Hayes 335 ARIKAREEAN (OLIGOCENE-MIOCENE) HERPETOTHERIUM (MARSUPIALIA, DIDELPHIDAE) FROM NEBRASKA AND FLORIDA F. Glynn Hayes1 North American Herpetotherium (Marsupialia, Didelphidae) is a small mouse-sized (<80g) opossum that ranged from the middle Eocene into the early Miocene (Uintan-Hemingfordian) of the Great Plains, Pacific Northwest, and Atlantic Southeast. New Arikareean (30-18.8Ma) Herpetotherium samples are described from the lower Arikaree Group of Nebraska (Dw-121 or Ridgeview Local Fauna, a locality south of Chadron, from localities of the Wildcat Ridge region, and the McCann Canyon Local Fauna) as well as from 4 localities in Florida (I-75, Brooksville 2, SB-1A, and White Springs). The samples from Dw-121 and Brooksville 2 are the first large (>25) populations of Herpetotherium described from the Arikareean age. Three species are recognized in the Arikareean: H. fugax, H. youngi, and H. merriami. Most early Arikareean samples are referred to H. fugax and are slightly smaller on average than older H. fugax. Later Arikareean Herpetotherium samples are referable to H. youngi on the basis of possessing a single central stylar cusp. A large species, H. merriami, from the early Arikareean John Day Formation in Oregon is similar to Herpetotherium molars described from the early Arikareean or late Whitneyan Florida I-75 locality. Taxonomy of Herpetotherium has previously been based on tooth size and upper molar stylar cusp morphology. H. fugax exhibits several configurations of stylar cusp morphology, separate to fused central stylar cusps, and a wide range of dental size variation. These ranges of variation are similar to those exhibited by the Chadronian species, H. valens. Key Words: Herpetotherium; Arikareean; taxonomic review; dental variation; Florida; Nebraska. INTRODUCTION Herpetotherium is a small (<40-80g) mouse-sized didelphid marsupial that ranges from the late Eocene to the early Miocene of North America (Rothecker & Storer 1996). It is a common component of Arikareean (30-18.8Ma) North American Land Mammal “Age” (NALMA) micro-mammal faunas from the Great Plains, Pacific Northwest, and Atlantic Southeast. How- ever, material usually consists of only scarce, fragmen- tary dental material and isolated teeth. Because of the lack of large samples, numerous described species have been based solely on stratigraphic occurrence or small differences in tooth morphology (particularly stylar cusp configuration in the upper molars), as well as small dif- ferences in dental size between species. In localities that produce significant numbers (>25) of specimens it is becoming increasingly evident that Herpetotherium populations show a wide range of dental variation (Hough 1961; Morton & Green 1976; Setoguchi 1978; Eberle & Storer 1995). In the most recent review of North American Ter- tiary marsupials, Korth (1994) reported no large samples of Arikareean Herpetotherium (>15) and many of those that were discussed had vague chronologic or strati- graphic resolution (Stock & Furlong 1922; J. Macdonald 1963, 1970; L. Macdonald 1972; Korth 1992). Much of the material discussed in this report from the Arikareean of Nebraska and Florida represents relatively large samples with more precise stratigraphic and biochronologic correlation. These samples provide more evidence that coexisting Herpetotherium populations exhibit a wide range of dental variation, both in size and morphology. I also include a review of the Florida record of pre-Pliocene (before Didelphis migrated from South America) marsupials. McGrew (1937) described the first Arikareean mar- supial in the Great Plains from Nebraska. At this time marsupials were well known from the late Eocene and Oligocene White River Group of Nebraska (Simpson 1928), but none from the younger Arikaree Group de- posits that overlie the White River. McGrew named a new species, “Peratherium” youngi, based on a single upper molar (UC 1544) collected from the upper Arikaree 1Department of Geosciences and State Museum of Natural History, University of Nebraska, Lincoln, NE 68588; Bull. Fla. Mus. Nat. Hist. (2005) 45(4): 335-353 336 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Group lower Harrison “beds” near Agate, Nebraska. The first North American Arikareean marsupial, “Peratherium” merriami had earlier been described on the basis of a single partial skull from the John Day For- mation of Oregon (Stock & Furlong 1922). The skull was collected from the lower Logan Butte volcaniclastic sediments that have been radiometrically dated (single crystal Ar/Ar) to be older than 28.8 Ma (Tedford et al., 2004). This would place H. merriami in the early Arikareean or possibly the late Whitneyan. The Picture Gorge Ignimbrite occurs at the top of the Logan Butte section and recent work by Robert Hunt (pers. comm., 2004) suggests that the lower part of the section is more likely Whitneyan in age. The third Arikareean “Peratherium”, P. spinderli, was named by Macdonald (1963) from the Wounded Knee faunas of South Da- kota. Korth (1992) described the only other Arikareean Nebraska Herpetotherium sample from the McCann Canyon LF. The McCann Canyon paper included most of the small mammals (except the carnivores) from a locality believed to be a Harrison Formation equivalent although the sediments cannot be directly correlated with the type Harrison of the Pine Ridge. Korth’s biochronologic analysis suggested an early late Arikareean age (24-22Ma, Tedford et al., 2004) for the fauna. In his paper Korth referred all later Tertiary North American didelphines