-5. 9 2.9 Female 9 5.5 0.14 5. 3-5.9 3.7 Table 1. — Continued Measurements and sex N ! Vlean ± 2 SE Range CV Length of anterior palatal foramen Male 11 4.7 ± 0.14 4. 2-5. 2 4.9 Female 9 4.6 ± 0. 19 Length of posterior palatal foramen 3.9-4. 9 6.2 Male 10 0.4 ± 0.07 0. 2-0.5 24.6 Female 9 0.4 ± 0.06 0.3-0.6 20.0 Length of diastema Male 11 6.5 ± 0.19 5.9-7.0 4.8 Female 8 6.2 ± 0.20 5. 7-6.5 4.6 Height of skull Male 6 8.7 ± 0.23 8.4-9. 1 3.3 Female 5 Length of mandibular toothrow 8.5 ± 0.24 8. 3-9.0 3.2 Male 11 3.5 ± 0.12 3. 2-3. 9 5.9 Female 9 Steatomys 3.4 ± 0.14 cuppedius 3. 0-3.7 6.2 Total length Male 24 109.7 ± 3.5 97-129 3.5 Female 9 106.3 ± 4.3 94-112 4.3 Length of tail Male 24 42.8 ± 1.4 38-50 8.0 Female 8 42.1 ± 1.9 37-45 6.3 Length of hindfoot Male 24 15.9 ± 0.18 15-17 2.8 Female 9 15.6 ± 0.48 14-16 4.7 Length of ear Male 24 12.7 ± 0.22 12-14 4.3 Female 9 12.4 ± 0.35 12-13 4.2 Weight Male 24 9.4 ± 0.90 6-14 23.4 Female 9 8.7 ± 1.25 5-10 21.6 Greatest length of skull Male 20 19.7 ± 0.30 18.4-21.4 3.5 Female 8 19.2 ± 0.63 17.5-20.3 4.6 Condylobasal length Male 21 18.5 ± 0.30 17.5-20.0 3.8 Female 8 18.0 ± 0.58 16.4-18.9 4.6 Zygomatic breadth Male 23 10.8 ± 0.14 10.3-11.5 3.0 Female 9 10.8 ± 0.26 10.2-11.5 3.5 Interorbital breadth Male 24 3.6 ± 0.05 3. 3-3.9 3.6 Female 9 3.7 ± 0.08 3. 5-3.9 3.4 Rostral breadth Male 24 3.7 ± 0.08 3.3-4. 1 5.6 Female 9 3.7 ± 0.12 3. 5-3. 9 4.6 1978 SWANEPOEL AND SCHLITTER— 57EA70ME5 SYSTEMATICS 57 Table I . — Continued Measurements and sex N 1 Mean ± 2 SE Range CV Oblique length of bulla Male 17 4.5 0.09 4.2-4.9 4.0 Female 6 4.5 0.25 4.0-4.9 6.9 Greatest length of bulla Male 15 6.8 ± 0.13 6.4-7. 3 3.8 Female 6 6.7 ± 0.26 6.2-7. 1 4.8 Length of maxillary toothrow Male 23 3.6 ± 0.07 3. 3-4.0 4.8 Female 9 3.7 0.12 3.4-3. 9 5.1 Breadth across upper molars Male 24 4.9 ± 0.09 4.4-5. 6 4.6 Female 9 4.9 0.13 4. 6-5. 2 4.0 Length of anterior palatal foramen Male 23 3.7 ± 0.09 3. 1-4.0 5.6 Female 9 3.6 0.19 3. 2-4.2 8.2 Length of posterior palatal foramen* Male 22 0.4 ± 0.03 0. 3-0.5 18.3 Female 9 0.5 ± 0.05 0.4-0. 6 15.9 Length of diastema* Male 24 4.6 ± 0.10 4.1-5. 0 5.3 Female 9 4.3 0.05 4. 2-4.4 1.8 Fteight of skull Male 18 7.6 ± 0.10 7. 2-7. 9 2.8 Female 6 7.5 ± 0.16 7. 2-7. 7 2.6 Length of mandibular toothrow Male 22 3.1 ± 0.06 2. 9-3. 4 4.3 Female 9 3.1 0.13 2. 7-3. 3 6.4 given in Table 1. In age category III, individuals from Pirisi, Ghana, females were found to be sig- nificantly larger than males only in rostral breadth (Table 2). Steatomys cuppedius . — No significant differ- ences in size between 17 males and nine females of age category II from Senegal were found. However, in age category II from Panisau, Nigeria, the length of the posterior palatal foramen was larger in fe- males, whereas the length of diastema was larger in males (Table 1). In the population from Panisau, Nigeria, males and females of age category III were found not to differ significantly in size (Table 2). Conclusions. — No significant differences in size betv/een males and females are obvious. Prom the comparisons of sexes of two age categories from five geographic samples, only two cranial measure- Table 2. — Secondary sexual variation in external and cranial measurements of age category III specimens of Steatomys caur- inus from Pirisi. Ghana, and S. cuppedius from Panisau. Ni- geria . Significance levels are the same as for Table 1 Measurements and sex N Mean 2 SE Range CV Steatomys caurinus Total length Male 14 158.2 ± 4.59 137-169 5.4 Female 7 157.6 ± 5.70 145-165 4.8 Length of tail Male 14 43.1 3.15 36-59 13.7 Female 7 41.4 ± 2.65 36-46 8.5 Length of hindfoot Male 14 19.7 0.39 18-21 3.7 Female 7 19.6 0.40 19-20 2.7 Length of ear Male 14 17.2 ± 0.43 16-18 4.7 Female 7 17.1 0.68 16-18 5.2 Weight Male 14 54.4 -F- 5.08 37-68 17.5 Female 7 53.3 8.31 34-66 20.6 Greatest length of skull Male 12 26.4 ± 0.47 24.9-27.8 3.1 Female 5 26.3 0.78 25.6-27.6 3.3 Condylobasal length Male 12 25.3 ± 0.53 23.5-26.6 3.6 Female 5 25.3 0.63 24.7-26.4 2.8 Zygomatic breadth Male 12 12.9 0.22 12.2-13.4 3.0 Female 4 13.0 0.43 12.7-13.6 3.3 Interorbital breadth Male 14 3.9 ± 0.10 3. 6-4. 2 4.7 Female 6 3.9 0.17 3. 6-4. 2 5.2 Rostral breadth* Male 14 4.8 ± 0.13 4.4-5. 1 5.2 Female 7 5.0 0.17 4. 7-5. 3 4.4 Oblique length of bulla Male 13 5.8 ± 0.12 5. 5-6.3 3.7 Female 4 6.0 0.17 5. 8-6. 2 2.9 Greatest length of bulla Male 10 8.8 ± 0.21 8.4-9. 2 3.8 Female 5 8.8 + 0.17 8. 5-9.0 2.2 Length of maxillary toothrow Male 14 4.0 0.09 3. 7-4. 3 4.1 Female 7 4.1 ± 0.17 3. 8-4. 4 5.6 Breadth across upper molars Male 14 5.9 0.11 5. 6-6.2 3.5 Female 7 6.1 ± 0.17 5. 8-6.4 3.8 58 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 2. — Continued Measurements and sex N Mean ± 2 SE Range CV Length of anterior palatal foramen Male 14 5.0 ±0.11 4. 7-5.4 4.0 Female 7 5.2 ± 0.24 4. 7-5. 6 6.2 Length of posterior palatal foramen Male 14 0.5 ± 0.09 0.3-0.8 31.0 Female 7 0.6 ± 0.16 0. 3-1.0 38.6 Length of diastema Male 14 7.2 ± 0.19 6. 5-7. 8 4.9 Female 7 7.3 ± 0.22 7. 0-7. 7 4.0 Fleight of skull Male 10 9.2 ± 0.25 8. 4-9.7 4.3 Female 4 9.3 ± 0.25 9.0-9.5 2.6 Length of mandibular toothrow Male 13 3.5 ± 0.13 3. 1-3.9 6.8 Female 7 3.4 ± 0.13 3. 2-3. 7 5.0 Steatomys cuppedius Total length Male 8 118.3 ± 5.82 107-133 7.0 Female 8 117.9 ± 4.86 105-125 5.8 Length of tail Male 8 44.0 ± 2.98 39-52 9.6 Female 8 43.3 ± 2.67 37-48 8.7 Length of hindfoot Male 8 16.0 ± 0.38 15-17 3.3 Female 8 16.0 ± 0.53 15-17 4.7 Length of ear Male 8 13.5 ± 0.38 13-14 4.0 Female 8 12.9 ± 0.59 12-14 6.5 Weight Male 8 12.1 ± 2.08 8-17 24.3 Female 8 13.3 ± 1.55 10-17 16.5 Greatest length of skull Male 8 20.5 ± 0.47 19.6-21.4 3.2 Female 8 20.8 ± 0.41 20.1-21.6 2.8 Condylobasal length Male 8 19.4 ± 0.51 18.1-20.2 3.7 Female 8 19.6 ± 0.37 19.1-20.6 2.7 Zygomatic breadth Male 8 III ± 0.26 10.5-11.5 3.4 Female 8 11.3 ± 0.25 10.7-11.8 3.2 Interorbital breadth Male 8 3.6 ± 0.13 3.4-4.0 5.1 Female 8 3.7 ± 0.07 3.6-3. 8 2.7 Rostral breadth Male 8 3.9 ± 0.14 3.6-4. 1 4.9 Female 8 3.9 ± 0.14 3. 7-4.3 5.0 Table 2. — Continued Measurements and sex N 1 Vlean ± 2 SE Range CV Oblique length of bulla Male 6 4.7 ± 0.08 4. 5-4.8 2.2 Female 8 4.7 ± 0.08 4. 5-4. 8 2.5 Greatest length of bulla Male 6 7.0 ± 0.12 6.7-7. 1 2.2 Female 7 7.1 ± 0.26 6.4-7. 4 4.8 Length of maxillary toothrow Male 8 3.6 ± 0.10 3.4-3. 9 4.1 Female 8 3.7 ± 0.13 3. 5-4.0 4.9 Breadth across upper molars Male 8 5.1 ± 0.16 4. 6-5. 3 4.4 Female 8 5.1 ± 0.12 4.9-5.4 3.4 Length of anterior palatal foramen Male 8 4.0 ± 0.24 3. 4-4. 4 8.5 Female 8 3.9 ± 0.15 3. 5-4. 2 5.4 Length of posterior palatal foramen Male 7 0.4 ± 0.07 0. 3-0.5 21.7 Female 8 0.4 ± 0.05 0. 3-0.5 18.9 Length of diastema Male 8 5.0 ± 0.14 4. 7-5. 2 4.0 Female 8 4.9 ± 0.12 4.6-5. 1 3.4 Height of skull Male 6 7.8 ± 0.16 7.6-8. 1 2.5 Female 8 7.8 ± 0.19 7. 5-8. 2 3.4 Length of mandibular toothrow Male 8 3.1 ± O.ll 2. 7-3. 2 5.2 Female 8 3.1 ± 0.10 2. 9-3. 3 4.4 ments were found to be significantly different — in individuals of S. cuppedius from Panisau, Nigeria. Therefore, in all subsequent analyses the sexes were pooled. Age Variation Age categories used in this study are shown in Eig. 1; these categories are usually referred to as follows: Age I, juveniles; Age II, subadults; Age III, young adults; Age IV, adults. Age I is not illustrated in Fig. 1; M^ in this age category has not erupted fully. These categories are arbitrary, based on den- tal wear, and do not reflect reproductive age. Stea- tomys cuppedius populations tested for age varia- tion are the Senegal population, which had sufficient sample sizes in categories I, II, and III, and the Panisau, Nigeria, sample, which had ade- quate numbers in categories I, HI, and IV. Those 1978 SWANEPOEL AND SCHLITTER— 57EA70ME5 SYSTEMATICS 59 Table 3. — Variation with age in external and cranial measure- ments of Steatomys caurinus from Yama, Ivory Coast, and S. cuppedius from Panisau, Nigeria. Age classes are listed In de- creasing order from the largest mean. Vertical lines to the right of each array of means connect maximally nonsignificant sub- sets at the 5% level. Groups of means of nonsignificant differ- ences are labelled "ns." Measurements SS- and age classes N Mean ± 2 SE Range CV STP Steatomys caurinus Total length IV 10 166.8 ± 8.00 155-200 7.5 ! III 13 147.8 ± 6.54 134-169 8,0 1 II 14 135.3 ± 2.64 121-156 6.5 1 Length of tail IV 10 52.7 ± 3.94 45-64 11.8 1 III 13 45.7 ± 3.16 38-58 12.5 II 44 44.1 ± 1.55 35-55 11.7 Length of hindfoot IV 10 19.0 ± 0.73 17-21 6.1 III 13 18.8 ± 0.51 18-20 4.9 ns II 44 18.5 ± 0.25 17-20 4.4 Length of ear IV 10 18.1 ± 0.55 17-20 4.8 j III 13 17.7 ± 0.53 16-20 5.4 1 II 44 16.7 ± 0.27 15-19 5.3 1 Weight IV 10 41.8 ± 4.66 35-58 17.6 1 III 13 32.0 ± 6.38 22-54 35.9 1 II 14 22.7 ± 1.86 12-38 27.2 1 Greatest length of skull IV 0 III 4 25.4 ±1.14 24.3-26.5 4.5 1 II II 23.7 ± 0.35 22.7-24.6 2.4 1 Condylobasal length IV 5 25.8 ± 0.79 24.6-26.7 3.4 1 III 9 23.9 ± 0.63 22.8-25.3 4.0 ] II 29 22.7 ± 0.33 20.8-23.9 3.9 1 Zygomatic breadth IV 4 13.2 ± 0.24 12.9-13.4 1.8 1 III 7 12.3 ± 0.53 11.3-13.0 5.7 II 23 11.7 ± 0.20 10.7-12.5 4.1 Interorbital breadth IV 9 4.1 ± 0.07 3. 9-4. 2 2.5 j III 12 3.9 ± 0.07 3.6-4. 1 3.3 1 II 38 3.8 ± 0.05 3.4-4. 1 4.4 Rostral breadth IV 9 5.1 ± 0.15 4. 9-5. 5 4.3 1 III 13 4.4 ±0.11 4. 1-4.7 4.5 1 II 39 4.2 ± 0.06 3. 9-4.6 4.3 Table 3. — Continued. Measurements SS- and age classes N Mean ± 2 SE Range CV STP Oblique length of bulla IV 7 5.8 ± 0.24 5.3-6. 1 5.6 III 10 5.4 ± 0.15 4. 9-5. 7 4.3 11 34 5.3 ± 0.08 4. 8-5. 7 4.2 Greatest length of bulla IV 6 9.0 ± 0.33 8. 6-9.5 4.5 1 111 8 8,4 ± 0.18 8. 0-8. 8 3.1 11 31 8.2 ± 0.09 7. 7-8. 7 3.2 Length of maxillary toothrow IV 10 4.1 ± 0.14 3. 7-4. 4 5.5 111 13 4.1 ± 0.10 3. 8-4.4 4.3 ns 11 42 4.0 ± 0.05 3. 6-4.4 4.0 Breadth across upper molars IV 8 6.1 ± 0.16 5. 8-6.5 3.8 1 III 12 5.7 ± 0.10 5.4-6. 1 3.1 11 37 5.5 ± 0.07 5. 1-6,1 3,6 Length of anterior palatal foramen IV 9 5.6 ± 0.13 5. 3-5. 9 3.4 1 111 12 4.9 ± 0.19 4. 4-5. 4 6.7 11 37 4.6 ±0.11 4. 2-5. 9 7.1 Length of posterior palatal foramen IV 7 0.5 ± 0.10 0.3-0. 7 26.2 111 II 0,5 ± 0.07 0. 3-0.7 23.1 ns 11 32 0.5 ± 0.04 0. 3-0.7 24.3 Length of diastema IV 9 7.2 ± 0.24 6. 8-8.0 5.0 1 III 13 6.3 ± 0.30 5. 6-7.4 8.7 1 11 40 6.0 ± 0.10 5. 2-6. 6 5.4 1 Height of skull IV 4 9.6 ± 0.15 9. 5-9. 8 1.6 1 III 9 9.0 ± 0.23 8. 6-9. 7 3.9 1 II 26 8.7 ± 0.10 8.3-9. 1 3.0 1 Length of mandibular toothrow IV 10 3.6 ± 0.18 3.I-4.I 8.1 III 13 3.5 ± 0.08 3. 2-3. 7 4.3 ns II 43 3.5 ± 0.05 3. 2-3. 9 4.9 Steatomys cuppedius Total length IV 6 132.7 ± 7.04 120-143 6.5 ! III 16 118. 1 ± 3.66 105-133 6.2 1 II 33 108.8 ± 2.85 94-129 7.5 ] Length of tail IV 6 48.7 ± 2.29 46-54 5.8 1 III 16 43.6 ± 1.94 37-52 8.9 11 32 42.6 ± 1. 14 37-50 7.5 Length of hindfoot IV 6 16.0 ± 0.52 15-17 4.0 III 16 16.0 ± 0.32 15-17 4.0 ns II 33 15.8 ± 0.19 14-17 3.5 60 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 3. — Continued. Table 3. — Continued. Measurements and age classes N Mean 2 SE Range CV Length of ear IV 6 13.7 ± 0.42 13-14 3.8 III 16 13.2 ± 0.38 12-14 5.7 II 33 12.6 ± 0.19 12-14 4.3 Weight IV 6 16.2 2.75 13-22 20.9 III 16 12.7 1.29 8-17 20.3 II 33 9.2 ± 0.74 5-14 23.0 Greatest length of skull IV 5 22.4 0.88 21.3-23.9 4.4 III 16 20.7 ± 0.31 19.6-21.6 3.0 II 28 19.5 ± 0.29 17.5-21.4 3.9 Condylobasal length IV 6 21.0 ± 0.76 20.1-22.5 4.5 III 16 19.5 0.31 18.1-20.6 3.2 II 29 18.3 0.28 16.4-20.0 4,1 Zygomatic breadth IV 6 11,8 -h 0.38 11.2-12.4 4.0 III 16 11.2 0.19 10.5-11.8 3.4 II 32 10.8 ± 0.12 10.2-1 1.5 3.1 Interorbital breadth IV 6 3.9 ± 0.16 3.6-4. 1 5.0 III 16 3.7 0.07 3. 4-4.0 4.0 II 33 3.6 ± 0.05 3. 3-3. 9 3.6 Rostral breadth IV 6 4.2 0.24 3. 9-4. 7 6.9 III 16 3,9 0.09 3. 6-4. 3 4.8 II 33 3.7 ± 0.07 3.3-4. 1 5.3 Oblique length of bulla IV 5 4.8 4: 0.23 4. 4-5.0 5.3 III 14 4.7 ± 0,06 4. 5-4. 8 2.3 II 23 4.5 ± 0.09 4. 0-4. 9 4.7 Greatest length of bulla IV 5 7.4 0.19 7. 1-7.7 2.9 III 13 7.1 0.15 6.4-7.4 4.0 II 21 6.8 0.12 6.2-7. 3 4.0 Length of maxillary toothrow IV 6 3.8 0.12 3. 6-4.0 3.9 III 16 3.7 0.09 3. 4-4.0 4.7 II 32 3.6 ± 0.06 3. 3-4.0 4.9 Breadth across upper IV molars 6 5.3 0.19 5. 1-5.7 4,4 III 16 5.1 ± 0.10 4. 6-5.4 3.8 II 33 4.9 ± 0.07 4.4-5. 6 4.3 Length of anterior palatal foramen IV 6 4.2 ± 0.18 4. 0-4. 6 5.3 III 16 4.0 ± 0.14 3.4-4. 4 6.9 II 32 3.7 0.08 3. 1-4.2 6.5 SS- STP Measurements and age classes N Mean ± 2 SE Range CV SS- STP Length of posterior palatal foramen IV 6 0.4 ± 0.08 0. 3-0.5 20.4 III 15 0.4 ± 0.04 0. 3-0.5 20.3 ns II 31 0.4 ± 0.03 0. 3-0.6 18.5 Length of diastema IV 6 5.3 ± 0.35 4.8-5. 9 8.0 1 III 16 5.0 ± 0.09 4. 6-5. 2 3.8 1 II 33 4.5 ± 0.08 4. 1-5.0 5.4 1 Height of skull IV 6 8.2 ± 0.18 8. 0-8. 5 2.7 1 III 14 7.8 ± 0.12 7. 5-8. 2 2.9 ! II 24 7.6 ± 0.08 7. 2-7. 9 2.7 Length of mandibular toothrow IV 6 3.4 ± 0.12 3. 2-3. 6 4.5 1 III 16 3.1 ± 0.07 2. 7-3. 3 4.7 II 31 3.1 ± 0.05 2. 7-3.4 4.9 populations of S. caiirinits tested were from Yama, Ivory Coast, with sufficient individuals of cate- gories II, III, and IV; Pirisi, Ghana (II and III); Gudi, Nigeria (II and III); and Celia, Upper Volta (II and III). For each of the 19 external and cranial measurements, samples of combined sexes were tested to determine if the differences between the means of the age categories were significantly dif- ferent {P ^ 0.05). If found to be significantly dif- ferent, the Sum of Squares Simultaneous Test Pro- cedure was used to find maximally nonsignificant subsets. The results of these tests in S. cuppedius from Panisau, Nigeria (age II, III, IV), and S. caiir- inits from Yama, Ivory Coast (age II, HL IV), are presented (Table 3). Steatomys caiiriniis. — In the population from Yama, Ivory Coast, nonoverlapping subsets among age categories II, III, and IV were found in six (total length, weight, condylobasal length, rostral breadth, diastemal length, and height of skull) of 19 mea- surements; no greatest length of skull measure- ments are available for age category IV in this sam- ple. Little or no variation in age exists in the following measurements: length of hindfoot; inter- orbital breadth; lengths of maxillary toothrow; pos- terior palatal foramen; mandibular toothrow. A sim- ilar trend was found in the other populations of S. canriniis tested but is not presented in tabular form. 1978 SWANEPOEL AND SCHLITTER— 57EA70MK5 SYSTEMATICS 61 ii o t/3 il> ^ :■= H -a o j o U C C/5 C/5 !>) C C -5 0> . — c/5 a ^ 2 - o. :s -O — >. c 2 u 'ij « 5: X V o u (U 1/ V o n CO V c/5 cy^ ^ ^ Ji2 1- U- ? CL O J= U C/5 iD 2,1 CT3 CJ •- L- 0> -C 0> CL ^ jC ^ o O -ZZ < > u CT3 (U c/5 £ ^ O ^ -D ^ c^ E ~ t: 5! E S X c/5 , 1> Q "'3 O ;= ^ o O “ s Ll, -o are reported from M’Bour, Senegal. 62 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 Steatomys cuppedius. — Three nonoverlapping subsets among age categories II, HI, and IV in nine (total length, weight, greatest length of skull, con- dylobasal length, zygomatic breadth, rostral breadth, greatest length of bulla, length of diastema, and height of skull) of 19 measurements were found in the population from Panisau, Nigeria (Table 3). Three characters (lengths of hindfoot, maxillary toothrow, and anterior palatal foramen) showed no significant differences among the different age cat- egories tested. In comparing measurements from the different age categories (I, II, and III) in the population from Senegal, a similar trend emerges, although it is not presented in tabular form. Conclusions. — From these results, it is apparent that none of the age categories can be pooled in either species for a study of geographic variation. It would be ideal if age IV (adult) specimens could be used for a study of geographic variation in these species. However, due to a lack of sufficient sam- ples of age IV specimens, only specimens belonging to age category HI were used. 7/ 1 divi dual V a ria tion The majority of measurements examined re- vealed a relatively low degree of individual varia- tion as expressed by the coefficient of variation. External measurements were generally more vari- able than the cranial measurements (Tables 1, 2, 3). This was, however, not as pronounced in Steato- mys cuppedius as in Steatomys caurinus. Of exter- nal measurements, weight showed the most varia- tion, with coefficients of variation above 20. Cranial measurements in both species usually had coeffi- cients of variation of less than 5, except in length of posterior palatal foramen where these values usually were above 15. Conclusions. — Due to the tendency of the exter- nal measurements, especially weight, to show rel- atively higher individual variation expressed as coefficients of variation than did the cranial mea- surements, all external measurements were exclud- ed from the multivariate analyses of geographic variation. Of the cranial measurements, length of posterior palatal foramen was excluded from this analysis due to its high level of individual variation. RELATIONSHIPS OF SPECIES All individuals from West Africa of age category HI, and a few specimens that were borderline be- tween category HI and IV in age, were grouped into 13 geographic samples. These geographic samples (OTUs) are as follows (see also Fig. 2): 1) Kouande, Benin; 2) Bangwon, Ghana; 3) Pirisi, Ghana; 4) Gudi, Nigeria; 5) Dio and Barga, Upper Volta; 6) Celia and Nayoure, Upper Volta; 7) Fo, Upper Vol- ta; 8) Wenchi, Ghana; 9) Tyenko, Ivory Coast; 10) Diali, Ivory Coast; 11) Sienso and Yama, Ivory Coast; 12) Diourbel, Kaffrine, Kaolack and Kough- eul, Senegal; and 13) Panisau, Nigeria. Steatomys caurinus was described by Thomas (1912) as being generally brownish fawn, but with a darker mid-dorsal area and more distinct color along the sides of the belly. Rosevear (1969) sum- marized the color of S. caurinus as being “medium warm reddish brown” (the tips of the hair were warm brown and the bases were dark gray). Ac- cording to the description by Hayman (1936), the dorsal pelage of S. jacksoni is plumbeous drab in color, with the flanks slightly paler, whereas Ro- sevear (1969) describes this species as being dark brownish-gray in color of pelage. A large number of the specimens examined in the present study have skins soiled by residual fats and oils, making an accurate definition of an already variable char- acter, color of pelage, difficult. The dorsal pelage color of S. cuppedius was described originally by Thomas and Hinton (1920) as being pale drab with pale sides, and more recently by Rosevear (1969) as being pale, slightly sandy gray. Specimens of this species generally did not have soiled skins such as was found for S. caurinus. Hayman (1936) stated in the original description of S. jacksoni that the braincase of the holotype was not sharply truncated as in other species of the genus he had examined, and that the interparietal was lengthened from front to back and noticeably widened anteroposteriorly in the middle part. We have found in the present study that the older ani- mals, such as age category IV individuals, exhibit relatively less truncated skulls than do the younger animals. Hayman (1936) gave a length of 4.5 mm for the interparietal of 5. jacksoni as compared to 2.5 mm inS. caurinus. The length of the interparietal bone of all specimens examined in the present study, with the exception of the holotype oiS.jack- 1978 SWANEPOEL AND SCHLlTTER—SrEATOMYS SYSTEMATICS 63 Table 4. — Geographic variation in external and cranial mea- surements of age category III individuals among 10 geographic samples (OTUs) of Steatomys caurinus and two geographic samples ofS. cuppedius. Results ofANOVA analysis indicating significance at the 5% level for S. cuppedius are indicated by an asterisk next to the character heading. See Fig. 2 and text for key to localities included in each OTU. Sam- pie N Mean ± 2 SE Range CV Steatomys caurinus Total length 1 1 167.0 2 2 129.0 121-137 3 21 158.0 ± 3.52 137-169 5.1 4 36 159.2 ± 3.93 126-178 7.4 5 2 142.5 142-143 6 18 161.2 ± 5.82 136-178 7.7 7 1 154.0 9 5 164.4 ±9.11 149-173 6.2 10 3 168.7 ± 27.1 142-186 13.9 11 15 148.6 ± 6.48 134-169 8.4 Length of tail 1 1 48.0 2 2 37.0 35-39 3 21 42.5 ± 2.26 36-59 12.2 4 36 45.9 ± 0.21 37-52 7.9 5 2 40.0 38-42 6 18 45.8 ± 2.09 40-53 9.7 7 1 51.0 9 5 55.4 ± 3.2 50-59 6.5 10 3 59.0 ± 12.5 47-68 18.3 11 15 46.6 ± 3.33 38-60 13.8 Length of hindfoot 1 1 18.0 2 2 17.5 17-18 3 21 19.7 ± 0.29 18-21 3.3 4 36 19.7 ± 0.21 18-21 3.2 5 2 19.0 19 6 18 18.8 ± 0.50 17-21 5.6 7 1 18.0 9 5 19.0 ± 1.10 17-20 6.4 10 3 19.0 ±1.15 18-20 5.3 11 15 18.7 ± 0.46 18-20 4.7 1 1 Length of ear 16.0 2 2 15.5 15-16 3 21 17.2 ± 0.36 16-18 4.7 4 36 17.6 ± 0.36 15-20 6.1 5 2 15.0 15 6 18 17.1 ± 0.57 14-19 7.1 7 1 18.0 9 5 17.4 ± 0.49 17-18 3.4 10 3 18.0 ± 2.0 16-19 9.6 11 15 17.7 ± 0.54 16-20 5.9 Weight 1 1 28.0 2 2 31.5 26-37 3 21 54.0 ± 4.26 34-68 18.1 Table 4. — Continued. Sam- ple N Mean ± 2 SE Range CV 4 35 48.2 ± 3.56 25-68 21.9 5 2 30.5 30-3 1 6 18 48.4 ± 5.35 23-70 23.4 7 1 17.0 9 5 39.6 ± 7.71 32-53 21.8 10 3 51.3 ± 24.0 29-70 40.4 11 15 31.7 ± 5.59 22-54 34.2 Greatest length of skull 1 2 1 2 3 17 26.4 ± 0.39 24.9-27.8 3.1 4 32 26.6 ± 0.38 24.2-28.4 4.0 5 1 23.7 6 9 26.6 ± 0.72 24.3-28.0 4.1 7 1 25.5 9 2 27.6 ± 1.00 27.1-28.1 2.6 10 3 — 11 5 25.2 ± 1.03 24.1-26.5 4.6 Condylobasal length 1 2 1 2 3 17 25.3 ± 0.41 23.4-26.6 3.3 4 26 25.3 ± 0.37 23.1-27.0 3.7 5 2 23.0 22.6-23.4 6 10 25.3 ± 0.87 22.8-27.0 5.5 7 1 24 2 9 5 25.6 ± 1.05 23.8-27.0 4.6 10 3 25.7 ± 2.60 23.2-27.5 8.8 11 1 1 23.9 ± 0.68 22.5-25.8 4.7 Zygomatic breadth 1 2 1 ') 3 16 13.0 ± 0.19 12.2-13.6 3.0 4 22 13.1 ± 0.23 12.2-13.8 4.1 5 1 12.0 6 12 12.8 ± 0.31 12.0-13.6 4.2 7 1 12.4 9 5 13.1 ± 0.61 12.1-13.8 5.2 10 3 13.5 ± 0.98 12.5-14.1 6.3 11 9 12.3 ± 0.51 1 1.3-13.5 6.2 Interorbital breadth 1 2 1 1 3.5 3 20 3.9 ± 0.08 3. 6-4. 2 4.7 4 34 3.9 ± 0.06 3. 5-4. 3 4.1 5 2 3.9 3.9 6 18 3.9 ± 0.06 3.7-4. 1 3.1 7 1 3.9 9 5 4.2 ± 0.19 3. 8-4.3 5.0 10 3 4.0 ± 0.29 3. 8-4. 3 6.2 11 14 3.9 ± 0.09 3. 6-4. 3 4.4 Rostral breadth 1 2 1 2 4.1 4.0-4.2 3 21 4.8 ± 0.12 4.4-5. 3 5.5 64 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 4. — Continued. Table 4. — Continued. Sam- Sam- pie N Mean ± 2 SE Range CV pie N Mean ± 2 SE Range CV 4 35 4.7 ±0.11 3. 9-5. 3 6.9 4 35 4.8 ± 0.08 4.4-5. 2 4.6 5 2 4.4 4. 0-4.7 5 2 4.3 4.3 6 17 4.9 ± 0.13 4. 5-5. 4 5.5 6 13 5.1 ± 0.21 4. 5-5. 5 7.4 7 1 4.6 7 1 5.0 9 5 4.8 ± 0.24 4.4-5. 1 5.6 9 5 5.3 ± 0.42 4.6-5. 9 8.9 10 3 4.9 ± 0.58 4.4-5.4 10.3 10 3 5.2 ± 0.70 4. 6-5. 8 11.6 11 15 4.5 ± 0.12 4. 1-4.9 5.1 11 14 4.9 ± 0.17 4.4-5. 4 6.5 Oblique length of bulla Length of posterior palatal foramen 1 2 1 1 5.2 1 2 1 2 0.5 0.4-0.5 3 17 5.9 ± 0.10 5. 5-6.3 3.5 3 21 0.5 ± 0.08 0.3-1. 0 33.0 4 31 5.8 ± 0.09 5. 2-6.5 4.3 4 30 0.5 ± 0.04 0.3-0. 7 24.2 5 2 6.0 5. 5-6.5 5 2 0.5 0.5 6 10 5.9 ± 0.19 5.6-6.5 5.0 6 17 0.6 ± 0.06 0.4-0. 8 17.4 7 1 5.7 7 1 0.4 9 5 5.7 ± 0.24 5.4-6.0 4.8 9 5 0.5 ± 0.07 0.4-0.6 16.1 10 3 6.0 ± 0.18 5.8-6. 1 2.5 10 3 0.6 ± 0.31 0.4-0.9 44.1 11 12 5.4 ± 0.15 4.9-5. 9 4.7 11 15 0.5 ± 0.07 0.3-0. 7 23.4 Greatest length of bulla Length of diastema 1 2 1 1 8.0 1 2 1 2 6.2 6.0-6.4 3 15 8.8 ± 0.15 8. 4-9. 2 3.3 3 21 7.2 ± 0.15 6. 5-7. 8 4.6 4 30 9.0 ± 0.13 8.0-9.7 3.8 4 36 7.0 ± 0.14 6.0-7. 6 5.9 5 2 8.2 8.0-8. 3 5 2 6.6 6. 3-6.8 6 9 8.7 ± 0.21 8. 2-9.3 3.6 6 17 7.0 ± 0.21 6.3-7. 7 6.2 7 1 8.5 7 1 6.8 9 5 9.0 ± 0.55 8. 1-9.6 6.8 9 5 6.9 ± 0.34 6.7-7.6 5.5 10 3 9.0 ± 0.80 8. 8-9.4 3.8 10 3 7.2 ± 1.00 6.3-8. 0 12.0 11 10 8.6 ± 0.28 8. 0-9.6 5.1 11 15 6.3 ± 0.28 5.8-7.4 5.6 Length of maxillary toothrow Height of skull 1 2 1 2 3.9 3.9 2 2 3 21 4. 1 ± 0.08 3. 7-4.4 4.6 3 14 9.2 ± 0.19 8.4-9. 7 3.8 4 36 4.2 ± 0.05 3. 8-4. 5 3.8 4 32 9.2 ± 0.17 7.7-10.1 5.2 5 2 4.0 3.8-4. 1 5 2 9.2 8. 7-9.6 6 17 4.3 ± 0.10 3.9-4. 6 4.9 6 11 9.3 ± 0.24 8.5-10.0 4.3 7 1 3.8 7 1 9.1 9 5 4.2 ± 0.12 4. 1-4.4 3.1 9 5 9.6 ± 0.41 8.9-10.0 4.7 10 3 4.1 ± 0.29 3.8-4. 3 6.1 10 3 9.9 ± 0.77 9.2-10.5 6.7 11 15 4.1 ± 0.08 3. 8-4. 4 4.0 11 11 9.0 ± 0.20 8. 6-9.2 3.8 Breadth across upper molars Length of mandibular toothrow 1 2 1 2 5.5 5.4-5. 5 1 2 1 2 3.3 3.3 3 21 5.9 ± 0.10 5. 6-6.4 3.7 3 20 3.5 ± 0.10 3. 1-3.9 6.2 4 34 6.0 ± 0.10 5.0-6.5 4.8 4 34 3.5 ± 0.06 3. 1-3.9 5.4 5 2 5.8 5. 7-5. 8 5 2 3.2 3.2 6 17 5.9 ± 0.17 5. 1-6.4 5.9 6 16 3.5 ± 0.07 3. 3-3. 8 4.2 7 1 6.0 7 1 3.2 9 5 6.0 ± 0.19 5. 7-6. 2 3.6 9 5 3.5 ± 0.12 3. 3-3. 6 3.9 10 3 6.0 ± 0.42 5.7-6.4 6.0 10 3 3.5 ± 0.29 3. 2-3. 7 7.2 11 14 5.7 ± 0.13 5. 4-6. 3 4.2 11 15 3.5 ± 0.08 3. 2-3. 7 4.2 Length of anterior palatal foramen Steatomys cuppedius 1 1 — Total length* 2 2 4.3 4. 0-4. 5 12 5 125.6 ± 2.58 122-130 2.3 3 21 5.1 ±0.11 4.7-5. 6 4.9 13 16 118.1 ± 3.66 105-133 6.2 1978 SWANEPOEL AND SCHLITTER— 57EA70MF5 SYSTEMATICS 65 Table 4. — Continued. Table 4. — Continued. Sam- Sam- ple N Mean ± 2 SE Range CV pie N Mean ± 2 SE Range CV 12 5 Length of tail 44.8 ± 1.60 43-47 4.0 13 16 43.6 ± 1.94 37-52 8.9 12 5 Length of hindfoot 15.8 ± 0.40 15-16 2.8 13 16 16.0 ± 0.32 15-17 4.0 12 5 Length of ear 12.8 ± 0.40 12-13 3.5 13 16 13.2 ± 0.38 12-14 5.7 12 5 Weight* 16.8 ± 5.31 7-23 35.3 13 16 12.7 ± 1.29 8-17 20.3 12 2 Greatest length of skull* 22.0 ± 0.40 21.8-22.2 1.3 13 16 20.7 ±0.31 19.6-21.6 3.0 12 3 Condylobasal length 20.2 ± 0.35 19.9-20.5 1.5 13 16 19.5 ± 0.31 18.1-20.6 3.2 12 2 Zygomatic breadth 11.6 ±0.30 11.4-11.7 1.8 13 16 11.2 ± 0.19 10.5-11.8 3.4 12 5 Interorbital breadth 3.6 ± 0.07 3. 5-3. 7 2.3 13 16 3.7 ± 0.07 3.4-4. 0 4.0 12 5 Rostral breadth 4.1 ±0.14 3. 9-4. 3 3.9 13 16 3.9 ± 0.09 3. 6-4. 3 4.8 12 3 Oblique length of bulla* 5.1 ±0.12 5.0-5. 2 2.0 13 14 4.7 ± 0.06 4. 5-4. 8 2.3 12 2 Greatest length of bulla 7.5 ±0.10 7.4-7. 5 0.9 13 13 7.1 ±0.15 6.4-7.4 4.0 12 5 Length of maxillary toothrow 3.6 ±0.16 3.4-3.9 5.0 13 16 3.7 ± 0.09 3.4-4.0 4.7 12 4 Breadth across upper molars 5.1 ± 0.08 5. 0-5. 2 1.6 13 16 5.1 ±0.10 4.6-5.4 3.8 12 3 Length of anterior palatal foramen 4.2 ± 0.07 4. 1-4.2 1.4 13 16 4.0 ±0.14 3.4-4.4 6.9 12 4 Length of posterior palatal foramen 0.4 ± 0.08 0. 3-0.5 20.4 13 15 0.4 ± 0.04 0.3-0.5 20.3 12 4 Length of diastema 5.1 ±0.13 4. 9-5, 2 2.5 13 16 5.0 ± 0.09 4.6-5. 2 3.8 Height of skull 12 4 8.1 ± 0.47 7.4-8.4 5.9 13 14 7.8 ±0.12 7. 5-8. 2 2.9 Length of mandibular toothrow 12 4 3.2 ±0.15 3. 1-3.4 4.7 13 16 3.1 ± 0.07 2. 7-3. 3 4.7 soni, measured less than 4 mm. The interparietal bones varied in shape from nearly triangular to quadrangular. This variation in shape occurred among specimens within the same age category. In his original description, Hayman (1936) con- cluded that S.jacksoni was not bigger than S. caiir- inus, but he compared only the external measure- ments of the two taxa. Thus the single most important character used by Hayman (1936) in de- scribing S. jacksoni was the size and shape of the interparietal bone. When the holotype was exam- ined, the interparietal bone did measure 4 mm in length and was clearly longer than any of the other specimens of Steatoinys examined. Based on this diagnostic character, none of the specimens we ex- amined could be assigned to S. jacksoni. Univariate analysis. — Results of standard uni- variate statistical analysis for individuals of Stea- tomys from the 13 geographic samples are given in Tables 4 and 5. In age category III specimens (Table 4), there is no overlap in the range of measurements between geographic samples 12 and 13 (S. cnppe- dius) and those of the remainder of the West Afri- can geographic samples in four of the 19 characters analyzed (greatest length of skull, condylobasal length, greatest length of bulla, and length of dia- stema). Comparison of these same analyses for age category IV individuals (Table 5) reveals that 12 of 19 characters (total length, length of ear, weight, greatest length of skull, condylobasal length, zy- gomatic breadth, rostral breadth, oblique length of bulla, greatest length of bulla, width across upper molars, length of anterior palatal foramen, length of diastema, and height of skull) show no overlap be- tween geographic samples 12 and 13 and the re- maining geographic samples analyzed. Steatoinys ciippedins is readily distinguishable from other West African fat mice in being markedly smaller in size, both externally and cranially, and much paler in color of pelage. 66 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 5. — Geographic variation in external ami crania! mea- Table 5. — Continued. surements of age category IV individuals among six geographic samples (OTUs) of Steatomys caurinus and two geographic Sam- samples of S. cuppedius. See Fig. 2 and text for key to localities p|e N Mean ± 2 SE Range CV included in each OTU. Sam- ple N Mean ± 2 SE Range CV 3 4 6 9 10 II 3 4 6 9 10 i I 3 4 6 9 10 1 1 3 4 6 9 10 II 3 4 6 9 10 11 3 4 6 9 10 11 3 4 6 9 10 1 1 Steatomys caurinus Total length 1 155 2 166 162-170 1 179 I 188 1 199 10 166.8 ± 8.0 155-200 7.5 Length of tail 1 41 2 43 41-45 1 47 1 65 10 52.7 ± 3.9 45-64 Length of hindfoot 1 18.0 2 19.5 19-20 1 18.0 1 20 1 21 10 19.0 ± 0.73 17-21 l.ength of ear 1 17.0 2 17.5 17-18 I 18.0 1 18.0 I 18.0 10 18.1 ±0.55 17-20 Weight 1 56.0 2 55.5 54-57 I 70.0 1 69.0 1 69.0 10 41.8 ± 4.66 35-58 Greatest length of skull I 27.2 I 29.3 0 1 28.5 0 0 Condylobasal length 1 25.6 1 28.0 I 27.2 I 27.1 1 28.3 5 25.8 ± 0.79 11.8 6.1 4.8 17.6 3.4 3 4 6 9 10 II 3 4 6 9 10 II 3 4 6 9 10 1 1 3 4 6 9 10 II 3 4 6 9 10 II 3 4 6 9 10 II 3 4 6 9 10 II 3 4 1 1 I 1 I 4 1 1 I 1 9 1 2 I I I 9 I I 0 I I 7 1 I 0 I I 6 1 T 10 I T I 1 8 Zygomatic breadth 13.1 15.1 13.9 14.2 14.5 13.2 ± 0.24 12.9-13.4 Interorbital breadth 4.2 4.1 3. 9-4. 2 4.1 4.0 4.2 4.1 ± 0.07 3. 9-4. 2 Rostral breadth 4.8 5.5 5.5 5.3 5.4 5. 3-5. 6 4. 9-5. 5 Oblique length of bulla 5.8 5.9 6.2 6.1 5.8 ± 0.33 8. 6-9.5 Greatest length of bulla 9.0 9.0 9.4 9.9 9.0 ± 0.33 8. 6-9.5 Length of maxillary toothrow 3.9 4.5 4.4-4.5 4.4 4.3 3.9 4.1 ±0.14 3. 7-4.4 Breadth across upper molars 6.0 6.4 5. 9-6.8 6.2 6.4 6.1 6.1 ±0.16 5. 8-6. 5 1.8 2.5 4.3 4.5 4.5 5.5 3.8 Length of anterior of anterior palatal foramen I 5.2 I 5.1 24.6-26.7 1978 SWANEPOEL AND SCHLITTER— 57EA70ME5 SYSTEMATICS 67 Table 5. — Continued. Table 5. — Continued. Sam- ple N Mean ± 2 SE Range CV Sam- ple N Mean ± 2 SE Range CV 6 1 5.6 Condylobasal length 9 1 5.2 12 2 20.3 19.9-20.7 10 1 6.0 13 6 21.0 ± 0.76 20.1-22.5 4.5 1 1 9 5 6 + 013 5 3-5.9 3.4 Zygomatic breadth Length of posterior palatal foramen 12 2 11.1 10.8-11.3 3 1 0.7 13 6 11.8 ±0.38 11.2-12.4 4.0 4 1 0 5 0 5 Interorbital breadth 6 1 0.6 9 1 0.6 12 2 3.7 3.7 10 1 0.5 13 6 3.9 ±0.16 3.6-4. 1 5.0 11 7 0.5 ±0.10 0. 3-0.7 26.2 Rostral breadth Length of diastema 12 1 4.1 3 1 7.3 13 6 4.2 ± 0.24 3.9-4. 7 6.9 4 2 7.5 6. 9-8.0 Oblique length of bulla 6 1 7.1 12 2 4.7 4. 5-4. 9 9 1 7.4 13 5 4.8 ± 0.23 4. 4-5.0 5.3 10 1 8.6 11 9 7.2 ± 0.24 6. 8-8.0 5.0 Greatest length of bulla 12 7.3 6. 9-7. 6 Height of skull 13 5 7.4 ±0.19 7. 1-7.7 2.9 3 1 9.3 4 1 10.6 Length of maxillary toothrow 6 1 10.1 12 2 3.7 3. 6-3. 8 9 1 10.5 13 6 3.8 ±0.12 3.6-4. 0 3.9 10 1 10. 1 Breadth across upper molars 11 4 9.6 + 0.15 9. 5-9.8 1.6 12 T 5.0 5.0 Length of mandibular toothrow 13 6 5.3 ±0.19 5. 1-5.7 4.4 3 0 Length of anterior palatal foramen 4 2 3.6 3. 5-3.6 12 2 3.8 3.5-4. 1 D 1 A.O 13 6 4.2 + 0.18 4. 0-4. 6 5.3 9 1 3.7 10 0 Length of posterior palatal foramen 11 10 3.6 ±0.18 3. 1-4.1 8.1 12 2 0.5 0.4-0.5 13 4 0.4 + 0.08 0.3-0. 5 : !0.4 Steatomvs ciippedius Total length Length of diastema 12 2 125.0 124-125 12 2 4.9 4. 7-5.0 13 6 132.7 ± 7.04 120-143 6.5 13 6 5.3 ± 0.35 4. 8-5. 9 8.0 Length of tail Height of skull 12 2 44.0 44 12 2 7.6 7. 5-7. 6 13 6 48.7 ± 2.29 46-54 5.8 13 6 8.2 ±0.18 8. 0-8. 5 2.7 Length of hindfoot Length of mandibular toothrow 12 2 15.0 15 12 2 3.1 3. 0-3. 2 13 6 16.0 ± 0.52 15-17 4.0 13 6 3.4 ±0.12 3. 2-3. 6 4.5 Length of ear 12 2 13.0 13 The holotype of 5. jacksoni is an individual that 13 6 13.7 ± 0.42 13-14 3.8 is a very old age category III or a young age : cate- Weight gory IV animal (Table 5, Sample 8). However , even 12 1 19.0 if placed as an age category 111 individual. it still 13 6 16.2 ± 2.75 13-22 20.9 falls within the normal variation of cranial mea- Greatest length of skull surements of S. caunnits. Zygomatic breadth of the 12 2 21.7 21.1-22.3 holotype appears to be relatively narrow even for 13 5 20.4 ± 0.88 21.3-23.9 4.5 an age III specimen. 68 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 3 0.2891 4 0.3208 6 0.5781 9 0.5030 10 0.9313 5 0.6513 7 0.7790 11 2.0873 12 0.5265 13 I 1 i 1 ^ 1 1 1 2.250 1.950 1.650 1.350 1.050 0.750 0.450 0.150 Fig. 3. — Distance phenogram of OTUs (see Fig. 2 and text for key to localities included in each OTU) of West African taxa of Steatomys computed from distance matrix based on standardized characters and clustered by unweighted pair-group method using arithmetic averages (UPGIVIA), OTUs consist of geographic samples of combined sexes. The cophenetic correlation for the phenogram is 0.94. Multivariate analysis. — Twelve cranial measure- ments of 10 geographic samples of age category III specimens of combined sexes were analyzed using the NT-SYS routines. Length of posterior palatal foramen and greatest length of skull were excluded from this analysis. Inclusion of greatest length of skull would have caused reduced geographic sam- ples (such as deletion of sample 10) and sample sizes because of missing data. A distance phenogram diagramming the relation- ship among 10 geographic samples (OTUs) of Stea- tomys is shown in Fig. 3. The cophenetic correla- tion coefficient for the phenogram is 0.946. The geographic samples fall into two major clusters. The upper group (OTUs 3, 4, 6, 9, 10, 5, 7, 11) repre- sents medium- to large-sized individuals presently assigned to 5. caurinus. The lower group represents small individuals from localities in Senegal (OTU 12) and Panisau, Nigeria (OTU 13) and corresponds to the presently recognized S. cuppedius. The first three principal components computed from the matrix of correlation among 12 cranial characters for the same 10 geographic samples (OTUs) are shown in Fig. 4. The proportion of the total phenetic variation accounted for in the first three components is 95.3%. The amounts of vari- ance assigned to each component are 86.9% for component I, 5.2% for component II, and 3.2% for component III. A factor matrix of character load- ings among the 12 cranial characters of the principal component analysis is given in Table 6. From the factor analysis it can be seen that the first principal component is heavily influenced by general overall cranial size. The second component is influenced most by length of mandibular toothrow, and the third by interorbital breadth. Examination of the three-dimensional plot (Fig. 4) reveals an expres- sion of phenetic variation similar to that shown in the distance phenogram (Fig. 3); overall cranial size is expressed as progressively larger from left to right. OTUs 12 (Senegal) and 13 (Panisau, Nigeria) cluster on the left, and all other samples cluster in a fairly loose group on the right. Subgroupings with- in the right cluster of OTUs correspond with those found in the distance phenogram. These subgroup- ings will be discussed in detail in the systematic accounts that follow. Taxonomic Conclusions We interpret the univariate and multivariate anal- yses to indicate that in West Africa the genus Stea- tomys is represented by three species. Geographi- cally, they are distributed as follows: Steatomys caurinus is known from Senegal, Ivory Coast, Up- per Volta, Ghana (excluding Wenchi), Togo, Benin, and central Nigeria; S. cuppedius from Senegal, northern Nigeria, and south central Niger; and S. jacksoni from Wenchi, Ashanti District, Ghana. The latter species is known only from the holotype. Steatomys cuppedius is clearly the smallest of the three. S. jacksoni and S. caurinus are similar in size; the only appreciable difference between them is the size of the interparietal bone, with S. jacksoni having a significantly larger one. Little variation in size and shape of this bone is present in all of the 1978 SWANEPOEL AND SCHLITTER— 57EA70M K5 SYSTEMATICS 69 Fig. 4. — Three-dimensional projection of iO OTUs of West African taxa of Steaioniys on to the first three principal components based upon a matrix of correlation among 12 cranial measurements. OTUs consist of geographic samples of combined sexes. See Fig. 2 and text for key to localities included in each OTU. Component I accounts for 86.Wc of the phenetic variation, component II for 5.2%, and component III for 3.2% for a combined expression of 95.3% of the total variation. Other material examined of the genus from West Africa. The possibility that the holotype of S.jack- soni might be an aberrant individual is recognized, but until additional specimens from the vicinity of the type locality of S. jacksoni are available for study, we prefer to retain this taxon as a valid spe- cies. Hubert et al. (1973) reported S. caiihnus from Senegal but presented no measurements. The genus Steatomys, with no species given, has also been reported from owl pellets collected from Senegal (Heim de Balsac, 1965, 1967), but once again no measurements were given. The material upon which these reports were based has not been ex- amined by us. We did not have available any S. caiuinus from Senegal and feel this record should be reexamined to see if the material might be S. cuppedius. SYSTEMATIC ACCOUNTS Within the following accounts, species and sub- species, if appropriate, are listed in alphabetical or- der. Steatomys caurinus Thomas, 1912 Geographic distribution of species. — Central Ni- geria, southern Niger, northern Benin, Togo, west- ern Ghana, Upper Volta, central to northern Ivory Coast, and Senegal (Fig. 2); most certainly more widespread and occurring in intervening areas. Diagnosis. — Large, both externally and cranial- ly, for genus in West Africa; pelage dark and indi- vidual hairs coarse; relatively small molars; inter- parietal bone short in length, usually less than 3.5 mm. Comparisons . — Steatomys caurinus can be dis- tinguished from S. cuppedius by a number of char- acters. Comparisons of greatest length of skull, con- dylobasal length, greatest length of bulla, and length of diastema show no overlap in measurements of age category 111 specimens between these two spe- cies. These same four characters, and an additional eight measurements, indicate no overlap between age category IV specimens of the two species. These eight are total length, length of ear, weight, zygomatic breadth, rostral breadth, oblique length of bulla, width across upper molars, length of an- terior palatal foramen, and height of skull. Externally, S. caurinus can be separated from S. cuppedius by dark color of pelage and coarse tex- ture of the hairs as opposed to the pale color of pelage and silky texture of the hairs in the latter species. From S. jacksoni, S. caurinus differs little in ex- 70 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 6, — Factor matrix of character loading of the first three principal components among 12 characters of West African taxa of Steatomys {left) and «/ Steatomys caurinus {right). Steatomys (three species) Steatomys caurinus Character Component I Component II Component III Component I Component II Component III Condylobasal length 0.996 0.007 -0.021 0.983 -0.076 0.036 Zygomatic breadth 0.958 0.089 -0.053 0.972 0.044 0.032 Interorbital breadth 0.849 -0.027 0.496 0.547 -0.133 -0.619 Rostral breadth 0.980 0.020 -0.089 0.941 0.046 0.146 Oblique length of bulla 0.874 -0.389 -0.244 0.283 0.810 0.424 Greatest length of bulla 0.985 0.043 -0.014 0.921 -0.236 -0.058 Length of maxillary toothrow 0.879 0.248 -0.092 0.568 -0.463 0.511 Breadth across upper molars 0.952 -0.242 0.051 0.692 0.497 -0.315 Length of anterior palatal foramen 0.921 0.154 0.157 0.804 -0.254 -0.315 Length of diastema 0.969 -0.191 -0.093 0.808 0.486 0.233 Height of skull 0.961 -0.169 0.063 0.752 0.289 -0.197 Length of mandibular toothrow 0.845 0.503 -0.139 0.719 -0.608 0.301 ternal or cranial dimensions nor, as far as we can tell, in color of pelage. The single character for dis- tinguishing these two species is the longer length of interparietal bone in S.jacksoni than in S. caurinus', a length of 4.5 mm in the former versus less than 3.5 mm, in fact except for a few individuals, less than 3.0 mm in the latter. Geographic Variation Univariate analysis. — In examining geographic variation in S. caurinus, five external and 14 cranial measurements of age category III individuals were examined and the following results were found (Ta- ble 4). The specimens of the combined geographic sample from Yama and Sienso (OTU 11) of north- western Ivory Coast are clearly smaller than are those of all of the other geographic samples tested in nine of the 19 characters examined. These nine characters are total length, greatest length of skull, condylobasal length, zygomatic breadth, rostral breadth, oblique length of bulla, width across upper molars, length of diastema, and height of skull. The geographic sample from Tyenko (OTU 9), which is less than 100 mi south of OTU 1 1 , and that from Diali (OTU 10), which is from still farther south and east in Ivory Coast than OTU 9, are both clearly larger than are the remaining geographic samples of S. caurinus examined in total length, length of tail, greatest length of skull, condylobasal length, length of anterior palatal foramen, and height of skull. OTUs 9 and 10 are even more strik- ingly large when compared with OTU 11, the geo- graphic sample exhibiting the smallest dimensions for the species. Three specimens were available for study from Diali (OTU 10); two of these specimens were large for the species. Diali and also the type locality of S. jacksoni (Wenchi, Ghana) are situated on the interface of the High forest and Invasive woodland in West Africa. The holotype of S.jack- soni and the two large specimens from Diali are comparable in size, except for length of interparietal. There is no indication of a trend in variation in hindfoot length although OTU 3 (Pirisi, Ghana) and OTU 4 (Gudi, Nigeria) tend to have relatively long hindfeet when compared to the other geographic samples. The same discordant pattern is apparent in the variation in the length of the ear, although there is no clear grouping of geographic areas into different size groups, OTU 1 1 displayed relatively long ears. In both instances, this discordance could be an artifact of different techniques used by prep- arators of skins. The cranial characters of interor- bital breadth, length of maxillary toothrow, length of posterior palatal foramen, and length of mandib- ular toothrow show little geographic variation. Considering the vast distances involved (about 1200 km), the geographic samples from OTU 3 in Ghana, OTU 6 in Upper Volta, and OTU 4 in Ni- geria show little geographic variation in the char- acters tested. Multivariate analysis. — Twelve cranial measure- ments of eight geographic samples of age category III specimens of combined sexes were analyzed us- ing the NT-SYS routines. Again the length of pos- terior palatal foramen and greatest length of skull were excluded from the analysis. A distance phe- 1978 SWANEPOEL AND SCHLITTER—57EA 70ME5 SYSTEMATICS 71 3 0 4887 6 0.6030 4 1.0093 9 0.8769 10 1.6658 5 1.1964 7 1.3670 11 I 1 1 1 1- - H 1 i 1.760 1.560 1.360 1.160 0.960 0.760 0.560 0.360 Fig. 5. — Distance phenogram of OTUs (see Fig. 2 and text for key to localities included in each OTU) of S tea toniys caiihiiits computed from distance matrix based on standardized characters and clustered by unweighted pair-group method using arithmetic averages (UPGMA). OTUs consist of geographic samples of combined sexes. The cophenetic correlation for the phenogram is 0.87. nogram showing the relationship among eight geo- graphic samples of Steatomys caurinus is illustrated in Eig. 5. The cophenetic correlation coefficient for the phenogram is 0.87. The OTUs separate into two major groups. The upper one (OTUs 3, 6, 4, 9, and 10) includes individuals of medium to large cranial dimensions. This group further clusters into two subgroups — OTUs 9 (Tyenko) and 10 (Diali) from the Ivory Coast are included in one subgroup, and OTUs 3 (Pirisi, Ghana), 4 (Gudi, Nigeria), and 6 (Celia, Upper Volta) in a second subgroup. Of these two subgroups, the former subgroup consists of the larger sized individuals of the two subgroups. Al- though OTUs 9 and 10 are paired together on the phenogram, they are well separated by phenetic distance. The lower major group consists of three OTUs paired together but well separated by phe- netic distance; OTUs 5 and 7 from Upper Volta, and OTU 11 from Yama and Sienso, Ivory Coast. The first three principal components computed from the matrix of correlation among 12 cranial characters for eight geographic samples (OTUs) of Steatomys caurinus are presented in Pig. 6. The amount of phenetic variation represented in each of the first three components is 60.1% for component 1, 16.3% for component II, and 10.2% for compo- nent 111, for a total expression of 86.6% of the total phenetic variation. A factor matrix of character loadings from correlation among the 12 cranial char- acters is given in Table 6. From Table 6, it can be seen that the first, and by far the most important, component is heavily influenced by general cranial size; especially those measurements expressing length of skull, breadth of zygoma and rostrum, and greatest length of bulla. Interorbital breadth. oblique length of bulla, and length of maxillary toothrow show relatively low values for component I. Component II is influenced positively by oblique length of bulla and negatively by length of mandib- ular toothrow, whereas component III has a high positive value for length of maxillary toothrow and a high negative value for interorbital breadth. Examination of the three-dimensional plot of the principal components (Fig. 6) reveals a pattern of variation in accordance with that shown in the dis- tance phenogram (Fig. 5). OTUs 5, 7, and 11 form one main group on the left, with each OTU distantly separated from each other. The other major group- ing, consisting of OTUs 3, 4, 6, 9, and 10 forms two subclusters with OTUs 3, 4, and 6 constituting the one and OTUs 9 and 10 the other. This projection plot indicates that OTU 1 1 con- sists of individuals with small overall cranial size, especially with short and narrow skulls (Table 6). Although the overall length of bulla, including mas- toidal portion, is long, the audital portion, as ex- pressed in component II, is shortened. Component II also indicates a relatively long mandibular tooth- row compared to OTUs 5 and 7. Interorbital breadth of individuals of OTU 1 1 is narrow as shown by the negative influence of this character on component III. Component III indicates a long maxillary toothrow for OTU 11. OTUs 5 and 7 from Upper Volta are small in overall size. Both are well separated from OTU 1 1 in components II and III by the characters men- tioned above. In spite of having generally short and narrow skulls, their scores for component III indi- cate a relatively broad interorbital region and short maxillary toothrow, the latter corresponding with 72 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Fig. 6. — Three-dimensional projection of eight samples of Steatomys caurinus on to the first three principal components based upon a matrix of correlation among 12 cranial measurements. OTUs consist of geographic samples of combined sexes. See Fig. 2 and text for key to localities included in each OTU. Component I accounts for 60.1% of the phenetic variation, component II for 16.3%, and component III for 10.2% for a combined expression of 86.6% of the total variation. their short mandibular toothrow as expressed on component II. The other geographic samples from Ivory Coast (OTUs 9 and 10), represented by a maximum of five and three specimens, respectively, contain large in- dividuals. These OTUs are essentially equal in skull dimensions as expressed in components I and III, but separate somewhat based on component II — an expression of length of auditory bulla and, inverse- ly, length of mandibular toothrow. The OTUs 3, 4, and 6 are all represented by large sample sizes (minimum of 18 specimens), and are morphologically identical even though 1,200 km separate OTU 3 from 4. Other than slight differ- ences in overall size of skull, and slightly narrower interorbital region and longer maxillary toothrow, OTUs 3, 4, and 6 differ little from OTUs 9 and 10. Taxonomic Conclusions Based on our assessment of geographic variation, we have separated 5. caurinus into two taxonomi- cally recognizable populations. The smallest indi- viduals in the species belong to the population from Yama and Sienso in northwestern Ivory Coast and are herein described as a new subspecies. A second subspecies, 5. caurinus caurinus Thomas, 1912, represented by mice of medium to large size, is known from Tyenko and Diali, Ivory Coast, Upper Volta, Ghana, Togo, Benin, and Nigeria. OTUs 2 and 6 from Upper Volta represent relatively small individuals, but, considering the small sample sizes and geographic distribution of the OTUs involved, is retained in S. c. caurinus. Eurther samples are required from within the areas between the Yama and Sienso (Ivory Coast) localities and Ghana 1978 SWANEPOEL AND SCH LITTER— 57EA70/W F5 SYSTEM ATICS 73 (OTU 2) and Upper Volta (OTUs 5 and 7) popula- tions in order to determine the distribution of the new subspecies and its relationships to the popu- lations at the latter localities. Also the taxonomic and distributional status of the individuals of large size from Tyenko (OTU 9) and Diali (OTU 10), Ivory Coast, is deserving of additional study. Steatomys caurinus caurinus Thomas, 1912 1912. Steatomys caurinus Thomas, Ann. Mag. Nat. Hist., ser. 9, 9:271, February. 1977. Steatomys pratensis caurinus, Coetzee, in Meester and Setzer (eds.). The mammals of Africa . . . , Smithsonian Inst. Press, 6-8: 1-4. Holotype. — Young adult female (age category III), skin and skull BMNH 12.1.16.24, from Pan- yam, 4,000 ft, Nigeria; obtained 13 July 1911 by Rev. G. T. Fox, original number 47. Skin in good condition; skull with left occipital region broken, both bullae separated from skull but repaired with glue, left zygoma broken. Measurements of holotype. — Total length, 164; length of tail, 54; length of hindfoot, 19; length of ear, 19; greatest length of skull, 27.3; interorbital breadth, 3.9; rostral breadth, 4.8; length of maxil- lary toothrow, 4.0; breadth across upper molars, 6.1; length of anterior palatal foramen, 5.0; length of diastema, 7.2; length of mandibular toothrow, 3.4. Distribution. — Known from Ivory Coast (Tyenko and Diali), Upper Volta, western and northern Ghana, northern Togo and Benin, and central Ni- geria. Literature records indicate its occurrence in Senegal as well. Comparisons. — For comparisons of Steatomys caurinus caurinus with other taxa from West Afri- ca, see each of the following accounts. Remarks. — The subspecies S. c. caurinus con- sists of individuals of medium to large size for the species and seems to occur over an extensive geo- graphic area from at least Ivory Coast east to Ni- geria, although originally known only from Nigeria (Thomas, 1912^, I9l2h). Within this area, speci- mens are generally uniform in size and in shape of skull. In the Ivory Coast, specimens of two sizes have been reported by Heim de Balsac ( 1967). For additional comments on this material, see the fol- lowing account. Steatomys caurinus was reported from Bandia, Senegal, by Hubert et al. ( 1973). Heim de Balsac (1967:219) reported a large and a small Steatomys from M’Bour, Senegal. We are accepting both lo- cality records as S. caurinus at this time but are unable to assign with certainty this material to sub- species. Because the relationship of the Senegalese specimens could well be to the populations of the Sudan Woodland of Upper Volta, we are listing the literature records under the nominate subspecies. Based on these published reports, it appears that the only locality of sympatry between 5. caurinus and S. cuppedius in West Africa would be M’Bour, Senegal. Specimens examined (196). — Benin: Kouande, 1 (USNM). Ghana: Bawku, 1 ,400 ft. I (BMNH); Bangwon, 5 (USNM); Pirisi, 43 (USNM); Sakpa, 1 (USNM). Ivory Coast: Bouna, I (USNM); Diali, 14 (USNM); Tyenko, 12 (USNM). Nigeria: 3 mi E Giidi, 52 (USNM); Panyam, 4,000 ft, II (BMNH). Toco: Dapango, I (USNM); I km N Celia, 48 (USNM); Dio, I (USNM); Fo, 1 (USNM); 36 mi SE Nayoure, 1 (USNM). Additional records. — Ivory Coast: Lamto ( Bellier, 1967; Heim de Balsac and Bellier, 1967:159; Bellier and Gatun, 1968;708, Heim de Balsac, 1967:215-219); Bouake (Heim de Bal- sac, 1967; Heim de Balsac and Bellier, 1967:159; Bellier and Gatun, 1968:708). Senegal: Bandia (Hubert et al., 1973:81), M’Bour (Heim de Balsac, 1967:219). Steatomys caurinus roseveari, new subspecies Holotype. — Adult male, skin and skull, USNM 467496, from Yama, Ivory Coast; obtained on 22 March 1969 by L. W. Robbins, original number 1 106. Skin in good condition, missing left front foot; skull in good condition, nasals damaged. Measurements of holotype. — Total length, 156; length of tail, 47; length of hindfoot, 18; length of ear, 17; weight, 42 g; condylobasal length, 25.3; zy- gomatic breadth, 13.0; interorbital breadth, 4. 1 ; ros- tral breadth, 5.0; oblique length of bulla, 5.5; great- est length of bulla, 9.1; length of maxillary toothrow, 4.2; breadth across upper molars, 5.8; length of anterior palatal foramen, 5.6; length of posterior palatal foramen, 0.4; height of skull, 9.9; and length of mandibular toothrow, 3.2. Additional measurements are listed in Tables 3, 4, and 5. Distribution . — At present known only from Sien- so and Yama in northwestern Ivory Coast. Comparisons. — Steatomys caurinus roseveari can be distinguished from S. c. caurinus by its smaller size, especially cranial dimensions. Most characteristic of this new subspecies are short and narrow skulls with relatively narrow interorbital re- gion and long maxillary and mandibular toothrows. From S . jacksoni and S. cuppedius, S. c. roseveari can be separated as for the species. Remarks. — Heim de Balsac (1967) considered the 74 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 taxonomic status of Steatomys in the Ivory Coast but failed to put specific epithets on his specimens. Heim de Balsac (1967:216-220) reported large and small specimens from Bouake and Lamto. Without having examined these specimens, we are unable to make a definitive judgment. It is possible these differences are attributable to nongeographic vari- ation such as age (certainly at least specimens A and D in figure 14 of Heim de Balsac, 1967, are young animals) or there might be two distinct taxa represented. S. caurinus roseveari is smaller than the nominate subspecies. The small form reported by Heim de Balsac (1967) might represent this new subspecies. If this were true, and it was sympatric with S. caurinus caurinus, then each must be rec- ognized as distinct species. In the geographic sam- ple from Tyenko (OTU 9) or Diali (OTU 10) we could not find any evidence of intergradation or of two distinct morphotypes being present. For now we consider these small individuals from Tyenko and Diali to be a subspecies of S. caurinus. Etymology. — This new subspecies is named for D. R. Rosev- ear who has contributed much, both in the field and in print, to the study of the mammals of West Africa. His monumental writ- ten contributions will serve for many years as a primary refer- ence for anyone interested in bats, rodents, or carnivores of this fascinating region of Africa. For his abundant contributions after his retirement and return to England from Nigeria, all who study small mammals in West Africa owe him a large amount of grat- itude. Specimen.'! examined (75). — Ivory Coast: Sienso, 6{USNM); Yama. 69 (USNM). Steatomys cuppedius Thomas and Hinton, 1920 Geographic distribution of species. — Known from northern Nigeria, southern Niger, and Sene- gal. Diagnosis. — Small, both externally and cranially, for the genus in West Africa; condylobasal length of skull less than 20.7 mm; length of maxillary toothrow usually less than 4.0 mm; pelage pale in color, with individual hairs soft. Comparisons. — From all other species of Stea- toniys occurring in West Africa, S. cuppedius can be separated by its small size, both externally and cranially. Skulls of S. cuppedius are less than 20.7 mm in condylobasal length in age category III in- dividuals. Geographic Variation Univariate analysis. — Five external and 14 cra- nial measurements of age category HI individuals of Steatomys cuppedius from Senegal and Nigeria were examined for geographic variation. The geo- graphic sample from Senegal is larger in size than is the one from Nigeria (Table 4). Of age category III individuals, four measurements (total length, weight, greatest length of skull, and oblique length of bulla) of the populations from Senegal were sig- nificantly larger than those from Nigeria. Nine other measurements averaged larger and four averaged smaller, whereas two were equal in size (Table 4). Only small samples of age category IV were avail- able for analysis of geographic variation (Table 5). A trend similar to that found in age category III is not indicated in such small samples. Taxonomic Conclusions Only three geographic samples of S. cuppedius were examined — a single individual from Niger and two geographic samples of sizeable proportions from Senegal and Nigeria. Considering the distance of approximately 2,500 km between the geographic samples from Senegal and Nigeria, the geographic variation present is not striking. Taxonomic separa- tion of these localities is not warranted; we con- sider Steatomys cuppedius to be a monotypic spe- cies. Steatomys cuppedius Thomas and Hinton, 1920 1920. Steatomys cuppedius Thomas and Hinton, Novit. Zool., 27:318, 15 June. 1977. Steatomys parvus cuppedius, Coetzee, In Meester and Setzer (eds.). The mammals of Africa .... Smithsonian Inst. Press, 6.8: 1-4. Holotype. — Adult female (age category IV), skin and skull, BMNH 21 .2. 1 1.85, from Farniso ( = Pan- isau), near Kano, 1,700 ft, Nigeria; obtained 29 De- cember 1919 by A. Buchanan, original number 70. Skin in good condition; skull in fair condition, left zygoma broken. Measurements of holotype. — Total length, 124; length of tail, 43; length of hindfoot, 14; length of ear, 14; greatest length of skull, 21.3; condylobasal length, 20.4; interorbital breadth, 3.6; rostral breadth, 4.2; greatest length of bulla, 7.7; breadth across upper molars, 5.2; length of anterior palatal foramen, 8.8; length of diastema, 5.2; height of skull, 8.0; length of mandibular toothrow, 3.0. Distribution. — This species is known from cen- tral Nigeria, Niger, and Senegal. Comparisons. — See comparisons for species. Remarks. — Originally described from material from northern Nigeria (Thomas and Hinton, 1920, 1921), Steatomys cuppedius has been uncommon in 1978 SWANEPOEL AND SCHLITTER— 57EA70MK5 SYSTEMATICS 75 collections of rodents from West Africa. Thomas (1925:194) reported a single specimen from Gan- gara, Niger, a locality across the Nigerian border to the northeast from the type locality. Although large numbers of Steatomys from Upper Volta were available, none proved to be S. cnppediiis. If the latter occurs in Upper Volta, it will probably be found in the extreme north and have a distribution similar to Gerhilliis nigeriae, as it seems to have in Nigeria. There seems to be some doubt as to the identity of the skulls reported from owl pellets by Heim de Balsac (1965). Later, Heim de Balsac and Bellier (1967) indicate that the material from '‘nord de Saint-Louis du Senegal” and ”au voisinage de M’Bour” was smaller than the material obtained in Ivory Coast. If this is true, it is likely this Sene- galese material belongs to 5. cuppedius although we are unable to identify it with certainty. We have included these records under the additional records of this species. Specimens examined (113). — Niger: Gangara, 1,400 ft, 1 (BMNH). Nigeria: Farniso, near Kano, 1,700 ft, 6 (BMNH); Panisau, 55 (USNM). Senegal: 17 km NE Kiourhel, 1 (USNM); 15 km N Karrfine, 16 (USNM); 6 km E Kaolack, 7 (USNM); Koungheul, 27 (USNM). Additional records. — Senegal; north of St. Louis, (Heim de Balsac, 1967;219; Heim de Balsac and Bellier, 1967; 159); M’Bour (Heim de Balsac and Bellier, 1967:159). Steatomys jacksoni Hayman, 1936 Geographic distribution of species. — Known only from the type locality, Wenchi, Ashanti [dis- trict], Ghana. Diagnosis. — Size large, both externally and cra- nially, for genus in West Africa; skull long and rel- atively narrow; interparietal bone long, 4.5 mm in length. Comparisons. — For comparisons with S. cuppe- dius and S. caurinus, see accounts of that species and for additional comments, see section “Rela- tionships of Species.” Steatomys jacksoni Hayman, 1936 1936. Steatomys jacksoni Hayman, Proc. Zool. Soc. London, for 1935, pp. 930, 10 January. 1977. Steatomys pratensis Jacksoni, Coetzee, in Meester and Setzer (eds.). The mammals of Africa .... Smithsonian Inst. Press, 6.8: 1-4. Holotype. — Young adult male (young age cate- gory IV), skin and skull BMNH 35.1.30.157, from Wenchi, Ashanti, Ghana; obtained 18 January 1934 by W. P. Lowe, original number 104. Skin and skull in good condition. Measurements of holotype. — Total length, 170; length of tail, 50; length of hindfoot, 18; length of ear, 18; greatest length of skull, 28.5; condylobasal length, 27.1; zygomatic breadth, 12.9; interorbital breadth, 4.0; rostral breadth, 5.3; oblique length of bulla, 5.8; greatest length of bulla, 9.6; length of maxillary toothrow, 4.0; breadth across upper mo- lars, 6.1; length of anterior palatal foramen, 5.7; length of diastema, 7.2; height of skull, 10.0; length of mandibular toothrow, 3.7. Distribution. — Same as given above. Comparisons. — See comparisons given above. Remarks. — We prefer to retain this taxon as a valid species based solely on the size and shape of the interparietal bone. Additional material from the type locality is required before a critical evaluation of variation in the interparietal bone can be made. Specimens examined (I). — Ghana: Wenchi, Ashanti, 1 (BMNH), ACKNOWLEDGMENTS We are indebted to the following institutions and curators who made available West African Steatomys specimens for study. Abbreviations preceding the names of institutions are used in the accounts above to identify the source of specimens. BMNH — British Museum (Natural History), London (I. Bish- op). USNM — United States National Museum of Natural History, Smithsonian Institution, Washington, D. C. (Henry W. Setzer). We are grateful to Hugh H. Genoways for helping us in many ways, besides critically reading the manuscript; Flora Gibson for typing early drafts of it, and Teresa Bona for typing and critically editing later drafts; Terry L. Yates for performing the NT-SYS computations on the Texas Tech University computers; Judy Schlitter for keypunching; Margaret Popovich for proofreading tables; Nancy Perkins for preparing the figures; and John Sutton for help with computer-facilitated statistical analyses. This research was performed while the senior author visited the Section of Mammals as a Resident Museum Specialist in the International Visitor Program of Carnegie Museum of Natural History. 76 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 GAZETTEER Names of geographic features listed below are those used in the text. The primary sources for spellings and coordinates of localities were the individual specimen tags and the gazetteers of the United States Board on Geographic Names (prepared by the Office of Geography, Department of Interior). Benin: Ghana: Ivory Coast: Kouande 10 20 N, 01 41 E Bangwon 10 58 N, 02 41 W Bawku 1 1 05 N, 00 11 W Pirisi 10 07 N, 02 27 W Sakpa 08 52 N, 02 21 W Wenchi 07 45 N, 02 02 W Bouake 07 41 N, 05 02 W Bouna 09 16 N, 03 00 W Diali 07 03 N, 05 37 W Katiola 08 08 N, 05 06 W Lamto 06 12 N, 04 58 W Ouango Fitini 09 34 N, 04 03 W Sienso 09 25 N, 07 31 W Tyenko 08 14 N, 07 24 W Yama 09 36 N, 06 19 W Niger: Gangara ca. 14 00 N, 09 00 E Nigeria: Farniso [ = Panisau] Gudi 08 54 N, 08 17 E Panisau (Panisa) 11 43 N, 07 32 E Senegal: Bandia 14 37 N, 17 02 W Diourbel 14 40 N, 16 15 W Kaffrine 14 06 N, 15 33 W Kaolack 14 06 N, 16 33 W KoungheuI 13 59 N, 14 48 W M’Bour 14 24 N, 16 58 W St. Louis 16 02 N, 16 30 W Togo: Dapango 10 52 N, 00 13 E Wulehe 08 40 N, 00 00 Upper Volta: Barga 13 51 N, 02 12 W Celia 11 38 N, 00 22 W Dio 13 20 N, 02 38 W Fo 11 53 N, 04 31 W Nayoure 12 15 N, 00 16 E LITERATURE CITED Allen, G.M. 1939. A checklist of African mammals. Bull. Mus. Comp. Zool., 83:1-763. Atchley. W. R. 1970. A biosystematic study of the subgenus Selfici oiCulicoides (Diptera: Ceratopogonidae). Univ. Kan- sas Sci. Bull., 49:181-336. Bellier, L. 1967. Recherches ecologiques dans la Savane de Lamto (Cote d'Ivoire): densities et biomasses de petits Mam- miferes. Terre et Vie, 1967:319-329. Bellier, L., and J.-C. Gautun. 1968. Note sur I’activite des Steatomys du groupe opimus en Cote-d’Ivoire. Mammalia, 32:208-209. CoETZEE, C. G. 1977. Part 6.8, Genus Steatomys. Pp. 1-4, in The mammals of Africa: an identification manual (J. Meester and H. W. Setzer, eds.), Smithsonian Inst. Press, Washing- ton, D. C. Ellerman, J. R. 1941. The families and genera of living rodents. Trustees British Mus. (Nat. Hist.), London, 2:xii + 1-690. Genoways, H. H. 1973. Systematics and evolutionary relation- ships of spiny pocket mice, genus L/o»ivi'. Spec. Publ. Mus., Texas Tech Univ., 5:1-368. Hayman.R.W. 1936. On a collection ofmammals from the Gold Coast (with a note on the leopards by R. I. Pocock). Proc. Zool. Soc. London, 1935:915-937. Heim de Balsac, H. 1965. Quelques enseignments d'order faunistique tires de I'etude du regime alimentaire de Tyto alba dans I'Ouest de I'Afrique. Alauda, 33:309-322. . 1967. Rongeurs de Lamto (Cote d’Ivoire) Faits nou- veaux d'ordre anatomique, taxonomique, et biogeogra- phique. Biolog. Gabon., 3:175-222. Heim DE Balsac, H., and L. Bellier. 1967. Liste preliminaire des rongeurs de Lamto (Cote-d’Ivoire). Mammalia, 31:156- 159. Hdberi , B., F. Adam, AND A. PouLET. 1973. Liste preliminaire des rongeurs du Senegal. Mammalia, 37:76-87. Power, D. M. 1970. Geographic variation of red-winged black- birds in central North America. Univ. Kansas Publ., Mus. Nat. Hist., 19: 1-83. Rosevear, D. R. 1965. The bats of West Africa. Trustees British Mus. (Nat. Hist.), London, xvii -I- 418 pp. . 1969. The rodents of West Africa. Trustees British Mus. (Nat. Hist.), London, xii + 604 pp. ScHNELL, G. D. 1970. A phenetic study of the suborder Lari ( Aves). I. Methods and results of principal component anal- yses. Syst. Zool., 19:35-57. Setzer. H. W. 1956. Mammals of the Anglo-Egyptian Sudan. Proc. U. S. Nat. Mus., 106:447-587. Sneath, P. H. a., AND R. R. SoKAL. 1973. Numerical taxon- omy. W. H. Freeman and Co., San Francisco, xv -I- 573 pp. SoKAL, R. R., AND P. H. A. Sneath. 1963. Principles of nu- merical taxonomy. W. H. Freeman and Co., San Francisco, xvi + 359 pp. Straney, D. O. 1978. Variance partitioning and nongeographic variation. J. Mamm., 59:1-11. Thomas, O. 1912^;. Mammals of the Panyam Plateau, northern Nigeria. — II. Ann. Mag. Nat. Hist., ser. 8, 9:269-274. . 19126. List of a third collection of mammals from Pan- yam, N. Nigeria, presented by the Rev. G. T. Fox. Ann. Mag. Nat. Hist., ser. 8, 9:683-686. . 1925. On the mammals (other than ruminants) collected by Captain Angus Buchanan during his Second Saharan Ex- pedition, and presented by him to the National Museum. Ann. Mag. Nat. Hist., ser. 9, 16:187-197. Thomas, O., and M. A. C. Hinton. 1920. Captain Angus Bu- chanan's Air expedition. I. On a series of small mammals from Kano. Novit. Zool., 27:315-320. . 1921. Captain Angus Buchanan’s Air expedition. II. On the mammals (other than ruminants) obtained during the ex- pedition to Air (Asben). Novit. Zool., 38: 1-13. A CRITICAL EXAMINATION OF ALLEGED SIBLING SPECIES IN THE LESSER THREE-TOED JERBOAS {SVBGENVS JACULUS) OF THE NORTH AFRICAN AND ARABIAN DESERTS DAVID L. HARRISON Harrison Zoological Museum, Bowerwood House, St. Botolph’s Road, Sevenoaks, Kent, England ABSTRACT Alleged sibling speciation in the lesser three-toed jerboas is amined. Jaculiis deserti Loche, 1867, is shown to be a junior considered. Variation in supposed diagnostic characters is ex- synonym of J. juculiis Linnaeus, 1758. INTRODUCTION Ellerman and Morrison-Scott (1951) in their com- prehensive review of Palaearctic mammals recog- nized only three species of this subgenus, namely Jaculiis jaciiliis Linnaeus, 1758, Jacnlus hlanfordi Murray, 1884, and Jaculiis orientalis Erxieben, 1777. The validity of these three taxa is not in doubt, but recently the suggestion has been made (Ranck, 1968) that two sibling species exist within the populations of the lesser three-toed jerboa, Ja- ciiliis jaculiis, occurring in northern Africa and the Arabian Peninsula. Eor one of these sibling species Ranck (1968) has employed an old taxon, Jaculiis deserti, based on Dipus deserti Loche, 1867. Ranck ( 1968) keyed these two species as follows: Dorsal color dark; two foramina on angular process of mandible; sole of hind foot and matatarsal area suffused with brownish hairs J. deserti Dorsal color pale; a single foramen on an- gular process of mandible; sole of hind foot and metatarsal area white or buff and lacking suffusion of brownish hairs J. Jaculiis Ranck (1968) gave the range of his J. deserti as Arabia, Iraq, Israel, Sinai, Egypt, Libya, and Al- geria and referred the following taxa as subspecies of it: J. deserti deserti Loche, 1867; J. deserti vas- tus Ranck, 1968; J. deserti rams Ranck, 1968; J. deserti fuscipes Ranck, 1968; J. deserti favilliis Setzer, 1955;/. deserti schluteri Nehring, 1901;/. deserti vocator Thomas, 1921; and/, deserti loftiisi Blanford, 1875. Lor Jaculiis Jaculiis Ranck (1968) gave the gen- eral range as Iraq, Syria, Lebanon, Israel, Jordan, Saudi Arabia, and North Africa south through the Sahara including Sudan, Chad, Niger, Mauritania, and Spanish Sahara. To this species Ranck referred the following taxa: Jaculiis Jaculiis Jaculiis Lin- naeus, \75^', J. J^K^'nlns arenaceous Ranck, 1968;/. Jaculiis collinsi Ranck, 1968; /. Jaculiis ciifreusis Ranck, \96^', J. jaculiis tripolitauiciis Ranck, 1968; J. Jaculiis whitcliiirchi Ranck, \968, J. Jaculiis sef- riiis Thomas and Hinton, \92\ \ J . Jaculiis centralis Thomas and Hinton, 1921 ; /. yV/c7////.s- hiitleri Thom- as, 1922. Other subspecies not specifically allocated by Ranck include /. Jacnlus airensis Thomas and Hinton, \92l; J. Jacnlus favoniciis Thomas, 1913; J . Jacnlus gordoni Thomas, 1903; and J . Jaculiis viil- tiirniis Thomas 1913. Harrison (1972), reviewing the material oi Jacii- lus from the Arabian Peninsula, has already thrown doubt on the validity of these “sibling species,” stating ”1 am quite unable to distinguish two small species oJ Jacnlus in the extensive Arabian material examined . . . individual variation in both color- ation of the soles and size and number of the an- gular foramina is so extensive and random in the Arabian jerboas that the definition of two species on this basis appears quite impossible.” As a preliminary part of an extended investiga- tion of geographical variation in Jacnlus, it ap- peared necessary to make a critical examination of Ranck’s sibling species criteria in North African as well as in the Arabian populations, in order to eval- uate the problem more fully and prevent further systematic confusion. That is the purpose of the present study, which is not intended in any other way as a revision of the numerous geographical forms listed above. Such a revision must clearly await far more detailed and extensive research. 77 78 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 R L HZM 57.6721 IN AMENAS ALGERIA HZM 29.3541 W. OF BADANAH SAUDI ARABIA Fig. I. — Examples of the angular processes of two specimens of Jaciiliis Jaculus from Algeria and Saudi Arabia showing variation in angular foramina. Angular formina scores are 1 - 2 above and 2 - I below. Abbreviations are R (right) and L (left); numbers are registry numbers for specimens in the Harrison Zoological Museum (HZM). METHODS Coloration Of the characters employed by Ranck to separate J. deserti and J. jaculits. both dorsal coloration and sole coloration are necessarily somewhat subjective, but have been found suscep- tible to the method of scoring listed below, in each specimen examined: Dorsal coloration 1. Pallid sandy Paler 2. Grayish sandy i 3. Brownish sandy Darker Sole coloration 1, White Paler 2. Buff i 3. Brown Darker A major source of confusion has been found to occur in the case of sole coloration in Jacidiis, namely discoloration of the sole and digital brushes by the substrate. A most striking ex- ample of this was seen in a series of skins from In Amenas, Algeria, in which the soles and digital brushes are bright or- ange— until washed with water, when the orange discoloration disappears, leaving a white condition. With experience it is usu- ally possible to distinguish discoloration from true pigmentation and washing is seldom required. Angular Foramina Fortunately, the angular foramina are easily evaluated and can be counted and recorded with stereomicroscopic drawings. Un- fortunately, the angular process is delicate and hence often bro- ken in preparation. It is pertinent to observe here that the angular foramina represent unossified areas in bone, which are occupied by membrane in life, a fact which no doubt accounts for their great variability, both in size and number. These criteria have been recorded for each specimen exam- ined and the results are summarized below in tabular form with all Arabian material and all North African material arranged in two separate groups. In these tables the scores for each speci- men are arranged with dorsal coloration, sole coloration, and foramen counts for the right and left mandible in series from left to right, thus a score of 2;2:2-2 would indicate an animal with intermediate coloration of the dorsum and soles and two angular foramina on each side. The results are simply expressed as the number of individuals from each group showing each numerical score (where a mandibular foramen is broken or absent it is scored as x). Only those specimens able to provide a combina- tion of dorsal and sole coloration with one or both angular fo- ramen counts are included in this study; skins without skulls and skulls without skins are excluded. Each score is listed as being compatible (COM), incompatible (NC), or intermediate (INT). 1978 HARRISON— SYSTEMATICS OF JACULUS 79 RESULTS AND DISCUSSION It is clear that if Ranch’s hypothesis of two sibling species is correct, based on the characters given in his key, then both in African and Arabian popula- tions two groups of scores ought to predominate in these results: Table 1. — Scores obtained in this study for dorsal coloration, sole coloration, and angidar foramina counts o/Jaculus. Num- ber of individuals with each score is indicated for each continent. COM = compatible, NC = not compatible, and INT = inter- mediate scores for Ranch's decision. Compatibility Score Arabian Peninsula Africa COM l;l;l - 1 4 2 COM I:I:X - 1 2 2 COM 1:1:1 - X 0 2 NC 1:1:1 - 2 2 0 NC 1:1:2 - 2 1 0 NC 1:1:2 - X 1 1 NC 1:1:X - 2 1 2 COM 1:2:1 - 1 1 19 COM 1:2:X - 1 0 4 COM 1:2:1 - X 0 1 NC 1:2:1 - 2 0 4 NC 1:2:2 - 1 0 3 INT 1:2:2 - 2 4 16 INT 1:2:2 - X 2 6 INT 1:2:X - 2 3 5 INT 1:2:2 - 3 0 3 INT 1:2:3 - 3 0 1 INT 1:2:2 - X 0 2 INT 1:3:2 - 2 0 2 INT 1:3:2 - X 1 2 NC 1:3:1 - X 1 0 INT 2:2:1 - 1 5 4 INT 2:2:X - 0 1 0 INT 2:2:X - 1 3 3 INT 2:2:1 - X 3 0 NC 2:2: 1 - 2 1 0 NC 2:2:1 - 3 0 2 INT 2:2:2 - 2 8 5 INT 2:2:2 - X 5 0 INT 2:2:X - 2 5 4 NC 2:2:2 - 1 2 2 INT 2:2:3 - 2 1 1 INT 2: 1:2 - 2 2 0 INT 2:1:2 - X 1 0 COM 2:3:2 - 2 2 0 COM 2:3:2 - X 2 0 COM 2:3:X - 2 2 0 NC 2:3:X - 1 1 0 COM 3:3:X - 2 0 1 NC 3:2:2 - 1 0 1 3 : 3 : 2 - 2 3 : 2 : 2 - 2 2 : 3 : 2 - 2 indicating the deserti species with dark or darkish dorsum and soles and two angular foramina. On the other hand, the following scores: 1 : 1 : 1 - 1 2 : 1 : 1 - 1 1 : 2 : 1 - 1 would indicate the jacidiis species with pallid or palish dorsum and soles but one angular foramen. On the other hand, scores including angular fo- ramen counts of 2 - 1 or 1 - 2 ought not to occur at all, indicating both species foramen count in one individual (Fig. 1), and scores of 1 : 1 : 2 - 2 3 : 3 : I - 1 similarly should not occur at all, whereas the oc- currence of many intermediate scores must cast fur- ther doubt on the validity of the hypothesis. In both African and Arabian jerboas, scores for the angular foramina of I - 2 and 2 - 1 occur with such frequency (Table 1, Fig. 1) that on this con- sideration alone the sibling species hypothesis as advanced by Ranck simply cannot be upheld. Fur- thermore, the predicted scores for the two species do not predominate; other incompatible scores also occur at an unacceptable high incidence and inter- mediate scores form the highest percentage in both populations. The number of individuals assessed and percentage occurrences of compatible, incom- patible, and intermediate scores are given in Table 1 Table 2. — Summary of compatibility results for Jaculus from Table I. Compatibility Arabia Africa Number Per- centage Number Per- centage Compatible 13 19 31 31 Incompatible 10 15 15 15 Intermediate 44 66 54 54 Total 67 100 80 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 CONCLUSIONS It is suggested that the number of angular foram- ina 'mJaculus is a matter of random individual vari- ation, often differing in the two mandibles of an individual. It is further suggested that pallid or dark- er dorsal coloration and sole brushes are in reality related to substrate coloration in Jaculus and rep- resent the effect of selection in producing “sub- strate races” more or less protected by cryptic re- semblance to the predominant substrate, as has been well demonstrated in the three-toed jerboa and many other desert rodents (Harrison, 1975). The concept of two sibling species within Jaculus ja- culus populations on the basis proposed by Ranch (1968) is rejected. ACKNOWLEDGMENTS The author is much indebted to J. Edwards Hill and staff of the Mammal Section, British Museum (Natural History), for kindly providing facilities to examine material of Jaculus in the National collection and to all those who have donated material of this species to the Harrison Zoological Museum, including T. Bickley, P. A. Clancey, M. D. Gallagher, D. Grant, J. G. Har- rison, H. Hovel, S. Howe, M. K. Lambert, M. Legg, R. E. Lewis, T. D. Rogers, G. B. Stafford, S. Seligman, and Sheikh Zaid bin Sultan. LITERATURE CITED Ellerman, J. R., AND T. C. S. MoRRisoN-Scon . 1951. Check- list of Palaearctic and Indian Mammals 1758 to 1946. Brit. Mus. (Nat. Hist.), London, 810 pp. Harrison, D. L. 1972. The mammals of Arabia: Lagomorpha, Rodentia. E. Benn Ltd., London, 3:xvii -f 384-670. . 1975. Desert coloration in rodents, Pp. 269-276, in Ro- dents in desert environments (1. Prakash and P. K. Ghosh, eds.), W. Junk, The Hague, xvi -I- 624 pp. Ranck, G. L. 1968. The rodents of Libya, taxonomy, ecology and zoogeographical relationship. Bull. U. S. Nat. Mus., 275:vii -r 1-264. ENERGETICS OE SURVIVAL IN HETEROCEPHALUS GLABER (RUPPELL), THE NAKED MOLE-RAT (RODENTIA: BATHYERGIDAE) J. U. M. JARVIS Zoology Department, University of Cape Town, Rondebosch, Cape Province 7700, Republic of South Africa ABSTRACT Heterocephalus glaber, a colonial mole-rat, has a highly struc- tured social system with a worker and non-worker class and a dominant breeding female. Recruitment rates for the colony and growth rates of the young are very low. Unusual features in its physiology (a high rate of thermal conductance, a low basal me- tabolic rate, a low body temperature, and poor ability to ther- moregulate) have in the past been regarded solely as adaptations to the high temperature and humidity of the burrows in which mole-rats live. It is here suggested that food is limiting to Hei- erocephalus and that many of its unusual features can be attrib- uted to the need to maintain a low energy budget for the colony. INTRODUCTION Heterocephalus glaber, the naked mole-rat is a small colonial rodent with an average weight of about 35 g. It occurs in arid regions of Kenya, Ethiopia, and Somalia — areas characterized by a low mean annual temperature amplitude, diurnal ambient temperatures of above 27°C, and an annual rainfall of less than 700 mm. Vegetation in these areas is sparse and many plants have swollen sub- terranean portions; these form the main food for Heterocephalus. Heterocephalus lives underground in a burrow system, which consists of extensive foraging bur- rows running at root or tuber level and a deeper nest area. The ground is very hard and the burrows semipermanent (Jarvis and Sale, 1971). The tubers and bulbs, on which they feed, are found by the energetically expensive method of extending the foraging burrows in an apparently random direction until food is encountered. From the burrow pattern, it appears that once a tuber has been located, the neighborhood is then searched thoroughly — seem- ingly in response to the fact that many tubers re- produce vegetatively and tend to occur in patches. Once located, large items of food such as tubers are left growing and are gradually hollowed out by the mole-rats, small food items are carried to the nest area and eaten there. The microclimate in all but the very superficial portions of the burrow system is extremely uni- form, with humidities (usually) above 90% and tem- peratures between 30 and 32°C. The mole-rats avoid extreme temperatures in the superficial burrows by restricting their burrowing activities to early morn- ing and late afternoon. Heterocephalus has a skin that is hairless, except for scattered sinusoidal hairs all over the body; it is well vascularized but lacks sweat glands and the normal mammalian layer of subcutaneous fat (Thig- pen, 1940). Consequently, these mole-rats have very high rates of thermal conductance (McNab, 1966). In laboratory studies on single naked mole- rats, McNab (1966, 1968) found them to have a me- tabolic rate that was less than 40% of that expected, a body temperature of about 32°C, and the poorest capacity for thermoregulation of any known mam- mal. My own findings confirm these observations (Withers and Jarvis, in preparation). McNab sug- gests that these modifications reduce the probability of overheating in an environment where evapora- tive and convective cooling are greatly reduced. All previous physiological studies on Hetero- cephalus (McNab, 1965, 1966, 1968) have been con- fined to isolated animals. However, my observa- tions on captive colonies of naked mole-rats have shown them to be highly social rodents. Members of the colony have specific roles in that there are worker and non-worker classes and a dominant breeding female. Furthermore, mutual contact and huddling are important to the mole-rats. I, there- fore, suggest that any explanation of their physiol- ogy, which does not consider their sociality as well as more aspects of their habitat than just tempera- ture, will at best only give partial answers to the questions raised. This paper attempts to summarize and synthesize our present knowledge of the ecol- ogy, physiology, and ethology of these mole-rats and to suggest ways in which these have been in- fluenced by the environmental pressures to which these mole-rats are subjected. More detailed ac- counts of their social structure and of the physiol- ogy of grouped and not single mole-rats will be pub- lished later. 