to Cope’s (1873) Herpetotherium which followed Crochet (1977), who separated Euro- pean Peratherium from North American “Peratherium” because Herpetotherium has a dominant central stylar cusp on upper molars whereas Peratherium species show a dominant stylar cusp B. Further, based on com- parison of the lower dentition of the material from McCann Canyon, Korth formally synonymized P. spindleri with H. youngi. Korth (1992) distinguished other species of Herpetotherium from H. youngi on the basis of size differences. In the Great Plains, Green and Martin (1976) re- viewed didelphids from the Oligocene and Miocene of South Dakota. In their analysis they found stylar cusp morphology of the upper molars to be variable and com- mented on the validity of diagnosing species on this ba- sis. They suggested that “P. spindleri” and “P. youngi” might be synonyms of “P. fugax” but the samples were too small to make a conclusive assignment. Their re- view also included description of a few teeth of “Peratherium sp.” from the “Hemingfordian” NALMA Black Bear Quarry II LF from the Rosebud Formation that was slightly smaller than older “Peratherium”. Later, Martin and Green (1984) described other small mam- mals from the fauna and extended the range of “Peratherium” sp. into the Hemingfordian Batesland Formation, but they did not describe this Batesland sample. In addition, they correlated the South Dakota Rosebud Formation with the “Marsland” (=“Upper Harrison” beds) of Nebraska which would place the small Rosebud marsupial sample in the late Arikareean (Tedford, et al. 1987, 2004). Reports have listed Oligocene and Miocene mar- supials from Florida, but none have discussed them in detail. Patton (1964) reported a didelphid from the I-75 LF. Wolff (1987) referred all then known Florida mar- supials from four localities to “Peratherium.” Later, Morgan and Pratt (1988) also listed “Peratherium” from the late Hemingfordian Brooks Sink locality and the Barstovian NALMA “Nichols Mine” fauna (now Bird Branch LF). I assign most of the Florida marsupials to Herpetotherium, except for those from Brooks Sink and Bird Branch which instead appear to be closely related to marmosine opossums. A partial didelphid lower tooth reported from the Barstovian of Texas (Slaughter 1978) is very similar to the marmosine-like Florida material and may not be a herpetotheriine as reported by Korth (1994). This would therefore restrict Herpetotherium to Florida localities no younger than the early Hemingfordian. Here, I describe all of the above Herpetotherium specimens, along with newly discov- ered specimens from the Brooksville 2 LF (Hayes 2000), from the White Springs LF (Morgan 1989), and the SB- 1A LF (Frailey 1978). Korth (1994) most recently reviewed all Chadronian through Hemingfordian North American marsupials. He recognized two families of marsupials during this time, Didelphidae (represented by Herpetotherium and Copedelphys) and Peradectidae (represented by Nanodelphys and Didelphidectes). Most post-Duchesnean NALMA Peratherium, except for P. titanelix and P. stevensoni (which Korth placed in a new genus, Copedelphys) were allocated to Herpetotherium as a genus distinct from Peratherium following his 1992 paper. Korth recognized five species of Herpetotherium (H. valens, H. fugax, H. merriami, H. youngi, and H. sp. from the Rosebud Formation) on the basis of size and stylar cusp differences. All of these species were thought to be derived from the Duchesnean Peratherium knighti. P. donahoei was synonymized with Herpetotherium valens, and occurs in the 337 Chadronian. Korth separated several of the small speci- mens from the Chadronian Pipestone Springs fauna as representing H. fugax, therefore extending the range of this species back to the middle Chadronian. H. fugax is the sole Orellan NALMA species. Korth did not recog- nize a Whitneyan NALMA species even though he had referred the probable (Emry et al. 1987) Cedar Ridge LF marsupial described by Setoguchi (1978) as “P. cf. spindleri” to H. fugax. By the Arikareean, Korth sug- gests Herpetotherium diversity increased to two or three species: (1) H. merriami, a relatively large species with pleisomorphic stylar cusp morphology; (2) H. youngi, a small species with a derived single central stylar cusp, and (3) H. sp., a very small species from the Rosebud Formation of South Dakota. After Korth’s review, Eberle and Storer (1995) reported on Herpetotherium from the Chadronian Calf Creek LF, Cypress Hills Forma- tion, Saskatchewan. Based on stylar cusp variation, two species of Herpetotherium were thought to exist in the Calf Creek LF. The authors’ analysis showed that, even though the sample did exhibit considerable differences in stylar cusp configuration and tooth size, this variation was normally and continuously distributed and therefore the material belonged to a single species, Herpetotherium valens. The following year, Rothecker and Storer (1996) described the marsupials from a stratigraphically lower fauna in the Cypress Hills Formation. In their analysis of the marsupials from the Duchesnean Lac Pelletier lower fauna, they proposed that all pre-Duchesnean North American “Peratherium” that Korth did not as- sign to Herpetotherium be so assigned, because the species from the Lac Pelletier lower fauna were transi- tional between older Uintan Herpetotherium and the Chadronian Herpetotherium valens. The samples showed an indistinguishable morphocline from earlier Peratherium-like species to later species of Herpetotherium. Although two species of