81 82 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 MATERIALS AND METHODS Naked mole-rats were caught by opening a foraging burrow and pushing a spade down behind any animal coming to inves- tigate the damaged burrow. Because of this method, only the worker class of mole-rats were caught. Monthly samples of ap- proximately 30 mole-rats were caught in this way at Mtito Andei, Kenya (240 km southeast of Nairobi), over a complete calendar year. From these animals data on body measurements, repro- ductive condition, and stomach contents were obtained. Fifty live mole-rats were collected in the same way at Mtito Andei in early 1974 and established in three "colonies” in the laboratory in Cape Town. After approximately one year each colony had established its social hierarchy and individuals from one colony were marked with subcutaneous injections of india ink, to enable a more detailed study of the social structure of a colony. The two other colonies were used in physiological stud- ies and left undisturbed for breeding. The preferred ambient temperature range of Heterocephalus was determined by establishing a temperature gradient (2I°C to 40°C) in a terrarium housing one of the colonies and observing the temperatures selected by single and huddling members of the colony. Observations were made at 10 min intervals and the number of mole-rats sleeping singly or huddled was noted. Emphasis in the physiological studies was placed on the re- sponses of the mole-rats to a temperature range that approxi- mated the normal burrow temperature. The temperature selec- tion studies, and also field measurements, suggested that the normal burrow temperature lies close to 32°C. McNab (1966) demonstrated that, below 20°C and above about 36°C, Hetero- cephalus almost completely abandons any attempt to thermo- regulate — again suggesting that these temperatures lie outside those normally experienced by the mole-rats. For these reasons, my investigations into the effect of temperature on the metabolic rate of Heterocephalus were limited to ambient temperatures lying between 20°C and 34°C. For the metabolic studies, mole-rats were placed in clear per- spex containers in a constant temperature chamber. Dried air was drawn through the containers at a known flow rate, which could be varied to suit the experimental temperature and the number of mole-rats in the container. Samples of air were drawn off the air stream and these were analyzed in a Beckman oxygen analyzer OM-IL. The activity of the mole-rats and the chamber temperature was monitored throughout the experiment. At the end of each experiment, the rectal body temperature of each mole-rat was taken using a Bailey Bat-4 laboratory thermometer with attached microprobe. The environmental temperatures used were 20°C, 25°C, 30°C, and 34°C, with the mole-rats in groups of four or singly. The mole-rats were not postabsorptive when placed in the chamber; the emphasis of this study was on deter- mining the responses of mole-rats under as normal a situation as possible. Mole-rats feed intermittently throughout the day and experimental animals were removed from the colony as re- quired. It is highly likely that the animals used in successive experiments were at about the same absorptive stage. RESULTS Social Organization The social structure of the established colony consisted of a single dominant female weighing 53 g, ten mole-rats of both sexes forming the working class and with an average weight of 32 g, and three non-working mole-rats (two males and one female) with an average weight of 38 g. A similar distribu- tion of numbers and sizes appeared to be the pattern in the other two captive colonies. The dominant female supresses breeding in all the other females in the colony. Before the social hi- erarchy was completely established in the captive colonies, females coming into estrus at the same time would fight each other, frequently resulting in death for one of them. In the established colonies, this situation never arises and high intensity aggres- sion is never seen. The breeding female is dominant over all other mole-rats in the colony. She initiates courtship and will solicit any male in the colony. However, it is not yet certain whether all males successfully copulate with her. Removal of the dominant female results in another mole-rat assum- ing the dominant role. This animal appears to come from the non-working mole-rats, but, as is evi- denced by the fact that my colonies were estab- lished from working class animals, the potential for breeding appears to lie latent in all the females of the colony. The litter size of mole-rats born in captivity ranged from three to 1 1 young. Death of a newborn litter may lead to the dominant female breeding again. However, if the young survive, the female appears to breed only once a year. One captive fe- male produced three litters and a total of 24 young within a period of 6 months, only the last litter sur- vived and she has not bred again for 11 months. In the field, breeding appears to be associated with the rainy season (Jarvis, 1969) and it is possible that in unusually good years a litter may be born in both the long and short rains. Growth rates of the young are exceptionally slow, with juveniles taking at least a year to attain adult size (Eig. 1). In spite of their slow maturation rates, the young mole-rats join the worker class at the age of 2 to 3 months. The non-working mole-rats are the next largest animals in the colony. They remain within the con- fines of the nest area and are only seen when they emerge to urinate or defecate. Their role in the col- 1978 JARVIS— NAKED MOLE-RAT ECOLOGY 83 AGE - MONTHS Fig. I. — The mean growth rate of Heterocephalus born and reared in captivity. Compiled from data from two litters (seven mole-rats). ony appears to be to huddle with the dominant fe- male. It is probable that they are also the most fit to reproduce and that only these males successfully copulate with the dominant female. It is also pos- sible that if the colony numbers are seriously de- pleted, or if the rains have been exceptionally good, the non-working female(s) may also breed. The working class mole-rats make up the remain- der of the colony, appear to be of an equal sex ratio, and comprise the smallest animals in the colony. The females are non-breeding, but many of the males appear to be potentially capable of breeding in that sperm are produced and they attempt to mount the dominant female when she is in estrus. In the monthly samples collected at Mtito Andei no pregnant or parous females were found in a total of over 150 animals, whereas many of the males had spermatozoa in their vasa deferentia, indicating that the situation found in the captive colony does re- flect field conditions. The working class mole-rats are responsible for the digging of the burrow system, location of the food, transport of food to the non-working animals, and for carrying nest materials to the nest. A high degree of cooperation is shown within this class when they are digging (Jarvis and Sale, 1971). Mu- tual coprophagy also occurs with the recipient mole-rat (frequently from the non-working class) begging feces from a donor mole-rat. The smaller size of the working class mole-rats is partly due to the fact that young animals are in- corporated into this class at an early age. However, my evidence suggests that the growth rate of these working animals is also retarded. All members of the colony studied were at least 3.5 years old and this size difference was still apparent. Eurthermore, the average weight of the whole captive colony is comparable to that of an almost complete colony of 39 mole-rats collected recently at Mtito Andei. These two facts suggest that this size difference is an inherent part of the colony structure. Another characteristic of the colony is the im- portance of mutual contact. This appears to be es- sential for the well-being of the mole-rats and, if forcibly separated from the remainder of the colo- ny, a mole-rat is restless and its overall condition deteriorates. Prior to sleeping, mole-rats frequently seek other members of the colony and, depending 84 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 AMBIENT TEMPERATURE Fig. 2. — Temperature selection in resting Heterocephalus. Observations were made at !0-min intervals and the number of animals resting singly or huddled at the different temperatures in the gradient are noted. on the ambient temperature, either huddle with them or lie close to them. The lower the ambient temperature, the greater the tendency to huddle. This sociality is in marked contrast to the majority of other species of mole-rats, which are solitary ag- gressive rodents. It should be stressed here that these observations on the social structure of Heterocephalus colonies should, at present, only be regarded as indications of what might be true for the species. More detailed laboratory studies on an almost complete colony of mole-rats captured in October 1977, are at present underway. Temperature Selection and Metabolic Rates From the temperature selection studies (Fig. 2), it can be seen that single resting mole-rats generally choose temperatures of between 32 and 38°C with a mean of 34°C. What is not shown in Fig. 2 is the duration of stay at these temperatures. Continuous observations on selected single animals demonstrat- ed that the visits to the higher temperatures tended to be fairly brief and once the mole-rat had “warmed-up” it selected a lower temperature for sleeping. Huddling mole-rats favored temperatures of between 25°C and 33°C in which to sleep (mean 30°C). It can be seen that the temperatures selected by Heterocephalus correspond closely to the bur- row temperatures of 30°C to 32°C measured in the field. Fig. 2 also clearly demonstrates the strong preference of Heterocephalus for huddling rather than solitary sleeping. From Fig. 3, it can be seen that physiologically, the optimum ambient temperatures also lie close to the normal burrow temperature range. There is a marked reduction in the metabolic rates of mole- rats at 30°C and 34°C and a sharp increase at lower temperatures. Also apparent from Fig. 3 is that throughout the temperature range studied, there is a significant dif- ference between the metabolic rates of mole-rats resting singly and those huddling. This difference is most marked at 25°C and 30°C where the meta- bolic rates of the groups of four huddling mole-rats are half of those of the single animals. At 34°C, the resting metabolic rate of the hud- dling mole-rats is close to the basal metabolic rate found by McNab (1966) and this temperature falls within the very narrow thermal neutral zone of these mole-rats. McNab (1966) has shown that 1978 JARVIS— NAKED MOLE-RAT ECOLOGY 85 AMBIENT TEMPERATURE ° C Fig. 3. — The metabolic rates and rectal body temperatures of Heterocephalus resting singly or huddling in groups of four, at different ambient temperatures. Standard deviation and standard error are indicated for the metabolic rates. The difference between the metabolic rates of single and huddling mole-rats is significant (P = .001 for 25°C and 30°C; P = .01 for 20°C and 34°C). above 34°C and below 31°C the energy cost to the animal increases very steeply. My own findings confirm this for the lower temperatures. The ability of Heterocephalus to thermoregulate is limited. The increased oxygen consumption with decreasing temperature and body temperatures el- DISCUSSION AND The temperature selection studies, environmental measurements, and the narrow ambient tempera- evated above ambient shown in Eig. 3 suggest that although these animals are thermolabile, they show some thermoregulatory ability within the tempera- ture range studied. Outside these temperatures, McNab (1966) found Heterocephalus to have a rap- idly diminishing ability to thermoregulate. CONCLUSIONS ture range over which Heterocephalus appears to thermoregulate and within which oxygen consump- 86 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 tion is minimal, all suggest that the naked mole-rat lives in an environment in which temperature fluc- tuations are minimal. They also suggest that al- though the burrow temperatures and humidities are fairly high, these mole-rats are rarely exposed to potentially lethal temperatures. Extremes of tem- perature are only encountered in very superficial burrows and these are avoided by Heterocephalus when temperatures are unfavorable. For these rea- sons, I feel it necessary to look for largely non-ther- moregulatory reasons for their unusual physiology and social structure. I would suggest that food is limiting to Hetero- cephalus. These mole-rats live in semidesert re- gions where most of their food is in the form of scattered tubers, which, unlike much annual vege- tation and grasses, are slow to respond to favorable rainfall. The food supply, although low, is at the same time fairly predictable and it appears that se- lection in both the physiology and social structure of Heterocephalus has favored features, which re- sult in a low but steady demand by the colony. The colony is so structured that the more ex- pendable mole-rats form the worker class and are involved in the energetically costly search for food. For maximal efficiency these animals do not breed. A limited few of the largest and probably the most genetically fit members of the colony reproduce and remain in the most protected part of the burrow system. These animals are waited on by the other members of the colony and they can therefore chan- nel all their energy into reproduction. They would also huddle with each other and further lower the metabolic cost of living. For the worker mole-rats, the largest expenditure of energy is in locating the patches of tubers. Once one tuber has been found, the chances of others occurring in the neighborhood are high and the whole area can be exploited with little additional cost to the colony. Evidence from the field suggests that burrowing activities (as indicated by an in- creased production of mole-hills) are heightened following the rains when the soil is softer and more easily worked than during the dry season. It there- fore appears that Heterocephalus concentrates its major food finding activities into the most energet- ically favorable times of the year. The tubers thus located are left growing and exploited as the colony has need of them. It seems probable that in semidesert areas a col- ony of mole-rats has a greater chance of surviving than a single animal. In the colony, the cost of bur- rowing can be shared by the working animals and the chance of locating clumps of tubers enhanced by having more than one animal searching the area. If the colony size is large enough, the search for food can be undertaken on several fronts at the same time. Obviously, there must be an optimum number of worker members to the colony; too many would rapidly exhaust the food found and too few would decrease the chances of finding the food sources. The burrow systems are extensive and frequently over 100 m long. Distances between tuber patches also appear to be considerable. When a long burrow system has to be dug, a small body size is advan- tageous in that burrow diameter can be kept to a minimum and the volume of soil displaced and the associated energy cost to the colony kept as low as possible. It is interesting therefore to find that Het- erocephalus is the smallest of the rodent moles. Unlike the majority of rodents of similar size, the number of naked mole-rats within the colony ap- pears to remain stable. This is evidenced by the low recruitment rates, limited breeding season, the low growth rates of the young, and the high longevity of the members of the colony (many of my captive animals are at least 5 years old). With this situation, it would appear that at no time of the year would there be a heavy demand on the energy reserves of the colony. Emphasis throughout is on a low but steady demand — a situation well suited to the steady but limited food resources available to the colony. A stable population size also lends itself to the development of a colony where a high degree of social structuring is possible. McNab (1966) suggests that because of a reduced potential for evaporative and convective cooling in the hot humid burrows in which Heterocephalus lives, there may be periods when there is consid- erable heat storage by the mole-rats, especially when digging. A low metabolic rate and high ther- mal conductance would reduce the probability of overheating. Although this may well be true, these features may also be linked to their limited energy resources. A metabolic rate that is less than 60% of the expected rate would considerably reduce the daily energy budget of the mole-rats and high rates of thermal conductance open the way to low-cost behavioral thermoregulation. The possibility that a low metabolic rate is linked to energy conservation was dismissed by McNab because he found that the lower limit of thermo- neutrality (31°C) in Heterocephalus lay above his 1978 JARVIS— NAKED MOLE-RAT ECOLOGY 87 mean burrow temperature measurements (30°C). Because of the small size of Heterocephahis and their high rates of thermal conductance, this would lead to a marked increase in energy expenditure above the basal level. He argued that if energy con- servation was important, there should have been a broad overlap between the burrow temperature and the zone of thermoneutrality. My measurements of burrow temperatures, dur- ing the height of the dry season, gave readings of above 31°C suggesting that, at certain times of the year, burrow temperatures do lie within the zone of thermoneutrality. My studies also suggest that, should the burrow temperature fall, Heterocepha- lus can reduce its rate of thermal conductance, and thus extend its lower limit of thermoneutrality, by huddling. Eurthermore, it is possible for Hetero- cephahis to utilize behavioral thermoregulation when temperatures are lower or higher than pre- ferred by simply moving to areas in the burrow sys- tem, which have a more favorable tem.perature. Thus, if heat stressed, the mole-rats could move to deeper and therefore cooler parts of the system where they can off-load heat passively. If too cold, movement to more superficial parts of the burrow during the daylight hours and the early evening would often bring the mole-rat to warmer parts of the system and the animal could again employ pas- sive means to raise its temperature. Therefore, this evidence suggests to me that here we have a rodent whose physiology, social organi- zation, and behavior have all been influenced by the harsh environment in which it lives — an environ- ment where temperatures are high and food appears to be a limiting factor. An understanding of the ethology and social organization of these mole-rats, and the linking of these with their physiology, sug- gests that both temperature and food resources must be considered when seeking an explanation for the unique features found in Heterocephahis . ACKNOWLEDGMENTS I am grateful to Dr. P. C. Withers and Mr. R. Bally for allow- ing me to draw on data on the physiology and temperature se- lection of Heterocephahis, to be published later, jointly. Dr. Peter and Mrs. Heidi Bally of Nairobi. Kenya, have given me invaluable assistance on numerous occasions both in sending me live Heterocephahis and in assisting me in my field work; my very sincere thanks go to both of them. My thanks also go to Mrs. S. Hardman for typing this paper. LITERATURE CITED Jarvis, J. U. M. 1969. The breeding season and litter size of African mole-rats. J. Reprod. Fert. Suppl., 6:237-248. Jarvis, J. U. M., AND J. B. Sale. 1971. Burrowing and burrow patterns of East African mole-rats Tachyoryctes, Heliophob- iiis dead Heterocephahis. J. Zook, London, 163:451-479. McNab, B. K. 1965. The adaptation of the naked mole-rat to its burrowing habits. Yearbook of Amer. Phil. Soc. Pp. 334-335. . 1966. The metabolism of fossorial rodents; a study of convergence. Ecology, 47:712-733. . 1968. The influence of fat deposits on the basal rate of metabolism in desert homoiotherms. Comp. Biochem. Phys- iol., 26:337-343. Thigpen, L.W. 1940. Histology ofthe skin of a normally hairless rodent. J. Mamm., 21:449-456. MODELING OF THE POPULATION CYCLES OF TWO RODENTS IN SENEGAL BERNARD HUBERT Laboratoire de Zoologie Appliquee, O.R.S.T.O.M., BP 1386, Dakar, Senegal E. ADAM Laboratoire de Zoologie Appliquee, O.R.S.T.O.M., BP 1386, Dakar, Senegal ALAIN R. POULET Laboratoire de Zoologie Appliquee, O.R.S.T.O.M., BP 1386, Dakar, Senegal ABSTRACT Models of adaptive strategy to periods of decreasing popula- ecological parameters of the two species account for their re- tions in two rodents in Senegal are given. Actual densities of spective adaptive success during periods of low population den- Mastomys erythroleucus and TateriUus gracilis are compared sities. mathematically to observed densities. Differences in various INTRODUCTION The two principal species of rodents — Mastomys erythroleucus (Temminck) (Rodentia, Muridae) and TateriUus gracilis Thomas (Rodentia, Gerbillidae) — present after the population outbreak of 1975-1976 in the Bandia region of Senegal ( I4°37'N, 17°01'W), were studied from November 1975 to August 1977. This study particularly involved the large popula- tion decrease after the 1975-1976 outbreak. Results are given for the different habitats together — a dry deciduous woodland, some areas of it cut for char- coal production and adjacent areas under cultiva- tion. Different soils are present in Bandia, but all the areas included in this study are on tropical lat- eritic soils. A further description of the area was presented by Hubert (1977). METHODS A large number of animals (more than 1,500 individuals) were caught in snap traps (60 traps during four nights/week in four different habitats) and in 500-m-long lines consisting of 50 live- traps of iron wire (type Manufrance) placed every 10 m. Speci- mens were autopsied to determine their sexual activity (partic- ularly the number of young in the litter of the females, which varies during the breeding period increasing at first and then decreasing. Table 1). The eye lenses were taken and dried for weighing to determine the age of the individuals collected, by comparison with a diagram established from rodents in captivity (Hubert and Adam, 1975), Thus the approximate dates of birth of each generation are known. From August 1976, an area of 600 m by 1,000 m was trapped twice each month with 160 traps in rotation and two plots of 4 ha each were trapped every 1.5 months by mark-and-release method, with a 10 m by 10 m grid of 441 live traps of iron wire (type Manufrance). The first trapping allowed determination of the density by CMR method during 10 days. The subsequent trapping (five nights every 1.5 months) allowed the monitoring of the marked population, estimation of the densities, and the distinction between migrations and mortality. Thus a monthly death rate was estimated for different periods; it varied accord- ing to the density and to possible epizootic disease being present. The death rate is calculated by the difference between the "load of living animals on the area” at one trapping period and at the following one, that is, the number of the formerly marked ani- mals increased by the newly marked, which will be recaptured later and an average number of “residents” animals represen- tative of the animals crossing the area during the trapping period. This loss could be interpreted as the death rate for a large enough area (where the number of entering rodents is equivalent to the departures) and when the calculation is made with the overall data for different environments taken together. A disease could have occurred from October to December 1976; in fact, a virus ("Bandia” virus, isolated from ticks and one Mastomys 10 years ago) was discovered again in January 1977 in four species present in Bandia, after a large population decrease. Its lethal effect has been demonstrated in the laboratory on Mastomys erythroleucus by the death of all the young in 10 days. Experiments are in progress for the other species. A mathematical formula has been adjusted for modeling the population cycle of rodents in terms of the following data, that is, number of young in each litter, mean date of birth of each generation, and monthly death rate for each period. Terms for the formula are as follows: P(t), the population at time t (in days); the population at time t = o; M, the monthly death rate (O =£ M 1); ni, the average number of young for the i'*’ litter; Ti, the date (in days) of the i'*’ litter. The sex ratio is supposed to be 1 .0. For Mastomys erythroleucus, a Hewlett Packard HP 65 com- 88 1978 HUBERT ET M..—MASTOMYS AND TATERILLUS POPULATION CYCLES 89 puter was employed, using the formula: P(t) = P„10 log( 1-M) 30 Pi,n, 10 log( 1-M) ,30 10 t-T| log( 1-M) 30 P„n2l0 T- log( 1-M) ,30 10 t-Tj log( 1-M) 30 Pi,n.>10' Tj log(l-M) 30 t-T, logU-M) •lo’" Tjlogd-M) + ’^[p„io"’ The formula is not simplified, as it was used for the programming of the HP 65. The evolution of the different generations is given by the same program, where fia = ng = n4 = 0, and so on. For TateriUus gracilis, it is not possible to use the same pro- gram because too many generations occur in the same year (nine from August 1975 to August 1976). In this case we used the following formula (example from May 1975 to August 1976): P(,n, 10 T^-T, log(l-M) 30 10 t-T^ log( I-M) 30 P(t) = P„F + p, f -h 7L (p,„, + p,,^,) F + P„F 2 2 2 + (Po + ^ Po«)'F + ^ (p„ + ^ P,„) -F + ^ (P« + Po.)-F + -^(Po + P,.)-F + ^[P» + P,« + (Pofi + P.uj • F + JitTp,, + JU p + JL (p,,,^ + p,,^^) + JLl pi .F 2 L 2 2 2 J + IffP" + ^ (P.« + P,D 2 L 2 2 P„ + _t_ log( l-M) where: F = 10^® Table I. — The main reproductive data for Mastomys erythro- leucus and TateriUus gracilis in the Bandia area during the 1975- 1976 and the 1976-1977 breeding periods. Species Approximate date of each generation Aver- age num- ber of young per litter Parti- cipa- tion of young ani- mals in breed- ing Mastomys 1 November 1975 8 - erythroleucus 1 December 1975 13 - 12 January 1976 13 - 28 February 1976 10 -f 1 October 1976 8 - 1 November 1976 10 - 1 December 1976 13 - TateriUus 10 May 1975 3 - gracilis 20 August 1975 3 - 15 September 1975 4 - 1 November 1975 4 - 5 December 1975 5 - 30 December 1975 5 - 25 January 1976 5 - 25 February 1976 3 - 25 March 1976 3 - 5 May 1976 1 -1- 20 August 1976 3 - 25 September 1976 4 - 25 October 1976 5 - 25 November 1976 5 - 20 December 1976 5 - 20 January 1977 5 - 30 March 1977 3 -t- P„B is the number of animals older than 6 months in P„; P„,, is the number of animals between 3 and 6 months old in P„. We know P(464) = 6; a sample of TateriUus caught at t = o gave the pop- ulation structure (dry crystalline lens weight), so we know also P„s = f(P„) and P„3 = f(P„) and now we can compute P„; P„ known, we can compute P(t) from t = 0 to 464 days. The ob- served density (by CMR method) of August 1976 is used as the basis of all the calculation for the two species. RESULTS The two graphs (Eigs. 1 and 2) present the fol- lowing data: The total fluctuations in the number of animals present per hectare at time t; the trapping population at time t, consisting of adults and re- cently weaned young; the appearance and growth of each litter until the disappearance of all its indi- viduals; the ratio of each age group in the popula- tion at time t. It is easy to see that if the actual densities of the two populations are now equivalent and close to the observed densities, they did not have the same pre- vious development. Mastomys erytliroleiiciis accounted for a large portion of the population outbreak of 1975-1976, and its densities were very high during the last year. The possible occurrence of an epizootic disease (or 90 BULLETIN CARNEGIE MUSEUM OE NATURAE HISTORY NO. 6 1975 1976 1977 Fig. I. — Fluctuations in the population level of Mastomys erythroleiiciis on 1 ha near Bandia, Senegal. P(t) 1978 HUBERT ET M..—MASTOMYS AND TATERILLUS POPULATION CYCLES 91 a different factor increasing the death rate) reduced the population considerably to the actual rate in spite of large reproduction. Mortality of the adults was high, and that of the young was such that the first litters, which had bred at the end of the breed- ing period in 1975-1976, could not do so in 1976- 1977. Although the densities of Taterilliis gracilis were relatively high in 1975-1976, there was no popula- tion outbreak of this species. Densities were almost unchanged and less subject to variations during that year because of a longer breeding period and a higher individual survival rate. However, the Ta- terilliis population was also affected by the disease, as the monthly death rate increased considerably and the population remained in the fields only be- cause of the continuation of the breeding period late in the year. DISCUSSION Once more the difference of adaptative strategy appeared between Mastomys and Taterilliis popu- lations as discussed below. Mastomys erythrolencus . — This species has a short breeding period, but with large litters (eight to 13 young per litter), allowing the population to reach a very high level. This large production of young animals permits colonization of new environ- ments, as described by Hubert (1977). They also possess resistance to various disasters (drought, diseases) and the ability to exploit the environment when the production of young is highest, as in the beginning of the dry season. These young animals supply the parental generation for the next year. Taterilliis gracilis. — The breeding period of this species continues for a longer time; it begins earlier in the wet season and continues later into the dry season, with the largest participation of the young animals. The fertility rate is lower than in Masto- mys-, three to five young are produced per litter ac- cording to the period of the breeding season. Pop- ulations are more regularly present in the fields than those of Mastomys and they resisted the disease by maintaining an almost standard breeding period in 1976-1977. The individuals of this species that live more than one year are more numerous than in Mastomys, thereby maintaining the population in large areas. For this computation, the death rate was sup- posed to be constant throughout the life of the an- imals, and all the females older than 6 months, or 3 months if the young females do participate to the breeding period, are supposed to be littering at each generation. These two hypotheses do not contradict the observed data. Using observed densities in Au- gust 1976 as the basis of calculation, the expected densities obtained for August 1977 are very close to the observed ones for the same period (that is, about two individuals per ha for each species). The resemblance between observed and comput- ed data allows us to do the same calculations on the fluctuations of the densities. This model can also be used for the calculation of productivity by estimat- ing the complete number of rodents produced, in- cluding the juveniles too young to be trapped. This work has been carried on with a financial support of the C.N.R.S.. contract no, 1 651-2294-ATP “Dynamiqiie des popu- lations.” LITERATURE CITED Hubert, B. 1977. Ecologie des populations de Rongeurs de Ban- Hubert, B., and F. Adam. 1975. Reproduction et croissance dia (Senegal) en zone sahelo-soudanienne. Terre et Vie, en elevage de quatre especes de Rongeurs Senegalais. Mam- 31:33-100. malia, 39:57-73. RADIO-TRACKING OF A SMALL RODENT, HYBOMYS UNIVITTATUS, IN AN AFRICAN EQUATORIAL FOREST HUGUETTE GENEST-VILLARD Museum National d’Histoire Naturelle, 55 rue de Buffon 75005 Paris, Erance ABSTRACT Preliminary studies of home range, activity periods, and females indicate males have larger home ranges and are generally movement in Hyhomys iinivittatus (Mammalia: Muridae) was more restricted to them than females. Both sexes are diurnal in studied by using radio-tracking. Comparisons between males and their periods of activity but females are less active than males. INTRODUCTION Most small rodents are easily caught in live-traps. However, it is often difficult to watch them in the field, especially in closed biotopes. Therefore, since 1975, I have used a radio-tracking technique (trans- mitter AVM, SMI type; telemetry receiver model LA 12, with Yagi antenna). I have been working for two years in the Central African Empire, near M’Baiki. These preliminary re- sults concern a diurnal murid, Hyhomys iinivittatus (Peters, 1876). This mouse measures 100 mm in length of head and body and weighs 50 to 70 g. The fur is soft. The dorsal color ranges from light yellowish brown to dark brown, with a black middorsal stripe that ex- tends from the nape to the base of the tail. The ventral color ranges from tawny to grayish white. This rodent species is only terrestrial. It lives in burrows in which it builds nests made of twigs. An individual may have three or four burrows but one is more often occupied than the others. Hyhomys is solitary in its burrows. METHODS Experimentation Ten Hyhomys, six males and four females of various ages, have been observed by the radio-tracking technique during De- cember 1975, January, April, and May 1976, and January, Feb- ruary, April, and May 1977; that is to say in the dry season and and the beginning of the rainy season. During these observations, one female aborted its young; two animals, male and female, lost their transmitters (the female was just leaving its home range); finally, a male died of a wound made by the antenna-collar that it had been carrying for 3 weeks. The transmitter weighs 12 g with its wrapping. Environment These observations were made in a dense equatorial forest in the process of secondary serai stage. There the underwood is dense and the ground is covered with branches and trunks of dead trees intermixed with creepers. An area of 10,000 square m has been gridded with narrow cross-trails at 20-m intervals, to compensate for the weak range of the transmitters. These cross-trails have been just roughly cleared, to reduce reluctance in the rodents to cross them. At each crossing a Sherman trap or a Saint-Etienne trap was placed. The rodents caught by this method have been marked by toe- clipping and thus it has been possible to watch them year round. The most assiduous visitors to the live-traps were selected to carry a transmitter; this selection increases the chance of re- covering the instrument later. RESULTS AND DISCUSSION Home Range of Hybomys Captures and recaptures gave us information about the home range of Hyhomys. From radio- tracking, it has been possible to delimit the areas more exactly. They vary from 4,500 to 6,100 square m for the cases of males and from 1,400 to 1,800 square m for females. A male spends its entire adult life (about 12 months) in the same home range. When the indi- vidual dies, its home range remains vacant for a while, but within a few weeks or months, another male or a female will take possession of the area. The limits of the new individual’s home range will not be exactly the same as those of the previous one, particularly if the new individual is a female. 92 1978 GENEST-VILLARD— RADIO-TRACKING HYBOMYS 93 Fig. 1. — Home ranges (in meters) of male and female Hyhomys univittatiis from January to May 1976. Four home ranges were calculated from results of the radio-tracking technique. Broken lines indicate the home ranges defined by the "capture-mark-and-re!ease" technique. The home ranges of neighboring males do not overlap. If one of the males dies, its neighbor can enlarge its own home range but this enlargement is always moderated. Only the young Hyhomys can move from one home range to another, contributing to the species dispersal. Eor instance, a young fe- male left its home range and lost its transmitter at 280 m from its burrow. This distance represents seven times the width of its home range. The animal certainly went farther than that because we never caught it again. Males do not tolerate other males in their home range. If two males come together, the owner gives chase to the other individual and bites it cruelly if owner overtakes the intruder. Some males share their home range with one or two females, which occupy smaller areas (see Fig. 1). It might happen that the female who lives with a male during some 94 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Fig. 2. — Movements of an individual male Hyhomys monitored by the radio-tracking technique. Daily movements are recorded for 4 days in January 1977. 1978 GENEST-VILLARD— RADIO-TRACKING //raOMFS 95 Fig. 3. — Movements of an individual female Hyhomys monitored over a 6-day period in February 1977. Numerous short movements around the burrows are not indicated. months chooses to go with a neighboring male. However, it can be seen that the coexistence of the two sexes is not the rule. The individuals do not sleep in the same burrow. Surely the overlapping of the home ranges favors mating in the breeding season. Hybomys Activity in the Forest All the activity of Hyhomys outside of the burrow is diurnal. Males and females do not have the same type of activity. Case of males. — The activity timetable of males is rather regular for an individual but it varies from one individual to another. Thus an early-rising male may leave its burrow near 7:20 A.M., whereas an- other may wait until 8:30 to 8:50 A.M. The individ- ual will run quickly a distance of 20 to 30 m to reach a part of its home range that it will be exploring carefully during the next several hours. It follows a different path each day and some of them do not visit the same part of their home range on two con- secutive days (see Fig. 2). In the morning, movement is done at top speed, but in the afternoon the animals are not so swift. During the day, we have observed short periods of decrease in movement, but they are not a true rest 96 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 period, because the animal feeds or performs its toilet. Besides these short periods of non-movement, there are two resting periods, one or two h long. The first is in the middle of the morning and the second in the course of the afternoon (for instance between 9:00 and 10:00 A.M. and between 4:30 and 5:30 P.M.). Males take their rest in their principal burrow (that is to say in the burrow where the an- imal spends most of its nights) or in some special areas that Hybomys like very much (for example, piles of twigs, dead branches, underpart of a blown down trunk, or old “termite houses” with many holes). Sometimes several species of rodents share the same “termiterium.” Hybomys returns to these places many times in the course of its walks. Each day, a male covers a distance of about 500 m. This distance has been evaluated by summing up all the recorded movements. I have not been able to take into account the numerous sporadic movements and small bolts that the animal does in quest of food; such activity is not perceptible with my method of telemetry. The mean speed of a male Hybomys varies be- tween 1 10 cm/min in the morning, to 80-90 cm/min in the afternoon, but it can reach a higher speed. For instance, an animal released at dusk at the place of its capture tries only to return to its resting place and it does that at a speed of 720 cm/min. It does not run directly to its burrow, but describes a large circle, as if trying to get its bearing, and it does not come into its hole without turning around first. A male usually spends the nights in the same bur- row. It regularly visits three or four other burrows in which it occasionally sleeps, and sometimes vis- its the burrow of a female, which may live in an overlapping home range. The last return to the principal burrow takes place between 6:00 P.M. and 6:05 P.M. or between 5:45 and 5:50 P.M., according to the individuals in- volved. If the males don’t begin their activity at the dawn, they cease their activity at full dark in the forest. Female activity. — Females are not as active as males. They move only short distances; they circle their burrows, entering into them often, even if there are no young in the nest. They repeat this frequent- ly and do not go further than 50 m from their bur- rows. Thus, females are sedentary, but, quite curi- ously, they have several burrows and can spend a night in each of them; those burrows are about 10 m apart (Fig. 3). In the course of a day, a female can spend long periods out of its burrow. It remains motionless un- der dead leaves, and returns to the same places for consecutive days between 10:00 A.M. and 4:00 P.M. At times, it takes a long rest in its burrow, for 40 min to 2 h between 10:00 A.M. and 4:00 P.M. Female activity begins between 7:45 and 8:30 A.M. and the activity ceases between 5:00 and 6:00 P.M. The time of activity of a female and of a male are not very different, but the females are not so reg- ular. The most striking difference between male and female activity concerns the intensity of this activ- ity and the length of the movement. It appears that females are not so strongly attached to their home range as males, since even as adults, they can leave one home range for another. A female, which lived in January on a part of the home range of a male, was caught in April, and watched during 3 weeks by the radio-tracking technique on another area. This latter was situated 70 m from the first area, overlapping the home range of another male slightly (see Fig. 1). In addition, females are not so “home- loving” since they scarcely spend two consecutive nights in the same burrow when they have no young. This paper was communicated hy Francis Fetter. SEASONAL POPULATION CHANGES IN RODENTS IN THE KENYA RIFT VALLEY M. J. DELANY Baharini Wildlife Sanctuary, Kenya, and Department of Biology, University of Southhampton, United Kingdom (present address: School of Environmental Science, University of Bradford, United Kingdom) C. J. ROBERTS Baharini Wildlife Sanctuary, Kenya, and Department of Biology, University of Southampton, United Kingdom (present address: Department of General Studies, Bradford College, United Kingdom) ABSTRACT Live trapping of small rodents was undertaken on three grids each of 81 traps in grassland and scrub vegetation in the middle of the wet and dry seasons in Nakuru National Park, Kenya. Snap trapping was carried out concurrently. The most abundant species was Arvicanthis niloticus; Praomys natalensis, Lemnis- comys striatus, Otomys angoniensis, and Rhahdomys pumilio were obtained in moderate number. There were relatively great- er changes in densities in the grassland than in the scrub at the two trapping occasions; densities were invariably highest in the dry season. Reproduction was maximal in the wet season and had terminated by the dry season. Populations of Arvicanthis contained animals of all ages at both periods, whereas Praomys had a much broader age spread in the wet than the dry season when few older animals were present. The scrub vegetation ap- parently forms population reservoirs at adverse times of the year when it appears that the grassland is unable to provide suitable conditions. INTRODUCTION In recent years several studies have been under- taken on the ecology of small rodents in the grass- lands of East Africa (Delany 1964, Delany and Neal, 1969; Neal, 1970; Cheeseman, 1975; Taylor and Green, 1976). With the exception of Taylor and Green’s (1976) work, which was undertaken in the same region as the present study, the remaining re- search has been carried out in western Uganda. These latter studies examined times of breeding, population dynamics, and other aspects of the ecol- ogy of animals in areas where there are typically two discrete rainy seasons. In Uganda it was found that the breeding of most species of small rodents commenced soon after the onset of the wet season and terminated shortly after it ended (Delany and Neal, 1969). Inevitably, these breeding patterns ex- erted their influence on population dynamics throughout the year (Cheeseman, 1975). In their study at Nakuru and Kitale in Kenya, Taylor and Green (1976) examined the relations be- tween reproduction, diet, and climate. They dem- onstrated that the seasonal effects of rainfall on the vegetation brought about variations in the quality and quantity of available food for rodents, some of which could be associated with the regulation of breeding. This valuable work was based upon ani- mals obtained by snap trapping and did not provide detailed information on seasonal changes in popu- lation densities and structure. The present study supplemented and extended the work of Taylor and Green (1976) by providing information on these as- pects of small rodent ecology in a typically grass- land locality having a single protracted rainy sea- son. This rainfall pattern prevails in many parts of the Kenya Rift Valley. Here the rains build up to a peak in March and April from the dry season in December and January. Typically, they continue into November without a break (Fig. 1). This atten- uated wet season is not commonly found in the sea- sonal tropics where the rains usually last for appre- ciably shorter periods. The field work in this study was undertaken at Baharini Wildlife Sanctuary of Nakuru National Park on the northern shores of Lake Nakuru, Kenya, during July, August, and December 1974 and January 1975. These periods coincided with the middle of the wet and dry sea- sons. 97 98 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Fig. 1. — Rainfall figures at Nakuru Railway Station. The histograms represent the mean figures from 1904 to 1970 and the continuous line 1974 figures (source: East African Meteorological Department). STUDY AREA The field research was undertaken in a relatively small area of grassland with occasional scrub to the north of Lake Nakuru and never more than 3 km from the lake shore. Hereabouts the lake is fringed by a narrow strip of olive bark acacia (Acacia xan- thophhea Benth.) woodland with an understory of shrubs, herbs, and grasses including Setaria palli- difusca (Schumach) C. E. Hubb, Erlangea cordi- folia (Benth. ex Oliv.), Tagetes miniita L.,Abiitilon mauritianum (Jacq.) Medic., and Solanum sp. Be- yond this woodland lies an extensive and level area of grassland, which to the east rises to a stony hill- side covered by small Acuc/u trees. Here extensive rock outcrops result in a poor covering of ground vegetation. Between the lakeside woods and the hillside, the extensive grasslands were interspersed with very small patches of scrub, and the occasional tree or small clump of trees. It was within an area approx- imately 3 km square of this grassland and scrub that the research was concentrated. This was at an al- titude of 1,870 to 1,950 m above sea level. Within this area live trapping was undertaken in five grids and snap trapping over the area as a whole within its typical vegetation types. As considerable refer- ence is subsequently made to the grid trappings, their vegetation is given in some detail. Grids 1 and 2 were placed in an extensive and typical area of grassland. Here in July much of the vegetation was 1 m or more tall and consisted main- ly of Hyparrhenia hirta (L.) Slapf. , Themeda trian- dra Eorsk. and Chloris gayana Kunth. Sporobolus pyramidalis Beauv., Setaria pallidifusca (Schu- mach) C. E. Hubb, A/;Ttn/a adoensis Hochst., and Harpachne schimperi A. Rich, were also present. The area was interspersed with a few small Acacia trees, the occasional compact bush of Lippia java- nica (Burm. f.) Spreng, small woody Hibiscus apo- neurus Sprague and Hutch., and scattered Solanum plants up to approximately 60 cm tall. There were numerous small herbs adding to the thickness of the ground cover. Over most of these grids the grass and herbage formed a dense cover although there were a few small areas of sparsely covered ground. In addition to the rodents obtained in this work, other mammals present included spring hares, Pe- detes capensis (Eorster), mole rats, Tachyoryctes splendens (Ruppell), ant bears, Orycteropus afer (Pallas), and steinbok, Rhaphicercus cainpestris (Thunberg). A herd of waterbuck, Kobus ellipsi- prymnus (Ogilby), were occasional visitors. Grid 3 contained slightly different vegetation to grids 1 and 2. Although there was a good grass cov- er of Hyparrhenia, Themeda, Setaria, Aristida, and Chloris on about three quarters of the grid, more herbs and shrubs interspersed with the grasses over 1978 the remainder. These included Erlangea, Lippia, Indigofera vohemarensis Baill. , Leo not is, and Tagetes. There was also a small clump of trees. The vegetational character of grid 4 and grid 5 was quite different from the preceding three grids. In view of the detailed studies undertaken on them, these two grids had their vegetation mapped. Whereas grids 1 to 3 were within grass-dominated vegetation, 4 and 5 had a much more shrub and bush character. On grid 4 grasses {Hyparrhenia, Themeda, Aristida, Chloris) were present but cov- ered little ground. Lippia was abundant and dense, attaining a stature of 50 to 150 cm over much of the grid. It was frequently accompanied by dense Cyn- odon aethiopicus Clayton and Harlan. Among the other plants present were Ocimum suave Willd., Alternanthera aspersa, Zehneria scabra (L. f. ) Sand., Bidens pilosa L., Riunex iisamharensis (Dammer), Tagetes, Leonotis, Solannm, and Lrlan- gea. Over much of this grid the thickness of the vegetation made it difficult to penetrate and neces- sitated the establishment of narrow footpaths. Grid 5 was similar to grid 4 in that it had the same shrubby character and contained large amounts of Lippia. There was more grass in grid 5 although the shrub-like character was much in evidence. It did not contain as great a variety of herbs and woody plants as grid 4 and was more readily penetrable. 99 Grids 4 and 5 were located 2 km from the track around the lake along the road to Lanet. They were situated 5 m to the west of this road and almost parallel to it. The two grids had the same compass bearings and were only 21 m apart. Because of their spatial relationship to the road, it was not possible to align them so that they had a continuous base line. Grid 5 was therefore displaced 10 m west of grid 4. These grids covered 40 square m. The veg- etation described for these grids was largely con- fined to the two small patches encompassed by them. They were surrounded by the grassland de- scribed for grids 1 and 2. The vegetation described in these 5 grids is typ- ical of the non-wooded areas to the north of Lake Nakuru. The removal trapping was undertaken in these types of vegetation. Neither cattle grazing nor burning had taken place in the area for some years prior to this study. In 1974, the rainfall at Nakuru (Fig. I) was similar to, if not quite the same as, the statistical average pattern. Higher than average rains fell from May to September and the mid wet season trough was ear- er than usual. However, April had its typically high rainfall and the total figure for 1974 of 887 mm was only 26 mm more than the mean. It would therefore appear that the figures for 1974 ap- proached the typical situation. DELANY AND ROBERTS— ECOLOGY OF KENYAN RODENTS METHODS Grid trapping was undertaken using 81 (9 by 9) of the larger Sherman live traps. On grids 1 and 3 the spacing was 10 m, on grid 2, 20 m with alternate trap sites of grid 1 forming the central core of this grid, and on grids 4 and 5, 5 m. Traps were examined twice daily in early morning and late afternoon. On capture, each animal was identified to species, weighed, sexed, marked by toe clipping, and the location of the trap in which it was caught recorded. For males it was noted if the testes were descended and for females if they could be recognized as pregnant and/or lactating. They were then released. Traps were set on the grids on the following dates: grid 1 — 11-14 July 1974, 3-8 January 1975; grid 2 — 15-17 July 1974; grid 3 — 1-3 August 1974; grid 4 — 5-11 August 1974, 28 December 1974 to 2 January 1975; grid 5 — 18-24 August 1974, 20-25 December 1974. To ascertain occu- pancy immediately peripheral to grid 4 at the August trapping two rows of traps 5 m apart were set along the two sides of this grid that were not bounded by the road or grid 5. They were in grassland and were set from 14 to 16 August 1974. The catch was handled in the same way as other live trappings. Snap trapping took place concurrently with the live trapping using 40 commercially produced rat traps. They were set in trap lines and moved at irregular intervals largely in response to their success. The animals caught by these traps provided more de- tailed information on reproduction and relative age than the live caught animals. They had the standard external measurements taken and were weighed. Reproductive and development con- ditions were recorded. These included the condition of the uterus (active, distended, inactive, undeveloped, number of placental scars), number of embryos, weight of reproductive tract of preg- nant females, lactation, descent and size of testes, and juvenile pelage. Relative ageing was undertaken using seven categories of dental attrition in the upper molar row. The categories, erect- ed for this study, ranged from the incompletely erupted row of Class I to the row displaying considerable dentine exposure in Class VII. A similar method was used by Delany (1971) and Cheeseman (1975). Throughout all the trappings a banana-flour paste was used as bait. RESULTS Numbers Caught August and December and January. These increas- In grids 1, 4, and 5 (Table 1), there was a consid- es ranged from double to fifty-fold in the case of erable increase in rodent numbers between July and grid 1. Arvicanthis and Praomys comprised just 100 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 1. — Numbers of rodents caught on grids i, 4 and 5. Species Grid 1 Grid 4 Grid 5 July January August December August December cJ $ cJ 9 cJ 9 d 9 550 mm in 1976. Rainfall normally occurs as sharp, localized showers. As is common in desert or subdesert regions, daily and annual temperature fluc- tuations are severe; temperatures can drop to -I0°C in winter (though seldom for more than a few nights per season) and reach 47°C in the shade in summer. MATERIAL AND METHODS A trapline of 37 stations, 20 m apart, and extending from the low dune, through the riverbed and plateau, and up the high eastern dune, was laid out in December 1970. Trapping sessions were generally six months apart, and during each session three traps (a Museum Special snap trap, a Victor or McGill rat snap trap, and an aluminum Sherman live trap) were put down in a 1 m radius of the trap station, and kept baited and set for 3 days and 4 nights. Taking each 12 h period as a “trapnight," this gave a total of 21 trapnights (7 “nights" x 3 traps) per station, or 777 trapnights per trap session for the whole trapline. All animals caught were removed and live captures released 3 km beyond the study area. As the width of the arbitrarily separated habitats (low dune, riverbed, plateau, high dune) differed along the tran- sect, 10 trap stations were within the boundaries of the low dune habitat, giving 210 trapnights/session; nine in the riverbed, giving 189 trapnights/session; II in the plateau, giving 231 trapnights/ 118 1978 NEL— KALAHARI MAMMAL COMMUNITIES 119 Nov May Jan Jul Jan | Jan Jul Dec Jul Jan Jul Fig. 1. — Total number, and number of seed-eating and leaf-eating rodents trapped at different periods in the Kalahari Gemsbok National Park. Number of species in parenthesis. session; seven in the high dune, giving 147 trapnights/session. For comparative purposes, captures were recalculated as per 100 trapnights, either for particular trap stations {see calculation of habitat-niche breadth below) or for a particular habitat. Without discussing in detail the best concept of a niche it should be noted that here it is regarded in the Flutchinsonian sense, that is being a n-dimensional hypervolume with quanti- fication possible by measuring resource utilization along several axes. In the discussion below the niche referred to is the realized niche. Data on distribution of the various species in the four habitats accrued from an analysis of captures at the trap stations, re- garded here as each representing a different microhabitat. Data on food preferences were obtained by analyzing stomach contents. Volumetric content and wet weights were noted, and several samples from each stomach were drawn off, shaken with water, and placed in a petri dish for subsequent determination of percentage occurrence, using a Wild M5 stereomicroscope with graticule eyepiece. In the absence of a reference collection of plant cuticles, vegetable matter was classified only as "white” (seeds, roots, and stems) or "green" (leaves). Insect material was not identified to taxonomic groups. Activity measurements were obtained in the laboratory for only three species, during late March-April. A four-chamber. activity-measuring cage was used (for details see Davis, 1972). Individual activity bouts were recorded on a moving strip chart in an Angus-Esterline event recorder, and subsequently ana- lyzed as to the number of bouts per hour. The above data were used to calculate diversity indices (and evenness of spread) for each habitat, and for the whole transect, using the Shannon-Wiener formula H' = — S Pi\og.,Pi in the form (after Lloyd and Ghelardi, 1963) H' = c[log„,^ - (^1 «,log.„n,)] where C = 3.3219 (a constant) to convert log, to log,,,. This mea- sure for diversity was selected in preference to others (for ex- ample, the Brouillin formula) as the data collected represent only samples of the total community (Peet, 1974). Evenness of spread E = //'///'max. where //'max 's loga (no. of species). Niche breadth, based on microhabitats occupied, or food taken, or times of activity, was calculated from Simpson’s index of diversity Ip" 120 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 where pi represents the proportion of the (th microhabitat (or food resource, or time period) actually used;i? varies from unity to n depending on the p, values, and is standardized by dividing by n. For calculating habitat-niche breadth values, and following Krebs and Wingate (1976), data were converted to captures/100 trapnights/trap station, thus/? = niche breadth,/?, = proportion of species’ total density in microhabitat ( = trap station) ,./? is defined from average density estimates: where = number of individuals per 100 trapnights in micro- habitat Standardized niche breadth for each species in any of the four habitats was then obtained by dividing by n, or the number of trap stations. Habitat-niche breadth and overlap (see below) was calculated for each of the four habitats separately, and then for the whole transect. Niche overlap between any two species, for a given niche dimension, was calculated following Pianka (1973): O,. = O. = S(Aa- Aa-) where 0,j = (9,, = niche overlap between species i and species j (range 0-1) Afc = proportion of species / numbers in resource A (micro- habitat, food type, time period) Aa- = proportion of species j numbers in resource A. RESULTS Table 1 lists the species present in the study area. Owing to the sampling technique some species were not collected in the regular trapline. Eor example. most of the Acacia erioloba trees close to the trap- line harbored tree rats Thallomys paediilcus, but they were virtually restricted to these “habitat is- Table 1. — Species of small mammals recorded in the study area 1970-1976. Approximate weight in grams, basic activity period, and relative density (no. capturesIlOO trapnights over II trapping periods in all four habitats) are given, as well as the habitats in which captured or seen. Abbreviation for habitats: LD = low dune; RB = riverbed: PL = plateau: HD = high dune. Relative Taxa Weight Activity density Habitats Order Insectivora Family Macroscelididae Eiephantuius intufi Family Soricidae 51.4 diurnal 0.13 LD, RB, HD Crocidura hirta 16.0 nocturnal 0.01 LD Order Rodentia Family Sciuridae Xerus inauris diurnal RB Family Pedetidae Pedetes capensis Family Cricetidae and Muridae Subfamily Gerbillinae nocturnal LD, RB Gerbillurus paeba 25.9 nocturnal 3.31 All DesmodiUus auricuiaris 46.1 nocturnal 0.37 All Patera brantsii 64.9 nocturnal 0.37 LD, PL, HD Subfamily Otomyinae Parotomys brantsii Subfamily Dendromurinae 80.0 diurnal LD Dendromus melanotis 6.3 nocturnal 0.08 LD, PL, HD Malacothrix typica Subfamily Murinae 13.0 nocturnal 0.01 PL Rhabdomys pumilio 32.0 diurnal 4.68 All Mus minutoides 4.7 crepuscular/diurnal 0.37 All Aethomys namaquensis 42.6 nocturnal 0.01 LD Zelotomys woosnami 62.4 nocturnal 0.08 LD, HD ' Thallomys paedulcus 75.0 crepuscular/nocturnal LD, RB Saccostomus campestris Family Bathyergidae 47.0 nocturnal 0.04 PL, HD Cryptomys hottentotus Family Hystricidae ? LD Hystrix africae-austraUs nocturnal LD 1978 NEL— KALAHARI MAMMAL COMMUNITIES 121 Table 2. — Relative density (no. of capturesHOO trapnights) of small mammals in each of four habitats in the Nossob River, Kalahari Gemsbok National Park. Census period Habitat Decem- ber 1970 May 1971 January 1972 July 1972 January 1973 January 1974 July 1974 Decem- ber 1974 July 1975 January 1976 July 1976 Low dune 2.38 1.91 21.91 38.57 6.19 30.00 8.57 8.57 3.33 26.67 Riverbed 1.06 7.41 15.34 25.93 1.05 2. 12 25.40 8.47 3.70 2 12 14.29 Plateau 0.43 — 4.76 24.24 — 0.43 9.09 — 0.43 0.87 7.36 High Dune 2.04 2.04 14.97 43.54 0.68 11.57 19.05 6.80 6.80 5.44 13.61 Overall 1.42 2.70 13.90 32.10 0.39 4.51 20.60 5.70 4.63 2.70 15.44 lands” and are therefore not reflected in the cap- tures. Similarly, although the trapline bisected a large Parotomys hrantsi colony during 1972, none were collected, probably due to the type of trap utilized (Nel and Rautenbach, 1976). On occasion small groups of ground squirrels crossed the census area, as did Pedetes capensis, but again they were not trapped. Otherwise the three trap-types used proved effective for collecting the other species present. Eig. 1 shows that over the 6-year period Decem- ber 1970 to December 1976 the total number of small mammals, as well as the number of and con- tribution by different species, fluctuated a great deal. Total numbers showed three peaks and three troughs, which relate fairly closely to rainfall during the previous 12 months. The low number of cap- tures (5) in December 1976, after the study area was accidentally completely burned 6 weeks previously, obviously cannot be used for comparative purpos- es. ‘ Between-habitat Variation Although the study area was arbitrarily divided into four habitats on the grounds of differing vege- tational aspects and soil types, pronounced differ- ences were apparent in the relative density of small mammals in each habitat, and therefore its contri- bution to total numbers (Table 2). The plateau al- ways had the lowest relative density, and contrib- uted least to the community sampled by the transect. Otherwise either the low dune or high dune habitat usually had the highest relative density and contributed most. In brief, an equal or nearly equal amount of rainfall has a different effect on the productivity of the different habitats, and their abil- ity to support different species. The species diversity (using the Shannon-Wie- ner index) was usually higher in the low and high dune than in the riverbed and plateau, but marked dominance by one species often resulted in very unequal distribution of total numbers among spe- cies, which resulted in a decrease of evenness in spread or E (Table 3). However, as the number of species (and therefore diversity) increased, in most cases, especially in the low and high dune habitats, an increase in evenness of numerical distribution between species resulted. Sometimes, however, no- tably from January to July 1976 in the high dune habitat, diversity remained virtually constant but evenness decreased markedly. Changes in Commitnity Strtictnre and Nnmhers Apart from the changes in relative density of an- imals in each of the four habitats, changes also oc- Table 3. — Changes in species diversity of small mammals of four habitats in the Kalahari Gemsbok National Park. For calcula- tions of H' and E, see methods. * = insufficient data. Trapping period Low dune River- bed Pla- teau High dune Over- all January 1972 H' 0.97 0.79 0.68 0.70 1.41 E 0.42 0.50 0.68 0.44 0.50 July 1972 H' 1.97 1.16 1.54 1.14 1.76 E 0.76 0.58 0.55 0.49 0.59 January 1974 H' 1.88 * * * * E 0.81 * * * * July 1974 H' 1.89 0.25 1.41 1.98 1.69 E 0.63 0.25 0.61 0.77 0.51 December 1974 H' 0.94 * * 1.49 1.09 E 0.60 * * 0.94 0.54 January 1976 H' 1.38 * * 1.41 1.49 E 0.87 * * 0.89 0.75 July 1976 H' 1.38 0.73 0.52 2.09 1.38 E 0.53 0.46 0.52 0.81 0.49 f 122 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 C|] Dec '70 (1) 0 May '71 (1) • Jan '72 (5) A Jul '72 (6) O Jan. 74 (5) □ Jul '74 (8) V Dec '74 (3) ■ Jul. '75 (2) A Jan '76 (3) ▼ Jul '76 (6) Fig. 2. — Contribution by different species to total numbers trapped at various times in the low dune habitat. Spacing of rank order of species, on horizontal axis, is arbitrary. G = Gerbillurus paeha; R = Rhabdomys piimilio; D = DesmodiUus auricidaris; T = Tatera brantsii', Z = Zelotomys woosnami; M = Miis minutoides', E = Elephantulus intufi', A = Aethomys namaquensis', De = Dendromiis nwlanotis; C = Crociduni hirta. curred in the proportion, which different species contributed to the community in a particular habitat at different periods (Figs. 2-4). On the low and high dunes, even though the number of species and the contribution by each in the community varied, the basic structure of the community remained very similar during the study period (Figs. 2 and 4). However, in the low dune community the propor- tions of total numbers contributed by the first and second ranking (in order of contribution) species were more fluid than in the high dune community. Over the study period, although the first or highest ranking species usually contributed a very large proportion to total numbers in either the low or high dunes, and thus dominated the particular commu- nity, this dominant species was not always the same. Figs. 2 and 4 also show that the period during which a particular species remained dominant was different in the low dune and high dune communi- ties; in the former, Gerbillurus paeba was dominant up to January 1972, then again from December 1974 to July 1975, and again in July 1976, whereas G. paeba remained dominant in the high dune com- munity up to January 1974. 1978 NEL— KALAHARI MAMMAL COMMUNITIES 123 Riverbed Plateau Fig. 3. — Contribution by different species, arranged in order of rank, to the riverbed and plateau communities. Ma = Malacothri.x typica; S = Saccostomus campestris. Other abbreviations and symbols as in Fig. 2. By contrast the communities in the riverbed and plateau showed a different structure. During most census periods the first-ranking species contributed nearly all the animals caught, and therefore com- pletely dominated the community; also species di- versity usually remained low (Table 3) and at times only one species was present. It is of interest to note that the communities in the low and high dunes and riverbed (and on oc- casion the plateau) were primarily composed of spe- cies in the middle range of weight classes. When species numbers rose, it was the scarcer species at the bottom or top of the weight range that appeared. The selective factors mitigating in favor of a weight of 20 to 40 g in this particular environment remains, however, obscure. Resource Utilizution — Niche Dimensions In this study, data were accumulated on only three facets of each species’ niche, so that resource partitioning can only be attempted on the spatial, trophic, and temporal (or habitat, food, and time of activity) levels. As the traps were placed on the ground only indirect assessment of vertical activity or foraging levels can be attempted. Also no quan- titative data are available for those species present in the study area but not trapped (see Table 1, and above). Vertical feeding levels are shown in Lig. 5, but no quantitative data are available to allow differ- ential utilization of various feeding levels to be as- sessed. Horizontal distribution is somewhat better documented, and standardized habitat-niche 124 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 j Jan. 73(1) Fig. 4. — Contribution by different species, arranged in rank order, to total numbers trapped at various times in the high dune habitat. Abbreviations and symbols as in Figs. 2 and 3. breadths, calculated as described in the methods section, are given in Table 4. As predicted by Lev- ine (1968), niche breadth tends to increase to an asymptote as numbers rise (Fig. 6 and Table 4). As is to be expected, the more common species at any level of total number of animals or species present, for example, G. paeba and R. pumilio, tend to have wider habitat-niche breadth (Table 4), which indi- cates that they are less specialized and were utiliz- ing the available habitat to a greater degree than others. The limiting factors in the habitat for any species remain conjectural, but subjectively cover seems to play a decisive role for at least some spe- cies; for example, Desmodillus favors very open and exposed areas, relying on acute hearing for per- ceiving predators (Lay, 1972), and, when rainfall and vegetative cover increases, as happened after 1972, they tend to decrease in numbers, being found in the general area only on places artificially cleared, such as campsites. Similarly, cover, in the broad sense, affected Rhabdomys abundance, whereas Thallomys, being here confined to large camelthorn Acacia erioloba trees, had a very patchy distribution but probably reasonably stable numbers. The amount of habitat-niche overlap, calculated as explained in the methods section, among the more common species at the higher levels of abun- dance (see also Table 2) are given in Tables 5 and 6. During periods of low numbers, species do not overlap in the habitats and/or microhabitats (distri- bution points) utilized. For example, in December 1970 Gerbilliiriis occurred only on the low and high dunes, Desmodillus in the riverbed, and Saccosto- mus on the plateau; in May 1971, Gerbillurus again occurred only on the low and high dunes, Rhab- 1978 NEL— KALAHARI MAMMAL COMMUNITIES !25 Table 4. — Standardized habitat niche breadth of five rodent genera trapped at 37 trap stations in four habitat types in the Nossob River, Kalahari Gemsbok National Park, from December 1970 to July 1976. LD = low dune, RB = riverbed, PI = plateau, HD = high dune. Overall values are for transect as a whole, f — ) Denotes absent from a particular habitat type during that trap session. Trap session Decern- Janu- Janu- Janu- Decern- Janu- ber May ary July ary ary July ber July ary July Taxa Habitat 1970 1971 1972 1972 1973 1974 1974 1974 1975 1976 1976 Gerbillurus LD 0.36 0.16 0.80 0.75 — 0.54 0.65 0.10 0.20 RB — — 0.30 0.44 — — — — — — 0.1! PI — — 0.49 0.71 — 0.09 0.18 — — — — HD 0.26 0.14 0.69 0.88 — 0.64 0.78 — — 0.38 0.14 Overall 0.14 0.06 0.48 0.60 — 0.23 0.36 0.03 — 0.07 0.10 Desmodillus LD — — 0.20 0.45 — 0.10 — — — 0.10 — RB 0.22 — 0.1 1 0.30 O.Il — — — — — — PI — — — 0.33 — — 0.09 — — — — HD — 0.14 0.14 0.43 — — — — — — — Overall 0.05 0.03 O.IO 0.36 0.03 0.03 0.03 — — 0.03 — Tatera LD RB — — 0.10 0.27 — 0.20 0.23 0.18 — 0.10 O.iO PI HD — — — 0.09 — — 0.29 0.40 — 0.14 0.14 Overall — — 0.03 0.10 — 0.05 0.10 0.12 — 0.05 0.04 Rhabdomys LD — — 0.27 0.57 — 0.20 0.73 0.65 0.57 0.27 0.87 RB — 0.42 0.50 0.77 0.11 0.30 0.84 0.43 0.26 0.30 0.53 PI — — — 0.32 — — 0.37 — — 0.09 0.22 HD — — — 0.47 — — 0.68 0.43 0.60 0.43 0.60 Overall — 0.10 0.16 0.47 0.03 0.12 0.57 0.30 0.33 0.24 0.51 Mus LD — — — 0.36 — 0.10 0.10 — — — 0.20 RB — — — — — — 0.22 — — — 0.33 PI — — 0.18 0.09 — — 0.09 — — — 0.18 HD — — — 0.43 0.14 — 0.26 — — — 0.26 Overall — — 0.05 0.19 0.03 0.03 0.11 — — — 0.23 domys only in the riverbed, and Desmodillus and Saccostomus on the high dune, but not at the same distribution (= trapping) points. During January 1974, Gerbillurus , Tatera, Rhab- domys, Desmodillus, and Mus occurred on the low dune; in the riverbed only Rhab domys occurred and on the plateau and high dune only Gerbillurus. Overlap values for the low dune are as follows: Gerbillurusi'Tatera = 0.24; Gerbillurus IRhabdomys = 0.24; Gerbiilurus/Mus = 0.33; Gerbillurus/Des- modillus = 0.33; Rhabdomysi'Tatera = 0.50. Oth- erwise, the species present did not overlap in dis- tribution points. During December 1974, the only overlap was be- tween Gerbillurus and Rhabdomys on the low dune (0.18), Tatera and Rhabdomys on the low dune and high dune (0.16 and 0.58, respectively), as well as Tatera and Zelotomys (0.67) and Rhabdomys and Zelotomys (0.58) on the high dune. Few small mammals (apart from Rhabdomys only one each of Dendromus and Elephantiilus) were captured in July 1975, and no overlap occurred. In January 1976, there was overlap between Ta- tera and Rhabdomys only on the low and high dunes (0.41 and 0.58, respectively), Tatera andDes- modillus (1.0) and Rhabdomys and Desmodillus (0.41) on the low dune, and between Gerbillurus and Rhabdomys (0.24) and Tatera and Rhabdomys (0.58) on the high dune. To relate amount of habitat-niche overlap to any degree of competition among species would prob- ably be fallacious. Obviously, a high amount of overlap between Rhabdomys (diurnal) and Gerbil- lurus (nocturnal) may not be competition for the same resource, as this is used at different times. Probably the same applies as well to all nocturnal species, which may utilize the same place, but at different times of the night. Where a high amount 126 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 Fig. 5. — Vertical foraging levels (open circles), or level at which nesting sites occur (open triangles). Not drawn to scale. of spatial overlap does exist (see Tables 5 and 6), different food sources in the same general area may well be utilized (Table 7). Niche overlap is perhaps better seen as an indication of the similarity of hab- itat (or microhabitat) favored by different species, and unless numbers reach critical heights and com- petitive exclusion can be demonstrated, available space should never be at a premium. In the absence of quantitative data on plant distribution, or occur- rence of areas of different foliage height diversity, however, it must remain an open question whether availability of different grades (or degrees) of cover might turn out to be a limiting factor for some spe- cies, in the presence of others. 1978 NEL— KALAHARI MAMMAL COMMUNITIES 127 Table 5. — Habitat niche overlap of five rodent genera based on occurrence at 37 microhabitats {trap stations) in the Nossob River, Kalahari Gemsbok National Park. Values above the hold face numbers are July 1972 data*, below are January 1972 data. Single figures refer to overall overlap: columns — top figure is overlap on low dune, then overlap on riverbed, plateau, and high dune (bottom figure) 0.00 denotes that species occur together in same habitat, but do not overlap at trapping points: — species do no co-occur in same habitat. Taxa Gerbillurus Desmodillus Patera Rhabdomys Mus .45 .24 .55 .47 Gerbillurus 1.00 0.44 .51 .49 .75 0.33 .47 .57 0.40 .59 .60 — .45 .40 .53 .12 .49 .57 Desmodillus 0.23 .00 1.00 .73 0.38 .36 .52 0.39 .32 .00 — .30 .33 .15 .00 .63 .39 Patera 0.12 — 0.00 — 1.00 0.33 .00 0.30 .00 .49 .00 .00 .45 Rhabdomys 0.10 .14 O.IO .27 0.00 ~ 1.00 0.31 .00 .45 Mus 0.04 .18 0.00 — 0.00 ~ 0.00 — 1.00 — — — — .14 .58 * Other overlap values for July 1972 are RhabdomyslElephantulus = 0.30 ; Dendromus/Gerbillurus on plateau only — .22, or 0.11 overall; Malacothri.xIGerbillurus on plateau only — .30, or 0.14 overall. -26 Food The food taken by the different species is given in Table 7; based on the three food-type categories, the standardized food-niche breadths are given in Table 8, and food-niche overlap in Table 9. As only three food categories were possible, data for food niche breadth and overlap should be viewed with caution, as these only reflect use among categories, and not within categories, which would perhaps be far more meaningful. These reservations of course apply equally to the figures for food-niche overlap. Nevertheless, as can be seen from Table 8, the niche breadth for most species increased during summer, when more insects were generally taken (Table 7). The very high niche overlap values (Table 9) between the majority of species probably reflects an artifact due to the tow number of food categories and is thus probably more apparent than real. As the overlap values are lower in summer than winter, however, it could suggest that competition for food might become a factor in the latter season. Time Activity Limited data on activity cycles are available for only three species (G. paeha, D. aiincnlans, and T. hrantsii). Based on this, temporal niche breadths are as follows: G. paeha, B = 0.89 (14 h activity); D. aiiricularis, B = 0.95 ( 14 h activity); and T. hrantsii, B = 0.94 (13 h activity). Temporal niche overlap values are GerhillitmsIDesmodillns = 0.94; Gerhilinrus/Tatera — 0.87; cindDesmodilliisITatera = 0.95. As these activity cycles were obtained in the laboratory it is arguable how well they relate to field conditions. CONCLUSIONS In general it was found that the number of species of animals, and that the leaf-eating species, as one present closely followed the trend in total number would expect, reacted more quickly to increased 128 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 6. — Habitat niche overlap values of six rodent genera based on occurrence at 37 microhabitats, in the Nossob River, Kalahari Gemsbok National Park, Values above the boldface numbers are for July 1974 data, below are July 1976* data. For other explanations, see legend to Table 5. Taxa Gerbillur 14 S Tatera Rhabdomys Mus Zelotomys Dendromus .64 .55 .00 .65 .37 Gerbillurus 1.00 0.50 — 0.35 — 0.22 — 0.45 — 0-29 ~ — .64 — — .00 .38 .74 .48 — .36 .00 .10 .00 .39 .22 Tatera 0.00 1.00 0.06 — 0.06 — 0.37 — 0.24 ~ .00 .00 .32 — .71 .25 .28 .16 .28 .16 Rhabdomvs 0.23 .66 0.20 — 1.00 0.27 .38 0.16 — 015 T, — — .41 — .54 .20 .20 .43 — .00 .50 .00 .25 .00 .00 .00 .55 Mus 0.34 0.00 0.47 .73 1.00 0.00 0.17 .45 .00 .82 — .45 nn Zelotomys 0.00 — 0.32 — 0.05 — 0.00 — 1.00 0.00 .00 1.0 .20 .00 — .50 .00 .40 .50 Dendromus 0.33 0.00 0.27 0.20 — 1.00 — — — — * In July 1976 ElephantulusIGerbillurus overlap on low dune by 0.32; ElephantulusIRhabdomys by 0.34 on low and 0.72 on high dune; and ElephantulusIMus by 0.63 on the high dune. rainfall than did the seedeaters. The discrepancy between the numbers of seedeaters in July 1972 and July 1974, in relation to rainfall, can be explained, albeit subjectively in the absence of quantitative data, by a marked increase in cover during 1974, which adversely affected the gerbils, in particular G. paeba and D. aitncidaris. The high numbers of small mammals during July 1972, compared to the rainfall, are more difficult to explain. However, two factors may account for this seeming anomaly — cloud cover, days of rainfall per month, tempera- ture, and amount of precipitation per shower vary Table 7. — Percentage of diet composed of white plant material, green plant material, or insects, of eight genera of mammals in the Kalahari Gemsbok National Park. Winter Summer Taxa No. seasons involved N % white % green % insects No. seasons involved N % white % green % insects Gerbillurus 2 II 76.4 23.6 0.0 3 43 39.3 43.8 16.8 Desmodillus 1 5 71.0 29.0 0.0 2 4 15.0 40.0 45.0 Tatera 3 12 67.1 27.5 5.4 3 22 34.5 59.9 5.6 Mus 2 10 81.0 2.0 17.0 1 1 100.0 0.0 0.0 Rhabdomys 2 62 31.0 58.0 11.0 2 5 18.0 72.0 10.0 Zelotomys 2 1 100% flesh 1 1 75.0 0.0 25.0 Dendromus 1 2 95.0 5.0 0.0 no data Elephantulus 1 3 8.3 0.0 91.6 no data 1978 NEL— KALAHARI MAMMAL COMMUNITIES 129 1,00 0.9 0.8 0,7 0.6 .c ■a CD 0) .Q 0, 5 90%) from the riverbed, they form only 20 to 45% of the volume of scats from the plateau (Nossob), probably reflecting local differences in availability of these prey items due to edaphic factors. Like- wise, in the "wet” season, although termites occur in most scats, their volumetric contribution falls, but the contribution by Grewia flava, a small berry, ris- es to >50%. In January 1977, marked differences were also found between the composition of the scats from the male and female of the pair inten- sively followed; the female’s scats were composed of 40 to 50% Grewia, 30 to 46% termites, and some scorpions, whereas that of the male consisted of 85% termites and about 2% Grewia. This discrep- ancy probably results from the difference in feeding range of the sexes (see below) during this period. The one and admittedly small sample of scats from the duneveld surrounding the river seems to indi- 132 1978 NEL— FOOD HABITS OF OTOCYON 133 Table 1. — Food items taken by hat-eared foxes in the southern Kalahari. Termites Ants Beetles Sunspiders Grasshoppers and crickets Moths and cutting worms ( = larval forms) Order Isoptera; Family Hodotermitidae; Hodotennes mossamhiciis Order Hymenoptera; Subfamily Campinotinae; Campenotiis fiilvopilosus and Campenotus spp. Subfamily Mirmicinae Order Coleoptera Family Carabidae Family Tenebrionidae Family Scarabaeidae Family Melolonthidae Order Solifugae; Family Solifugidae; Solpiiga spp. Order Orthoptera Family Arcridiidae Family Gryllidae Order Lepidoptera Family Noctuidae Millipedes Order Myriapoda Scorpions Order Scorpionida cate a somewhat different diet, again probably re- flecting between-habitat differences in food avail- ability. In contrast to the findings of Bothma (1971) and Smithers (1971) very few vertebrate remains were found in the feces. It would appear therefore that termites form an important part of the diet of bat- eared foxes in the southwestern Kalahari, being taken throughout the year. In winter, however, ants form the bulk of the foxes’ diet, whereas Grewia berries, available in the "wet” season, are partic- ularly favored. Indeed the female, intensively ob- Table 2. — Occurrence of food items in bat-eared fox feces from the southern Kalahari. The number of samples is indicated below each date. Figures in parentheses are percentages. Food items Winter- Riverbed Nossob July 1976 98 -Dry Dunes Nossob July 1976 7 Summer — Dry Plateau Riverbed Cubitje Nossob Quap December December 1976 1976 94 54 Riverbed Nossob Early January 1977 54 Summer — Riverbed Nossob Middle January 1977 30 ‘•Wet" Plateau Nossob Middle January 1976 32 Plateau Cubitje Quap Early January 1977 13 Invertebrate Termites 76 (77.6) 7 (100) 94 (100) 54 (100) 54 ( 100) 29 (96.7) 32 (100) 13 (100) Ants 85 (86.7) 51 (54.3) 49 (90.7) 43 (79.6) 14 (46.7) 10 (31.3) 13 (100) Beetles 63 (64.3) 7 (100) 84 (89.4) 54 (100) 52 (96.3) 19 (63.3) 23 (76.7) 7 (53.9) Beetle larvae 2 (2.0) 21 (22.3) 53 (98.2) 27 (50.0) 16 (53.3) 15 (46.9) 8 (61.5) Sunspiders 1 (1.0) 7 (100) 8 (8.5) 6(11.1) 6(11.1) 3 (10.0) 5 (15.6) 6 (46.2) Grasshoppers 33 (33.7) 7 (100) 5 (5.3) 3 (5.6) 3 (5.6) 4 (13.3) 2 (6.3) Moths/worms 17 (17.4) 7 (7.5) 13 (24.1) 1 (3.1) Scorpions 1 (1.9) 2 (6.7) 1 (7.7) Millipedes 1 (1.9) 2 (6.7) 2 (15.4) Vertebrate Hair 26 (26.5) 7 (100) 14 (14.9) 3 (5.6) 1 (1.9) 5 (16.7) 2 (6.3) Bones 9 (9.2) 7 (100) 7 (7.5) Feathers 4 (4.1) 5 (5.3) Vegetable Grass 90 (91.8) 7 (100) 93 (98.9) 54 (100) 54 (100) 26 (86.7) 32 (100) 13 (100) Seeds 74 (75.5) 9(9.6) 8 (14.8) 4 (13.3) Grewia flava 23 (42.6) 18 (60.0) 11 (84.6) 134 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 served, would (in January) head early each evening straight for some Grewia bushes, browse for a while, and then continue foraging. Foraging Time, Speed and Area Activity (mostly spent on foraging and feeding) varies through the year and is perhaps related to thermoregulation. In winter, bat-eared fox groups in the riverbed are active by day, from approxi- mately 0600 h up to 2000 h with a peak in foraging from 1200 h to about 1700 h. All groups flushed at night were lying down and presumably sleeping. In midsummer (December-January) on the other hand, the activity cycle was reversed, with individ- uals active from about 2000 h to about 0830 h, and slightly later if the day was overcast. Occasionally individuals were seen active as late as 1045 h, but this is exceptional. Especially in the late afternoons (1700 h) individuals would occasionally venture from dens, forage desultorily with evident signs of overheating for periods up to 15 min, and then re- turn either to shade or the den itself. By mid-Feb- ruary foraging starts at 1600 h, and continues until about 0900 h, so it would seem that seasons impose a gradual shift in the activity cycle one way or the other. The difference in the time spent foraging by a male and female with cubs, during December and January, was marked. The female would suckle the young before commencing foraging, and would then forage constantly for about 9.5 to 10 h before re- turning to the den to suckle the young; during this time away from the den she might spend about 15 to 30 min resting or grooming. After suckling the young in the morning she would forage for another 1 to 2 h, usually close to the den. The male, by contrast, forages for 3 to 4 h, and always close to the den; when not foraging, he lies at the den en- trance. When the young first start foraging the male would forage alone for an hour or so, from 2000 h onwards, return to the den and accompany the cubs for 1 or 2 foraging trips, with a total time of around 3 h (data for early January). It seems probable therefore that as the cubs grow the time spent for- aging by the male would gradually increase as the cubs increase their own foraging and rely less on suckling. Although data is available from only three dens it seems probable that this trait is common to bat-eared foxes in this region — the male guards the young and initiates them into foraging. Apart from suckling the young, all the female's time is taken up foraging, imposed by the rigors of nursing and the necessity to obtain enough nourishment for her- self. No regurgitation of food for the pair-mate at the den or the cubs was ever witnessed, nor was any food caching observed. Suckling time also seems to decrease as the cubs grow older — a pair of very small cubs suckled for 7 min at Grootkolk on 3 December 1976; at the spe- cific den we observed constantly near Nossob camp suckling continued for 7 min on 12 December 1976, 4 min 24.5 s on 4 January 1977, and for 3 min 50 s on 6 January 1977. Whether suckling in the den by day occurs is unknown; certainly when the cubs are very small this is the rule. As they grow suckling, at least in the late afternoon and early morning, takes place at the entrance to the den with the fe- male standing up, hind legs slightly splayed, while the cubs more or less hang from the nipples as they suckle. Suckling from at least two different nipples by an individual cub during a single feeding bout is common. Suckling, with the female lying down, was never observed; if the cubs were still very small and tried to suckle while the female was recumbent, she would get up and disappear down the den, fol- lowed by the cubs. Limited data suggest that the foraging range is to some extent dependent on group size; a group of 10 individuals followed for 3 days in February ranged over 1.5 to 2 square km, whereas the female closely observed in December/January had a foraging range of <1.0 square km, with that of her mate consid- erably less. Foraging areas tend to shift slightly over time, and those of neighbouring groups over- lap widely. No defense or marking of foraging areas was ever seen; by contrast, up to 15 individuals (from four groups) have been seen foraging close together in an area of less than 0.5 square km. For- aging "speed” varies greatly, obviously depending on food availability; during February 1976, for a group of 10, this varied between 0.5 to 1.2 km/h. The female already referred to above foraged about 0.5 to 2 km/h, and usually moved about 12 km per night during the course of her foraging. Feeding Boats Observations indicated (see below) that when ac- tually feeding the position of the ears changes, from being directed forward to being pulled back. The time-interval between the flipping back-flipping for- ward again was taken as the duration of a particular feeding bout. It was often not possible to see an animal actually chewing, especially at night or when 1978 N EL— FOOD HABITS OF OTOCYON 135 Table 3. — Number and duration of feeding bouts of bat-eared foxes in the southern Kalahari. Season Time N No. bouts/ 15 min Mean duration (s) Range (s) Prey Winter “Dry” 1500 h to 1700 h 7 81.5 7.6 1-46.0 Termites Summer “Dry” 2000 h to 0415 h 9 50.6 7.1 1-58.1 Mainly termites Summer “Wet” 2000 h to 0835 h 13 40.0 3.5 1-22.9 7 the mouth was obscured by vegetation, but the ears were always visible. Table 3 summarizes the limited data on number, duration, and range of duration of feeding bouts by bat-eared foxes at various periods and times of day. The only continuous recording of feeding in a particular animal was in December 1976 on a known female; for that particular night (6-7 December 1976) it appeared that after the initial in- tensive feeding a somewhat more slack period in feeding happens near midnight; during this time ex- tensive self-grooming took place. Thereafter both the number and duration of feeding bouts increased, but not to the previous level. However, a great deal of variation in number and duration of feeding bouts at a particular time period, but on different days, occurs in the same animal. Whether the decline in feeding bouts is related to increasingly less activity of prey items as the night progresses, in summer at any rate, is unknown. The known female did however show changes in the number of digging bouts/ 15 min through the night as follows: from 2000 to 2200 h on average 4.3 bouts/15 min were performed; from 2200 to 0500 h none at all; thereafter the incidence of digging in- creased sharply up to about 9 bouts/ 15 min after 0800 h. The male tended to have more digging bouts, perhaps necessitated by his small foraging range during the time cubs were at their den. Foraging Behavior When leaving the den to forage, bat-eared foxes would do so without any preliminary scouting around for possible predators. Before setting off as a group, some members may indulge in bouts of allogrooming, and this was usually the case be- tween the male and three cubs referred to above. Large groups tend to split up into smaller groups (often pairs) during the course of their foraging, and such subgroups may be separated by up to 200 m, although they move in the same general direction. Whether the composition of these subgroups re- mains constant is not known. Small family groups (two to five individuals) normally stay within five to 10 m of each other, often feeding close (<1 m) together. The initial part of a foraging route can remain constant for a few days at a time, and this portion usually contains several deposits of feces added to every day. This may relate to wind direction; move- ment from the den is normally first downwind, and then for long distances (up to 1 km) crosswind. Dur- ing foraging, movements tend to be erratic; straight- line movements between feeding stations of up to 50 m or more are interspersed with S-shaped runs, the latter especially evident when moving up- or downwind. When feeding stations are close togeth- er, movement is slow with frequent casting about, with often just the head and forequarters being moved as prey is taken first in front, then on the left or right side of the animal; complete turns can then occur. The fox’s head is then seldom lifted, the nose remaining close to the ground surface. Movement, although in a general direction, tends to be erratic with frequent twists and turns. When foraging, the head is held low and the ears pointed forward at about 45°. If the fox had been walking fast or running, the food source may be overshot, and, as the animal turns, the head is low- ered abruptly. Otherwise the head is simply low- ered even further. In all cases, when the prey is captured, the ears flip back from their front-facing position, and immediately as chewing or swallowing starts the ears point forward again. Sometimes the head is lifted when chewing or swallowing, but this is not the rule. Usually, after a few paces with the head held close to the ground and ears cocked for- ward, the head is lifted, back straightens, and the animal looks sideways. On occasion, when moving at a run from one feeding station to the next, the impression is gained that the new prey source was first either heard or smelled, and then reacted to, or else it could be a known source of high prey (termites in all cases investigated) density. Certainly the movement 136 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 seems purposeful, and serves to attract partner(s) to the same spot. Perhaps due to the high density of termites in the study area little digging, or signs of it, was seen. Although bat-eared foxes are said to take birds, I observed no interest by foxes in birds flushed or alighting near them. On the other hand, grasshop- pers or even moths are actively pursued, and foxes jump into the air to catch them. The “pounce” common to other canids, for example black-backed jackal, was seen only once in a pair foraging around a grass-covered fallen tree stump. Such patches of habitat normally contain numbers of striped mice, Rhabdomys pumilio, and these could have been the quarry of the foraging foxes. A more attenuated form of the pounce is seen when sunspiders are chased. Little strife was ever noticed between pair or group members during foraging. Cubs are seldom snapped at by adults when foraging together, and this only happens when a cub would try to take a morsel away from the adult. Only on one occasion was frequent agonistic behavior seen between members of a pair; the prey items could not be iden- tified with certainty (perhaps grasshoppers or bee- tles) and, while they were foraging with noses close together, one snapped at the other when the latter tried to take its prey from it. These interactions lasted about 1 to 2 s, with 12 occurring in 18 min. DISCUSSION The limited data available suggest that in the southwestern Kalahari bat-eared foxes subsist throughout the year almost exclusively on inverte- brate prey. Although differences between the diets of individual bat-eared foxes are common, and de- pend to some extent on the particular habitat (for example, riverbed, plateau) occupied, seasonal trends in the diet of foxes in a particular habitat can be discerned. This results from changes in avail- ability of prey items, and because the activity pe- riods of the foxes shift with the changing seasons. In common with other canids, and probably also because of the rather stable home ranges, food se- lection is opportunistic. However, certain items are certainly preferred, especially the ripe berries of Grewia flava, if these occur in the home range of a particular fox. The bat-eared fox differs from other canids however in that food caching is absent — per- haps not unexpected, due to the nature of their prey. The opportunistic nature of feeding would en- hance social tolerance, as pointed out by Kleiman and Eisenberg (1973). Certainly the home ranges of bat-eared fox pairs, or groups, overlap widely and no defense of ranges was ever witnessed. In fact, many groups may forage in close proximity, utiliz- ing the local abundance of particular prey items, such as termites. This is in contrast to the other two co-occurring canids, black-backed jackal, Canis mesomelas, and silver fox, Vulpes chanm, in which mutually exclusive home ranges are found, and whose diets consist mainly of vertebrates, especially rodents. The foraging range of these canids overlaps and the diet of the bat-eared fox in the study area perhaps reflects an adaptation to food-niche sepa- ration, in order to lessen possible competition for the same resource. The pair-bond in bat-eared foxes, common also to other canids (Kleiman and Eisenberg, 1973), is long lasting and the male plays a major role in guarding and rearing the young. The dispersed na- ture of the prey items, each with low individual nu- tritional value, forces the female, when suckling, to forage far and for long priods. By contrast, until the cubs are old enough to accompany first the male, and then both parents, the male stays in close prox- imity to the den and forages for short periods only. During the present study bat-eared fox density in the study area was high, perhaps reflecting higher insect density resulting from 3 years of abnormally high rainfall. Increased prey density would result in the female being able to find enough food to sustain suckling entire litters, and also that the male could find enough food close to the den. Conversely, in drought years both the female and male may have trouble finding enough nourishment, in the case of the female to suckle the entire litter, and in the case of the male to be able to remain close enough to the den for effective defense of the young. Increased mortality in cubs could in such cases result from either lack of food (milk), or being taken by pred- ators, which are common in the area, or a combi- nation of both. 1978 NEL— FOOD HABITS OF OTOCYON 137 ACKNOWFEDGMENTS This paper was written while I was a Senior Visitor to the Animal Ecology Research Group, University of Oxford, made possible by generous grants from the Council of Scientific and Industrial Research. I wish to thank Malcolm Coe, my host at Oxford, for many courtesies extended me during my stay, and to the University of Pretoria for providing funds supporting the fieldwork. Discussions with Hans Kruuk aided in the preparation of this paper, and his comments and those of Devra Kleiman on the draft manuscript are greatly appreciated. FITERATURE CITED Bothma, J. DU P. 1966. Notes on the stomach contents of cer- tain Carnivora (Mammalia) from the Kalahari Gemsbok Park. Koedoe, 9:37-39. . 