Herpetotherium (H. sp., cf. marsupium and H. sp., cf. knighti) were described, based on distinct size and mi- nor morphology differences, stylar cusp variation in po- sition, twinning, and separation of cusps was noted for both. eral localities in the lower Arikaree Group (including the Gering Formation) of the Wildcat Ridge region of west- ern Nebraska (Fig.1). As mentioned above, Korth (1992) described the marsupials from University of Nebraska State Museum localities Cr-125 and Cr-117, collectively termed the McCann Canyon LF. I am only adding addi- tional newly collected material to the McCann Canyon LF. The other samples have not previously been de- scribed. UNSM Dw-121. A new Arikareean Herpetother- ium sample was recovered from a locality south of Chadron, Nebraska, discovered by Nebraska highway salvage paleontologist Bruce Bailey. This locality, desig- nated UNSM Dw-121 (Bailey 1992, 1999, 2004) was found in fluvial sediments of the basal Arikaree Group incised into the fine-grained volcaniclastics of the White River Group along the axis of the Pine Ridge paleovalley (northern Gering paleovalley of Swinehart et al. 1985). Bailey (2004) recently listed the fauna from Dw-121 and named it the Ridgeview LF. It is one of the stratigraphically lowest localities producing microvertebrates (the other is Wagner Quarry [Hayes, 2004]) of the Arikaree Group in the Pine Ridge region. A diverse fauna has been recovered from the sediments representing at least 26 mammalian species, including the hedgehog Ocajila and the rodent Geringia. Faunal comparison suggests an early Arikareean age for the site (Bailey, 2004, Tedford et al., 2004). The marsupials recovered from UNSM Dw-121 are represented by over 300 specimens of Herpetotherium and over 50 specimens of an undescribed species of Nanodelphys. The two are sepa- rable on the basis of size and stylar cusp morphology. Nanodelphys is considerably smaller and all the upper molars have an ectoflexus and lack a central stylar cusp unlike Herpetotherium which has a central stylar cusp or cusps and an ectoflexus only on M3. Gering Formation. Marsupials were collected from the “brown siltstone”, Gering Formation and other undifferentiated lower Arikaree Group deposits (= “Mon- roe Creek”) along the Wildcat Ridge of western Ne- braska by Swisher (1982). These sediments are consid- ered early early Arikareean, ~30-28 ma (Tedford et al. 1987, 1996, 2004) and unconformably overlie rocks of the earlier Oligocene White River Group, Brule Forma- tion. The age is also supported by several radiometric dates (Tedford et al. 1996) taken from ash deposits within the rocks. Swisher (1982) collected at 27 localities of the Wildcat Ridge, of which eight (Swisher 1982 locali- MATERIALS AND METHODS NEBRASKA LOCALITIES Arikareean-age Herpetotherium described in this paper were collected from a locality in the Pine Ridge region (UNSM Dw-121= Ridgeview LF) and from sev- HAYES: Arikareean Herpetotherium From Nebraska and Florida 3 3 8 C E N O Z O IC V E R T E B R A T E S : P apers to H onor S . D avid W ebb Figure 1. Age relationship and location of Nebraska and Florida Herpetotherium localities. 339 ties: 1- in the “brown siltstone”, 2, 3, 11, 21, 23- in the Gering Formation, and 27- in the undifferentiated lower Arikaree). Eight localities produced some Herpetotherium material, predominately lower teeth and jaw fragments that he referred to “Peratherium spindleri.” FLORIDA LOCALITIES Herpetotherium occurs at four localities in Florida that range in age from late Oligocene to early Miocene (Fig. 1). Prior to this report, marsupials were reported from four localities in the state: I-75, Brooks Sink, Bird Branch, and Thomas Farm. Herpetotherium is found in only the I-75 locality and three new localities: Brooksville 2, SB-1A, and White Springs. The Brooks Sink and Bird Branch marsupials are not herpetotheriines and will be reported on elsewhere. The well-known Hemingfordian Thomas Farm site is removed from the list because I believe that the single reported didelphid tooth (UF 97364; Wolff 1987; Marshall et al 1990) is a contami- nant introduced from another locality. I recognized it as an upper molar of Nesophontes derived from prior screen washing of sediment from a Haitian site. Samples of marsupials from Florida sites vary from single specimens (e.g., SB-1A) to more than fifty teeth (Brooksville 2 LF). As most of the sites are sinkhole or fissure-fill deposits, they can only be dated by biochronologic correlation. However, a few sites were collected from sediments that inter-tongue with or are superposed by marine strata which permits aging by stable isotopic means. Biochronologic analyses by Albright (1999) and Hayes (2000) have led to a better understanding of the temporal relationships between the late Oligocene and early Miocene land mammal sites in Florida. I-75 LF. Patton (1969) recorded the I-75 LF from a small sinkhole deposit uncovered by construction of Interstate 75 just west of Gainesville, Florida (Alachua County). It is a small but diverse sample that includes taxa known previously from only Western North Ameri- can localities. Patton (1969) tentatively identified its age as Whitneyan (late Oligocene) because of the presence of Mesohippus, a paleolagine rabbit, Nanotragulus, two small oreodonts, and Paleogale. Current evidence sup- ports either a Whitneyan or possibly early Arikareean land mammal age for the I-75 LF (Hayes 2000). White Springs LF. The White Springs LF was recovered from ancient shoreline sediments of the Parachula Formation exposed along the Suwannee River lying unconformably on the Oligocene Suwannee Lime- stone. Morgan (1989) placed the site in the early Mi- ocene (22-21 Ma) on the basis of correlation of the marine sediments to the upper N4 and lower N5 sub- tropical foraminferal zones. However, Jones et al. (1993) obtained an 87Sr/ 86Sr date of 24.4 +/- 0.5-1.0 Ma from mollusk shells at the locality. Recent faunal analysis has supported this older date (Albright 1998; Hayes 2000). Screen washing of sediment from the site White Springs 3B produced a single specimen of marsupial, the trigo- nid of a lower molar. SB-1A (= Live Oak) LF. The SB-1A LF (also known as the Live Oak LF) includes Mammacyon cf. obtusidens, Paroligobunis frazieri, Nothokemas waldropi, and Protosciurus. Frailey (1978) described most of the larger taxa and assigned SB-1A to a “middle” Arikareean age. Frailey characterized SB-1A as an unstratified conglomeratic sequence above the Suwannee Limestone (Oligocene). The vertebrates are all terrestrial and the sediments were poorly sorted clasts of Suwannee Limestone in a matrix of clay derived from weathering of the limestone. On this basis, Frailey sug- gested that the fossil-bearing deposit originated as slopewash and slumping in a fissure. SB-1A produced a single marsupial specimen, a lower molar. Brooksville 2 LF. The late Oligocene age of Brooksville dates from Patton (1967), who described four specimens that he believed showed affinities to Great Plains taxa of that time. This fauna was desig- nated as the Brooksville 1 LF (Hayes 2000), because in 1994, fissure fills in the Suwannee Limestone at a dif- ferent location near Brooksville produced an abundant and diverse assemblage of vertebrates designated the Brooksville 2 LF. This new fauna includes the largest sample of Florida marsupials. Brooksville 2 contains Miohippus, an entoptychine rodent, Palaeogale, Megalagus, Agnotocastor, and Nanotragulus. These taxa, in comparison to those of the Great Plains, suggest an early middle Arikareean age (Hayes 2000). MEASUREMENTS AND TECHNOLOGY A Gertner measuring microscope was used to measure to the nearest 0.01 mm the same dental pa- rameters established by Korth (1994:fig. 1) to enable direct comparison with the specimens of his study. Den- tal terminology used is shown in Figure 2. In tables and text, a lower case letter indicates lower teeth, while an upper case letter indicates upper teeth. Cheek teeth are described as P3, M1, M2, M3, and M4, instead of M1, HAYES: Arikareean Herpetotherium From Nebraska and Florida 340 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb M2, M3, M4, and M5 (Marshall 1990). The position of a tooth in the toothrow used in the study of variation was determined by association in a maxilla or jaw with the exception of M3 which, because of its distinct morphol- ogy, could be easily separated from other molars. All measurements are in millimeters unless otherwise stated. Quotation marks are used where informal or obsolete nomenclature is retained for purposes of discussion. Abbreviations.—(others defined in text) ap, ante- rior to posterior measurement, length; apt, anterior to posterior measurement of trigonid on lower molars, length; CV, coefficient of variation; F, Fauna; LF, Local Fauna; lf, local fauna; M, mean; m# or M#, molar; N, number of specimens; OR, observed range of variation; p# or P#, premolar; S, standard deviation; SE, standard error of the mean; tr, transverse measurement, width; tra, transverse measurement of anterior trigonid width in lower molars; trp, transverse measurement of poste- rior talonid width in lower molars. Institutional abbreviations.—AMNH, American Museum of Natural History, New York; UC, University of Chicago collections, Field Museum, Chicago, Illinois; UCMP, University of California Museum of Paleontol- ogy, Berkeley, California; UF, University of Florida col- lections, Florida Museum of Natural History, Gainesville; UNSM: University of Nebraska State Museum, Lincoln. GENERAL UPPER MOLAR DESCRIPTION OF HERPETOTHERIUM The upper molars of most species of Herpetotherium are characterized by a large central stylar cusp or cusps and equal, reduced, or absent stylar cusp B (Figs. 2-3). In the oldest (Uintan NALMA) species, stylar cusp B may be larger than the central cusps. Peratherium has a reduced central cusp and enlarged stylar cusp B (for other comparisons see Korth 1994). Lower molars dif- fer only in size from species to species of Herpetotherium (Korth 1994). Upper molars have the metacone higher than pro- tocone. The centrocrista is ‘V’-shaped. The M1 is longer than wide with slight to no flexion in the central buccal margin. The protocone is situated along the anterior margin of the molar. The metaconule and protoconule are usually present with metaconule larger and protoconule quickly worn away. An anterior cingulum runs from stylar cusp A, which can be distinct or inte- grated into the cingulum, and joins the preprotocrista. A narrow posterior cingulum connects with the postprotocrista and ends below the metacone. Stylar cusp B joins the paracrista but is usually distinct in unworn teeth and can be as large as the central stylar cusp/ cusps in Herpetotherium valens and older species. The central buccal margin contains one or two distinct cusps, C and D, in varying degrees of separation or fusion with one another. Stylar cusp D is generally the largest cen- tral stylar cusp with C becoming more reduced in younger samples, and possibly completely fusing with D in the youngest samples. The position of D varies from