1971. Food habits of some Carnivora (Mammalia) from southern Africa. Ann. Transvaal Mus., 27:15-26. Kleiman, D. G., and J. F. Eisenberg. 1973. Comparisons of canid and felid social systems from an evolutionary perspec- tive. Anim. Behav., 21:637-659. Smiihers, R. H. N. 1971. The mammals of Botswana. Mus. Mem., Nat. Mus. Rhodesia, 4:1-340. COEXISTENCE IN TRANSVAAL CARNIVORA I. L. RAUTENBACH Transvaal Museum, Box 413, Pretoria, Republic of South Africa 0001 J. A. J. NET Mammal Research Institute, Department of Zoology, University of Pretoria, Pretoria, Republic of South Africa 0002 ABSTRACT How coexisting carnivore species avoid interspecific compe- tition is examined by consideration of their more prominent physical and behavioral characteristics. An attempt is made to explain coexistence of the 33 Transvaal carnivore species. The behavioral characteristics, which are considered here in various combinations, are daily activity regimen, food preference, hab- itat preference, geographical distribution, and social structure. The mean species body weight as an indicator of the size of prey on which a carnivore exists is also incorporated. Eighty-two % of the carnivores are shown to form a trend ranging from a noc- turnal/solitary mode of life to an entirely diurnal/gregarious ex- istence. INTRODUCTION Some two decades ago this paper might well have been titled “Niche occupation by Transvaal carni- vores.” The concept that each species fulfills a unique functional role in a specific place dates back to Grinnell (1924) and Elton (1927), and has served a useful function in subjectively describing the niche of an animal. However, it never really ex- plained in detail how each animal fills its particular niche. Modern study of the niche and niche theory flows from Hutchinson's (1957) landmark paper, and allows quantification of the role each animal plays, from measurements of the amounts of var- ious resources (axes in a hypervolume) utilized. Prior to this. Cause’s (1934) experiments led to the idea that competition serves to separate species, and therefore the niches they fill. Competition through evolutionary time therefore led to separa- tion of resource utilization in coexisting species, and the niches species occupy are therefore as much an outcome of evolution as, for example, their physical characteristics. On the other hand, al- though the physical characteristics of a particular species may be fairly constant over much of its dis- tributional range, the exact niche it occupies (not in the descriptive Eltonian sense but in the analyt- ical Hutchinsonian one) usually varies, depending on the habitat it occupies and the nature of other species in the community. The mammal fauna of the Transvaal has been in- tensively surveyed over the last five years (Rauten- bach, in preparation). This Province possesses a particularly rich mammal fauna, consisting of 175 species of which 33 are carnivores. Adaptation and radiation has led to different parts of resources being utilized, especially food and activity periods by different members of this assemblage of carni- vores. It is of interest to note how resources are shared and competition lessened, and coexistence enhanced. Rather than work out niche occupation by the various carnivores, which to be meaningful would involve quantifying resource utilization in various axes by carnivores in specific communities, the ap- proach here taken is to look at various attributes of co-occurring species, and then to see where and if competition may come into force. This is done by considering in combination average species mass, basic food preference, daily activity regimen, hab- itat selection, distribution patterns, and specific so- cial characteristics. Trends in adaptations are also considered, especially the advantage of differential body size in coexisting carnivores preying on the same food types, as variation in body size could affect prey size taken (Rosenzweig, 1966). It was also necessary to categorize behavior, in the full realization that the behavioral scope of each species may well be wider than the particular category to which it is designated. 138 1978 RAUTENBACH AND NEL— TRANSVAAL CARNIVORES 139 Table 1. — The 33 species of carnivores occurring in the Transvaal. Average body weight expressed in kg, the log. value of the mean body weight in grams, as well as the daily activity, social structure, and basic feeding categories to which each species is assigned, are indicated. See text for further explanations. I = Insectivorous, P = Predatory, O = Omnivorous, and S = Scavenging. Species Average weight (kg) N Log. weight (g) Daily activity regimen Social structure Basic feeding adaptation Otocyon megalotis 3.4 (7) 3.53 iii 3 I Lycaon pictus 22.0 (12) 4.34 V 5 P Vulpes chama 2.9 (22) 3.46 i 1 P Canis adustus 10.0 (5) 4.00 ii 1 0 Cards mesomelas 7.8 (48) 3.89 ii 2 O Aonyx capensis 12.1 (4) 4.08 iii 3 I Lutra maculicollis 4.5 (1) 3.65 ii 3 1 Mellivora capensis 8.9 (5) 3.95 ii 2 I Poecilogale albinucha 0.4 (4) 2.60 ii 3 P Ictonyx striatus 1.1 (10) 3.04 i 2 I Viverra civetta 12.4 (5) 4.09 i 1 O Genetta genetta 1.9 (15) 3.28 i 2 P Genetta tigrina 1.9 (24) 3.28 i 2 P Suricata suricatta 0.7 (19) 2.85 V 5 I Paracyrnctus selousi 1.6 (39) 3.20 i 2 1 Cynictis penicillata 0.8 (20) 2.90 iv 3 I Herpestes ichneumon 3.1 (14) 3.49 V 3 P Herpestes sanguineus 0.5 (25) 2.70 V 1 P Rhynchogale melleri 2.8 (1) 3.45 ii I 0 Ichneumia alhicauda 3.6 (1) 3.56 i 2 p Atilax paludinosus 4.3 ■ (5) 3.63 i 1 I Mungos mungo 1.3 (7) 3.11 V 5 I Helogale parvula 0.2 (13) 2.30 V 5 1 Proteles cristatus 9.9 (14) 4.00 i 1 I Hyaena brunnea 36.1 (7) 4.56 ii 2 s Crocuta crocuta 69.7 (8) 4.84 ii 4 p Acinonyx jubatus 35.1 (3) 4.55 iv 2 p Panthera pardus 41.7 (4) 4.62 ii 1 p Panthera leo 204.1 (4) 5.31 ii 4 p Felis nigripes 1.5 (8) 3.18 ii 1 p Felis serval 9.6 (5) 3.98 i 1 p Felis caracal 10.5 (10) 4.02 ii 1 p Felis libyca 4.7 (58) 3.67 ii 1 p METHODS Table 1 lists the 33 carnivore species occurring within the Transvaal, with average weight, expressed in kg of both sexes combined, indicated for each species. Weight data are based on Transvaal Museum records, supplemented by relevant infor- mation from Smithers (1971). Samples sizes (N) are indicated. The logarithmic values for the means of species weights as ex- pressed in g were calculated and are also given. Based upon personal observations and unpublished data (Rau- tenbach, in preparation; Nel, in preparation), as well as published information (see Smithers, 1971 ; Rowe-Rowe, 1977a, 1971 h), an integral numerical value has been assigned to the daily activity regime of each species. These range from exclusively nocturnal with a Roman numerical value of i, to exclusively diurnal with a numerical value of v (Table 1). Categories ii and iv denote noc- turnal species with some diurnal activity, and diurnal species with occasional nocturnal activities, respectively. Similarly, integral Arabic numerical values 1 through 5 have been designated for the solitary to gregarious behavioral range, ranked from very solitary with a numerical value of 1, through to very gregarious with a numerical value of 5. The various species were each assigned to one of these five social category values on the grounds of average social grouping, allowing for other situations mentioned in the lit- erature. The integral values assigned to these two behavioral patterns considered (activity and social groupings) are only arbitrary points spaced along a continuum, and each represents an average categorized value considered most typical for the species. Judg- ment herein was subjective. We could not use more than five sub- divisions with any accuracy, but in spite of this the resulting di- visions are found to be both convenient and meaningful. Hunting behavior is adapted to basic food preference. Diet and the mode of acquiring nourishment are other important aspects of the adaptive behavioral makeup of a species’ accompanying avoidance of competition. Also considered in this study, then, are 140 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 the four basic feeding methods or food types of carnivores, that is scavenging, omnivorous, insectivorous (denoting a diet of any invertebrate), and predacious. In assigning each species listed in Table 1 to a feeding category, it must be stressed that carnivores are opportunistic with regard to food items taken, especially un- der low interspecific competitive conditions. Only what is con- sidered to be the primary or optimum feeding trait of a species when under more intense interspecific competition is considered here. In Fig. 1 the integral values of the activity regimen and the social structure are plotted against each other for each species. Intra- specific social interrelationships are presented on the horizontal axis, and the activity regimen on the vertical axis. In Fig. 2 the four basic feeding categories are presented by vertical columns, each of which is divided into diurnal and nocturnal subsections. The nocturnal subsections are stippled. Each species was as- signed to its appropriate column with regard to its basic feeding behavior and characteristic daily activity cycle. Position against the vertical axis was assigned by the logarithmic value of the av- erage adult body mass, expressed in g. The principle is that clus- tering of species indicates possible interspecific competition, and vice versa. This is based on the correlation between the size of the predator and the size of the prey it can effectively handle, or usually catches. It has been calculated that the maximum mass of prey that can be handled with efficiency by an individual true predator is 1 .5 times that of the predator itself. Group cooperation accounts for a higher ratio between the individual predator and the prey. It conversely follows that a big carnivore could not ex- clusively hunt very small prey because the energy gain herein would not warrant the investment in such an energy expenditure. RESULTS AND DISCUSSION Those species falling within the limits of behav- ioral values li, lii, 2i, and 2ii in Fig. 1, are all noc- turnal and solitary, and represent the majority (58%) of the Transvaal Carnivora. The lines in Fig. 1 con- nect the upper values for both variables of this noc- turnal/solitary block, with the upper values of the very gregarious and exclusively diurnal group (value 5v). All species falling between these two lines are considered to represent a trend from a solitary and a nocturnal existence to an entirely gregarious and diurnal mode of life. No less than 82% of all carni- vores in the Transvaal follow this trend. L. maciili- cotlis, P. albinucha, O. megalotis, and especially A. capensis are behaviorally intermediate between the two extremes within this trend. It is within this trend that interspecific competition is potentially the high- est, as will be elaborated below. Three of the four species at the extreme diurnal/gregarious end of the trend (Fig. 1) are small insectivores and thus poten- tially in direct competition. Eighteen % of the carnivore species under consid- eration do not conform to this trend, and have adopt- ed a strategy, which seems to minimize possible competition. However, where four species have ra- diated toward a diurnal/solitary mode of life {H. san- guineus very successfully), only two species radiat- ed a short distance toward a nocturnal/gregarious existence. There are no extremely nocturnal/gregarious spe- cies (value 5i), although the lion and the spotted hyena are approaching this condition. A possible explanation for the poor radiation toward an ex- treme nocturnal/gregarious behavioral range could be the difficulty of maintaining group structure in the dark. Smaller gregarious species are mostly in- sectivorous and diurnal and finding food in the dark may also present difficulties, apart from the diffi- culty in locating predators in time. Schaller and Lowther (1969) consider the lion, in contrast to the wild dog, as incompletely adapted to a social life because lions frequently quarrel over the proceeds of a hunt. If their interpretation is correct, the true position of the lion on the graph in Fig. 1 may be more toward the left, and consequently even closer to the general trend. C. crocuta is basically a nocturnal animal, but may also be active during the day. According to Kruuk (1966, 1972) the species tends to scavenge by day, and to become efficient pack hunters and killers by night. The spotted hyena has a complex matriarchal social system, with the females physi- cally bigger than the males and dominating them. C. crocuta thus has radiated successfully some dis- tance away from the trend, toward a nocturnal/gre- garious existence. Otocyon, although regarded by most as a noctur- nal species, has a diurnal mode of life in undis- turbed areas during winter. In settled areas, how- ever, it becomes exclusively nocturnal. In discussing the eastward range extension of the species in the Transvaal, Pienaar (1970) mentions that it is exclu- sively nocturnal in the Kruger National Park, and ascribes this to a form of protective behavior of colonists in a new territory. In the Transvaal as a whole the species is almost entirely nocturnal, but on the other hand it occurs for the most part in this Province only in settled areas. Studies elsewhere (Nel, 1978) show that activity is perhaps correlated to the need to thermoregulate efficiently. Most ob- servations on the bat-eared fox in the Transvaal are 1978 RAUTENBACH AND NEL— TRANSVAAL CARNIVORES 141 Fig. 1. — Graphical presentation of species separation by plotting the categoric values assigned to intraspecific social relations against the categoric values of daily activity cycles. See text for further explanations. of solitary or small groups of animals, but again this would depend on the time of year of observations (Nel, 1978). This species is thus plotted in the po- sition 3iii within the trend, although it could be ar- gued that the Transvaal population should be plot- ted together with the lion just outside the trend. It thus would appear that a vacuum exists at the nocturnal/gregarious end of the behavioral range, but that carnivores in the Transvaal do not utilize it, for reasons at present not fully understood. H. sanguineus is the most predacious of the Her- pestinae in the Transvaal, being an efficient killer of vertebrate prey. It is furthermore solitary and diurnal, in contrast to the general tendency for the more predatory small carnivores to be solitary and nocturnal (see Ewer, 1973:277). This seeming anomaly could result from an adaptive radiation to utilizing resources (especially habitat and food) with a low utilization pressure. H. ichneumon and C. penicillata are only partly social species. When hunting for food both species are solitary and in this respect they are reminiscent of H. sanguineus . H. ichneumon is predatory, whereas C. penicillata is insectivorous. The distri- butional ranges of these two species furthermore do not overlap at all. C. pencillata is unique in the sense that when actively seeking food it is a solitary insectivore, in contrast to the other diurnal insec- tivores, which are social species. The cheetah displays the three basic felid hunting techniques — stalking, utilization of the forepaws to fell its prey, and an oriented neck or choking throat bite according to the size of the prey. However, the cheetah atypically (for a felid) outruns its quarry LOG.,0 BODY WEIGHT (gram) 142 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 BASIC FEEDING BEHAVIOR Fig. 2. — Graphical presentation of niche occupation. The four basic feeding categories are presented as vertical columns, each sub- divided by a stippled column denoting nocturnal activity and an unstippled column denoting daylight activity. Species are assigned to their appropriate columns and are vertically spaced against the X-axis representing the log. value of the mean body weight in grams. and possesses distinctive anatomical adaptations for this particular way of hunting, which can best be performed in daylight. There appears to be very little need for group participation. The cheetah thus clearly acquired behavioral and physical adapta- tions to enable it to radiate adaptively into a less competitive area. Of the four carnivores above the trend illustrated in Fig. 1, the cheetah utilizes a dif- 1978 RAUTENBACH AND NEL— TRANSVAAL CARNIVORES 143 ferent trophic level as a result of its larger size. Yet the survival of the cheetah is threatened. Perhaps the reason for its precarious conservation status in the Transvaal should be sought in its low ranking position in the predator hierarchy. Cheetahs are often robbed of their prey by lions, leopards, and hyenas, and are even preyed upon by these more powerful predators (Schaller and Lowther, 1969; Pienaar, 1969). A strong bias towards the insectivorous and pre- dacious modes of life is evident (Pig. 2). The ratio of species between the four feeding classes is 1:4:12:16. Forty-eight % of Transvaal carnivores are predacious, which is considered to be the pri- mary feeding trait of the Order. The remaining 52% have radiated away from a true predacious exis- tence toward utilization of other protein resources, and have behaviorally adapted themselves to pro- curing them. Furthermore, no less than 75% of all species are predominantly nocturnal. The mean weight of the species in the omnivorous category is 8.25 kg, that of the insectivores is 4.07 kg, and the predators 25.82 kg. We agree with Skinner ( 1976) that H. hrrmnea is basically a scavenger. This is further substantiated by the special dental and cranial adaptations ac- quired to cope with a scavenging way of life. Such a life style is for several reasons an uncertain ex- istence, with chance playing no minor role. This is reflected in the single species represented in this category as well as the fact that it is primarily sol- itary, presumably in order to avoid excessive intra- specific competition for limited resources. Consid- ering the apparent hardships of a scavenging life style, a lower mean weight may be an appropriate manner of reducing the energy requirements of the species. However, all indications are that the brown hyena is in all aspects primarily adapted towards capitalizing on the proceeds of the hunting endeav- ors of the larger predators. An omnivorous life style is seen as the most op- portunisitic of all, and can include as food items vertebrates (which are actively hunted), insects, carrion, and vegetable matter, especially fruit. The concept of a smaller body size as a means of re- ducing the energy requirements of the species with such a precarious existence can be illustrated by the fact that the mean species weight in the omni- vorous category is only 8.3 kg, as opposed to the mean of 25.8 kg of the predatory category and mean of 36. 1 kg of H. bntnnea in the scavenging category. R. melleri is much smaller than the other three species in the omnivorous category, and from this it is concluded that overlap in feeding interests is small. V. civetta is ecologically separated from C. mesomelas and C. adustus. The latter two species are inhabitants of the open plains and avoid forests. Like Smithers’ (1971) findings, our own observa- tions on V. civetta indicate a close association with riverine and subriverine woodlands. C. adustus is limited in range to the eastern Transvaal lowveld and a small area north of Pretoria. C. mesomelas ranges throughout the Transvaal. The two species are thus partly sympatric, and as is suggested in Fig. 1 may be in conflict here. Although Shortridge (1934) and Smithers (1971) speculate that C. me- somelas is being gradually replaced by C. adustus in the overlapping zone, this could not be demon- strated in the Transvaal. According to Pienaar (1963) C. mesomelas is numerically the more suc- cessful species in the Kruger National Park. C. adustus is however slightly larger than C. meso- melas, and indications are that it relies less on veg- etable matter as a food source. The insectivorous feeding category has the low- est mean body weight. This is considered as a sig- nificant adaptation to the small size of the individual prey, and the quantity and effort required on the part of the carnivore to fulfill its energy require- ments. There are three clusters in this category that warrant closer scrutiny (see Fig. 1). A. capensis is the biggest member of the insec- tivorous group. It is an aquatic mammal subsisting almost entirely on crabs (Rowe-Rowe, \911a, 19776). The terrestrial P. cristatus is the biggest carnivore living on Insecta, namely almost exclu- sively termites (especially Triner vi tenues). It is not well equipped to dig out subterranean termites. M. capensis is also terrestrial and overlaps in range with the aardwolf. It however hunts invertebrates bigger than termites, especially spiders. The honey- badger is particularly well adapted to procuring this subterranean prey. L. maculicollis, A. paludinosis, and O. megalotis also form a cluster in Fig. 1. The latter species is however a terrestrial inhabitant of the open plains, whereas the former two are to varying degrees sem- iaquatic. The spotted-necked otter and the marsh mongoose appear to be in conflict as they both rely heavily on crustaceans in their respective diets, and furthermore overlap in geographic range and habitat requirements. A. paludinosus is however a more versatile animal because it is more mobile on land. It wanders greater distances away from water and 144 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 utilizes a wider spectrum of food resources. It is furthermore believed to hunt for aquatic prey only in the shallows, as opposed to L. maculicollis . C. penicillata and 5. suricatta also overlap in dis- tributional range. Where the suricate is very gre- garious and almost exclusively insectivorous, the yellow mongoose is a solitary hunter, which takes vertebrate prey as well as invertebrates. The predatory category is the true domain of the Felidae, and no felid has radiated away from it. They are specialist killers, the only group capable of handling prey larger than themselves singlehand- ed. This is achieved mostly by means of a lethal well-directed single neckbite, or derivations there- of. Felidae are, in general, also expert stalkers. Of the nonfelids in this feeding category, the mus- telidP. alhiniicha is an exception, in that it behaves very similarly to the Felidae with regard to killing efficiency and the size of prey that it can handle. The remainder, that is the viverrids, canids, and Crocuta, all belong conditionally to the predatory category. C. crocuta and L. pictus rely on group cooperation to kill, and are relatively inefficient predators when alone. The remainder of the non- felids rely on the other food sources already dis- cussed, and when they kill, it is mostly prey much smaller than themselves (excluding domestic stock). Very little is known of the serval, but from the information that is available, it would appear not to be in conflict with the caracal, as is indicated in Fig. A behavioral trend is indicated in carnivores, which ranges from a direct correlation between a nocturnal/solitary mode of life, to an entirely diur- nal/gregarious existence. We conclude that 82% of the Transvaal carnivores fall within this trend. Pre- sumed adaptive radiation away from this trend is restricted to six species. Carnivores are considered incapable of adapting to an entirely nocturnal/gre- garious life style. In the majority of coexisting species interspecific 1. The serval appears to be restricted to areas with permanent surface water and its associated forests, and preys mostly on rodents. The caracal, on the other hand, does not prefer forests and is a true predator of prey more equal in size to itself. The geographic ranges of V. chama and /. albi- cauda overlap only peripherally in the Transvaal. F. libyca, on the other hand, is widely distributed and overlaps with the ranges of both the former species. F. libyca and V. chama are separated in size to the extent that they presumably avoid con- flict by means of differential choice in prey size. I. cdbicauda is restricted to riverine forests, whereas F. libyca has a wide habitat tolerance. The latter species therefore appears to be a universalist, the former a specialist extremely well adapted to its particular narrow niche. In the zone of contact be- tween these two species, it can be postulated that /. albicauda has the edge in a competitive situation. The two species of genets are partly sympatric. Our own experience agrees with that of Smithers ( 1 97 1 ) in that these two species are ecologically sep- arated. G. tigrina prefers a habitat close to water, whereas G. genetta exists away from it. The range of F. nigripes overlaps partially with that of G. ge- netta, and not at all with G. tigrina. However, so little is known about the general biology of the black-footed cat, that no suggestions can be offered as to how it avoids conflict with the small-spotted genet. competition is avoided, primarily through different food sources, differences in size of food items (cor- related to different body size of the carnivores), or differential use of habitat types. However, in the instances of the two jackal species, L. maculicollis and A. paludinosiis, as well as F. nigripes and G. genetta, at least partial interspecific competition is suspected. A more intimate knowledge of the gen- eral biology of these six species may in time show more subtle mechanisms of avoiding conflict. ACKNOWLEDGMENTS This paper is based on a long-term intensive mammal survey in the Transvaal, a project financed jointly by the S.A. Council for Scientific and Industrial Research and the Transvaal Mu- seum. Presentation of this paper at the Colloquium was in the case of Rautenbach made possible by funds from the Transvaal Museum and the Department of National Education. Travel ex- penses for Nel were provided by the University of Pretoria and the Mammal Research Institute of this University. We express our deep gratitude to these institutions for their assistance. The manuscript was typed by Mrs. E. du Plooy. 1978 RAUTENBACH AND NEL— TRANSVAAL CARNIVORES 145 LITERATURE CITED Elton, C. S. 1927. Animal ecology. Sidgwick & Jackson, Lon- don, XX -I- 209 pp. Ewer, R. F. 1973. The carnivores. Cornell Univ. Press, New York, XV -I- 494 pp. Cause, G. F. 1934. Experimental analysis of Vito Volterra’s mathematical theory of the struggle for existence. Science, 79:16-17. Grinnell, J. 1924. Geography and evolution. Ecology, 5:225- 229. Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harbor Symp. Quart. Biol., 22:415-427. Kruuk, H. 1966. Clan-system and feeding habits of spotted hyaenas (Crociita crocuta Erxl.). Nature, 209: 1257-1258. . 1972. The spotted hyaena: a study of predation and social behavior. Univ. Chicago Press, Chicago, xvi -I- 335 pp. Nel, J. a. J. 1978. Notes on the food and foraging behavior of the bat-eared fox, Otocyon megalotis, Bull. Carnegie Mus. Nat. Hist., 6:132-137. " Pienaar, U. de V. 1963. The large mammals of the Kruger Na- tional Park — their distribution and present day status. Koe- doe, 6:1-37. . 1969. Predator-prey relationships amongst the larger mammals of the Kruger National Park. Koedoe, 12: 108-176. . 1970. A note on the occurrence of bat-eared fox Otoc vou megalotis (Desmarest) in the Kruger National Park. Koedoe, 13:23-27. Rosenzweig, M. L. 1966. Community structure in sympatric Carnivora. J. Mamm., 47:602-612. Rowe-Rowe, D. T. 1977a. Prey capture and feeding behavior of South African otters. Lammergeyer, 23:13-21. . 1977b. Variations in the predatory behaviour of the claw- less otter. Lammergeyer, 23:22-27. ScHALLER, G. B., AND G. R. LowTHER. 1969. The relevance of carnivore behaviour to the study of early hominids. South- west. J. Anthropol., 25:307-341. Shortridge, G. C. 1934. The mammals of South West Africa. William Heineman, Ltd., London, l:xxv -i- 1-438; 2:ix -i- 439-779. Skinner, J. D. 1976. Ecology of the brown hyaena Hyaena briuinea in the Transvaal with a distribution map for southern Africa. S. African J. Sci., 72:262-269. Smtihers, R. H. N. 1971. The mammals of Botswana. Mus. Mem., Nat. Mus. Rhodesia, 4:1-340. SYSTEMATICS OF THE HYRACOIDEA; TOWARD A CLARIFICATION H. N. HOECK Max-Planck-Institut fiir Verhaltensphysiologie, Seewiesen, West Germany, and Serengeti Research Institute, Tanzania National Parks, Tanzania (present address: Charles Darwin Research Station, Casilla 58-39, Guayaquil, Ecuador) ABSTRACT Opinion is divided on whether the Order of the Hyracoidea Dendrohyrax. The anatomical and behavioral features here pre- contains three genera, P/-(HY7v;V/, //erew/!yra.v and Dt’/n/ro//yru,v, sented show that there is a roughly equal, definite distinction or whether Heterohyrax should be regarded as a subgenus of between all three, justifying a differentiation into three genera. INTRODUCTION The systematic relationships of the Hyracoidea are still open to many questions. The classification at the species level is not at all clear, and there are still divergent opinions about the number of genera in this Order. Hahn (1935, 1959), in his revision of the Hyra- coidea, distinguished three genera — the tree hyrax, Dendrohyrax, the bush hyrax, Heterohyrax, and the rock hyrax, Procavia. In more recent reviews, Bothma (1971) and Kingdon (1971) agree with Hahn, whereas Hayman (letter dated 29 October 1964 to C. R. S. Pitman, on file in British Museum [Natural History], London) and Roche (1972) main- tain that there are only two genera, Procavia and Dendrohydrax, Heterohyrax being a subgenus of the latter. Roche even thinks it possible that Den- drohyrax and Heterohyrax may be merely different species; the only basis for his assumption is the small extent of cranial and especially dental dis- tinction between these two. The molars are of brachydont structure in both, whereas Pracm’/V/ has hypsodont dentition. A new perspective may settle the controversy, and this paper therefore describes some anatomical features and behavioral aspects of Procavia john- stoni inatschiei Neumann 1900, Heterohyrax hrucei dieseneri Brauer 1917, and Dendrohyrax arboreus stuhhnanni (Matschie) 1892, as representatives for each genus. METHODS While studying the ecology and social behavior of the rock hyrax P. johnstoni and the bush hyrax H. hnictd (Hoeck 1975, 1977 (i, and in preparation) in the Serengeti National Park, Tan- zania, in 1971-1973 and 1975-1976, over 350 animals were trapped, examined, and basic body measurements were taken before release. Eight D. arboreus. living on fig trees {Ficus natalensis), in the Ngorongoro Crater floor were also trapped, measured, and ob- served for 10 nights. Total body length and anus-preputial opening distance were measured with the animal lying stretched on its back. The anus- preputial opening measurement of one adult Dendrohyrax vali- dus male from the West Kilimanjaro Forest was kindly supplied by Mr. P. Fox. RESULTS Anatomical Differences Aniis-prepntial opening. — This measurement al- lows a clear distinction between males of the three species (Table 1). H. hrucei males have twice the anus-preputial opening distance of P. johnstoni males, and over three times that of D. arboreus and D. validns. Table I includes, for comparative purposes, the body weight and length of adult males (over 16 months). H. brncei and D. arboreus have identical measurements, whereas P. Johnstoni is heavier and larger. Penis structure. — There is a striking difference of penis structure in D. arboreus, H. hrucei, and P. johnstoni (Eig. 1). D. arboreus males have a short, simply-built penis that is slightly curved. No evi- dent difference was observed in the external penis anatomy of D. arboreus and D. validus. The penis of H. hrucei is complex; on the penis end, and aris- ing within a cup-like glans penis, is a short, thin 146 1978 HOECK— SYSTEMATICS OF HYRACOIDEA 147 CD DENDRO- HETERO- PROCAVIA HYRAX HYRAX Fig. 1. — Diagram of external penis anatomy in adult males of Dendrohyrax arhoreus, D. validus, Heterohyrax hrucei, and Procuviu johnstoni, showing approximate size relationship. Top, penis as viewed ventrally; middle, viewed from above the supine animal; bottom, dorsal view. appendage, which has the penis opening. Fully erected, the little appendage also stiffens, the penis measures over 6 cm (Fig. 2). P. johnstoni have a short, simply-built penis with a slightly elliptical cross-section, the diameter increasing slightly to- ward the tip. More detailed features of the anatomy and histology of the penis of Heterohyrax and Pro- cavia are given by Glover and Sale (1968). Foot color. — The skin of the foot pads of P. john- stoni and H. hrucei is black, whereas in D. arhoreus it is pink. D. arhoreus ruwenzorii, however, living among rocks in the Ruwenzori Mountains, Uganda, have black skin on these pads (T. Struhsacker, per- sonal communication). Behavioral Differences Activity patterns and feeding behavior. — The long-term observations in the Serengeti showed that both P. johnstoni and H. hrucei are diurnal. Al- though feeding times are identical, feeding behavior differs. P. johnstoni is mainly a grazer; H . hrucei feeds almost exclusively on browse material. For further details see Hoeck ( 1975, \976a, 19776)- The brief observations in the Ngorongoro Crater clearly established D. arhoreus as nocturnal, browsing al- most exclusively on Ficus natalensis and Acacia alhida trees. Mating behavior. — The following differences were observed in the mating behavior of P. john- 148 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 Fig. 2. — Heterohyrax hrucei male and female just before copu- lation. Notice the length of the male’s nearly-erected penis. stoni and H. brucei, based on 14 and 21 observed copulations, respectively. In P. johnstoni, usually after an initial mating call, the male executes weaving head movements, the penis is erected, and the dorsal hairs raised. The female presses her rump against the male's flank or breast. After mounting, the male grasps the sides of the female vigorously with his forelegs, makes several thrusting motions, the last being a short jerk, and then jumps down. Copulation lasts only a few seconds (Hoeck, \916b). In H. brucei, the male approaches the female, giving a short shrill call almost inaudible to humans, and both perform a short “dance” during which the male smells the vagina. He then mounts, holding the female’s sides with his forelegs. With fully erected penis the male makes several thrusting mo- tions, while swinging the head from side to side and sometimes opening his mouth, probably calling. The penis is not introduced, but pressed against the vagina. The position being maintained, the penis slackens; after 20 to 30 seconds several renewed thrusts are made with fully erected penis, but still without introduction. After some 3 to 5 min the pen- is is completely introduced with a sudden violent jerk, whereupon the female jumps, bites, and chas- es the male (Hoeck, 1977c). Mating was not observed in D. arboreiis. The territorial call. — Adult males of H. brucei, D. arboreus, ?LnA P. johnstoni have very distinctive calls, as shown in the sonogram (Fig. 3). The call of H. brucei is shrill and long, lasting about 1.5 seconds. It is given repeatedly for up to 5 min. The calls of D. arboreus start with several crack- ing sounds, which are followed by a loud scream, repeated several times. The sonogram shows the transition between the cracking sounds and the scream. Several short cracking sounds follow im- mediately after each of the first few screams, whereas in the later part of the sequence the scream occurs alone. The call of P. johnstoni is a repetitious bark, be- coming longer and louder toward the end of the sequence, the last barks ending with gutteral noises. One of these last barks with the following gutteral can be seen in the sonogram. Calls are loud in all three species (audible for sev- eral hundred meters, depending on wind condi- tions), and a calling sequence may last up to 5 min. In H. brucei and D. arboreus only the adult males were observed to produce these calls, whereas in P. johnstoni on rare occasions adult females made similar calls. The territorial calls became more fre- quent in P. johnstoni and H. brucei toward the mat- Table 1. — Anus-preputial opening distance, body weight, and total body length for adult males. Lengths were measured with the animals lying stretched on their backs. Species Distance anus-preputial opening (cm) Body weight (kg) Body length (cm) N Mean SD N Mean SD N Mean SD Procavia johnstoni 41 3.5 0.91 66 2.95 0.72 43 48.68 4.47 Heterohyrax brucei 31 8.0 0.81 57 1.75 0.17 28 43.39 1.78 Dendrohyrax arboreus 2 1.7 2 1.62 2 43.85 Dendrohyrax validus 1 2.5 1978 HOECK— SYSTEMATICS OF HYRACOIDEA 149 kHz 7 - kHz 7 - Fig. 3. — Sonograms of male territorial calls for Heterohyrax brucei, Dendrohyrax arhoreus, and Procavia johnstoni. A calling se- quence may last up to 5 min. Representative sounds from sequences are given for all three species. ing season, and were usually made by territorial males (Hoeck, in preparation). The call of one ter- ritorial male may stimulate all others in the vicinity to call. Each animals’s calls are so distinctive that they allow individual recognition by the human ob- server. DISCUSSION The anatomical and behavioral differences pre- opening distance, and the territorial call, show sented, namely penis anatomy, anus-preputial brucei to be distinct from both D. arboreus and 73 ic 150 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 johnstoni. As a browser, H. hrucei resembles D. arhoreiis, whereas its activity pattern is identical with that of P. johnstoni. Lonneberg (1916) and Hahn ( 1959) suggested that the variation in hyrax molar structure could be a dietetic adaptation; as grass is a relatively coarse material, grazers should be expected to have hyp- sodont dentition (high crowns with relatively short roots), whereas browsers, consuming softer food, should have brachydont dentition (short crowns with relatively long roots). Observation of the feed- ing behavior shows exactly this state of affairs. P. johnstoni, mainly a grazer, has hypsodont molars, whereas the browsers H. brucei and D. arborens have brachydont dentition. If our studies are con- fined to a comparison of molar structure and feeding behavior, a very close relationship between Hete- rohyra.x and Dendrohyrax seems to be established. But for a revision of the systematics of an animal group, as many taxonomic criteria as possible should be examined not only anatomical, but also genetical, ecological and behavioral parameters, to ensure accurate comparison and exact taxonomic grouping. In a group of species the distribution of behav- ioral similarities and differences tend to be related with phylogenetic relationships within the group (Brown, 1975). Behavior is a preeminently suitable field for adaptation; it is regularly a pacemaker in evolution, that is, it precedes adaptive anatomical I would like to express my gratitude to the Tanzania National Parks Trustees, the Serengeti Research Institute, and the Ngo- rongoro Conservation Authorities for all their kind support; to Prof. J. Jacobs and Prof. W. Wickler for their valuable help and change (Mayr, 1958; Wickler, 1972). In selecting a behavior pattern for comparative analysis, it should ideally be highly stereotyped within a species, but variable across species. Such a species character- istic behavior pattern is, for example, the copula- tory behavior as shown in the comparative study of muroid rodents by Dewsbury (1975). The territorial call is so distinct in character that it can confidently be used as a simple method of locating and recognizing different species. By this means a population of Dendrohyrax validus was discovered and first reported for the Kenya coast (Seibt et al., 1977). The distance between the anus and the preputial opening, first noted by Coetzee (1966), and the structure of the penis, are important and very useful features for a rapid taxonomic identification. The females reproductive tract has not so far been ana- lyzed, but this may also show a characteristic struc- ture. These observations suggest that Heterohyrax is as far apart from Dendrohyrax as from Procavia and that one can differentiate three groups in the Hyracoidea no matter on which taxonomical level. Unless further studies, for example, genetic analy- ses, provide other evidence, I suggest that Hahn’s (1935, 1959) and Bothma’s (1966) recognition of three genera in the Order Hyracoidea should be re- tained. advice. Many thanks also to Mr. P. Fox for his cooperation, to Mr. C. Nyaole Gagah for helping observe D. arhoreiis, and to Mrs. P. Rechten for revising the English. Serengeti Research Institute Publication No. 220. LITERATURE CITED Bothma, J. DU P. 1971. Part 12, Order Hyracoidea. Pp. l-8,/7; The mammals of Africa: an identification manual (J. Meester and H. W. Setzer, eds.), Smithsonian Inst. Press, Washing- ton, D.C. Brown, J. L. 1975. The evolution of behavior. W. W. Norton & Company, New York. Coetzee, C.G. 1976. The relative position of the penis in South- ern African dassies (Hyracoidea) as a character of taxonomic importance. Zoologica Afr., 2:223-224. Dewsbury, D. A. 1975. Diversity and adaptation in rodent cop- ulatory behavior. Science, 190:947-954. Glover, T. D., AND J. B. Sale. 1968. The reproductive system of male rock hyrax (Procavia and Heterohyrax). J. Zool., London 156:351-362. Hahn, H. 1935. Die Familie der Procaviidae. Ph.D. Thesis, Friedrich- Wilhelms-Universitat, Berlin. . 1959. Von Baum-, Busch- und Klippschliefern, den klei- nen Verwandten der Seekiihe und Elefanten. In, Die Neue Brehm-Biicherei, Wittenberg Lutherstadt, Ziemsen. Hoeck, H. N. 1975. Differential feeding behaviour of the sym- patric hyrax Procavia johnstoni and Heterohyrax brucei. Oecologia, 22:15-47. . 1976i/. Procavia johnstoni (Procaviidae) feeding behav- iour. Film E 2177 of the Inst. Wiss. Eilm, Gottingen. . \976b. Procavia johnstoni (Procaviidea) Mating behav- iour. Eilm E 2178 of the Inst. Wiss. Eilm, Gottingen. . 1971a. “Teat order” in hyrax (Procavia johnstoni and Heterohyrax brucei). Z. Saugetierkunde, 42:112-115. 1978 HOECK— SYSTEMATICS OE HYRACOIDEA 151 . 1977/?. Heterohyrax brucei (Procaviidea) feeding behav- iour. Film of the Inst. Wiss. Film, Gottingen, in press. . 1977c. Heterohyrax brucei (Procaviidea) mating behav- ior. Film of the Inst. Wiss. Film, Gottingen, in press. Kingdon,J. 1971. East African mammals: an atlas of evolution in Africa. Academic Press, New York-London, 1 :x -r 1-446. Lonneberg, E. 1916. Mammals collected by H. R. H. Prince Vilhelm's Expedition to British East Africa 1914. Ark. Zoll. 12:1919-1929. Ma'i R, E. 1958. Behavior and systematics. Pp. 341-362, Be- havior and evolution (A. Roe and G. G. Simpson, eds.), Yale Univ. Press, New Haven, Connecticut, viii -t 557 pp. Roche, J. 1972. Systematique du genre Procavia et des damans en general. Mammalia, 36:22-49. Seibi , U., H. N. Hoeck, AND W. WiCKLER. 