being in line with the premetacrista to being well posterior and in line with the metacone. Stylar cusp E is minute, joins with the metacrista, and disappears quickly with wear. The M2 is similar to the M1 in width and length and the buccal edge has a slightly more well-defined ectoflexus. Stylar cusp morphology resembles the M1. Stylar cusps C and D are centrally located and can vary in the same manner as observed on the M1. The M3 is much wider than long with a pronounced ectoflexus (Fig.3D). The protocone is central rather than anterior as in the M1 and M2. Stylar cusp B can be reduced or absent in younger samples, but is usually present and can be of equal height to the central cusp/ cusps in unworn teeth of older species. The M3 typi- cally exhibits only one central stylar cusp which has been Figure 2. Herpetotherium dental nomenclature for a right upper molar. 341 Figure 3. Teeth of Herpetotherium from Florida, occlusal and buccal views, scale bar = 1 mm. I-75, (A) UF 97361 L M2; (B) UF 97362 L M2. Brooksville 2: (C) UF 156255 R M4; (D) UF 156292 R M3; (E) UF 156262 R M2; (F) UF 163554 R M1; (G) UF 156294 R p2-p3, m1; (H) UF 156283 lower cheek tooth; (I) UF 156284 lower cheek tooth; (J) UF 163574 L m4; (K) UF 163572 L m4. HAYES: Arikareean Herpetotherium From Nebraska and Florida 342 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb cusp C with D completely absent. However, both au- thors have noted M3s that exhibit a central stylar cusp with twinned apices or even M3s with two separate central cusps that suggest that D is not absent. The new material described here also shows these variations. In teeth that have two cusps the larger is often the more posterior of the central cusps. The dominant central cusp appears to be a fusion of C with a larger D, rather than the loss of D. The M3 is therefore similar to the M1 and M2 in trends of stylar cusp morphology. The process of fusion may have progressed more in the M3 of older species. This evidence, that the central cusp of Herpetotherium M3s and M4s may be stylar cusp D and C fused together, suggests that the diagnoses of Korth (1994) and Eberle and Storer (1995) should be emended. The M4 is very compressed anteroposteriorly, yet is almost as wide as M3 (Fig. 3C). The metacone is small at the posterior margin and not connected to the paracone by a crista. The anterior cingulum begins from stylar cusp A, thickens anterolingually to paracrista, and is continuous with preprotocrista. Protocone and para- cone are centrally aligned. There is a small reduced single central stylar cusp that could represent C, D, or both of them fused together. SYSTEMATIC PALEONTOLOGY Class MAMMALIA Linnaeus 1758 Subclass THERIA Parker and Haswell 1897 Infraclass MARSUPIALIA Illiger 1811 Order DIDELPHIMORPHA Gill 1872 Family DIDELPHIDAE Gray 1821 Subfamily HERPETOTHERIINAE Trouessart 1879 Genus HERPETOTHERIUM Cope 1873 HERPETOTHERIUM FUGAX (Cope 1873) Holotype.—AMNH 5254, R M1-M4. Type Locality.—Cedar Creek Beds, White River Formation, Logan County, Colorado. Revised Range.—Middle Chadronian through early Arikareean of the Great Plains, Montana, Wyo- ming, and Colorado. Revised Diagnosis.—Smaller on average than H. valens and H. merriami. Similar in size to H. youngi. Slight decrease in average size from Chadronian to early Arikareean. Stylar cusp B reduced in comparison to H. valens and larger than H. youngi. Central stylar cusps variable. C and D may be separate to completely fused into a single central stylar cusp. SPECIMENS FROM DW-121= RIDGEVIEW LF, NEBRASKA Referred Specimens.—UNSM 77041-77360, UNSM 77041 LM2-4, 77042 RP2-M1, 77043 LM2-4, 77044 LM1-4, 77045 RM2-4, 77046 RM2-3, 77048 RM1- 4, 77049 RM2-4, 77050 LP3-M1, 77051 LM2-3, 77053 RM-2, 77057 LP3-M1, 77059 LM1-2, 77061 RP3-M4, 77062 LM1-3, 77065 LM1-2, 77067 RM3-4, 77072 RM1- 2; UNSM 77064, 77066, 77067, 77071, 77089-77095, 77111, 77117, 77133-77145, 77148-77150, 77152-77154, 77157-77158, 77181, 77197, 77198, 77202, 77203, 77210, 77212, 77215, 77225, 77229: 77219, 77220, 77221, 77224, 77228, M1s; UNSM 77063, 77098, 77101, 77102, 77104, 77133, 77135-77139, 77196, 77200, 77211, 77217, 77227, 77366, 77367, M2s; UNSM 77068 p3, 77070 m2-3, 77073 m2, 77074 p2-3, 77076 m1-3, 77077 m2, 77082 p2-3, UNSM 77070, 77075, 77076, 77140, 77141, 77143, 77144, 77146, 77151, 77154, 77155, 77164, 77167, 77168, 77185, 77194, 77205, 77364, m4; UNSM 77068, 77074, 77076, 77077, 77080, 77082, m1. Description.—Dental measurements presented in Table 1. The M1 central stylar cusp/cusps are always larger than stylar cusp B in the DW-121 sample. There are four configurations of the central stylar region. In smaller teeth, there is a single conical cusp, a complete fusion of C and D (Fig. 4F). Slightly larger teeth show a single elongated blade-like central cusp (Fig. 4I). In large teeth, C forms a distinct small cusp on the anterior slope of D (Fig. 4L). The largest teeth have C and D separate with D the dominant cusp (Fig. 4N). The most common variations are a single blade cusp or the reduced C (Table 2). Stylar cusp E is occasionally present as a slight rise on the buccal edge at the metacrista. In M2s there can be a single central cusp that is either conical (Fig. 4E) or blade-like (Fig. 4H). If two central cusps are present, D is dominant with C moved slightly lingual from the buccal margin (Fig. 4K). Stylar cusp B is usually smaller that the central cusps except in large M2s. Single and reduced C are the most common configurations (Table 2). Similar to the M1, M3 exhibits four central stylar cusp morphologies based on the size of the teeth: A small single cusp (Fig 4D). A large elongate single cusp (Fig. 4G). A twinned central cusp with a distinct furrow in the center that separates the cusp into equal parts (Fig. 4J). The fourth morphology is a doubled central cusp with D dominant and C slightly smaller and equal to B (Fig 4M). Almost all the M3s that were relatively unworn exhib- ited a twinned central cusp (Table 2). Discussion.—The sample from Dw-121 is the sec- 343 Table 1. Dental measurements of Herpetotherium fugax from Dw-121, Nebraska. Individual measurements in all tables that do not have catalog numbers directly above them are taken from a single dentary/maxillary fragment with more than one tooth whose catalog number can be found above the first measurement of the specimen. UNSM77082 UNSM77074 UNSM77079 p2 a-p 1.39 1.37 1.43 tr 0.56 0.61 0.55 UNSM77080 UNSM77068 p3 a-p 1.31 1.27 1.30 1.37 tr 0.65 0.68 0.62 0.68 N M OR R SD SE CV m1 a-p 8 1.66 1.53-1.78 0.25 0.09 0.03 5.4 tr 8 0.97 0.88-1.05 0.17 0.04 0.02 4.1 UNSM77070 UNSM77077 UNSM77073 m2 a-p 1.77 1.62 1.73 tr 1.02 1.04 1.07 UNSM77076 UNSM77075 m3 a-p 1.66 1.53 1.79 tr 1.10 0.94 1.14 N M OR R SD SE CV m4 a-p 19 1.77 1.49-1.97 0.48 0.12 0.03 6.8 tr 19 1.00 0.90-1.14 0.24 0.07 0.02 7.0 UNSM77042 UNSM77057 UNSM77050 P3 a-p 1.59 1.46 1.86 tr 0.80 0.82 0.92 N M OR R SD SE CV M1 a-p 27 1.7 1.45-1.93 0.48 0.11 0.02 6.4 tr 27 1.58 1.39-1.81 0.42 0.09 0.02 5.7 M2 a-p 15 1.74 1.48-1.87 0.39 0.12 0.03 6.9 tr 15 1.85 1.52-2.11 0.59 0.15 0.04 8.1 M3 a-p 55 1.8 1.58-2.01 0.43 0.11 0.01 6.1 tr 55 2.11 1.92-2.36 0.44 0.11 0.01 5.2 M4 a-p 7 1.02 0.87-1.14 0.27 0.10 0.04 9.8 tr 7 1.95 1.79-2.07 0.28 0.10 0.04 5.1 HAYES: Arikareean Herpetotherium From Nebraska and Florida 344 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Figure 4. Variation in Arikareean Herpetotherium stylar cusp arrangement, x12. Buccal view shown. Left molars have been reversed in view. (A) H. merriami, UCMP 24240 R M1-M3. “Gering”: (B) UNSM 81321 R M3; (C) UNSM 11576, R M1-M2. Dw- 121: (D) UNSM 77210 R M3; (E) UNSM 77121 L M2; (F) UNSM 77099 L M1; (G) UNSM 77133 R M3; (H) UNSM 77204 L M2; (I) 77101 R M1; (J) UNSM 77135 L M3; (K) UNSM 77190 R M2; (L) UNSM 77134 L M1; (M) UNSM 77145 R M3; (N) UNSM 77100 R M1. Brooksville 2: (O) UF 156242 R M3; (P) UF 163555 R M2; (Q) UF 163555 R M1; (R) UF 156262 R M2; (S), UF 156291 L M1. Cr- 125 (McCann Canyon): (T) UNSM 4197 R M3; (U) UNSM 24200 L M2; (V) UNSM 77172 R M1. 345 described by Korth (1994) and Eberle and Storer (1995) in their diagnoses of Herpetotherium as representing the second largest sample of Herpetotherium (N= 300+) known from a single locality (the first is the aforemen- tioned Calf Creek LF). The Dw-121 teeth are on aver- age slightly smaller than those of Whitneyan and older Herpetotherium (Fig. 5). The Dw-121 sample has a continuous distribution of stylar cusp characters and size similar to Herpetotherium valens from the Calf Creek LF sample (Eberle & Storer 1995). In general, the teeth show that stylar cusp variation is a function of size. The smaller teeth have a single central cusp and the largest teeth exhibit two distinct cusps, those of intermediate sizes exhibit partial fusion of the central cusps. Although there is not a clear distinction between size groups, all the variations overlap. The retention of variable single to double central stylar cusps in comparison to Herpetotherium youngi, which was diagnosed by Korth (1994) as having a single central cusp, suggests the early Arikareean samples from Nebraska should be assigned to Herpetotherium fugax. Because the fluvial sediments at Dw-121 incise the White River Group, the question arises as to the possibility of reworking of older Herpetotherium mate- rial into the sample. However, based on the associated taxa and the preservation of the specimens, there is no evidence of reworking of material from the underlying Brule Formation (Bailey 1992). SPECIMENS FROM BROOKSVILLE 2 LF, FLORIDA Referred Specimens.—UF 156251-156293 isolated cheek teeth; UF 156294 Rp2, p3, m2; UF 156295 RM1- 4; UF 156296-156304 isolated cheek teeth; UF 163551- 163560 isolated teeth (Fig. 3C-K). Description.— Dental measurements presented in Table 3. Stylar cusp B joins the paracrista but is distinct and can be almost as large as C in unworn teeth. Stylar cusp E joins with the metacrista. Stylar cusp D is the largest stylar cusp, usually fused with C, but on some specimens C is distinguishable from D as a cuspule or furrow on the anterior slope (Fig. 4S). The position of D varies from being in line with the pre-metacrista to be- ing posterior and in line with the metacone. The M2 has stylar cusp B distinct but connected to the paracrista. Stylar cusp A is a minute cusp beside B and forms a small lateral flange, which thickens lin- gually into the anterior cingulum. Stylar cusp D is cen- trally located and the largest cusp with C occasionally forming a cuspule on the anterior slope (Fig. 4R). Stylar cusp B is present but reduced on the M3. The central dominant cusp is usually twinned with a fur- row at the apex that equally divides the cusp (Fig. 4O). Stylar cusp E is absent or quickly reduced with light wear. Discussion.