1977. Dendrohyrax validus True, 1890, in Kenia. Z. Saugetierkunde 42: 1 15-1 18. WicKLER, W. 1972. Verhalten und Umwelt. Verlag Hoffmann und Campe, Hamburg. PATTERNS OF SPECIATION IN AFRICAN MAMMALS PETER GRUBB Department of Zoology, P. O. Box 67, University of Ghana, Legon, Accra, Ghana ABSTRACT The distribution and speciation of African mammals can be interpreted in terms of both isolation and dispersal associated with Quaternary climatic and vegetational changes. Patterns of faunal diversity and endemism suggest the former existence of refugia in Forest Regions and dispersal between these Regions. Isolation has also been significant in savanna, but dispersal pat- terns are less easily assessed here. In certain superspecies, the directions involved in dispersals can be ascertained from the distribution of primitive and derivative character states, which in turn imply the cotemporal existence of ancestral and descen- dant taxa. Long term "ecological translation" in speciation ap- pears to have occurred down the faunal diversity gradient, es- pecially across the forest-savanna boundary, but also from savanna to arid zones. Different mammalian taxa have speciated in different ways and there are other impediments to the recon- struction of speciation patterns. INTRODUCTION Currently much attention is focusing on the his- torical geography and speciation of tropical biota. This paper, based on the literature and on my own taxonomic and faunistic studies mostly still in prog- ress, draws attention to some of the main problems involved in assessing the speciation of African mammals during the fluctuating climatic conditions of the Quaternary. The study of mammalian speciation seeks to de- scribe the dispersion and geographic variation of species accurately, to identify hybrid zones and other types of apparent secondary contact, and to recognize primitive and derivative characters in the species being studied. It then attempts to infer lo- calities of refugia occupied under adverse climatic conditions and to identify directions taken during periods of dispersal. It may be possible to provide a relative chronology of events (Table 7); hypoth- eses of this kind gain conviction if several mammals appear to have had similar histories. Einally, it may be possible to provide more absolute chronologies, with the assistance of other disciplines, and to trace the histories of regional faunas. Many difficulties are involved in this exercise. TERMINOLOGY Superspecies are "monophyletic groups of entirely or essen- tially allopatric species too different to be included in a single species" (Mayr, 1963:44). In the plural, the term is used (Bi- galke, 1972) to embrace "monospecific” superspecies as well. Allospecies are species which are members of superspecies. Semispecies are either highly distinctive subspecies or full spe- cies whose status is debatable. Species groups are monophyletic assemblages of closely allied or sibling species, some of which may be allopatric, so that the term superspecies is inappropriate (Hall and Moreau, 1970). Subspecies groups are monophyletic groups of subspecies, which may approach species status. SPECIATION AND ZOOGEOGRAPHY Previous Studies Interest in the evolutionary geography of African mammals dates from the early part of this century, with the publication of general interpretations of present faunal distributions in terms of Tertiary and Quaternary phenomena (Schwarz, 1924, 1926o; Lonnberg, 1929; Braestrup, 1935) and evolutionary studies on particular taxa, including the African buffalo, Syncerus coffer (Christy, 1924rt, 1924^, 1929; Malbrant, 1935), monkeys, Cercopithecus and Colohus (Schwarz, \926b, 1928, 1929), sun- squirrels, Heliosciurus (Ingoldby, 1927) and harte- beest, Alcelophus (Ruxton and Schwarz, 1929). Two of these papers (Schwarz, 1928; Ruxton and Schwarz, 1929) were presented as the first parts of a series on the speciation of African mammals, but there was no sequel and after such a promising be- ginning, no progress in the subject was made for 20 years. The works mentioned had so little impact that only Lonnberg’s and Braestrup’s gained pass- ing mention by Moreau (1952), and none is dis- cussed by Moreau (1963, 1966) or Hamilton (1976). 152 1978 GRUBB— AFRICAN MAMMAL SPECIATION 153 Nevertheless, Lonnberg (1929) had already pro- posed that Pleistocene climatic cycles, by breaking up and uniting habitat types, were important in in- itiating the speciation of African mammals. It was not until after the Second World War that a gradual reawakening of interest in mammalian his- torical biogeography developed. Booth contributed studies on evolutionary geography of West African mammals (Booth 1954, 1958ri, 1958/?) and a reap- praisal of Schwarz’s (1928) work on speciation in the mona monkeys. Blancou (1954) and Grubb (1971) reassessed Christy’s (1929) and Malbrant’s (1935) contributions on speciation in the African buffalo. Brain and Meester (1964) analyzed specia- tion in southern African Myosorex, Groves (1971) the evolutionary dispersal of the gorilla, and Die- terlen ( 1971) speciation of Dendromus. Jotterand (1972), following the work of Matthey, discussed chromosome evolution in relation to the speciation of Mus. Eisentraut (1973) investigated the Pleisto- cene history of Eernando Po and the Cameroon highlands. Most recently, Kingdon (1971, 1974 C. polykonios polykomos (W) ^ C. p. vellerosns (W) ^ C. gnereza (N. forest border to Ethiopia, Tanzania) (h) Philantoniba maxwelli (W) — * P. nionticola (C, E) (i) Cephalophns spadix (E) — > C. silvicnltor (W, C) (j) Neotragns pygniaens (W) ^ N. batesi (WC, EC) ^ N. moschatns inoschatns (E) ^ N. m. livingstonianns (E) 2. Savanna and forest/savanna mammals; SS = Southern Savanna, NS = Northern Savanna, SA = Somali Arid, SWA = Southwest Arid (a) Cercopithecns (aethiops) pygerythrns C. (a.) aethiops —> C. (a.) sabaens (b) Papio cynocephalns (SS) — * P. annbis (NS) P. nrsinns (SWA); P. papio (NS) ^ P. hamadryas (SA) (c) Kerns erythropns (NS) ^ A', inanris. A', princeps (SWA) (d) Cricetoniys einini (forest) ^ C. gambianns ansorgei (SS) ^ C. g. gambianns (NS) (e) Procavia habessinica and other northern Procavia spp. — ► P. capensis, P. welwitschii (SWA) (0 Tragelaphns scriptns scriptns (forest) ^ T. s. ornatus (SS) ^ T. s. .'iylvaticns (SS) (g) Tanrotragns derbianns (NS) ^ T. oryx (SS) (h) Syncerns caffer nanns (forest) ^ S. c. bracbyceros (NS) A. f. caffer (NS, SS) (i) Danudiscns korrignm (NS), D. Innatns (SS) D. dor c as (SS) (j) Alcelaphns bnselaphus tora. swaynei (SA) — > A. b. cokei (SA) ^ A. b. bnselaphus (N. Africa) — ► A. h. major (NS) —* A. h. lehvel (NS) —> A. b. caania (SWA) (k) Madoqua saltlana (SA) ^ M . kirkli (SA, SWA) M. gnentheri (SA) (l) Gazella soemerringi (SA) ^ G. granti (SA) (m) Equns grevyi (SA) ^ E. zebra (SWA) ^ E. bnrchelli (SS) E. qnagga (SS) es, montane forest or, for that matter, non-forest biota may have dispersed into lowland environ- ments more than at present. This happens today to some extent, as some highland forest species pen- 160 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 Table 7. — Hypothetical sequence of events in the speciation of Colobus monkeys. 1. Dry Phase — Colobus verus in (West) Africa. 2. Moist Phase — C. verus disperses to Central Africa. 3. Dry — C. ibaJius) foae differentiates in E.C. refuge; acquires reddish pelage, greater sexual dimorphism; cephalic hair whorls diverge and are reduced. 4. Moist — C. (b.)foai disperses to West Africa. 5. Dry — C. (b.) badius differentiates in W. refuge; acquires blacker pelage, cephalic hair whorls lost. 6. Moist — C. (b.) badius disperses eastward to W.C. Forest Region. 7. Dry — C. satanas differentiates in W.C. Refuge; loses red pelage, acquires enlarged larynx, long mantle, pallid juvenile pelage (?). 8. Moist — C. satanas disperses to C. Regions. 9. Dry — C. angolensis differentiates in C. Regions; acquires white pelage, loses sexual swellings, becomes more tolerant of dry forest. 10. Moist — C. angolensis disperses to W. Region along route north of present forest boundary. 11. Dry — C. polykomos differentiates in W. Refuge — during an extreme dry period? 12. Moist — C. polykomos disperses eastward, principally along forest galleries. 13. Dry — C. guereza diverges in isolation as even more a species of dry forest (see Clutton-Brock, 1974). 14. Moist — C. guereza disperses widely into East Africa and Ethiopia during period which leads to present day. etrate low latitude lowland forest (Cercopithecus hamlyni and C . Ihoesti itself, Rahm, 1970) and other highland species descend to lowland forest at higher latitudes (Brain and Meester, 1964, for Myosorex, and Misonne, 1963, for Otomys). Dispersal Direction The direction of dispersal along a particular route is more difficult to assess than the route itself and inferences can be drawn only from the general faun- istic premise that small faunas, especially if they lack endemics, are derivatives of larger ones, and from the study of primitive and derivative character states of species and subspecies. In a number of forest mammals, especially the primates, it is pos- sible to identify sequences of successively more derivative subspecies or species, suggesting that a population has dispersed and formed a new isolate, which has itself given rise to another and so on, with each climatic cycle, so that a chain of ances- tors and descendants are formed, which are yet contemporaries (Table 6). Schwarz’s conclusions (1926/?, 1928, 1929) were similar, though he did not always distinguish primitive and derived character states. From Table 6 it is possible to interpret the spe- ciation of colobus monkeys in terms of seven chief dispersals between the Western and Central Forest Regions in association with seven phases of isola- tion and differentiation (Table 7). The last putative eastward dispersal of colobus and the succeeding phase of isolation (commencing ca. 35,000 years BP?) can readily be related to inferred eastward dispersal in some other monkeys and a squirrel (Table 8). The hypothesis is based on Table 6 with extrapolation to other species where primitive-de- rivative character states are not yet identified, and is supported by certain zoogeographical details — for instance, none of the eastward-invasive animals have managed to reach Fernando Po, whereas the “resident” species are well represented there. Booth (1958/?) interpreted these sequences in Ta- ble 8 as “stepped dines” reflecting the breakup of more continuous populations in a series of Pleisto- cene refugia, of which he recognized two in the Western Forest Region to account for primate sub- speciation there. He did not explain how the eastern subspecies are in some cases intermediate between the western one and derivative taxa occurring out- side the Region (Table 6: b,e,g). A succession of isolations and dispersals provides, I think, a better explanation. This hypothesis can be correlated with others concerning approximately contemporaneous evo- lutionary events — namely the derivation of the South-Central Forest Zone primates, speciation among West-Central Forest mammals (Table 5), or the concept of a dispersal corridor between West- ern and Central Forest north of present forest lim- its— to provide a preliminary reconstruction of re- cent forest faunistic history. Nevertheless, the level of supposition in the reconstruction remains very apparent. Ecological Translation across the Forest-savanna Boundary Direction in dispersal and speciation may be con- sidered not only in a strictly geographic sense, but also from the ecological aspect. Most African mam- mals are probably confined to a major biome — for- est, mesic savanna, arid savanna, or desert — though the extent to which this is true has not been critically examined (see distribution maps in Smith- ers, 1971; Schouteden, 1944-1946; Davis, 1974). The boundary between forest and savanna is a particu- larly clear one, and few species occur unequivo- cally in both habitats even though forest species 1978 GRUBB— AFRICAN MAMMAL SPECIATION 161 Table 8. — Suggested eastward dispersal of western forest monkeys and a squirrel. Genera Primary Western Refuge ^ Secondary Refuge Nigeria, W of Niger River West- Central Region ^ East- Central Region — ► South- Central Region Colobus p. polykomos p. vellerosus p. vellerosus guereza guereza Cercocebus a. atys a. lunulatus torquatus torquatus Cercopithecus d. diana d. roloway Cercopithecus p. buttikoferi p. petaurista erythrogaster Cercopithecus c. campbelli c. lowei mono mona denti wolf Colobus b. badius b. waldroni b. preussi Funisciurus lemniscatus lemniscatus raptorum raptorum akka pyrrhopus group group group group group Cercopithecus nictitans stampflii Colobus verus verus verus penetrate savanna along forest galleries and savan- na species enter the forest when it is opened up by farms and roads (Rahm, 1972). Successively stronger levels of differentiation be- tween forest and savanna vicariants can be recog- nized in at least 30 superspecies and species-groups (Table 9). Leaving these species aside, there are still at least 26 genera, which include both forest and savanna species. When the distribution and tax- onomy of these animals are better known, it will probably be possible to segregate some further evo- lutionary units as species groups or superspecies here as well. Five families whose species have not yet been considered, and another nine which al- ready have, contain separate forest and savanna genera. A total of 28 families occurs in both biomes, as compared with three in the forest alone and 15, of which eight are rodents, only outside the forest. The taxonomic levels of differentiation between forest and savanna mammals suggest there has been a long-term exchange between the faunas of the biomes through infraspecific dispersal from one to the other and eventually through speciation of vi- cariant populations, because eurytopic species oc- curring in both habitats are few. Hypothetically, faunal exchange between biomes could be equally balanced, but the evidence, such as it is, suggests that it has been predominantly from forest to savanna. Schwarz (1924) and Lonnberg (1929) believed that plains mammals had a forest origin, and Lonn- berg indeed supposed that some of the savanna mammals of Africa had originated from the strand- ing of forest species during a long-term recession of the Great Hylea. Understandably, this explanation was rejected by Chapin (1932) and Moreau ( 1952) but its supporting evidence was not reinterpreted by either. Forest origins have been proposed or im- plied by various authors for Soricidae (Heim de Balsac and Lamotte, 1956, 1957), Cercopithecidae (Lonnberg, 1929; Napier, 1970), Procaviidae and Sciuridae (Lonnberg, 1929), Giraffidae (Harris, 1976), Bovidae (Estes, 1974), Nycteridae (Braes- trup, 1935; Koopman, 1975), and Hystricidae (Kingdon, 1974/?), while the most derived species in several of these families and in Manidae, Hip- popotamidae, Suidae, Herpestinae (see Pocock, 1919; Taylor, 1974, 1976), and Eelidae are clearly savanna taxa. These citations do not in themselves contribute evidence for a forest-savanna ecological translation, because the literature is diffuse and di- verse, few authors having even considered the hypothesis’ plausibility. The shortage of pan-African taxonomic studies and more especially of evolu- tionary studies partly explains this deficiency. For families other than those mentioned above, no statement of forest origin has been made, nor can the original biome be regarded as self-evident, even though as in the case of the Pteropodidae, one is inclined to suspect it on zoogeographic grounds. Moreoever, some mammalian families (for exam- ple, Gerbillidae) have had a very long history of speciation in non-forest environments and may have originated there so that their ultimate ancestral habitat cannot readily be determined. Forest origins are accepted in interpreting the phylogeny of horse, elephant, or man, but proposals for forest origins in African mammals have not all remained undis- puted. Thus, Kingdon (1971) regards the forest hab- itat as secondary in hyraxes, not noting contrary evidence presented by Lonnberg (1929) for a forest origin, and Lonnberg himself had to make a forest 162 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 lable 9. — Mammalian taxa occurring in both forest and savanna. 1. Species with weakly differentiated forest and savanna populations Crocidara occidentalis Taphozous mauritiamis Hipposideros commersoni Pipistrellns nanus Heliosciuriis gambianus Mellivora capensis Lutra macullicollis Genetta pardina Viverra civetta Atilax paludinosus Panthera pardiis Tragelaphus spekii 2. Species with strongly differentiated forest and savanna subspecies Funisciurus congicus Orycteropus afer Loxodonta africuna Potamochoeriis porciis Tragelaphus scriptus Syncerus caffer 3. Superspecies with forest (listed first) and savanna allospecies Tadarida congica, T. trevori Manis gigantea, M. temmincki Aonyx congicus, A. capensis Galago inustus, G. senegalensis Paraxerus paUiatus, P. cepapi, P. ochraceus Cricetomys emini, C. gambianus Manis gigantea, M. temmincki Aonyx congicus, A. capensis 4. Species groups with forest and savanna allospecies Hipposideros ruber, H. cafer Rhinolophus alcyone, R. landeri Heliosciurus gambianus, H. mutablis, H. rufobrachium Thamnomys rutilans, T. cometes, T. dolichurus Lemniscomys species-complex 5. Genera with separate forest and savanna species (a) Species additional to those already listed in the following genera Crocidura, Taphozous, Hipposideros, Pipistrellns, Rhinolophus, Galago, Genetta, Tragelaphus (b) Species in the following genera Sylvisorex, Epomops, Rousettus, Nycteris, Eptesicus, Glauconycteris, Scotophilus, Kerivoula, Myopterus, Tadarida, Cercopithecus, Dendromus, Aethomys, Praomys, Graphiurus, Bdeogale, Felis 6. Families with different forest and savanna genera Soricidae, Macroscelididae, Cercopithecidae, Pteropodidae, Sciuridae, Cricetomyidae, Muridae, including Oenomys- Thamnomys group, Malacomys group, Acomys group and Arvicanthis group, Procaviidae, Hystricidae, Hippopotamidae, Suidae, Giraffidae, Viverridae, Bovidae habitat secondary for bovids, because he was mis- takenly forced to believe that they had radiated out- side the continent. Kortland’s dehumanization hy- pothesis suggests a savanna origin for the chimpanzee (Kortland and Kooij, 1963; Kortland and Van Zon, 1968); Buettner-Janusch (1966:267) proposed a forest invasion by Cercopithecus; and Kingdon’s (1971) speculation on Galago required speciation out of, but then back into, the forest. Misonne (1969) considers murids to be primarily savanna animals in Africa, yet he regards forest as a marginal refuge habitat (1969:167, 171) for certain more primitive species. From his conclusions, it seems quite possible that apart from the Praomys and Mz/5 groups, all African murids could represent a single radiation within the Continent; there seems to be no very compelling reason for supposing that the occurrence of certain primitive genera and spe- cies in the forest should not indicate a forest origin for this radiation. Few other statements to the effect that African forest mammals were derived from savanna species have been made. The problem of ultimate biome origin is nevertheless not an easy one to handle, for speciation across a major ecotone can be interpret- ed in several ways from primitive-derivative char- acters. With a colonist in biome B and its cotem- poral primitive ancestor in biome A, there has been a dispersive speciation from A to B, but, if the ancestor is itself replaced in A by a third and even more advanced form, while the taxon in B retains primitive features, dispersal could be misinterpret- ed as having occurred from B to A. Animals like Sylvicapra grimmia and Cercopithecus aethiops, for example, retain very conservative color patterns and hair-banding characters, with respect to their congeners in the ancestral forest habitat, yet a wealth of evidence suggests that ecological drift in Bovidae and Cercopithecidae has indeed been from forest to savanna. Evidence for relatively recent dispersal into the savanna by forest mammals comes from a study of certain eurytopic species or species groups occur- ring in both biomes. These mammals exhibit a mor- phological continuum between populations, yet have areas of parapatry, sympatry, or secondary intergradation between extreme phenotypes where demes have converged on each other in dispersing from different areas. Primitive characters occur in the forest populations, more derivative ones in the savanna, whereas the reverse is not usually evident. 1978 GRUBB— AFRICAN MAMMAL SPECIATION 163 The buffalo, Syncenis caffer, provides one of the best examples (Grubb, 1971, and in preparation), so great is the contrast between the forest and savanna forms and so clear are the derivative characters of the latter. This species and the elephant, Loxodonta africana, apparently dispersed into the savanna in West Africa and continued their dispersal through the Northern Savanna and around the forest block into Angola and South Africa. In East Africa, both replaced other species in the late Pleistocene — the more primitive Homoioceros and the more deriva- tive E’/ep/lai' recki (Maglio, 1973). More species appear to have crossed the south- ern forest-savanna border, and in Angola there is clinal variation between forest and savanna popu- lations of Hipposideros {caffer group, Koopman, 1975), Cricetomys, and Heliosciums, which in other parts of Africa are sympatric. The Hipposideros caffer group, the giant rat Cricetomys, the bush- buck, Tragelaphiis scriptus, and the bushpig, Po- tamochoerus percus, all have dispersed from the forest through the Southern Savanna northward into East Africa. The bushbuck did not reach far- ther than southern Sudan in this northward pro- gression, other subspecies having colonized Ethio- pia and the Northern Savanna, but the bushpig did get to the Ethiopian highlands. Cricetomys gambi- aniis and Hipposideros caffer have spread farther, throughout the Northern Savanna, the latter coex- isting with its forest congener H. ruber where this species penetrates savanna. A third route has been adopted by the squirrel Funisciurus congicus. Species of this genus reach into the savanna along riverain forest (F. pyrrho- piis, F. anerythriis, F. substriatus, F. bayoni) but F. congicus extends much further, from the forest of the Cuvette Central south through western An- gola into the dry mopane woodland of South West Africa, in one continuous clinal sequence. Speciation and Dispersal in Savanna and Arid Biomes Endemism or vicariantism of non-forest mam- mals tends to be related to the two Savanna Zones or three Arid Zones of Davis ( 1962) with, for in- stance, Taurotragus derbianus in the Northern Sa- vanna, T. oryx in the Southern, or Oryx dammah in the Sudanese, O. beisa in the Somali, O. gazella in the South West Arid (Ansell, 1972). These Zones have hence been centers of isolation, but it is un- likely that they were ever reduced to small refugia, contra Moreau (1966:98). Other regions have had a less prominent role in speciation. In the Northern Savanna, a number of endemics are found only east (for example, Helo- gale dybowskii) or west (for example, Genetta thier- ryi) of the Lake Chad and Chari Basin area, where- as certain taxa are represented by divergent allospecies or subspecies in these regions (Schwarz, 1926u; Grubb, 1971; Groves, 1975). The Pleistocene lake Megachad may have been the instrumental bar- rier to dispersal here. A small faunal element is lo- cated in the Senegal area, including endemic semi- species (particularly Papio papio and Sylvicapra grimmia coronata) and isolates of otherwise widely distributed species. The factor encouraging isola- tion is not known. The Nile swamps and Ethiopian Highlands were probably also refuge centers for sa- vanna mammals in arid periods. In the Southern Savanna, the Muchinga scarp is involved in separating semispecies of baboons, blue monkeys, sun-squirrels, and waterbuck (Ansell, 1960) and may have accounted for speciation be- tween roan and sable antelopes, for it lies central to their area of sympatry. The Rufiji basin (King- don, 1974^/) and the Escarpment Zone of Angola (Cabral, 1966) have also been proposed as barriers promoting isolation in mammals, whereas others such as the Nyasa Rift were probably involved. Endemism in southern Africa is discussed by Rau- tenbach ( 1978) and speciation in Elepluintulus, Oto- myinae, and Chrysochloridae, for example, must have been complex in this part of the Continent. Speciation in Saharan mammals, particulary ro- dents, is perhaps best linked with evolutionary studies of the whole Eremian fauna. There is a com- plex situation in the Somali Arid Zone, with spe- ciation or secondary contact among large- and small-nosed dikdik, giraffe, larger and smaller ga- zelles, hartebeest, and zebras (Keast, 1965). The mammals making up the famous East African Plains Fauna, which occurs within the Zone, are of diverse zoogeographical affinity and wildebeest, gazelle, and perhaps hartebeest and rhinoceros (see Brooks, 1961; Groves, 1967) have been influenced in their subspeciation by the Kenya Rift. The discontinuous distribution of species and superspecies in the So- mali and South West Arid zones has been exten- sively discussed and explained by dispersal along a former Arid Corridor during a dry climatic phase (Benson and White, 1960; Ansell et al., 1962; An- sell, 1960; Bigalke, 1972). The dry-savanna mam- mals of the Rukwa valley and isolated populations of wildebeest, giraffe, and tsessebe in Zambia (An- 164 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 sell, 1960) perhaps represent a vestige of this dis- persal. Dispersal directions are not always easily in- ferred in non-forest mammals but they may have involved complex patterns of replacement with very wideranging dispersal. For example, in their northern distributions, as well as in southern Afri- ca, neither hartebeest nor baboons have their most closely allied allospecies in contact. The Ethiopian hartebeest, tora, is phylogenetically nearer the West African major or the extinct North African huselaphus than the very different Sudanese lelwel with which it forms a hybrid zone (Ruxton and Schwarz, 1929). And Papio papio of Gambia and Senegal is nearer P. hamadryas of Ethiopia and Somalia than P. anubis of the intervening Northern Savanna (Hill, 1970). A notable problem is the absence of any indica- tion of dispersal through Angola and Congo to the Central African Empire, where a route should have been available when forest was reduced to small refugia (see Carcasson, 1964; Hamilton, 1976). The significance of this has yet to be assessed. A number of superspecies are not restricted to arid or savanna zones, but show more complex dis- tribution patterns, with some allospecies in both mesic and arid zones and others in only one or the other. This group includes mammals where the Northern Savanna form is either restricted to the Guinea Savanna or is more eurytopic, reaching into the Sudanese Arid Zone, whereas the nearest ally is a stenotopic Southern Arid species. Alcelaphiis (huselaphus) lelwel, Papio anubis, Xerus erythro- pus, and Procavia latastei in the Northern Savanna are represented by A. (h.) caama, P. ursinus, X. inauris, andP. capensis in the Southern Arid Zone, to which they are restricted. Related trends are seen in Sylvicapra grimmia, where derivative subspecies have penetrated the Southern Arid, yet conserva- tive subspecies have not dispersed into the Su- danese Arid (Groves, personal communication), and the eland, where the primitive Taurotragus der- bianus is restricted to Guinea savanna, yet the de- rivative T. oryx is much more habitat-tolerant and extends into the Kalahari. The giraffe, white rhi- noceros, and hedgehog are also in part more eco- logically restricted in their southern distributions. These distribution patterns suggest an ecological translation from mesic to arid habitats, a view sup- ported by the derivative character of all the South- ern Arid allospecies (Table 6). Other Southern Arid or Grassland species (Oryx gazella, Antidorcas marsupialis , Damaliscus dor c as, Connochaetes gnou, Phacoclioerus aethiopicus aethiopicus, Equus quagga) are also derivative. And the paucity and strongly derivative adaptations of desert mammals suggest that they too have originated from other habitats. Ecological Direction in Dispersive Speciation From this discussion and earlier remarks about forest origins of savanna species, it would appear that if speciation involves an ecological translation from one biome to another, then it is predominantly from mesic, predictable and less seasonal habitats toward arid, less predictable and more seasonal en- vironments, and hence involves a descent of the species-diversity gradient. The slope of this gra- dient in Africa is considerable; it is best measured in areas where habitats are clearly stratified, as in West Africa. Rosevear (1949) presents data from which it can be calculated. The gradient in number of mammal species per biome is 130, 60, 47, 37 from forest through Guinea, Sudan, and Sahel savannas in Nigeria. These data are provisional; I obtain 141, 106, 86, 49 and 22 (desert) for the same transect west of the Volta River. Implications for theories of faunal steady states (MacArthur, 1972; 174; Roz- enzweig, 1975), or faunal exchange (Wilson, 1965) arise if dispersive speciation is indeed predomi- nantly unidirectional along this gradient, and the possibility that biomes are faunistically imbalanced necessarily follows. CONCLUSION An important impediment to the reconstruction of speciation in African mammals is the scarcity of taxa with apparently significant and interpretable geographic variation and dispersion. It is impossible to recognize centers of dispersal or refugia when variation is clinal and low-key, and it is impossible to trace any dispersal track when primitive-deriva- tive characters have not been recognized. And these may be misleading, in suggesting a dispersal direction opposite to the one which occurred. There is in any case some danger in assuming that primi- tive-derivative morphological series always repre- 1978 GRUBB— AFRICAN MAMMAL SPECIATION 165 sent ancestors and descendants, though this may be nearer the truth and at least no more mistaken than suggesting that two contemporaries share an extinct common ancestor that was phenetically equidistant between them. However, in many cases, the tem- poral divergence in morphology that does occur be- tween primitive and derivative stocks once gene flow between them has ceased must now obscure the relationship. Very few reconstructions of species histories are likely to be possible, therefore, even with the de- gree of surmise entailed in the Colobiis model (Ta- ble 7). The absolute time span involved in a case like this is hard to assess, for we do not know what scale of paleoclimatic events were needed to drive the isolation-dispersal system, and therefore wheth- er speciation took a longer or shorter time. Even relative time-scaling presents problems. Taxa that have diverged to different degrees, for instance, may have speciated at different rates or over dif- ferent periods of time, and there may now be no way of distinguishing such equally parsimonious hypotheses. Even with a single superspecies, events may get telescoped — a succession of diver- gences may be regarded as cotemporal (Booth, 1958/j). What we clearly need are more evolutionary data for developing more rigorous models of speciation. We may expect further paleontological studies to do this for the larger mammals and karyological work for the smaller species, as it has already con- tributed so much outside Africa. Logistic problems may allow faunistic and taxonomic studies to hold sway for some time and we certainly need many more accurate distribution maps. The classification of distributions is an important first step in recon- structing faunal histories. Some of the hypotheses presented in this paper are supported by evidence from African birds, reptiles, and butterflies and a less sectarian approach may in the future prove more productive. ACKNOWLEDGMENTS I am grateful to the American Philosophical Society, the Royal versity of Ghana for grant support or for financial assistance in Society, the Carnegie Museum of Natural History and the Uni- attending the Colloquium. LITERATURE CITED Ansell, W. F. H. 1960. Mammals of Northern Rhodesia. Gov- ernment Printer, Lusaka, xxxi -f 155 ■+■ 24 pp. . 1972. Part 15, Order Artiodactyla. Pp. 1-84, //( The mam- mals of Africa: an identification manual (J. Meester and H. W. Setzer, eds,), Smithsonian Inst. Press, Washington, D.C. Ansell, W. F. H., C. W. Benson, and B. I. Mitchell. 1962. Notes on some mammals from Nyasaland and adjacent areas. Nyasaland J., 15:38-54. Azzaroli, I., AND A. M. SiMONETTA. 1966. Camivori della So- malia exitaliana. Monitore Zoo!. Ital., 74 (suppl.): 102-195. Backhaus, D. 1958. Sind Kobiis defassa (Ruppell, 1835) und Kobiis ellipsiprymnus (Ogilby, 1833) verschiedene Arten? Saugtierk. Mitt., 6: 174. Benson, C. W., and C. M. N. White. 1960. Discontinuous dis- tributions (Aves). Proceedings 1st Federated Science Con- gress, pp. 195-216. Bere, R. M. 1962. The wild mammals of Uganda. Longmans, London, 148 pp. Bigalke, R. C. 1972. The contemporary mammal fauna of Af- rica. Pp. 141-194, in Evolution, mammals and southern con- tinents (A. Keast, F. C. Erk, and B. Glass, eds.). State Uni- versity of New York Press, Albany, 543 pp. Blancou, L. 1954. Buffles d'Afrique. Zooleo, 27:425-434. Booth, A. H. 1954. The Dahomey Gap and the mammalian fau- na of the West African forests. Rev. Zool. Bot. Afr. 50:305- 314. . 1955. Speciation in the mona monkeys. J. Mamm., 36:434-449. . 1958a. The Niger, the Volta and the Dahomey Gap as geographic barriers. Evolution, 12:48-62. . 19586. The zoogeography of West African primates: a review. Bull. Inst. Fr. Afr. Noire, 20:587-622. Booth, C. P. 1968. Taxonomic studies of Cercopithecus mitis Wolf (East Africa). National Geographic Society Research Reports, 1963 Projects, pp. 37-51. Braestrup, F. W. 1935. Remarks on climatic change and faunal evolution in Africa. Zoogeographica, 2:484-494. Brain, C. K., and J. Meester. 1964. Past climatic changes as biological isolating mechanisms in southern Africa. Mono- graphic Biologicae, 14:333-340. Brooks, A. C. 1961. A study of the Thomson s gazelle (Gace//a thomsonii Gunther) in Tanganyika. H. M. Stationery Office, London. Buettner-Janusch, j. 1966. Origins of man. Physical anthro- pology. John Wiley and Sons, Inc., New York, xxvii -i- 674 pp. Cabral, J. C. DE M. 1966. Some new data on Angolan Muridae. Zool. Africana, 2:193-203. 166 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 . 1969. As genetas de Angola. Bol. Inst. Invest. Cien. Angola, 6:7-40. Carcasson, R. H. 1964. A preliminary survey of the zoogeog- raphy of African butterflies. E. African Wildlife J., 2:122- 157. Chapin, J. P. 1932. The birds of the Belgian Congo, part I. Bull. Amer. Mus. Nat. Hist., 65:1-756. Chrisi y, C. 1924u. African forest, bush and buffaloes. Bull. New York Zool. Soc., 27:130-142. . 19247>. Biggame and pygmies. Macmillan, London, xxxi -H 325 pp. . 1929. The African buffaloes. Proc. Zool. Soc. London, pp, 445-462. Clutton-Brock, T. H. 1974. Primate social organisation and ecology. Nature, 250:539-542. CoETZEE, J. A., AND E. M. Van ZiNDEREN Barker. 1970. Pa- laeoecological problems of the Quaternary of Africa. S. Af- rican J. Sci., 66:78-84. Corbet, G. B. 1974. Part 1.5, Family Macroscelididae. Pp. 1-6, ill The mammals of Africa: an identification manual (J. Mees- ter and H. W. Setzer, eds.), Smithsonian Inst. Press, Wash- ington, D.C. Corbet, G. B., and J. Hanks. 1968. A revision of the elephant shrews, family Macroscelididae. Bull. British Mus. Nat. Hist. (Zool.), 16:47-111. Corbet, G. B., and D. W. Yalden. 1972. Recent records of mammals (other than bats) from Ethiopia. Bull. British Mus. Nat. Hist. (Zool.), 22:213-252. Dandelot, P. 1959. Note sur la classification des cercopit- heques du groupe aethiops. Mammalia, 23:357-368. Davis, D. H. S. 1962. Distribution of southern African Muridae, with notes on some of their fossil antecedents. Ann. Cape Prov. Mus., 2:56-76. . 1966. Contribution to the revision of the genus Talent in Africa. Ann. Mus. Roy. Afr. Centr., 144:47-65. . 1974. The distribution of some small southern African mammals (Mammalia: Insectivora, Rodentia). Ann. Trans- vaal Mus., 29:135-184. Dieterlen, F. 1971. Beitrage zur Systematik, Okologie und Biologic der Gattung Dendronius (Dendromurinae, Criceti- dae, Rodentia) ins besondere ihrerzentral afrikanischen For- men. Saugetierk. Mitt. 19:97-132. Eisentraut, M. 1963. Die Wirbeltiere des Kamerungebirges. Parey, Hamburg, 353 pp. . 1973. Die Wirbelthiere fauna von Fernando Poo und Westkamerun. Bonner Zool. Monog., 3:1-428. Estes, R. D. 1974. Social organization of the African Bovidae. Pp. 167-205, in The behaviour of ungulates and its relation to management, vol. 1 (V. Geist and F. Walther, eds.), ILfCN Publ., New Ser., 24: 1-51 1. Freedman, L. 1963. A biometric study of Papio cynocephalus skulls from northern Rhodesia and Nyasaland. J. Mamm., 44:24-43. Gabow, S. a. 1975. Behavioral stabilization of a baboon hybrid zone. Amer. Nat., 109:701-712. Groves, C. P. 1967. Geographic variation in the black rhinoc- eros Diceros hicornis (L., 1758). Z. Saugetierk., 32:267-276. . 1971. Distribution and place of origin of the gorilla. Man, n.s., 6:44-51. . 1975. Notes on the gazelles. 1 . G«ze//u and the zoogeography of Central African Bovidae. Z. Saugetierk., 40:308-319. Grubb, P. 1971. Variation and incipient speciation in the African buffalo. Z. Saugetierk., 37:121-144. Hall, B. P., and R. E. Moreau. 1970. An atlas of speciation in African passerine birds. British Museum (Natural Histo- ry), London, xv + 423 pp. Hall, J. B., AND M. D. SwAiNE. 1976. Classification and ecol- ogy of closed-canopy forest in Ghana. J. Ecol., 64:913-951. Hamilton, A. 1976. The significance of patterns of distribution shown by forest plants and animals in tropical Africa for the reconstruction of Upper Pleistocene palaeoenvironments: a review. Palaeoecology of Africa, 9:63-97. Harris, J. M. 1976. Pleistocene Giraffidae (Mammalia, Artio- dactyla) from East Rudolf, Kenya. Pp. 283-332, in Fossil Vertebrates of Africa (R. J. G. Savage and S. C. Coryndon, eds.). Academic Press, London, 4:xi + 1-338. Heim de Balsac, H., and M. Lamotte. 1956. Evolution et phylogenie des soricides africains. I. La lignee Myosorex- Snrdisorex. Mammalia, 20:140-167. . 1957. Evolutionetphylogeniedessoricidesafricains.il. La \ignee Sylvisorex-Siincns-Crocidiira. Mammalia, 21:15- 49. Hill, W. C. O. 1970. Primates. Comparative anatomy and tax- onomy. 8. Cynopithecinae. Cercopithecinae. Edinburgh University Press, Edinburgh, xix + 680 pp. Hollister, N. 1924. East African mammals in the U.S. National Museum. 3. Primates, Artiodactyla, Perissodactyla, Probos- cidea and Hyracoidea. Bull. U.S. Nat. Mus., 99:1-151. Ingoldby, C. M. 1927. Some notes on the African squirrels of the genus Heliosciimis. Proc. Zool. Soc. London, pp. 471- 487. Jotterand, M. 1972. Le polymorphisme chromosomique des Mus (Leggadas) africains. Cytogenetique, zoogeographie, evolution. Rev. Suisse Zool., 79:287-359. Keast, a. 1965. Interrelationships of two zebra species in an overlap zone. J. Mamm., 46:53-66. Kingdon,J. 1971. East African mammals: an atlas of evolution in Africa. Academic Press, London, Lx -I- 1-446. . 1974. The Niger, the Volta and the Dahomey Gap as geographic barriers. Evolution, 12:48-62. Clausen, H. S. 1964. Notes on the distribution of West African cyprinodonts, with special reference to a hitherto unnoticed ichthyogeographical boundary. Vidensk. Midd. fra Dansk Naturh. Foren., 126:323-326. Dekeyser, P. L. 1955. Les mammiferes de I'Afrique Noire Fran- caise. Inst. Franc. Afr. Noire Init. Afr., 1:1-426. DeVree, F., a. DeRoo, and W. N. Verheyen. 1969. Contri- bution a I'etude des Chiropteres de la Republique du Togo. Rev. Zool. Bot. Afr., 80:200-207. DeVree, F., J. Hulselmans, and W. N. Verheyen. 1970. Contribution a I'etude des Chiropteres de la Republique du Togo 2. Liste preliminaire des Chiropteres recoltes par la deuxieme mission zoologique beige au Togo. Rev. Zool. Bot. Afr., 82:41^6. DeVree, F., and E. Van der Straeten. 1971. Contribution a I'etude des Chiropteres de la Republique du Togo 3. Liste preliminaire des Chiropteres recoltes par la troisieme mission zoologique beige au Togo. Rev. Zool. Bot. Afr., 83: 159-164. FitzPatrick, E. a. 1971. Pedology: A systematic approach to soil science. Oliver and Boyd, Edinburgh, Scotland, xvii -I- 306 pp. Livingstone, D. A. 1975. Late Quaternary climatic change in Africa. Ann. Rev. Ecol. Syst., 6:249-280. Marchant, S. 1954. The relation of the southern Nigerian avi- fauna to those of the Upper and Lower Guinea. Ibis, 96:371- 379. Meester, J., AND H. W. Setzer (EDS.). 1971. The mammals of Africa: an identification manual. Smithsonian Inst. Press, Washington, D.C., vii -f 480 pp. Moreau, R. E. 1969. Climatic changes and the distribution of forest vertebrates in West Africa. J. Zool., London, 158:39- 61. Richards, P. W. 1966. The tropical rainforest, an ecological study. Univ. Press, Cambridge, England, xvii -I- 450 pp. Robbins, C. B. 1978. Small mammals of Togo and Dahomey I. Chiroptera. Mammalia, in press. Rosevear, D. R. 1965. The bats of West Africa. British Mus. (Nat. Hist.), London, xvii -I- 418 pp. ScHiOTZ, A. 1967. The treefrogs (Rhacophoridae) of West Af- rica. Spolia Zool. Mus. Haun., 25:1-345. The Atlas of Africa. 1973. France, Inst. Geog. Nat., Paris, The Free Press (MacMillan Publ. Co., New York), 335 pp. A NUMERICAL RE-APPRAISAL OE THE SOUTHERN AFRICAN BIOTIC ZONES I. L. RAUTENBACH Transvaal Museum, Box 413, Pretoria 0001, Republic of South Africa ABSTRACT The acceptability and credibility of the empirically derived is analyzed. A total of six biotic zones, including three that biotic zones of southern Africa are mathematically tested by were formerly regarded as subzones, are found to be viable bio- means of Duellman's (1965) Faunal Resemblance Factor analy- geographic entities of full zonal status, sis. The distribution of 275 southern African mammalian species INTRODUCTION Zoogeography has been defined as . . the sci- entific study of the distribution of animals on earth . . (Udvardy, 1969:1). Through the years a num- ber of attempts have been made to classify animal life into meaningful distributional units, and the field has been subdivided in diverse ways toward different ends. Perhaps the suggested subdivision most pertinent to this paper is that proposed by Darlington (1957:11), who distinguishes three pos- sible levels of approach: 1) geographical distribu- tion over the entire earth; 2) regional distribution over selected segments of the earth; 3) local distri- bution, including species geography (“the geo- graphical distribution of species in relation to each other and to ecology and evolution”). This study will consider a statistical analysis of the distributional trends of southern African mam- mals, in an effort to reevaluate the validity and credibility of the empirically derived biotic zones of the southern subcontinent. It is primarily aimed at the second level of Darlington’s zoogeographical approach, and deals with the regional distribution of the mammals of southern Africa. Unfortunately regional studies such as this are often bound to po- litical, rather than natural, areas. Udvardy (1969:6) distinguishes static faunistic and regional zoogeography from dynamic causal zoogeography. He furthermore distinguishes spe- cifically between zoogeography and ecology. Zoo- geography in its purest sense concerns itself with the reasons for the arrival and settling of a species in a certain area. A study of why and how a species is able to live in that particular area is an ecological problem. Similarly, Simpson (1965:71-73) recog- nizes three levels of zoogeography — geographical, ecological, and historical. Both the zoogeographical and ecological attributes of distribution are the product of evolutionary processes during the course of time; hence, truly explanatory models can only be framed on a historical basis. An accurate and detailed knowledge of subspe- ciation is essential in most modern computations aimed at causal zoogeography. The reverse is, how- ever, also true; a consideration of the biogeography of taxa is important when studying subspeciation. Zoogeography and taxonomy are thus interdepen- dent. The subspecific status of the majority of southern African mammals is, in the modern con- text, unsatisfactorily resolved, and thus severely hampers any detailed and accurate biogeographical analysis. Available analytical procedures based on subspeciation have therefore not been considered in this study pending further detailed survey work and subsequent taxonomic studies on the subspe- cies level. Consequently, this study is essentially limited to Udvardy’s static faunistic and regional zoogeographical approach, and is thus primarily ecological in context. The major biogeographical zones or provinces currently accepted for Africa stem from Sclater (1896). He subdivided Africa into four subregions — the Sahara, West Africa, Cape, and Malagasy. Re- cently, the Malagasy subregion was upgraded to re- gional status (Darlington, 1957). Hence the current concept of the Ethiopian region is Africa south of the Sahara. Chapin (1923, 1932), working on the avifauna of Zaire (formerly Belgian Congo), combined former approaches (Wallace, 1876; Sclater, 1896; Reichen- ow, 1900; Sharpe, 1893) with his own knowledge of the birds of tropical Africa. He divided Zaire into distinct avifaunal regions, which he based on veg- etation types best fitting the distribution of birds. Chapin then attempted to follow these avifaunal re- 175 176 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 gions into adjacent countries, eventually arriving at a subdivision of the Ethiopian region into biogeo- graphical districts. His West African subregion, as well as his East and South African subregions, cor- respond closely to the subregions of Sclater (1896). Both approaches were essentially aimed at a broad separation of tropical forests from savannas and deserts. In Chapin’s (1932) treatise of Africa, southern Africa was subdivided into only two faunal dis- tricts— an eastern and a western-arid district. Chap- in’s work was soon accepted (see Bates, 1924; Lynes, 1924). However, Chapin (1932) himself comments that the least satisfactory portion of his zoogeographical map of Africa is the southern Af- rican district. He considers further subdivision nec- essary here, especially in order to accommodate the highveld grassland and the woodland savanna, as well as tropical montane and coastal forests. Moreau (1952) collates and critically discusses the Tertiary geology and climate of Africa. In this light, he analyzes the distribution of passerine avi- fauna of Africa, firstly by biomes, and secondly with respect to its affinities with the avifaunas of Europe and Asia. He geographically subdivides these main biomes into smaller biotic zones. He considers affinities both on generic and specific levels, and found differences between these derived biotic zones to be great. With regard to southern Africa, Moreau (1952) retains Chapin’s (1923) South West Arid district more or less unaltered, as a biotic zone. However, Moreau introduces the concept of the Southern Sa- vanna biotic zone, which combines Chapin’s east- ern, climatically moderate, woodland districts. As suggested by Chapin (1932), Moreau now also rec- ognizes montane forests as distinct on a biotic zone level. He furthermore recognizes the small, but flo- ristically very rich and distinct winter rainfall area around Cape Town, with its Mediterranean climate, as a separate biotic zone. Moreau (1952) uses both the terms “biotic zone’’ and “biome.” The first term has a definite zoogeo- graphical connotation, the second ecological. Ac- cording to Smith (1966), the biotic province (zone) concept “. . . embraces a continuous geographic area that contains ecological associations distin- guishable from those of adjacent provinces (zones), especially at the species and subspecies level . . . .’’ The biome, on the other hand, is a major eco- system, and is seen by Smith as ”... a broad eco- logical unit characterized by the distinetive life forms of the climax species, plant or animal ....’’ Southern African biotic zones can in reality also be seen as biomes, except that as such they are only parts of the major biomes of Africa. Whatever the case, Moreau can be credited to be one of the first to employ the correct terminology in an African zo- ogeographical treatise, with consideration to con- cepts and terms developed in related fields such as ecology. This distinction between “biotic zone’’ and “biome” is recognized and applied in this re- port. Davis (1962) employs the southern African por- tion of Moreau’s (1952) biogeographical map in an analysis of distribution patterns of the local Mu- ridae. He agrees with Moreau in the validity of the South Western Cape as a biotic zone. However, Davis’ (1962) main zoogeographical contribution lies in the fact that for the first time the biotic zones, which were founded on avifaunal distributional data, are analyzed from a mammalian point of view, albeit on only one family. Davis slightly alters the borders of the biotic zones to conform with the veg- etation map of Keay (1959). In a discussion of the origins of the southern Af- rican mammalian fauna, Meester (1965) accepts Davis’ modified version of Moreau’s biotic zones. Although Davis’ generalized attempts to subdivide the biotic zones went unnoticed, Meester’s definite recognition of the Namib as a subzone of the South West Arid, and the Grassland as a subzone of the Southern Savanna, was soon accepted. This ap- proach to the recognition of biotic zones is even more compatible with the biome concept. There are other proposed systems for subdividing the subcontinent into major biogeographic units, especially those of Liversidge (1962) and Winter- bottom (1962). However, the biotic zone concept as outlined above has become commonly accepted as relevant from an ecological viewpoint, especially with regard to higher vertebrates. It is also, to my knowledge, the only zoogeographical system con- sidered in recent years for work on mammal distri- bution (see Davis, 1962; Meester, 1965). It is there- fore appropriate that biotic zones should receive closer scrutiny here, especially because no less than four currently recognized major biotic zones (two with two subzones each) are represented in south- ern Africa. Meester’s (1965) refined version of Mo- reau’s (1952) biotic zones is analyzed in this report. 1978 RAUTENBACH— SOUTHERN AFRICAN BIOTIC ZONES 177 METHODS The African biotic zones and subzones have been empirically derived by considering main vegetation types and how they best fit the distribution of species, initially of birds and later of mam- mals. The zones are thus largely subjective. A number of species may be confined to a single biotic zone (endemics), but very few have ranges coinciding entirely with the boundaries of the par- ticular zone in which they occur. Generally their ranges are more restricted. Such endemic species are few in number, yet serve as the main argument to justify the recognition of the biotic zone. The majority of species occurs over several biotic zones, be- cause the distributions are limited by factors more generalized than those governing the vegetation types on which the biotic zones are primarily based. These widespread species apparently formerly served no role in justifying the recognition of biotic zones. Duellman (1965:677) proposes a statistical analysis to express the validity of biogeographical subdivisions, based on the known distribution of all species in the entire area. He termed it the “Faunal Resemblance Factor," which is statistically expressed as FRF = 2C/N I -I- N2, where C equals the number of species in common between the two zones compared, N1 equals the number of species in the first zone, and N2 equals the number of species in the second zone. An index value of 0.000 would indicate no taxonomic resemblance between two zonal faunas, and an index of 1.000 would indicate complete identity. A value of 0.500 would indicate that one-half of the species in each of the two zonal faunas are held in common, provided that they are of equal size. In the case of unequal-sized faunas, both dissim- ilarity in species composition and relative equality in species density are expressed. Duellman’s ( 1965) formula is a simplified, yet equally effective, derivate of the Burt coefficient (Burt, 1958). Both formulas take the average of the two samples as the denominator (contrary to the Simpson and Jaccard coefficients, see Simpson, I960), in an effort to reduce the effect of difference in size between them. However, the influence of differential fau- nal sizes is not entirely eliminated, and is yet another factor expressing similarity or dissimilarity between zones. These for- mulas are furthermore designed for taxa of which the geograph- ical distributions of species are not well known. Only the pres- ence or absence of taxa is of great importance. The Duellman coefficient is therefore ideally suited for this analysis, and was decided upon as being the simplest of the two mentioned here. RESULTS The distributions of 275 species of southern Af- rican mammals are given in Table 1. This list was compiled from updated but unpublished distribution maps kept in the Transvaal Museum for curatorial purposes, as well as from the literature, particularly Smithers (1971), Meester and Setzer (1971-1977), Davis (1974), Pringle (1974), Lynch (1975), and Smithers and Lobao Tello (1976). The taxonomic treatise of Meester and Setzer (1971) was followed. In the calculation of FRF indices, the Southern Sa- vanna Grassland and Woodland subzones and the Namib subzone were treated as hypothetically valid zones, as indicated in Table 1. The list excludes feral and exotic species, as well as poorly known endemics of doubtful taxonomic status. As far as possible, the natural (historic) ranges of species were considered, thus compensating for human im- pact. Species with extremely limited ranges, or known from only a few localities, were considered as representative of the biotic zone in which they occur. Where the majority of localities for a species fall within a given zone, with only a few isolated instances falling just inside an adjacent zone, these were considered as typical only of the zone where the distribution is concentrated, and not as a con- stituent of the mammal fauna of the second zone. However, if such scattered localities are deep into the second zone, they were considered typical of that zonal fauna as well. Judgment was subjective. Typical Forest zone species occurring outside that zone, but restricted to riverine forests, were con- sidered as pure forest zone species. However, the influence of dispersal corridors, such as the Kuiseb and Orange rivers, was not taken into account. The distribution of bats as a group is particularly poorly documented, which may adversely influence the results of this analysis. Excluding bats was con- sidered. However, certain mammalian taxa, as well as nutritionally and ecologically adapted groups, demonstrate diverse latitudinal clinal trends in com- position and densities (Nel, 1975). Thus it was de- cided to include the meager information on bats in an effort to retain a more balanced image of trends in overall mammalian ecological distribution. A simple matrix of similarity, indicating the de- gree of interrelationships of mammalian faunas of southern African biotic zones, is given in Table 2. Absolute numbers of species in common are indi- cated below the diagonal. Italic numerals on the diagonal indicate the number of species in each zone, and the bracketed numerals underneath these denote the known number of endemic species. Above the diagonal is an index of faunal resem- blance, calculated after Duellman (1965). 178 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 1 . — Distribution of southern African mammals according to Biotic Zones. Taxa Biotic Zones Namib South West Arid South West Cape Southern Savanna Woodland Southern Savanna Grassland Forest Petrodromus tetradactylus — — — X X Macroscelides prohoscideus X X — — — — Elephantulus intufi X X — X — — Elephantulus rupestris — X — — — — Elephantulus myurus — — — X X — Elephantulus edwardi — X — — — — Elephantulus hrachyrhynchus — — — X X — Erinaceus frontalis — X — X X — Myosorex varius — — X X X X Myosorex cafer — — — X X X Suncus lixus — — — X — — Suncus gracilis — - — — X — Sylvisorex megalura — — — — — X Crocidura occidentalis — — — X — X Crocidura flavescens — — X — X X Crocidura luna — — — — — X Crocidura mariquensis — — — X X — Crocidura hirta — X — X X X Crocidura silacea — — — X X X Crocidura cyanea X X — X X X Crocidura maquassiensis — — — X X — Crocidura bicolor — X — X — — Chrysospalax trevelyani — — — X — X Chrysospalax villosus — — — X X X Cryptochloris wintoni — X — — — — Cryptochloris zyli — — X — — — Chrysochloris asiatica — X — — — — C hrysochloris visagiei — X — — — — Eremitalpa granti X X X — — — Chlorotalpa sclateri — X — — X — Chlorotalpa duthiae — — — — — X Chlorotalpa arendsi — — — — — X Galcochloris obtusirostris — — — X — — Amblysomus gunningi — — — — — X Amblysomus hottentotus — — — X X X Amblysomus iris — — — X X X Amblysomus julianae — — — X X — Eidolon helvum — X — X X — Epomophorus wahlbergi — — — X — X Epomophorus gambianus — — — X — — Epomophorus crypturus — — — X — — Epomophorus angolensis — — — X — — Rousettus aegyptiacus — — — X — X Rousettus angolensis — — — — — X Taphozotis mauritianus — — — X — — Taphozous perforatus — — — X — — Coleura afra — — — X — — Nycteris hispida — — — — — X Nycteris grandis — — — — — X Nycteris macrotis — — — X — X Nycteris woodi — — — X — X Nycteris thebaica X X — X — X Rhinolophus hildebrandti — — — X — X Rhinolophus fumigatus — X — X — — Rhinolophus clivosus — — X X X X 1978 RAUTENBACH— SOUTHERN AFRICAN BIOTIC ZONES 179 Table 1. — Continued. Taxa Biotic Zones Namib South West Arid South West Cape Southern Savanna Woodland Southern Savanna Grassland Forest Rhinolophus darlingi X X — X — — Rhinolophus landeri — — — X — — Rhinolophus hlasii — — — X — — Rhinolophus capensis — X X X — — Rhinolophus simulator — — — X — — Rhinolophus denti — X — X — — Rhinolophus swinnyi — — — X — — Hipposideros comrnersoni — X — X — — Hipposideros caffer — X — X — X Triaenops persicus — — — X — — Cloeotis percivali — — — X — — Myotis welwitschii — — — X X — Myotis seabrai — X — — — — Myotis lesueuri — — X — — — Myotis tricolor — X — X X — Myotis bocagei — — — X — — Nycticeius schlieffeni — X — X — — Pipistrellus nanus — — — X — X Pipistrellus kuhli — — — X — X Pipistrellus rusticus — — - X — X Pipistrellus rueppelli — — — X — — Eptesicus rendalli — — — X — — Eptesicus hottentotus X X — X — — Eptesicus melckorum — — X — — — Eptesicus zuluensis X X — X — — Eptesicus somalicus — — — X — — Eptesicus capensis — X — X X — Eptesicus notius — — X — — — Glauconycterus variegata — — — X — — Laephotis wintoni — — — X — — Scotophilus gigas — — — X — — Scotophilus nigrita — — — X — — Scotophilus leucogaster — X — X — — Kerivoula argentata — — — X — — Kerivoula harrisoni — — — X — — Kerivoula lanosa — — — X — X Miniopterus fraterculus — — — X — — Miniopterus schreihersi X X — X X X Olomops martiensseni — — — X — — Sauromys petrophilus — X — X — — Tadarida acetabulosus — — — X — — Tadarida inidas — X — X — — Tadarida niveiventer — — — X — — Tadarida condylura — — — X — X Tadarida nigeriae — X — X — — Tadarida chapini — — — X — — Tadarida puinila — — — X — — Tadarida fulminans — — — X — — Tadarida aegyptiaca — X X X X X Tadarida ansorgei — — — X — — Galago crassicaudatus — — — X — X Galago senegalensis — X — X — X Papio cynocephalus — — — X — — Papio ursinus X X X X X X Cercopithecus mitis — — — X — X 180 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Table 1. — Continued. Taxa Biotic Zones Namib South West Arid South West Cape Southern Savanna Woodland Southern Savanna Grassland Forest Cercopithecus aethiops — X — X — X Manis temmincki — X — X X — Otocyon megalotis — X — X — — Vulpes chama X X — X X — Canis mesontelas X X — X X — Canis adustus — — — X — — Lycaon pictus — X — X — — Ictonyx striatus X X — X — — Poecilogale athinucha — X — X X — Mellivora capensis — X — X — — Lutra macuHcollis — X — X X — Aonyx capensis — X — X X — Nandinia hinotata — — — — — X Viverra civetta — — — X — — Genetta genetta X X X X X — Genetta tigrina — — X X — X Genetta ruhiginosa — — — X — — Genetta mossambica — — — X — — Suricata suricatta X X — — X — Paracynictis selousi — — — X — — Bdeogale crassicauda — — — X — — Cynictis penicillata — X — X X — Herpestes ichenumon — — — X — — Herpestes pulverulentus — X — — X — Herpestes sanquineus — X — X X X Herpestes ratlamuchi — X — X — — Rhynchogale melleri — — — X — X Ichneumia alhicauda — — — X — X Atilax paludinosus — X X X X X Mangos mango — X — X — — Helogale parvuta — X — X — — Proteles cristatus X X — X X — Hyaena hrannea X X — X X — Crocata crocata X X — X X — Felis libyca X X X X X — Felis nigripes — X — — X — Felis serval — — X X — — Felis caracal X X — X X — Panthera pardus X X X X X X Panthera leo — X X X X — Acinonyx jahatus — X — X — — Orycteropus afer — X — X X — Loxodonta africana — X — X — — Procavia capensis X X X X X X Procavia welwitschii X X — — — — Heterohyrax hrucei — — — X — — Dendrohyrax arboreas — — — X — — Diceros bicornis — X X X — — Ceratotheriam simum — X — X X — Equus zebra X X — — — — Eqaas barchelli — X X X X — Potamochoerus porcas — — X X — X Phacochoerus aethiopicus — X — X — — Hippopotamus amphibius — X X X X — Giraffa Camelopardalis — X — X — — 1978 RAUTENBACH— SOUTHERN AFRICAN BIOTIC ZONES 181 Table 1. — Continued. Taxa Biotic Zones Namib South West Arid South West Cape Southern Savanna Woodland Southern Savanna Grassland Forest Cephalophus natalensis — — — X — X Cephalophus monticola — — — X — X Sylvicapra grimmia — X X X X — Raphicerus campestris — X X X X — Raphicenis melanotis — — X X — — Raphicerus sharpei — — — X — — Ourebia ourebi — — — X X — Neotragus moschatus — — — X — — Oreotragus oreotragus X X — X — — Madoqua kirki — X — — — — Pelea capreolus — X X X X — Redunca arundinum — — — X — — Redunca fulvorufula — — — X X — Kobus ellipsiprymnus — X — X — — Kobus vardoni — — — X — — Kobus leche — — — X — — Aepyceros melampus — X — X — — Aepyceros petersi — X — X — — Antidorcas marsupialis X X — — X — Oryx gazelki X X — X — — Hippotragus leucophaeus — — X — — — Hippotragus niger — — — X — — Hippotragus equinus — — — X — — Damaliscus lunatus — — — X — — Damaliscus dorcas dorcas — — X — — — Damaliscus dorcas phillipsi — X — — X — Alcelaphus buselaphus — X — X X — Alcelaphus lichtensteini — X — X — — Connochaetes taurinus — X — X — — Connochaetes gnou — — — — X — Tragelaphus scriptus — — X X — — Tragelaphus spekei — — — X — — Tragelaphus angasi — — — X — — Tragelaphus strepsiceros — X — X — — Taurotragus oryx — X X X X — Syncerus caffer — — — X — — Lepus capensis — X — X X — Lepus saxatilis — X — X X — Bunolagus monticularis — X — — — — Pronolagus crassicaudatus — — — X — — Pronolagus rupestris — X — X X — Pronolagus randensis — X — X X — Bathyergiis Janetta — X — — — — Bathyergus suillus — — X — — X Georychus capensis — X X — X — Cryptomys damarensis — X — — — — Cryptomys hottentotus — X — X X X Hystrix africaeaustralis X X — X X X Petromus typicus X X — — — — Thryonomys swinderianus — — — X — — Thryonomys gregorianus — — — X — — Xerus inauris — X — — X — Xerus princeps — X — — — — Heliosciurus rufobrachium — — — X — X Funisciurus congicus — X — — — — 182 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 Table 1. — Continued. Taxa Biotic Zones Namib South West Arid South West Cape Southern Savanna Woodland Southern Savanna Grassland Forest Para.xerus palliatus — - — — — X Paraxerus cepapi — — — X — — Pedetes capensis X X — X X — Graphiurus ocularis — X X — — — Graphiurus platyops — X — X X — Graphiurus murinus — X — X X X Cricetomys ganibianus — — — X — X Dendromus nyikae — — — — — X Dendromus melanotis — X X X X X Dendromus mesomelas — — X X X X Dendromus mystacalis — — — X X X Malacothrix typica — X — — X — Mystrornys albicaudatus — — X — X — Petromysciis monticularis — X — — — — Petromyscus collinus X X — — — — Saccostomus campestris — X X X X — Steatomys pratensis — X — X — — Steatomys krebsi — X — — X — Steatomys minutus — X — X X — Acomys spinosissimus — — — X — X Acomys subspinosus — — X — — X Aethomys granti — X — — — — Aethomys namaquensis X X X X X — Aethomys chrysophilus — X — X — — Aethomys nyikae — — — X — — Dasymus incomtus — — X X X — Mas indutus — — — X X — Mus minutoides X X X X X X Lemniscomys griselda — X — X — — Pelomys fallax — — — X — X Praomys natalensis — X X X X X Praomys shortridgei — — — X — — Praomys rerreauxi — — X — — X Rhabdornys pumilio X X X X X X Thallomys paedulcus X X — X — — Thamnomys cometes — — — — — X Thamnomys dolichurus — — — X — X Zelotomys woosnami — X — — — — Parotomys brantsi X X X - — — Parotomys littledalei X X — — — — Otomys laminatus — — — X X X Otomys angoniensis — X — X X — Otomys saundersiae — — X — X — Otomys irroratus — X X X X X Otomys sloggetti — — — — X — Otomys unisulcatus — X — — — — Desmodillus auricularis X X — — — — Gerbillurus vallinus X X — X — — Gerbillurus tytonis X — — — — — Gerbillurus paeba X X X X — — Gerbillurus setzeri X — — — — — Patera leucogaster X X — X — — Patera afra — — X — — — Patera brantsii — X — X X — Patera inclusa — — — X — — 1978 RAUTENBACH— SOUTHERN AFRICAN BIOTIC ZONES 183 Table 1. — Continued. Biotic Zones Southern Southern South West South West Savanna Savanna Taxa Namib Arid Cape Woodland Grassland Forest Total (275 species) 43 136 50 209 91 73 Percentage of total fauna 15.6 49.5 18.2 76.0 33.1 26.6 Total no. of endemic species Percentage endemics to 2 16 7 60 3 12 total zonal fauna 4.65 11.76 14.00 28.71 3.30 16.44 DISCUSSIONS It must be stressed that subcontinental distribu- tional data is as yet incomplete for the majority of species, particularly so in the Cape Province and South West Africa. Furthermore, the accuracy of this analysis will be greatly enhanced if conducted on the subspecies level, rather than on a species level. This ideal will be delayed for many years as a result of the unsatisfactory status of the knowl- edge of subspeciation in southern African mam- mals. On the other hand, a more intimate knowledge of the distribution patterns of species does not nec- essarily imply a high incidence of range extensions into biotic zones where they have previously been unrecorded. When species’ geographic ranges are better known and the occurrence of not too many species are recorded in new zones, the results of this analysis will not change dramatically. A more accurate FRF analysis facilitated by subspecies consideration will probably only enhance the find- ings of this treatment because a higher degree of endemism is expected. Whatever the case, the fol- lowing points are pertinent from Table 2 and war- rant further comment here, especially with regard to my aim to assess the validity of biotic zones as viable biogeographical areas. Superficially, the FRF indices of all zones under consideration are low enough to warrant their con- sideration as distinct zones (Table 2). Closer scru- tiny is however essential. The Namib is most closely related to the South West Arid, albeit with a FRF index as low as 0.458. The Namib’s FRF indices when calculated against the other zones are, however, much lower, which confirms distinctness from these. The Namib pos- sesses only two endemic species (Table 1), namely G. tytonis and G. setzeri . However, by far the great- est majority (41) of the Namib’s total mammal fauna (43) consists of a faction of the bigger South West Arid fauna (some species also occur elsewhere). Table 2. — Resemblance of mammalian faunas of the six southern African Biotic Zones and Suhzones. (See text for explanation; italic numerals on diagonal indicate total number of species in zone, the numerals in brackets underneath these denote the known number of endemic species). Biotic Zones Southern Savanna Woodland South West Arid Southern Savanna Grassland Forest South West Cape Namib Southern Savanna Woodland 209 (60) .580 .500 .404 .247 .238 South West Arid 100 136 (16) .573 .201 .290 .458 Southern Savanna Grassland 75 65 91 (3) .341 .411 .299 Forest 57 21 28 73 (12) .309 .138 South West Cape 32 27 29 19 50 (7) .237 Namib 30 41 20 8 11 43 (2) 184 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Consequently it can be considered as merely a de- pauperate fauna of the latter, resulting from the in- hospitable nature of the Namib. The Namib’s bio- geographical uniqueness thus lies not so much in its typical endemic fauna, or its faunal composition for that matter, but rather in the fauna it does not pos- sess. The Namib is therefore considered here as a biotic zone of full rank. Detailed analysis has shown that the Namib can be further subdivided, on the basis of the sand dunes being faunistically more de- pauperate than the gravel plains (see Coetzee, 1969). In spite of its tremendous floral diversity, the South Western Cape is also very depauperate in mammalian fauna. However, it must be pointed out that this is, in terms of intensive mammal survey- ing, the most neglected biotic zone of all. It has seven endemic taxa as far as is known — C. zyli, M. leseuri, E. melckonun, E. notiiis, D. d. dorcas, T. afra, and the extinct H. leucophaeus (see Table 1). The remainder of the faunal element is made up of mammalian species shared with other biotic zones. This zone shares 32 species with the Woodland zone, and 29 with the Grassland zone. However, as a result of the enormous differential species diver- sity between the Woodland and the South West Cape, the FRF analysis indicates a closer resem- blance between the latter zone and the less diver- sified Grassland zone, with an index of 0.411. It is interesting to note that the South West Cape and the Grassland zones are unconnected. The Forest zone is also quite distinct from the others. It is faunistically most closely related to the Southern Savanna Woodland subzone with a FRF index of 0.401. It possesses 12 endemic species (see Table 1). The Woodland and Grassland zones and the South West Arid zone are the three areas related more closely to each other than any other combi- nation of zones. Their individual FRF indices in relation to each other are however considered low enough to warrant their individual recognition. Be- cause the Grassland has been considered a subzone of the Southern Savanna biotic zone, closest resem- blance is expected between it and the related Wood- land subzone. This is, however, not the case. Both in terms of absolute number of species in common and FRF index, the Woodland and South West Arid are faunistically most closely related (100 species in common; FRF index 0.580). This is followed by a closer resemblance between Grassland and South West Arid in terms of FRF indices (0.573), but in terms of number of species in common, a closer resemblance between Grassland and Woodland (75 species). This inconsistency can be ascribed to the disproportionate sizes of the three zonal faunas and as compensated for by Duellman’s formula, espe- cially designed for such instances. The Southern Savanna Woodland has by far the richest mamma- lian fauna; 209 species, representing 76.0% of the total southern African mammalian fauna, occur here, including 60 endemics (predominantly bats). This is followed by the South West Arid, with a total di- versity of 136 species, that is 49.5% of the total of 275 southern African species, with 16 endemics. The Savanna Woodland undoubtedly offers the highest variety of habitats, being ecologically more diversified both horizontally and vertically. Its rich species diversity could be related to this fact more than any other. The temptation is great to assume that the re- spective faunal elements of other zones have orig- inated by a radiation of Woodland-adapted species. Undoubtedly this is true in many instances, espe- cially in the case of species, which do not rely on trees as an integral element in their habitat require- ments. On the other hand, the high number of en- demics typical of the Woodland and South West Arid areas combined (76 species) can be interpreted as a faunal element specialized towards a depend- ence on woodland in some manner or other. The fact that such a large portion (100 species) of the non-endemic fauna of the Woodland apparently ra- diated adaptively into the South West Arid is re- flected by the highest FRF index of all (0.580). Mostly due to lower average annual precipitation, the latter zone has a less developed woodland flora, and consequently a less diversified mammalian fau- na. The same situation could also be demonstrated with Grassland-adapted species finding suitable habitat in adjacent Woodland Savanna (FRF index 0.500) and South West Arid (FRF index 0.573). Forest is scattered through three zones (Southern Savanna Woodland, Southern Savanna Grassland, and South West Cape), and has higher FRF indices with these than with the non-adjacent Namib and South West Arid. This trend of a relatively higher FRF index reflecting a sharing of species between adjacent zones, numerically radiating clinally from the Woodland Savanna, appears to be the rule. There is one exception, that is South West Cape being faunistically closest to the non-adjacent Grassland, with a FRF index of 0.423. 1978 RAUTENBACH— SOUTHERN AERICAN BIOTIC ZONES 185 Nel (1975) found an almost linear correlation be- tween number of species and mean annual precip- itation in a latitudinal direction in southern Africa. The result is a low-to-high gradient in species den- sities from west to east, as mean annual rainfall increases. This is particularly the case with bats. Nel could also find no real correlation between spe- cies density and altitude. The altitudinal profile of southern Africa is relatively low, which probably explains this phenomenon. This, however, needs closer study to confirm its validity. It would appear from the results of this analysis that a low-to-high gradient in species densities could also be demonstrated in a south to north di- rection. Species densities increase from 50 in the South West Cape, to 91 in the Grassland, to 136 in the South West Arid, to 209 in the Woodland. Al- though rainfall again undoubtedly plays some role in this trend, other causal factors such as decreasing latitude, temperature, faunal origin, and dispersion, will have to be considered in a more detailed anal- ysis. CONCLUSIONS 1) Six biotic zones are recognized as viable bio- geographical entities, as deduced from this analysis. Where the Grassland and the Woodland have for- merly been regarded as subzones of the then South- ern Savanna biotic zone, terminology may hence be confusing when referring to these as biotic zones of full rank. In order to retain the Pan-African impli- cations and perspective of the term Southern Sa- vanna, I suggest that these two biotic zones be known as the Southern Savanna Woodland and the Southern Savanna Grassland biotic zones. This sug- gestion is made in the full realization that in the latter case, the definition of a savanna is stretched to the limit. Terminology for the Namib biotic zone remains unchanged, indicating its elevated zonal status. 2) Biotic zones are here regarded as the largest biogeographic units in which southern Africa could be subdivided, that is Southern Savanna Woodland, Southern Savanna Grassland, Eorest, Namib, South West Arid, and South West Cape. 3) Very few species have such a wide habitat tol- erance that they occur in all biotic zones. Ende- mism is, on the other hand, equally as unusual. In the majority of instances, species are shared be- tween various combinations of zones, and the unique feature of the FRF analysis is to take this into account, apart from endemism. Therefore, a high FRF index indicates a high incidence of shared species and therefore closer faunal similarity. 4) The FRF analysis in fact takes three charac- teristics into consideration when expressing the fau- nal distinctness of a zone: the respective species densities of the two zones under consideration; the number of species in common; and indirectly the number of distinctive species of each zone. Based on the results of the FRF analysis, the Namib zone is deduced to be fully distinct from the South West Arid. It is considered a bona fide biotic zone in full realization of the fact that it has a very small dis- tinctive fauna. The Grassland is similarly consid- ered to be a distinct biotic zone, rather than a sub- zone. In both these instances, one of the previous considerations for their recognition as zones of low- er rank was the low degree of endemism. 5) The six biotic zones recognized here as bio- geographical entities, correlate very well with what I regard as major ecological biomes in southern Africa. 6) The FRF indices of the South West Arid, the Woodland, and the Grassland, as compared with each other, are all over 0.500. There is no estab- lished value over which a zone cannot be consid- ered statistically valid, and judgment is therefore subjective. The FRF value of these zones in ques- tion is here considered low enough to warrant their recognition as valid biotic zones. In compari- son Armstrong (1972) considers Merriam’s (1890) life zones, which the former author tested with Duellman’s FRF analysis, as valid with indices as high as 0.847. In the present analysis, the generally lower FRF indices could also be ascribed to dispro- portionate faunal densities between certain zones. These differences between the sizes of zonal faunas are here regarded as valid criteria in considering the rank of a particular zone. 7) With the exception of seven species, the re- mainder of the 68 southern African bat species are all recorded from the Southern Savanna Woodland, among other zones. The presence of the bat fauna in the other biotic zones is dramatically less (five in the Namib; 20 in the South West Arid; six in the South West Cape; seven in the Grassland; and 18 in the Forest). The Chiroptera is the least known group of mammals in southern Africa, and although Duellman’s (1965) formula partly compensates for 186 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 this shortcoming, it has an undetermined bearing on the accuracy of the above observation. The infer- ence is, however, that as a group bats have a re- markable attachment to Woodland Savanna. 8) Biotic zones are empirically derived by con- sideration of major vegetation types. Although the former are proved to be significant from a faunal point of view, it does not necessarily reflect the best way to describe faunal distribution patterns, espe- cially because individual geographic ranges of spe- cies seldom overlap entirely with any biotic zone. Biotic zones as significant biogeographical entities should rather, from a faunal point of view, be seen as illustrating the gross direct relationship of the animal to its floral environment, and to a lesser ex- This paper is introductory to a detailed zoogeographical anal- ysis of distribution patterns of Transvaal mammals. It is aimed at establishing the credibility of biotic zones as major biogeo- graphical entities within this Province in order to facilitate fur- ther subdivision into community-types. The research on the dis- tribution of Transvaal mammals has been jointly sponsored by the Transvaal Museum and the South African Council for Sci- entific and Industrial Research, to whom I express my gratitude. tent, to the physical environment. In the context of biotic zones, faunal interrelationships should there- fore be interpreted from an ecological point of view. 9) Continued intensive mammal surveying in southern Africa is considered essential for a better understanding of both subspeciation and zoogeo- graphieal interpretation, through more detailed analysis. Especially the Chiroptera throughout the subcontinent, and the faunas of South West Arid and South West Cape biotic zones, need intensive attention in terms of surveying. It is, however, not expected that a more intimate knowledge of these aspects will dramatically change the results and im- plications of this analysis. I am also grateful to Prof. J. Meester and Prof. J. A. J. Nel for critically reading the manuscript and making helpful suggestions. An invitation by the Carnegie Museum of Natural History to attend a colloquium on the taxonomy and ecology of the small mammals of Africa, where this paper was presented, was grate- fully accepted. I also thank the S.A. Department of National Education and the Transvaal Museum for travel grants to and from Pittsburgh, as well as Mrs. E. du Plooy for typing the manuscript. LITERATURE CITED Armstrong, D. M. 1972. Distribution of mammals in Colorado. Monogr. Mus. Nat. Hist., Univ. Kansas, 3:x -I- 1-415. Bates, G. L. 1924. On the birds collected in north-western and northern Cameroon and parts of northern Nigeria. Ibis, 6: 1- 45. Burt, W. H. 1958. The history and affinities of the Recent land mammals of western North America. Pp. 131-154, in Zoo- geography (C. L. Hubbs, ed.), Publ. Amer. Assoc. Advance- ment Sci., 51:x -I- 1-509. Chapin, J. P. 1923. Ecological aspects of bird distribution in trop- ical Africa. Amer. Nat., 57:106-125. . 1932. Birds of the Belgian Congo. Bull. Amer. Mus. Nat. Hist., 65:1-756. CoETZEE, C. G. 1969. The distribution of mammals in the Namib desert and adjoining inland escarpment. Scient. Papers Na- mib Desert Res. Station, 40:23-36. Darlington, P. J. 1957. Zoogeography: the geographical distri- bution of animals. John Wiley & Sons, New York, xi -I- 675 pp. Davis, D. H. S. 1962. Distribution patterns of southern African Muridae, with notes on some of their fossil antecendants. Ann. Cape Prov. Mus., 2:56-76. . 1974. The distribution of some smaller southern African mammals {Mammalia: Insectivora, Rodentia). Ann. Trans- vaal Mus., 29:136-184. Duellman. W. E. 1965. A biogeographic account in the herpe- tofauna of Michoacan, Mexico. Univ. Kansas Publ., Mus. Nat. Hist., 15:627-709. Keay, R. W. j. 1959. Vegetation map of Africa south of the Tropic of Cancer. L' Association pour I'Etude Taxonomique de la Flore d'Afrique Tropicale, Oxford Univ. Press, 24 pp., map. Liversidge, R. 1962. Distribution of birds in relation to vegeta- tion. Ann. Cape Prov. Mus., 2:143-151. Lynch, C. D. 1975. The distribution of mammals in the Orange Free State, South Africa. Navorsinge van die Nasionale Mu- seum, Bloemfontein, 3:109-139. Lynes, H. 1924. On the birds of north and central Darfur, with notes on the west-central Kordofan and north Nuba prov- inces of British Sudan. Ibis, 6:399-446. Meester, J. 1965. The origins of the southern African mammal fauna. Zool. Afr., 1:87-95. Meester, J., and H. W. Setzer. 1971-1977. The mammals of Africa: an identification manual. Smithsonian Inst. Press, Washington, D.C. Merriam, C. H. 1890. Results of a biological survey of the San Francisco mountain region and desert of the little Colorado in Arizona, I. General results . . . with special reference to the distribution of species. N. Amer. Fauna, 3:5-34. Moreau, R.E. 1952. Africa since the Mesozoic: with particular reference to certain biological problems. Proc. Zool. Soc. London, 121:869-913. Nel, j. a. J. 1975. Species density and Ecological diversity of 1978 RAUTENBACH— SOUTHERN AFRICAN BIOTIC ZONES 187 South African mammal communities. South African J. Sci., 71:168-170. Pringle, J. A. 1974. The distribution of mammals in Natal. Part I. Primates, Hyracoidea, Lagomorpha (except Phol- idota and Tubilidentata. Ann. Natal Mus., 22:173-186. Reichenow, a. 1900-1905. Die vogel Afrikas. 3 Volumes and Atlas. ScLATER, W. L. 1896. The geography of mammals. IV. The Ethiopian region. Geogr. J., 7:282-296. Sharpe, R. B. 1893. II. On the zoo-geographical areas of the world, illustrating the distribution of birds. Nat. Sci., 3:100- 108. Simpson, G. G. 1960. Notes on the measurements of faunal re- semblance. Amer. J. Sci., 258:300-311. . 1965. The geography of evolution, collected essays. Capricorn Books, New York, 249 pp. Smith, R. L. 1966. Ecology and field biology. Harper & Row, New York and London, xiv -I- 686 pp. Smithers, R. H. N. 1971. The mammals of Botswana. Mus. Mem., Nat. Mus. Rhodesia, 4:1-340. Smithers, R. H. N., and J. L. P. Lobao Tello. 1976. Check list and atlas of the mammals of Mocambique. Mus. Mem., Nat. Mus. and Monuments Rhodesia, 8:viii -i- 1-184. Udvardy, M. D. F. 1969. Dynamic zoogeography, with special reference to land animals. Illustrated by Charles S. Papp. Van Nostrand Reinhold Co., New York, 445 pp. Wallace, A. R. 1876. The geographical distribution of animals. Harper, New York, 1 :xxiii -i- 1-503; 2: vi -i- 1-553 (reprinted 1962 by Hafner, New York and London). WiNTERBOTTOM, J. M. 1962. A note on zoogeographical limits in South-East Africa, as suggested by the avifauna. Ann. Cape Prov. Mus., 2:152-154. KARYOTYPIC DATA FOR AFRICAN MAMMALS, WITH A DESCRIPTION OF AN IN VIVO BONE MARROW TECHNIQUE LYNN W. ROBBINS Museum of the High Plains, Fort Hays State University, Hays, Kansas 67601 U.S.A. ROBERT J. BAKER The Museum, Texas Tech University, Lubbock, Texas 79409 U.S.A. ABSTRACT Basic information, which should be included in any publica- view of the literature concerning karyotypes of African mammals tion on the chromosomes of mammals, is given. A field-tested was done and these karyotypic data are listed for 292 species bone marrow-/« vivo method of karyotyping is presented. A re- and subspecies of African mammals. INTRODUCTION Karyotypes have proven to be valuable data for evolutionary and systematic studies. A summary of chromosomal data for African mammals is pre- sented in Appendix I. Eor maximum value, any publication on the chromosomes of a species should contain the following information: 1. A photomicrograph of the karyotype. This is necessary if this is the first report for a species or if your karyotypic data differs from that pre- viously published for the species. 2. Diploid Number. 3. “Nombre fundamental” or number of arms of the autosomal complement. 4. Morphology of sex chromosomes. 5. Sex of specimens examined. 6. Number of specimens examined. 7. Geographic origin of specimen examined. 8. Museum where voucher specimens are deposit- ed (with museum numbers, if possible). 9. Minimum number of spreads examined from any specimen included. METHODS AND TECHNIQUES Preparation of somatic chromosomes is a simple process which can be conducted in the field. For the bone marrow-m vivo technique described be- low, live animals are required. The following tech- nique is modified after Baker (1970). 1. Inject the live animal intraperitoneally with a 0.03% Vinblastine (Velban of Eli Lilly & Co.) or colchicine solution at 0.01 ml per gram of body weight. 2. After two hours sacrifice the animal and re- move a long bone, such as the femur in rodents or the humerus in bats, without damaging the proximal end. Remove the flesh and a chip of bone from the proximal end to expose the red bone marrow cav- ity. Flush the shaft with 3 ml of a 1.0% sodium citrate solution. Pipette vigorously to break up any cell clumps. The sodium citrate solution will sup- port bacterial growth and should be prepared daily under field conditions. 3. Let the resultant cell-suspension set for about 10 min. 4. Centrifuge the suspension at 1,500 RPM for 4 min. 5. Discard as much of the supernatant fluid as possible, being careful to leave the button of cells undisturbed. Add 3 ml of freshly prepared Carnoy's fixative (3 parts absolute methanol: I part glacial acetic acid). Floating material and lipids may be removed at this stage. Disrupt the cell button with a pipette until the cell suspension is homogeneous. Allow cells to fix for about 10 min. 6. Centrifuge for 4 min and decant supernate. Resuspend cells in 1.0 ml of fixative and centrifuge as before. This step is repeated at least three times. After final washing, cells are resuspended in 1.0 ml of fixative. 