—The Brooksville 2 sample is the larg- est sample of marsupials in Florida. The size of the up- per molars falls within the smaller measurements for Herpetotherium fugax and encompasses the measure- ments for Herpetotherium youngi (Tables 1, 3; Fig. 5). The sizes of the lower teeth are similar to H. youngi and H. fugax. The Brooksville 2 sample does show a greater degree of size variation than Dw-121, particu- larly the M2 and M3, and the CV values were higher than 10 for the M2 and M4 (Table 3). This could indi- cate that the Florida species is highly dimorphic in the posterior molars in comparison to Dw-121, that more than one species is present in the fauna, or that there was significant time in the deposition of the site to accu- mulate a greater range of non-coexisting Herpetotherium populations that occupied the diverse Florida landscape of the time in comparison to the Great Plains (Hayes 2000). Varying stylar cusps in comparison to Herpetotherium youngi suggests that Brooksville 2, like the early Arikareean samples from Nebraska, be as- signed to Herpetotherium fugax. Brooksville 2 does generally show a greater degree of fusion between cusps C and D relative to the Dw-121 sample. There were few teeth found that have the two cusps completely sepa- rated. The early “medial” Arikareean Brooksville 2 mar- supials show transitional features between H. fugax and the “later” Arikareean H. youngi. There was one exception to the general Herpetotherium stylar cusp morphology. UF 163554 is an M1 with a posteriorly placed cusp D, a very small Single Blade Side Double M1 3 8 8 2 M2 7 3 8 1 M3 7 3 0 10 Table 2. Variation of central stylar cusp in upper molars from Dw-121, Nebraska. Each box indicates the number of teeth showing the particular variation. Variation states defined in text and Figure 4. HAYES: Arikareean Herpetotherium From Nebraska and Florida 346 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Figure 5. Comparison of Herpetotherium upper molar measurements from Duchesnean through Arikareean NALMAs. Compiled from author’s measurements; Stock and Furlong 1922 (H. merriami); Green and Martin 1976 (H. sp.- Rosebud); Setoguchi 1978 (H. fugax- Cedar Ridge); Korth 1994, 1992 (H. fugax, H. youngi-McCann Canyon); Eberle and Storer 1995 (H. valens- Calf Creek); Rothecker and Storer 1996 (H. sp., cf. H. marsupium, H. sp., cf. H. knighti). 347 Figure 5. (cont.) anteriorly placed cusp C, and an enlarged cusp B that is equal in size to D (Fig. 3F). This stylar cusp configura- tion is closest to the oldest Herpetotherium species ex- cept for the smaller cusp C. With only a single tooth showing this condition, it is probably an aberrant condi- tion or as stated above a less probable indication that a second but similar species is present. HERPETOTHERIUM YOUNGI (McGrew 1937) Holotype.—UC 1544, L M2. Revised Range.—Middle to latest Arikareean of South Dakota and Nebraska. Revised Diagnosis.—Similar in size to early Arikareean and Whitneyan Herpetotherium fugax. Smaller than Herpetotherium valens and Herpetotherium merriami. Stylar cusp B on all molars reduced in comparison to H. fugax. Single central sty- lar cusp on upper molars (total fusion of C and D). Older species have varying degrees of central stylar cusp sepa- ration. Referred Specimens from Cr-125 and Cr-117.— UNSM 24200 LM2-3; UNSM 24186, 24187, 24189, 24191, 24195 jaws with lower cheek teeth; UNSM 24188, 24190, isolated lower cheek teeth. UNSM 24197 RM3, 24201 RM3, and UNSM 24202 RM1. Additional material: UNSM 4197 RM3, UNSM 24800 jaw with lower cheek teeth, UNSM 77172 RM1, 77173, LM1, UNSM 77179 LM2. Description.—Upper molars with a single conical central stylar cusp (Fig. 4T-V). Stylar cusp B is smaller than the central cusp on M1 and M2 in lightly worn teeth and disappears with moderate wear. Stylar cusp B on M3 is incorporated into the paracrista and almost equal to the central cusp in height in the little worn teeth, but not as massive. Stylar cusp E is absent on all molars. Discussion.—In size, although slightly larger on average than early Arikareean Herpetotherium (Fig. 5), which is perhaps due to the small sample, the mate- rial from the McCann Canyon LF as well as the other material referred to Herpetotherium youngi by Korth (1994) falls well within the range of the Dw-121 and Brooksville 2 Herpetotherium. It also is similar in size to the Cedar Ridge Herpetotherium (Setoguchi 1978) that Korth (1994) referred to Herpetotherium fugax (Fig. 5). The new material from the McCann Canyon LF provides additional evidence that younger Herpetotherium has a fused cusp C and D as the cen- HAYES: Arikareean Herpetotherium From Nebraska and Florida 348 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb UF 156294 UF 163552 p2 a-p 1.18 1.21 tr 0.52 0.53 UF 156294 UF 163553 p3 a-p 1.12 1.13 tr 0.57 0.53 UF 156282 UF 156284 UF 156286 m1 a-p 1.73 1.75 1.66 apt 0.65 0.7 0.65 tra 0.92 0.93 0.97 trp 0.96 1.03 1.07 UF 156279 UF 156294 UF 156287 m2 a-p 1.81 1.56 1.70 apt 0.80 0.84 0.77 tra 0.89 0.86 0.95 trp 0.90 0.96 1.00 UF 156283 UF 156288 m3 a-p 1.83 1.85 apt 0.90 1.01 tra 1.15 1.06 trp 0.99 0.91 N M OR SD CV M1 a-p 9 1.68 1.51-1.81 0.11 6.5 tr 9 1.52 1.22-1.78 0.16 10.5 M2 a-p 9 1.69 1.38-1.91 0.19 11.2 tr 9 1.88 1.62-2.14 0.19 10.1 M3 a-p 15 1.70 1.40-1.90 0.14 8.2 tr 15 2.10 1.75-2.29 0.21 10.0 M4 a-p 4 1.08 0.93-1.24 0.14 12.9 tr 4 1.96 1.83-2.12 0.12 6.1 Table 3. Dental measurements of Herpetotherium fugax from Brooksville 2 LF, Florida. 