7. Place two or three drops of cell suspension on a clean slide and ignite. When the fire extinguishes itself, the residue is promptly slung from the slide. Four slides from each specimen are usually made. 8. Dry slides are stained for 12 min in a 2% Giem- sa stain (1 ml of Giemsa’s stock solution in 50 ml 188 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 189 of buffer). Buffer is made by mixing 0.469 g of NaH2P04 and 0.937 g of Na2HP04 in 1,000 ml of distilled water. The buffered stain can be used to stain three or four sets of slides. If the buffer so- lution is unavailable, then staining can be by other methods such as one part Giemsa’s stock solution to eight parts distilled water for 15 min. If staining with the latter solution is poor, heating the stain to near 50°C will often help. This distilled water base stain can be used to stain only one set of slides and then new stain must be mixed. Metacentric is a biarmed element that has arms of equal length (ratio is not greater than 1:1.1). Sub- metacentric is a biarmed element that has an arm ratio greater than 1:1.1 but less than 1:2. Subtelo- centric is a biarmed element that has an arm ratio greater than 1:2, but a second short arm is clearly visible. Acrocentric (^telocentric for practical pur- poses) is an element that appears to be uniarmed when viewed with a light microscope. When cal- 9. When a slide is removed from the stain it must be quickly rinsed with distilled water or a film of stain will cover the slide. Slide should be dry before covering with balsam or permount and a 22 by 40 mm coverslip. Voucher specimens, with accurate collection data, should be deposited in a reputable museum. The tag on the voucher specimen should show that this specimen was karyotyped and microscope slides should be cross referenced to the voucher specimen. culating the nombre fundamental or number of arms of the autosomal complement, each metacentric, submetacentric, or subtelocentric is given a value of 2, whereas each acrocentric is given a value of 1. The nombre fundamental (NF) of Matthey in- cludes the arms of the sex elements and auto, whereas the “number of arms of the autosomal complement” (AA Appendix 1) does not include the sex elements. ACKNOWLEDGMENTS We thank R. Laurie Robbins and Ginger Smith for assistance in preparation and editing. We thank Duane A. Schlitter and Hugh H. Genoways for organizing this symposium on African small mammals. Without their encouragement, this manuscript would never have been produced. Supported in part by the In- stitute of Museum Research. LITERATURE CITED Allenbach, M. 1964. Formule chromosomique de Malacothrix typica Smith (Rodentia, Muridae, Dendromyinae). Bull. Soc. Vaud. Sci. Nat., 68:493-494. Baker, R. J. 1970. The role of karyotypes in phylogenetic stud- ies of bats. Pp. 303-312, in About Bats (B. H. Slaughter and D. W. Walton, eds.). Southern Methodist Univ. Press, Dal- las, Texas, vii -f 339 pp. Baker, R. J., B. L. Davis, R. J. Jordan, and A. Binous. 1975. Karyotypic and morphometric studies of Tunisian mammals: bats. Mammalia, 38:697-710. Bender, M. A., and E. H. Y. Chu. 1963. The chromosomes of primates. Pp. 261-310,//; Evolutionary and genetic biology of primates. Academic Press, New York, l:xiv -f 1-327. Bender, M. A., and L. E. Mettler. 1958. Chromosome stud- ies of primates. Science, 128: 186-190. Benirschke, K. 1967. The chromosomes of the family Equidae. Mamm. Chrom. Newsl., 8:231-232. Benirschke, K., and L. E. Brownhill. 1963. Mitotic chro- mosomes in some species of marmosets. Mamm. Chrom. Newsl., 13:7. Benirschke, K., L. E. Brownhill, and R. A. McFeeley. 1963. The chromosomes of the Grant Zebra, Ecptus qiiufiga hoehmi. Mamm. Chrom. Newsl., 10:82. Benirschke, K., AND N. Malouf. 1965. Chromosome studies of Equidae. Equus, 1:253-284. Benirschke, K., D. H. Wurster, R. J. Low, and N. B. Ai- KiN. 1970. The chromosome complement of the aardvark, Orycteropiis afer. Chromosoma, 31:68. Benirschke, K., L. E. Brownhill, and M. E. Beaih. 1962. Somatic chromosomes of the horse, the donkey and their hybrids, the mule and the hinny. J. Reprod. Fertil., 4:319- 326. Bovey, R. 1949. Les chromosomes des Chiropteres et des In- sectivores. Rev. Suisse Zool., 56:372-458. Capanna, E., and M. V. CiviiELLi. 1966. 1 Cromosomi del Pipistrello Albolimbato. Caryologia, 19:231-240. Capanna, E., M. V. Civitelli, and L. Conti. 1967. 1 cro- mosomi somatici del Pipistrello “Eerro de cavallo minore" (Mammalia-Chiroptera). Acci. Acad. NAZ. Lined Rc., 63:125-128. 190 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 Chandra, H. S., D. A. Hungerford, and J. Wagner. 1967. Chromosomes of five artiodactyl mammals. Chromosoma, 21:211. Chiarelli, B. 1961. Chromosomes of the orang-utan (Pongo pygmaeus). Nature, 4799:285. . 1962rj. Some new data on the chromosomes of Catar- rhina. Experientia, 18:405-406. . 19627>. Survey on the caryology of the genus Cercopi- theciis. Mamm. Chrom. Newsl., 22:209. . 1963. Primi resultati de ricerche di genetica e cariologica comparata in primati el oro interesse evolutive. Rev. An- throp., 50:87-124. . 1966. Caryology and taxonomy of the catarrhine mon- keys. Amer. J. Phys. Anthropol., 24:155. . 1974. The chromosomes of prosimians. Pp. 871-880, in Prosimian biology (R. D. Martin, G. A. Doyle, and A. C. Walker, eds.), G. Duckworth and Co. Ltd., London, xxi + 983 pp. Chu, E. H. Y., and M. a. Bender. 1961. Chromosome cytol- ogy and evolution in primates. Science, 133:1399-1405. . 1962. Cytogenetics and evolution of primates. Ann. New York Acad. Sci., 102:253-266. Chu, E. H. Y., and N. M. Giles. 1957. A study of primate chro- mosome complements. Amer. Nat., 91:273-282. CocKRUM, E. L., T. C. Vaughan, and P. J. Vaughan. 1976. A review of North African short-tailed gerbils (Dipodillns) with description of a new taxon from Tunisia. Mammalia, 40:313-326. . 1977. Gerhilliis andersoni de Winton, a species new to Tunisia. Mammalia, 40:467-473. Darlingion, C. D., and A. Hague. 1955. Chromosomes of monkeys and men. Nature, 175:32. DE Boer, L. E. M. 1970. Cytology of the Cercopithecidae with special reference to the karyotype of Cercopithecus hamlyni Pocock, 1907. Pp. 121-126, in Proc. Third Internal. Con- gress Primatology. Vol. 2. Neurobiology, immunology, cytology (J. Biegert and W. Leutenegger, eds.), S. Karger, Basel. . 1972a. The karyotype of Galago alleni Waterhouse 1837 (2n = 40) compared with that of Galago senegalensis Geoffroy 1796 (2n = 38) (Primates, Prosimii: Galagidae). Genetica, 43:183-189. . 1972b. The karyotype of Galago demidovii Eischer 1808 (2n = 58) (Primates, Prosimii: Galagidae). Genenen Phaenen, 15:19-22. deHondt,H.A. 1974. Karyological studies of two insectivores of Egypt. Proc. Egyptian Acad. Sci., 25:171-174. Dieterlen, E. 1976. Die afrikanisch muridengattung Lop/»/ro- mys Peters, 1874. Stuttgarter Beitr. Naturk., ser. A, 285:1- 96. Dulic, B., AND F. A. Mutere. 1973u. Comparative study of the chromosomes of some molossid bats from eastern Africa. Period. Biol., 75:61-65. . 19736. Les chromosomes de trois especes de Megachi- ropteres (Mammalia, Chiroptera) d’Afrique oriental. Cary- ologia, 26:389-396. . 1974. The chromosomes of two bats from East Africa: Rhinoloplms clivosns Cretzschma 1928 and Hipposideros caffer (Sundevall, 1946). Period. Biol., 76:31-34. . 1977. Chromosomes of some East African bats. Sauge- tierk. Mitt., 25:231-233. Egozcue, j. 1970. The chromosomes of the lesser bushbaby {Galago senegalensis) and the greater bushbaby {Galago crassicandatus). Folia Primatol., 12:236-240. Egozcue, J., AND M. ViLARASAU DE Egozcue. 1966. The chro- mosomes of the angwatibo {Arctocehus calabarensis). Mamm. Chrom. Newsl., 20:53. Eisentraut, M. 1968. Beitrag zur Saugatierfauna von Kame- run. Bonner. Zool. Beitrage, 19:1-14. . 1969. Die verbreitung der muriden-gattung //v/o/ny,sc«s auf Fernando Po und in Westkamerun. Z. Saugetierk., 34:296-307. Ford, C. E., AND J. L. Hamerton. 1956. Chromosomes of five rodent species. Nature, 177:140-141. Fredga, K. 1972. Comparative chromosome studies in mon- gooses (Carnivora, Viverrvidae). Hereditas, 71:1-74. Genest, H., and F. Petter. 1973. Les Taterillus de Republic Centrafricaine (Rongeurs, Gerbillides). Mammalia, 37:66- 75. Gropp, A., D. Giers, AND U. Tettenborn. 1969u. Das Chro- mosomenkomplement des Wildschweins (Sits scrofa). Ex- perientia, 25:778. Gropp, A., P. CiTOLER, AND M. Geisler. 19696. Karyotypvari- ation und Heterochromatinmuster bei Igeln {Erinaceus and Hemiechinus). Chromosoma, 27:288-307. Gropp, A., and A. T. Natarajan. 1972. Karyotype and het- erochromatin pattern of the Algerian hedgehog. Cytogenet- ics, 11:259-269. Hamerton, J. L., M. Fraccaro, L. DeCarli, F. Nuzzo, H. P. Klinger, L. Hulliger, A. Taylor, and E. M. Lang. 1961. Somatic chromosomes of the gorilla. Nature, 192:225. Hamerton, J. L., H. P. Klinger, D. E. Mutton, and E. M. Lang. 1963. The somatic chromosome of the Hominoidea. Cytogenetics, 2:240-263. Hard, W. L. 1968. The karyotype of a male cheetah, Ac/no/;yx jnhiatiis jnhiatus. Mamm. Chrom. Newsl., 9:16. . 1969. The chromosomes of duikers. Mamm. Chrom. Newsl., 10:216. Hauschteck-Jungen, E., and R. Meili. 1967. Vergleich der chromosomensatze von Steinwild {Capra ibex) und Haus- ziege {Capra hiriiis). Chromosoma, 21:198-210. Heck, H., D. Wurster, and K. Benirschke. 1968. Chromo- some study of members of the subfamilies Caprinae and Bov- inae, family Bovidae: the musk ox, ibex, aoudad, Congo buf- falo, and gaur. Z. Saugetierk., 33:172. Heinichen, I. G. 1969. Karyotype of the South West African plains zebra and mountain zebra: their cytogenetic relation- ship to Chapman's zebra and the Cradock mountain zebra. Madoqua, 1:47. . 1970. Karyological studies on southern African Peris- sodactyla. Koedoe, 13:51-108. Hosli, P., and E. M. Lang. 1970. A preliminary note on the chromosomes of the Giraffidae: Giraffa Camelopardalis and Okapia johnstoni . Mamm. Chrom. Newsl., 1 1:109-1 10. Hosli, P., and D. Thurig. 1966. The karyotype of the male African elephant, Loxodonta africana . Cytogenetics, 5:243. Hsu, T. C. 1960. Chromosomes of black panther and jaguar. Mamm. Chrom. Newsl., 3:4. Hsu, T. C., and F. E. Arrighi. 1966. The karyotype of 13 car- nivores. Mamm. Chrom. Newsl., 21:155-160. Hsu, T. C., AND K. Benirschke. 1967-1971. An atlas of mam- malian chromosomes. Springer-Verlag, New York. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 191 Hsu, T. C., H. H. Rearden, AND G. F. Luquette. 1963. Kary- ological studies of nine species of Felidae. Amer. Nat., 97:225-234. Huang, C. C., and L. C. S i rong. 1961. Chromosomes of the African’ mouse, J. Hered., 52:95. Hubert, B. 1978u. Revision of the genus Saccostomiis (Ro- dentia, Cricetomyinae), with new morphological and chromosomal data from specimens from the Lower Omo Valley, Ethiopia. Bull. Carnegie Mus. Nat. Hist., 6:48-52. . 1978^. Caryotype de Gerbillus pulviiuitiis Rhoads, \H96 (Rongeurs, Gerbillides) de la vallee I'Omo (Ethiopie), Mammalia, 42:225-228. Hubert, B., F. Adam, AND A. PouLET. 1973. Liste preliminaire des Ronguers du Senegal. Mammalia, 37:76-87. Hubert, B., AND W. Bohme. 1978. Karyotype ofGerhillus pyr- amicliim I. Geoffroy (Rodentia, Gerbillidae) from Senegal. Bull. Carnegie Mus. Nat. Hist., 6:38-40. Hungerford, D. a., H. S. Chandra, and R. L. Snyder. 1967. Somatic chromosomes of a black rhinoceros. Amer. Nat., 101:357-358. Hungerford, D. A., H. S. Chandra, R. L. Snyder, and F. A. Ulmer, Jr. 1966. Chromosomes of three elephants, two Asian {Eleplws maxi mus) and one African (Loxodonta afri- cana). Cytogenetics, 5:243-246. Hungerford, D. A., H. S. Chandra, R. L. Snyder, and F. A. Ulmer. 1970. Chromosomes of three elephants, two Asian (Elephas maximiis) and one African (Loxodonta africana). Mamm. Chrom. Newsl., 11:110. Hungerford, D. A., AND R. L. Snyder. 1969. Chromosomes of the rock hyrax, Procavia capensis (Pallas). Experientia, 25:870. Ismail, H., and P. V. Tobias. 1956. In Chromosomes, sex-cells and evolution in a mammal (P. V. Tobias, ed.), Lund & Co., London. Jordan, R. G., B. L. Davis, and H. Baccar. 1975. Karyotypes and morphometric studies of Tunisian Gerhillus. Mammalia, 38:667-680. Jotterand, M. R. 1971. La formula chromosomique de quatre especes de Felidae. Rev. Suisse Zool., 78:1248-1249. . 1972. Le polymorphism chromosomique de Mas (Leg- gada) africanns. Cytogenetique, zoogeographie, evolution. Rev. Suisse Zool., 79:287-359. Koulischer, L., J. Tijskens, and J. Mortelmans. 1967. Mammalian cytogenetics. I. The chromosomes of three species of Bovidae: Bos tanrns, Bison bonasus and Cephal- ophus grimmi. Acta Zool. Path., 43:135-141. Kuhn,H.J. 1967. //! Progress in Primatology. Fischer U., Stutt- gart. Lay, D. M., K. Agerson, and C. F. Nadler. 1975. Chromo- somes of some species of Gerbillus (Mammalia, Rodentia). Z. Saugetierk., 40:141-150. Lay, D. M., and C. F. Nadler. 1972. Cytogenetics and origin of North African spalax (Rodentia: Spalacidae). Cytogenet- ics, 11:279-285. Lyons, N. F., C. A. Green, D. H. Gordan, and C. R. Wal- ters. 1977. G-banded chromosome analysis of Praomys natalensis (Smith) (Rodentia, Muridae) from Rhodesia. I. 36 Chromosome population. Heredity, 38:197-200. Making, S. 1947. Notes on the chromosomes of four species of small mammals (Chromosomal studies in domestic mammals V). J. Fac. Sci. Hokkaido Univ., 4:345-357. Malouf, H., AND T. G. Schneider. 1965. Karyotype of Fe /A aurata. Mamm. Chrom. Newsl., 15:107. Manfreddi Romanini, M. G., C. Pellicciari, F. Bolchi, and E. Capanna. 1975. Donnees nouvelles sur le contenu en ADN des noyaux postkinetiques ches les chiropteres. Mam- malia, 39:675-683. Manna, G. K., AND M. Talukdar. 1965. Somatic chromosome number in twenty species of mammals from India. Mamm. Chrom. Newsl., 17:77. Matthey, R. 1936. La formule chromosomiale et les hetero- chromosomes ches \es Apodemas europeans . Z. Zellf, Mikr. Anat., 25:501. . 1952. Chromosomes sexuels multiples chez un ronger: Gerbillus pyramidum (Geoffroy). Arch. Julius Klaus-Stift. Vererb-Forsch., 27:163-166. . 1953. Les chromosomes des Muridae: revision critique et materiaux nouveaux pour servir a I'histoire de I'evolution chromosomique ches les rongeurs. Rev. Suisse Zool., 60:225-283. . 1954ri. Les chromosomes de Macroscelides rozeti Du- vernois (Mammalia — Insectivora). Existe-t-il une serie poly- ploide chez les Macroscelidae? Rev. Suisse Zool., 61:669- 677. . 19547>. Chromosomes et systematiques des Canides. Mammalia, 18:225-230. . I954r-. Nouvelles recherches sur les chromosomes des Muridae. Caryologia, 6:1^4. . 1954< rx so ON "O c Cd so 3^2 J= O u u. C eg e ■o o o a -o c ' c O C/5 \0 E CD Cl. S' S' 2 -2 l-'i S3 ^ eg o U eg O U o X eg o U 00 X < < eg eg M eg C/J ^ c/5 X X > 2 2 ^ o U sO so r- r- ON ON ON o 0> 0> s 2 s c/5 c/5 eg eg O O U U O S a 13 13 ao o > p o U ^ i; ? ' c 73 ti 73 :^3 g u G C) -i; p G t o i: 0> c/5 ■ o ON •TO c eg 2 -p ■§ ^ eg 1“ $5 o S ck: < S < < . , CO X . , s s CO CO X X X X • Oo '3 ^ S g i U <0 Co 'a* ^ -C s -2 a -2 <25 a, H Appendix I. — Continued. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 195 u O C E -S <1^ > O X a o 6D a z cC o Oil cn Z cd z a z ur ^ CQ 3 S Q C/5 I/"! “O S -C cd r S S s s CO CO CO C/3 c/5 CO X X X 1 1 1 X 1 1 X 00 1 00 1 1 ^ 1 1 1 r^j 1 1 «-*N m 1 ro 1 00 00 1 1 ^ 00 1 00 00 NO r- ! ! NO NO 1 NO NO NO !3- s: <3 & c s S o OB cd Z • "O c cd u tr (U On Cn (X ■3 ^ ^ Cl ^ liS a o ^ Cl. a: u Uh lU 3 S 13 -O ^ -C C 13 c« •U ^ -o — 3 “ Q o> On c/5 ^ U . O o ^ ca S u cn t) >- < < < f- U CO H H OB • • X H s CO CO CO CO X X X X X X 1 X rj- o o NO NO 1 Tf — rn 1 ri o O n o rj o o NO NO NO NO NO NO NO NO Appendix I. — Continued. 196 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 o n E ^ w OJ 03 2^ ^ p X .£ ry^ ^ (1 j X D, u O E ° i S 2: o. (U c<3 CU U ^ 1/^ o> r-- t- - o\ „• I 5 o ® CQ S , c c 15 ° !C ° u u ^ w 0> o. D "2 's § I — <>i w <3 u V Cl Q s ^ 2 cd 03 S c OJ cd o -5 c i- c 3 3 3 H W H 3 s c/5 s CO s s CO CO S c/2 s on >< >< X X X 1 1 X 0 r-i 0 1 ^ oc 1 1 00 0 00 48 50 50 50 0 0 £ 03 .— (U — *o c o 3 ^ H ct: < X s X 2 ^ OJ *o QJ ON pa - 3: !_' -a 3: y 2i o § Ji g 2 CQ 3 .2 M J3 u -<= J= W U -o U U CO X N3 Ci. 0 >) 2 03 'S C; Oo S3 •2 •2 k. 1x3 "S 1 0 "5 Ci 0 CO t 1' £; £; CO (/i c 15 X) o o s CO X s CO X o o Oo Oo -2 -2 *2 "2 O O Appendix I. — Continued. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 197 ■2 u ^ O 2: >> i! .c -o U CQ U G. a 03 H 2 2 U e 3 o U ^ c c ^ •u (U 03 ^ >. U) 0 1 X Q. o e s c/5 < E c/5 E C/5 X X X X c/ X S . r . ^ . ^ E, ^ c/ s S S s W) 2 b E s X c/5 c/5 c/5 c/1 2 S c/ •a X X 1 X 1 ! X X > X X b Z •S a 2- o o Oo Oo -S -5 *2 *2 T3 ^ 73 ~Ci ^ -Ci ^ Qu ^ U C> ? O O ^ w O ^ ^ ^ C) c u ? -c if '= Ci u u S ^ 'c: 5 2 ■= O tj 1.J tj O ^ ^ ^ I ^ ? if CJj Cb u u ^ ^ u u u u -Ci ^ ■ II 2 a C( ^ ■ c ■>5 <>5 c) 2 2:: ’S. 2 S, ‘>5 S. C> 5 g 0 b. 5 b. U u Appendix I. — Continued. 198 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 2 u C'i c/3 (N .S — ■a ' c a c . as CQ ^ 3 00 a X *u c a 3 -5 U L3 *D y § .2 3 JZ JZ u u c3 X r-1 (N fN ''O CO -- X) o H ■3 G cd U X — =5 u u u u.“in-s *a\2 Ic X X X £ O U U U Q ro OS .b w E Cd V3 K X • - OS c (L) OS ffi - c ' SD X OS d — 00 U .3 os ^ C SO ^ a> Os CQ "O •-' — C = 3 “ 2 J K u ■o e d O > S X o U u c E ^ 5 0. -o 5 0 o ^ C 1 -j c o Appendix I. — Continued. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 199 o c E OJ ^ ^ G X .S r/^ c (11 • = ^ c oo ON U - pa ^ -o , o c ~ •£ $ S X a s 2- c o u 0> 00 E ^ ^ 00 o^ ^ — a- o^ ;^-g C ■ t; ON "O ^ O S Uh ^ ON ON "S ^ ^ ON ON N — — C OS ^ JZ ^ "" «« ON . ON s] 2: o> s g W C w a c B cd o S S ^ ^ S: U cd cd ^ ^ ON N ^ ^ Cd ^ I ^ c £ s ^ ^ P;: ^ - 2 ON id (U C ■S ^ w c/5 -a cd cd u. S S: 5 o U 3 O C/D o c/5 £ S UJ c/5 3 O C/5 CL O. >> >^ c OD Of) 3 UU PJ H a >. 0£) UP CL . >. Ofl UP [ < ^ 2 Z C3 D. >. C 00 3 ttJ H >> 00 o < X c/5 C/2 ;L =« X c/i X X X H S ^ X X X o o CD — I Dt o I I -tv I NO Tj- T+ T+ I ’^0+^0 ri r4 Tf Tf TJ- Tt ^ 2 o X cd H 00 £:■ c i >. e; -a .a a s'-* aa S fc t 5 Ci O ^5 L. L. CO s: s: 3 ^ ^ Q Q 5: Cl O C‘ S I H ? ^ s: ^ 52 5 5 I III 1 5 -5 -ci ■'Cj -S sr^ ^ ^ ^ o •cr -s ^ -Ci -«Si II O Oo 0/3 ei e) Gerbillus henleyi 52 63-65 — — — — Egypt Wassiff et al., 1969 Gerbillus nanus 52 66-68 — — — — — Wahrman and Zahavi, 1955 52 — — — — — Egypt Wassiff et al., 1969 200 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 ON r iP 00 1/-J o (U a; O CO X. cd "O c/5_5 ^ -z: ^ >s c/3 m ^ .b »— J >/-. -J c r- ON ON (U ON . c ^ c ™ ^ 1) u ON ON ON d) CO ^ CQ o r- ON Ov ON Pu Cu (X *0-0-0 c c c On NO r- On On ON c 3 H 03 CS s ^ *0 *0 3 ^ CQ j5 Ct3 c/3 S Z S Z X rJ CL >. c OC 3 UJ < tU H C/5 W H ^ < (u , oc S .2 X X X =« X O rj o 00 o o Oj- \D t3- -Ot •2 .2 I i g. S a ,a« i; ~= ^ £ “i. 5. >3 'o -5 -5 -2 IS -5 IS i£ L. ^ L. k, ^ d) o e> o o O C5 >, -a 1) c S B ^ S X *o 0) d> c ^ JC 'S « 3 S S X 30 w cd (3 3 S S X < < 3 3 3 O O W C/5 C/5 C/5 '' iXi c/3 0) 3 w w "5 3 3 3-0 S S S c§ o > §:< D U s c/2 s c/5 S S X CO 0> 015 u *3 s S 05 00 u < E C/3 X =^J 1 1 I X X 1 X 1 1 >< X 1 X 1 1 1 1 1 1 X 1 x o 1 ^ 1 1 1 1 1 1 1 1 1 1 1 1 1 ro 1 1 1 ri 1 1 1 1 j 1 1 1 I 1 1 r4 1 1 1 1 1 1 1 1 I 1 1 1 NO 1 r- 1 1 1 NO 1 1 1 1 NO 1 1 o 1 1 1 NO r<3 1 1 1 1 r- '7111 1 1 7" 1 1 1 r- 1 1 1 1 o o 1 1 1 1 \6 1 o ' 1 NO r- 44 44 44 54 48 48 44 44 52 50 48 36 48 46 42 52 30 C3 Cl ^ 2 X, X cs llw "CS C5 Cl .5= CS -Ci Oo Oo C5 >3 CS !3 Q C5 ^ ^ L. k,. k,. lu. dt ^ ^ d> ^ a ^ a a ^ In. ^ Jk. S». dj ^ dl di ^ ^ C — . !s 2 . ^ E c >. to • - Jj (U u -2? ■£ g- ^ ^ s Appendix I. — Continued. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 201 03 u C 3 O u o c S ^ ^ ^ ^ B X c/5 <[} >» W) X D. Pu z o X a o\ ^ r- os C T3 2 2 « t? .£ “ ■S 'c W w 3 i2 S ■§ 3 S H O o< S u w < S < u u u S s c/2 'U "CS 60 g .2 I X X (y 0> -a c cd , c/5 >> C/D (U rsi C dJ C ■ }=i <— -r- -5 ^ £5 Os X) 'S XJ _ . . _0 X) _D O So So Qi ^ QC ^ 0^ C/5 ' ;■ ' ' ^0’^0+fQO^DO+ so \o m czi • — dj X C X) U O a Cii Os 13 OS o OS cQ cd v Si u dJ cd CD > L. » .2 *3 cd 'a o w cd >. O o L- dj 3 a> c/5 a a D c d> c/5 a CL X cd JC O o > E o c/5 ■£ c2 < u c d) E cd U 60 Z s s'" c/2 M ^ u- CJ iS E X X c/2 'u ^ 00 I iS S X X I X X E S S 3 _ 5 00 « 00 g S C/l S E _ 3 Cd "O E M C/5 C CH^D^^O w w ^ w S 00 OS I m ri r4 C'J u Ci- o r ^ Cl c OX, d) E c 2i cdcdcdcdcdcd^'^ S 2 S S S S X < j= CiD z 3 O C/5 3 UJ C/5 , >. cd cd •u D ^ is •£ £ c c w W rn os I 3 ^ 3 >. 2 w is £ e , u >. u >^ ^ D 0) i: <3-> 4^ £ "^ "E £ "^ “E £ 3 § > I S > I s> so r-- os ^ 00 OS ^ >^ >% - 4) 4) h ^ X2 -C .2 ^ S 6 ^ OS ^ £Q — *3 ^ M ^ «/o m r-- r^ . os ^ os Os OS « OS ^ ^ ^ --.r . >v >^ tc >> e S ^ .s f2 e 3 €d cd SSh S S a J S X s s o u < u o U o p^ O 0-* 3 o o Pk a o C/3 C/D o “O a o TO e O u. QJ c o 'O G o © c Cd O U Cd o U cd E V Uh cd e 4> C Cd U cd 1 E ‘ 4) PP c cd U 1 - 1 o > 1 s > re! O u >. o o o U U S' S' o o > > ou b si £ < O N c/5 “D (D -O o .b c »S cd 4) P^ N c^ U * ^ MM ^ c E Ji 4) cd ^ ^ P X .5 GO = dj S-) £ e g. a < eJ5 © £ & o E < P3 < X X Sx >-< X =%J ^ X X s cfl s X s C/D 00 < < ^ w s X 5 , 2 euQ r M ^ S X 1 X X M 1 X X X 1 1 1 X 1 1 ® ! 11 1 1 2 1 1 1 1 li z $; QO O X cd H I 5 o S V C Q -C U Q a II S; ^ S 5 S s s: E E 4j ^ s o Si«. §■ ‘-:i S S § ■ii 3 § 3 o 5 o 'S P “ds o -sd s ■? s g' o = ? 5 3 e 3 ■2 §-^ §• - ►3 Kj § Ps . E § 0 ^ ^ II 1 s -s; s 2 ds ^ mC, do do do do ^ <4) 3 3 § § 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 203 O C - » X >» Oi) ^ ^ ^ 1 E o 1 >. < o 3 u O < o u O C/5 > d^ > c/5 ^ u > « r^t I I \0 ri o > >. cd >» ■B = x: h -c x: >> >> dJ dj x: X cd^cd'^cdcdcdcd SuStxSSSS ON « X On On On ON ON , , , ^ — r o; qj 0^ (i> ^■Se-5-S cd cd cd cd (X S S S S H H H (X CX 0. 3 O O c/) cd O u a o <4i: < c/3 Cd o U c/3 Cd O U • W PJ pci ui . >> pci X ui >a E cd 1 < < < 1 s < 1 1 1 < o o < u u u u > u c/5 u > ^ CJ AX o X a s c/5 < ; Y:A Urn o s X JJ X E X ^ 1 1 s X .2 ’C cd > s' S C/5 X 1 1 X X 1 1 1 X < X S S 2 c/5 c/5 c/2 XXX cy XXX ri ri ri s: ^3 OO : Oo ■ — • “§ cio '2: Sf I' C 5! ^ a ^ a gcuacbT3a:a«^a S,-« oo-a g S®5 “« ?«■=; cwD'g-S Oo0o0«~ ao — "''S>~Jax^a‘~J-2K^E ~ >^ s o> o >> 16 O y N £ s ^ Pi 'c t I ^ >. © u 3 G I O OS © 5^ "O On I I ^ X .2 ’q GO >. i3 Q. £i£) W 3 •S O c © On *© 3 c3 _ 3 ^ .« ON £ ^ pa • ’V G a ^ ^ .£ © (D ■§ s o ■ "O . 3 ' 'g o N 3 2 ^ c$ U 3 w §:-t 5 ? S 3 > >. >H X S 00 C/5 OP w X X X X j 1 1 X m o i 1 i 1 ?; O r'l 'c 1 1 1 1 ® r-- r-- r- ! ! 1 1 3 — ^ 3 fU 2 C^ -S "© © 5 g 3 ^ Appendix I. — Continued. 1978 ROBBINS AND BAKER— KARYOLOGY OF AFRICAN MAMMALS 205 o c E ^ K 1) c3 Ji o. e X .E X CJ « s C 00 pa D rj >0 T_, 1^ ON ^ ON On ■ X X X XX K~ U ON 3 — ON 3 C 4J ON CQ — ^ § S-TO' 3 3 0£ OX) . •o -o .3 Os C NO (U ON 00 E S c/5 nO — < -C o c/5 >, s s c/5 2 0/3 O — 'o "cS i *3 I p ^ s < 3 E < a E GO 3 c/5 > C/5 C-- s X o E X S > s ’"O *3 P S C c/5 s s c/5 > s CO X s CO 1 ! X X X 1 >< 1 X X X 1 1 1 1 1 X < r/ r/ o 1 1 00 00 1 1 1 1 i ^ CQ 1 m r/N r-/ 1 1 r4 r4 1 1 1 1 1 ^ < j O 00 00 00 00 1 ® 1 ^ oo oo 1 1 1 1 1 < I NO NO NO *y~) 1 r- \ NO NO NO 1 1 1 1 1 ON c o X 3 H ci^ <;j Qu ? ? "S 3:' 3:'ci: Co O o. 5 c/5 3 Ci. 5; E > qq ■> Ci. OC Oc G u o' CO o 206 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 •— c/5 i _ u c E ^ V ^ ^ ^ o. c X .£ ^ e o U a a < 2 X o. u O E D3 00 w) VO T3 u ON £ ^ cd W) "O sy -N. a: a: 3 — vo vO vD ON ON ^ ^ TO -U 3 2 'O -o O O ^ O CH- ri ^ 'Tf Tj- Tf =§ •2 i Ci. ; C (U vO CQ 2 ■o . 3 ;or H H X 53 S S •i CO “S ^ ^ Cj C T- ^ Co On (U ' -C _: cj a> c r^ 2^ m " -C T3 C -- H D3 X u. 00 !3 a ^ 5 c I § a o (U 0) NO CQ Ov T3 . C/5 X ■o c (/5 l/^ U NO •3 ON s CO H s < e c/5 < £ oo £ £ . r. £ s s C/5 s CO C/5 CO X X >< X x 'ct 00 o 00 r tU W CQ X CQ E c z, X — c 0 1 s s X >> -2 X j= o o ^ - ON g u — *o *- e cs X- w ^ x: o o ?; .b s CQ ■E O NO 03 2 c c 2-i CQ I II “O II aj u Ofi • = c .E X X C> U Cj ^ ^ ^ ^ ^ 5 S <\) 5! *«5r ^ -C) ^ Oc J3 o “S X o a X a cj c/5 • - o c r- 0> On CQ - «- -Ci o> nJ .2^ 00 o r- 1/5 NO nO I- ON = ON 3 — 03 — O CQ O NO 1 X ssli M \D « _ 1- 0\ 3 „ ^ I <3 a CT3 _U o _o .2 c/5 0> t: ’ 3 - cn c« c/5 c/5 > < s s 00 X s X S S s S c/5 C/5 OX) g >- s < X 1 X 1 X X X X X X 1 X X 1 X 1 1 X NO 1 1 1 1 1 ri NO Tf 1 1 1 1 1 1 ^ r*S 1 1 1 1 1 ri rj 1 1 1 1 1 1 00 1 1 1 1 1 NO r-- NO NO l/~i NO «/~t 1 1 1 1 1 1 208 BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY NO. 6 u o c £ ^ ^ c o U a a < >- o X a o e •b - CQ CU ON ■§ pa m ON ON On ON ON ro ^ nO nC nO nD \D On On On On On ^ ^ ^ ^ ^ C "O "O "D *0 „ U La u. u. u 3 CQ c/5 X DC K K X K < X 1 X 1 1 X X X >< o 1 ^ 1 1 ^ sO so ri 00 1 I 1 00 «/-i 1 1 1 1 o ^ >- X < < X E T3 o o c C 03 2 o < ^ • ' ^ < s X i: I j: ^ 0+ O rn ^ O - ri r^i S: c- •S- A % C' ac S- -Cj o X -Cl X -c o X >. Cl. C! O X ' h, X 10 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 = ON = a OD c a — c/5 :0 X a c/5 U. -b in c r-- ^ vO u 'O c ON CQ ON <73 ^ « “ = os = X 3 — - c/5 ^ X c o U Q. a < o c E O (U eg Cl c X ,E CO = D W) _o >- O X a o E CO CUD 2 p iH S CU c» o Cl. a -j; o 1 1 a 5 PROBLEMS AND PRIORITIES OF RESEARCH ON THE TAXONOMY AND ECOLOGY OF AFRICAN SMALL MAMMALS DUANE A. SCHLITTER Section of Mammals, Carnegie Museum of Natural History, 4400 Eorbes Avenue, Pittsburgh, Pennsylvania 15213 U.S.A. Africa is still one of the research frontiers for studies of small mammal ecology and taxonomy. Yet for specialists from Europe and North America, transportation costs for persons and equipment is expensive and logistical support in Africa is costly. African scientists are not faced with such high ex- penses to begin or carry on research, but rather have to compete with other priorities in developing countries for funds that might be available for re- search. African authorities formulating resource utiliza- tion plans are caught between factors of rapidly in- creasing populations, economic priorities, and po- tential short-term exploitive policies in attempting to implement long-term integrated land use policies which will allow for and include preservation of var- ious representative ecosystems. Probably the de- struction of entire ecosystems in Africa precipitates the greatest concern on the part of specialists study- ing the ecology and taxonomy of African small mammals. Time is growing critical if studies of many species are to be done adequately. In addi- tion, changes required to insure adequate food sup- plies and protection from diseases add to the con- cern. Forest ecosystems, especially relict high forest, are becoming smaller and more isolated. Deserts of Africa are expanding. Ecological and behavioral studies of small mam- mals in Africa are hampered by lack of financial support. Most ecological data have been collected opportunistically during other studies. Most popu- lation ecology studies are of short duration or sea- sonal in nature. Long-term monitoring studies of small mammal populations are few in number or include only short periods of trapping during the year. Long-term studies of resource division and habitat utilization have seldom been done. In all instances, quantitative analyses of data are marked by small sample sizes. Sufficient financial support must be made available to allow for long-range stud- ies of up to 10 years’ duration. This support must include salaries as well as logistical expenses. Long-range field studies of behavior are needed for most small mammals. Some of those most ob- viously offering opportunity for field observations are Eidolon, Hypsignathus, Thryonomys, Criceto- mys, Pedetes, and various viverrid genera. Recent advances in night vision equipment make it possible to observe many of the nocturnally active genera. Careful consideration should be given in selecting sites for ecological and behavioral studies. Among those in Africa offering the most pristine environ- mental conditions are the National Parks and Re- serves. These areas, in addition to being more pris- tine, are staffed by people who have a vested interest in such studies. Laboratory or some other type of sheltered work space is usually available and security for laboratory and field equipment is usually better in these areas than elsewhere. In most instances, African scientists doing re- search on small mammals are employed in other roles. Thus, they may only be able to conduct re- search during their spare time. Whenever possible, these scientists and their students should be includ- ed in research projects originating in Europe and North America. Such participation may be the only opportunity for them, especially the students, to learn specialized methodologies and techniques. In recent years, a more active phase of interest has developed in studies of taxonomy of African small mammals. This upsurge of interest has been precipitated by the need for better taxonomic treat- ments of various groups of mammals on the part of ecologists, conservationists, agriculturalists, and medical zoologists. Even with the introduction of karyology in taxonomic studies in Africa, method- ology still lags far behind some other areas of the world. Commitment of logistical support is needed so that all extant specimens, at least those in the major collections, can be included in taxonomic re- visions. There must also be commitments of addi- tional support so that those areas from which crit- ical material is lacking can be visited in order to obtain any necessary missing data. Individuals must be encouraged to use the best and strongest statis- tical tests available to them in their taxonomic stud- ies. Univariate statistical tests, including analyses of variance, must be a part of any taxonomic study 211 212 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 whenever possible. Ideally, multivariate statistical analyses should be included as well. Numerical taxonomic treatments of data in generic and specific revisions of African mammals must be encouraged. Those individuals who are involved with taxo- nomic studies of African small mammals are well aware of which groups should be receiving the high- est priorities in taxonomic research. However, a review of these groups will indicate the enormous task that remains to be done. Among the traditional insectivores, one genus stands out in difficulty well above most other Af- rican small mammals. The genus Crocidura is un- doubtedly the most complex of the insectivores and may be of any African mammal (Heim de Balsac and Meester, 1977). Sibling species are common, often sympatric in distribution, and study speci- mens are usually inadequate in number and widely distributed. Heim de Balsac and Meester (1977) list 85 species, but admit that this is a tentative compilation. Other insectivores deserving of treat- ment are the hedgehogs and golden moles at both generic and specific levels. Recent taxonomic ar- rangements of the latter by Simonetta (1968) and Meester (1974) indicate the differences of opinion that exist on the number and relationships of genera and species. Other genera, such as Myosorex, Syl- visorex, Simcus, and Elephanttdus are deserving of additional attention but less critically so at this time. Within the bats, a large number of genera are in critical need of taxonomic revision. Chief among these are Epomops, Epomophorus, Rousettus, Ta- phozous, Rhinolophus, Hipposideros, Myotis, Nyc- ticeius (sensu lata), Pipistrellus, Eptesiciis, Scoto- philus, Glaucouycteris (sensu s trie to), Kerivoida, Miniopterus, and Tadarida. Some genera, such as Rousettus, Pteropus, Emballonura, Plecotus, Nyc- talus, Miniopterus, and Otomops, have major Eur- asian affinities as well, and taxonomic studies of these genera must consider both continents. Oth- ers, such as Rhinolophus, Hipposideros, Pipistrel- lus, Eptesicus and Tadarida, seem to separate into species groups on a continental basis. The genus Lepus is in critical need of taxonomic treatment over the whole of Africa and Eurasia. In addition, a monographic study of both Recent and fossil genera of lagomorphs, including all continents of occurrence, should be encouraged. The group showing the highest radiation of spe- cies in Africa is the rodents. Within this group, the squirrels contain the genera Eunisciurus and Par- axerus, which are critically in need of taxonomic revisions. Additionally, Heliosciurus is deserving of attention, with both H. rufobrachium and H. gam- bianus in need of studies of geographic variation. Studies of generic relationships are needed for both sciurid and anomalurid squirrels. Among the genera of small rodents, the following African ones are in critical need of taxonomic re- visions, utilizing every available method of estab- lishing specific relationships: Meriones, Gerbillus, Tatera, Taterillus, Gerbillurus, Mus, Aethomys, Acomys, Arvicanthis, Hybomys, Lenmiscomys, Lo- phuromys, Pelomys, Mastomys, Praomys (sensu stricto), Hyloniyscus, Grammomys, Cryptomys , Tachyoryctes, Steatomys, Dendromus, Otomys, and Graphiurus. Attempts should be made to in- clude in these studies other data than those tradi- tionally included, for example, differences in life history, morphology of glans penes, sperm mor- phology, and karyotypes. Additionally, studies of geographic and non-geo- graphic variation are required for a number of cur- rently recognized monotypic genera and some gen- era with only a few species. Without doubt, Ellerman (1940, 1941) and Ellerman et al. (1953) oversynonomyzed certain genera of rodents. Hu- bert (1978) has shown good evidence for two spe- cies in the genus Saccostomus . Other monotypic genera, which should receive critical review of geo- graphic variation are Dasymys, Oenomys, Rhab- domys, Thallomys, Uranomys, Malacothrix, Pe- detes, Desmodillus, and Psanunomys. Unfortunately, in many taxonomic revisions of African genera of rodents published in the past, in- sufficient attention has been given to non-geograph- ic variation. Secondary sexual and age variation are present in nearly all rodents. Care must be exer- cised to compare only rodents of the same age in these taxonomic studies as these animals invariably continue to grow throughout most of their life. In the remaining groups of small mammals, taxo- nomic studies are needed at the generic and specific levels for the prosimians, particularly Galago, the viverrids, especially Genetta, and the antelope gen- era Cephalophus (sensu lato), Madoqua, and Ga- zetla. Hoeck (1978) has introduced a new approach in the study of the generic relationships of hyraxes and indicates as well the need for a comprehensive study of the genus Procavia. Until better taxonomic treatments of many gen- era are available, statistical studies of the zoogeo- graphic relationships of African mammals on a con- 1978 SCHLITTER— PROBLEMS AND PRIORITIES 213 tinental basis will be inadequate. Unfortunately, many zoogeographical studies, even regional ones, must make concessions for areas lacking adequate faunal surveys and for taxonomic confusion in cer- tain genera. Faunal surveys of African countries are becom- ing less common, primarily due to difficulties in ob- taining necessary funding for such field work and maintenance of resulting collections. Yet, non- mammalogists are voicing an increasing desire for reports of the mammals of a given country. Even so, mammal surveys of certain regions would still be a valuable undertaking. The Ogoue Basin, the central Sahara, the western Sahara in the region formerly included in Spanish Sahara, the horn of Africa including Somalia, Ethiopia, and particularly Eritrea Province, the massifs on either side of the Rift in Central Africa, southern Zaire, and eastern Angola are in need of additional surveys for mam- mals. Attention to endangered species of African mam- mals has focused mainly on large mammals and the primates. Many species of small mammals are so poorly known it is impossible to judge their status. Many obviously good species are known by less than 12 specimens, for example, Leimacomys buttneri, Dendroprionomys rousseloti, Zinkerella insignis, Hylomysciis baeri, and Glauconycteris su- perba . Permits to conduct ecological and taxonomic studies in Africa are not required in all countries. Before issuing permits, some others require a com- plete and detailed proposal of the research to be submitted well in advance of the starting date; Ken- ya, for example, requires this information six months in advance. Without such permission, re- search conducted in these countries is illegal. For most countries, export of specimens requires a spe- cial permit or at least a letter of waiver. Unless all of the specialists cooperate in obtaining the neces- sary permission to conduct research and subse- quent permits, requirements will become more re- strictive and permission more difficult to obtain. Individuals performing any type of research in Africa are well aware of the problems associated with economic growth versus conservation efforts. In dealing with the problems in this regard, the na- tions of Africa, although almost all are in a devel- oping state, are no different from those of Europe and North America. However, the participants of the colloquium felt they would be remiss in their professional responsibility if they did not consider the destruction of various habitats, particularly for- est ecosystems, and the mammals occurring there. The recommendations below received unanimous approval from the participants. RECOMMENDATIONS OF THE PARTICIPANTS IN THE COLLOQUIUM ON ECOLOGY AND TAXONOMY OF AFRICAN SMALL MAMMALS SPONSORED BY CARNEGIE MUSEUM OF NATURAL HISTORY, 19-22 SEPTEMBER 1977 I. The participants of the Colloquium on Ecology and Taxonomy of African Small Mammals rec- ognize that these animals play an important ecological role in many natural and man-made habitats. Some species can be pests of agricul- ture and forestry as well as disease vectors in many parts of the continent. The Colloquium fully recognizes the high priority to be given to these problems. II. However, all of these species are vital to the functioning and maintenance of natural ecosys- tems. Furthermore, the role of many is totally unknown although probably very important in most instances. In view of the rapid decimation of natural habitats in Africa, particularly all types of natural forests, due to human en- croachment and development of exotic tree plantations, the participants of the Colloquium on Ecology and Taxonomy recommend that Institutions of all African countries and the In- ternational agencies concerned with these countries: A. Make every possible effort to conserve large and representative examples of all mature forest types, particularly montane and rain forests, which exist within their respective countries or areas of interest. These forest parks or nature reserves must be large enough (circa. 2,000 square km) to conserve viable populations of all animals and plants comprising the ecosystem, and must be completely protected against all forms of destructive exploitation. In some circumstances the remaining relict forest blocks are much smaller than the recom- mended size and in these cases an area as 214 BULLETIN CARNEGIE MUSEUM OE NATURAL HISTORY NO. 6 large as possible should be conserved. Where possible, boundaries of existing for- est parks and nature reserves should be ex- panded to include sufficiently large and rep- resentative areas; B. Make every possible effort to develop ed- ucational programs in their schools and mass media for their general public which explain the value and necessity of forest conservation, and the importance of land- use programs based on long-term ecological goals rather than short-term exploitation; C. Make every possible effort to develop in- tegrated land-utilization policies based on sound scientific and long-term ecological principles, to overcome the competitive na- ture which usually prevails in the interac- tions of the various government depart- ments concerned with land-utilization problems; D. Make every possible effort to have all pro- posals for land, agricultural, and industrial development schemes reviewed by a board of ecologists before these proposals are im- plemented; E. Encourage basic and long-term biological research on forest ecosystems. This re- search should include not only projects of theoretical importance, but those of rele- vance to the practical problems in the coun- tries concerned, such as the control of pest species through the conservation of natural and mature forest habitats. III. The participants further concurred that inte- grated land-use policies based on long-term in- terests provide a more sound economic base than the short-term exploitive policies current- ly prevailing in many countries. IV. Finally, and of primary importance, it was rec- ognized that many of the problems associated with the implementation of these recommen- dations are directly related to the human pop- ulation explosion. LITERATURE CITED Ellerman, J. R. 1940. The families and genera of living rodents. I. Rodents other than Muridae. Trustees of the British Mu- seum, London, xxvii + 689 pp. . 1941. The families and genera of living rodents. II. Fam- ily Muridae. Trustees of the British Museum, London, xii + 690 pp. Ellerman, J. R., T. C. S. Morrison-Scoti , and R. W. Hay- man. 1953. Southern African mammals, 1758 to 1951: a re- classification. Trustees of the British Museum, London, 363 pp. Heim DE Balsac, H., AND J. Meester. 1977. Part 1, Order In- sectivora. Pp. 1-29, in The mammals of Africa: an identifi- cation manual (J. Meester and H. W. Setzer, eds.), Smith- sonian Inst. Press, Washington, D. C. Hoeck, H. N. 1978. Systematics of the Hyracoidea: toward a clarification. Bull. Carnegie Mus. Nat. Hist., 6:146-151. Hubert, B. 1978. Revision of the genus Saccostomus (Ro- dentia, Cricetomyinae), with new morphological and chro- mosomal data from specimens from the Lower Omo Val- ley, Ethiopia. Bull. Carnegie Mus. Nat. Hist., 6:48-52. Meester, J. 1974. Part 1.3, Family Chrysochloridae. Pp. 1-7, in The mammals of Africa: an identification manual (J. Mees- ter and H. W. Setzer, eds.), Smithsonian Inst. Press, Wash- ington, D. C. Simonetta, a. M. 1968. A new golden mole from Somalia with an appendix on the taxonomy of the family Chrysochloridae (Mammalia, Insectivora). Mont. Zool. Ital., n.s., 2(suppl.):27- 55. ) ) w\ , •r4 I 1 » Copies of the following Bulletins of Carnegie Museum of Natural History may be obtained at the prices listed from the Publications Secretary, Carnegie Museum of Natural History, 4400 Forbes Avenue, Pitts- burgh, Pennsylvania 15213. 1. Krishtalka, L. 1976. Early Tertiary Adapisoricidae and Erinaceidae (Mammalia, Insectivora) of North America. 40 pp., 13 figs $2.50 2. Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977. The Clark’s Cave bone deposit and the late Pleistocene paleoecology of the central Appalachian Mountains of Virginia. 88 pp., 21 figs. $12.00 3. Wetzel, R. M. 1977. The Chacoan peccary, Catagonus wagneri (Rusconi). 36 pp., 10 figs. . . $6.00 4. Coombs, M. C. 1978. Reevaluation of early Miocene North American Moropus (Perissodactyla, Chalicotheriidae, Schizotheriinae). 62 pp., 28 figs $5.00 5. Clench, M. H., and R. C. Leberman. 1978. Weights of 151 species of Pennsylvania birds analyzed by month, age, and sex. 87 pp $5.00 cr < 5 5 B RAR I ES^SMITHSONIAN”‘lNSTITUTION^NOIiniliSNl“'NVlNOSHimS SBiyVdail LIBRARIES SMITHSONIAN INSTIT 2 f“ z r- X, z r* Z :o i > 33 m ^>VAs><>^ m <■ ^ m ^ m (s> — E (f> — CO £ CO liniliSNl“NVINOSHilWS SBiavaail libraries SMITHSONIAN institution NOliniliSNI NVINOSHilWS S3 lav Z c/3 Z >. ^ Z ^ r« i g %% g . X CO C/3 CO > z BRARIES SMITHSONIAN INSTITUTION MOIiniliSNI_NVINOSHillMS S3 I H VH 9 IT_L I B R A R I E S SMITHSONIAN_INSTITl , O _ z iniliSNrNVINOSHillAIS S3iaVH9n LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS S3I9V z r* Z r* . z m ^ xouu^3ix m CO ± CO t: BRARIEs"sMITHSONIAN"lNSTITUTION'"^0'-LniliSNI^NVINOSHimS^^S3iaVaan^LIBRARIES^SMITHSONIAN^INSTIT -r 'f/yj. Xs-^. T /^/f^ ' /-X > iliniil’sNI^NVINOSHllWs‘^S3 lava a 11^ LI BRARIES SMITHSONIAN_INSTITUTION NOliniliSNI NVINOSHIIWS S3ia\ CO BRARlES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIINS S3iavaan LIBRARIES SMITHSONIAN INSTIT O h- 3 )linillSNI~NVINOSHllWS S 3 I a V a 9 11 ® ^ ^ ^ ^ ^ SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS S3ia\ >■ 5 ■ ^ > B RAR I ES^'sMITHSONIAN^INSTITUTION ‘^f^OlinillSNI_ NVmOSHllWS”S3 1 H Va 9 il“L I B RAR I ES SMITHSONIAN_INSTIT 5 “ X.h;a^ g 5 )iinillSNI^NVINOSHlllNS^S3 I a V a 9 11^*- 1 B R A R I ES _SMITHSONIAN_INSTITUTION __NOIinillSNI_NVINOSHlllAiS S3 1 a' P - 00 o > ' ^ __ m ~ rn XOm dcJ>^ ^ m ^ (/) — — V (/) ‘ — BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHIIIAIS S3iaVaail L I B R A R I E S SMITHSONIAN INSTT W Z: .V E — r33T— ^ < .V*" ^ 5 .xN .x\Vxs, ■X z < Q DL_^ _ o " — o ^"1 L)|^ _ -V//- o linillSNCNVmOSHimS^SB I dvy an^*’-! BRAR I ES^SMITHS0NIAN^INSTiTUT10N_^N0linillSNI^NVIN0SHillAiS^S3 I dV m "v^osv'ix 'll m x4iyos>i>^ ^ m CO — c/5 S CO — ^ — BRARIES SMITHSONIAN INSTITUTION NOIlfUIiSNI NVINOSHiUNS SaiHVMan L I B R A R I E S SMITHSONIAN INSTITl <0 2: , CO 2 " ^ 2: E I I - _j 2 /w!^91^4\ —I 2 \.^s. jSv —I /'rk^TrJV.^-' — i ‘ i 1'^ ^ I I I % iiniiiSNi NviNosHiiws^^sa I ava a n u b rar 1 es^smithsonian institution NoiiniiiSNi_NviNosHiiws S3iav -, \ c/5 ~ CO c/5 ^ 2 C/5 o 5 "• 5 Z ' ■ _1 2 _j 2 -j BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHimS S3iavaail LIBRARIES SMITHSONIAN INSTITl z ^ ^ 2 r- z Z ^ S ~ ^ S^*sov>x ? . i-n o xCvtsvJiTv. _ ... O ,^rav'5v,trx m _ • m m ' N^ovTUyiV' m ^ xo/m^iX m 'X^vdc^' — m CO £ — w \ — CO ± CO iiiniiiSNi NviNOSHims S3iavaan libraries Smithsonian institution NoiiniiiSNi nvinoshiiiais S3iav 2 c/5 Z ,.. 2v-CO z CO BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHi^S S3iavaan LIBRARIES SMITHSONIAN_ INSTIT CO 2 CO . c/5 ^ CO ^ .CO m ^ CO ^ . W5 o DC^ ~ X^/Iis^ O — NiOiu^iX o Z o iliniliSNl“'NVINOSHillNS S3iavaai1 U BRAR l ES^SMITHS0NIAN~‘|NSTITUTI0N NOIiniliSNl“'NVINOSHimS S3iav ~ 2 r- 2 r- 2 r- •, 2 m 5; X.jMsviiX m m xovosvi^ x^vAivj^' m W' ^ Xjvas CO ' — CO _ CO 5 CO BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillAJS SSiavaaiT LIBRARIES SMITHSONIAN INSTITl ^ ^ ^ z X CO 2 CO — "* - < \v S ^ »7>*w s ^ < iliniliSNI_NVINOSHillNS S3 I a Va 8 13 LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillNS S3iaV CO_— /n — cn , CO o Z' o — x-2;a^ o Z _J 2 -J ^ BRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillNS S3iavaai3 LIBRARIES SMITHSONIAN INSTIT z r z r- 2 «" z “ 2 . rm o ~ .. o xr^w^rx — xv^sovx, 9 . rr, CD > w- rn ""^^x ^ m m ^'•>2^ CO ^ £ (f, £ — CO — CO DliniliSNI NVINOSHillNS S3iaVaai3 LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHillNS S3 I a\( 2 CO 2 CO 2 X- CO 2 CO ,.5 X^S^x 1 x«^xc ? xc«?^3rx < x^x. S < Xs^^^ 1