349 tral sharp stylar cusp. Herpetotherium youngi was di- agnosed (Korth 1994) as a species with a single central cusp on all upper molars. The validity of H. youngi on this basis is questionable since the sample of upper mo- lars available for comparison is so small, however the diagnosis cannot be ruled out. I therefore retain H. youngi as a later Arikareean species separable from Herpetotherium fugax because it has a single central cusp and a reduced cusp B. Herpetotherium sp. from the Rosebud (Morton and Green 1976) also has a domi- nant single central stylar cusp and it falls within the size range for H. youngi for the M1, but has a smaller M2 and M3 than H. youngi. Again, this might be due to limited sample size of both H. youngi and Herpetotherium sp. Additional material might show that the Rosebud Herpetotherium could be referred to H. youngi. HERPETOTHERIUM CF. MERRIAMI (Stock and Furlong 1922) Referred Specimens.—UF 97361 LM2, UF 97362 LM2, both from the I-75 LF, Alachua County, Florida. Description.—Both specimens from this locality appear to be left M2s. Measurements for UF 97361 are: a-p =1.98, tr =1.91; for UF97362: a-p =2.02, tr =2.11. There is a very slight flexion at stylar cusp C and an otherwise straight buccal margin. Stylar cusps C and D fused at the base, but distinguishable as separate cusps with D more dominant and C forming a small cuspule on the anterior slope of D. Stylar cusp E is a minute cusp that joins the metacrista. Stylar cusp B is connected to paracrista. Stylar cusp A forms a minute flange anterior to B. There is a barely visible anterior cingulum begin- ning at the base of the preprotocrista and ending before the paracone. The protocone is on the anterior margin. Discussion.—The size of these teeth is larger than those of Herpetotherium youngi, but falls within the size range for Herpetotherium valens and Herpetotherium merriami (Figs. 3A-B, 5). H. merriami has a relatively large M1 and M2 (entirely outside the range of size variation) in comparison to other Arikareean Herpetotherium teeth, yet the M3 length falls within the upper measurements for Dw-121 (Fig. 5). This ap- parent reduction of the M3 may be a general character- istic of H. merriami and therefore useful for diagnosis of the species if more material is recovered from the John Day region. The cusp arrangement of the I-75 specimens is also similar to H. valens and H. merriami. Korth (1994) notes that H. valens has a larger cusp B than H. merriami and differentiates the two species on that basis. Eberle and Storer (1995) agreed that H. valens is similar in size to H. merriami with a larger more strongly developed stylar cusp B than other Herpetotherium species. Based on age, the I-75 teeth are closer to H. merriami. However, the degree of wear on these specimens makes it difficult to accurately as- sess the size of stylar cusp B and only a tentative refer- ral can be made based on size differences. H. merriami and the I-75 species may represent coastal variants of Herpetotherium outside of the Great Plains. HERPETOTHERIUM sp. “BROWN SILTSTONE”, GERING FORMATION, AND UNDIFFERENTIATED LOWER ARIKAREE GROUP, NEBRASKA Referred Specimens.—UNSM 81321 LM3-4, 11576 LM1-2, 14951 Rm1-3, 11684 R m1,m3, 14947 Lm1-3, 11593 Rm3-4, 11632 Rm3-4, 11652 RM1-4, 11685 Rp2-m3, 11505 Lm2, 11584 RM3-4, 11576 RM1- 2. Description.—The two M1s that are not completely worn have a single elongate central stylar cusp. Stylar cusp B is distinct and smaller than the central cusp. Sty- lar cusp E is either not present or worn away. The M2 that is relatively unworn has two separate central cusps with C larger than B and slightly offset from the buccal margin (Fig. 4C). All M3s were considerably worn and had a single elongate central stylar cusp (Fig. 4B). Lower molars are typical of Herpetotherium. Discussion.—The “Gering” marsupials are repre- sented by neither large samples nor relatively unworn specimens of upper molars, so an accurate determina- tion of species is not possible. Although most of the teeth preserve only a single central cusp, the morphology of UNSM 11576 (a relatively large M2 with doubled cen- tral cusps) suggests that variation similar to other Herpetotherium populations is present in the Gering marsupials. However, the possibility of reworking from older White River sediments cannot be excluded. Ex- cept for the M2 mentioned above, the teeth fall within the size range of the Dw-121 sample (Table 4, Fig. 5). SB-1A LF, FLORIDA Referred Specimen.—UF 97363 L m4. Description.—The left m4 from SB-1A shows typi- cal Herpetotherium characteristics. The cusp of the protoconid is broken off. The base of the valley formed by the protoconid and metaconid is higher than the one formed by the protoconid and paraconid. There is a very strong anterior cingulid present. The talonid is much re- duced and quite narrow compared to the trigonid. The HAYES: Arikareean Herpetotherium From Nebraska and Florida 350 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb hypoflexid is not pronounced, instead there is a wide shelf formed that begins at the base of the protoconid and curves to the base of the hypoconid and merges into the talonid. The cristid obliqua is low and joins the trigo- nid wall at the base of the metacone. The hypoconulid extends posteriorly from a worn entoconid. The tooth measures: a-p =1.67, tra =1.01, trp =0.62, and apt =0.78. Discussion.—The SB-1A tooth is comparable in size to Herpetotherium youngi and the Brooksville 2 sample. This tooth does compare well with a m4 of H. youngi figured in Green and Martin (1976:fig. 3e). Older species such as Herpetotherium valens appear to have a more pronounced hypoflexid in the m4. WHITE SPRINGS LOCAL FAUNA, FLORIDA Referred Specimen.—UF 125547 Left trigonid. Description.—Trigonid morphology is typical of Herpetotherium. Protoconid is the largest cusp, followed by the metaconid, and then the paraconid. There is a strong anterior cingulid. The valley formed between pro- toconid and metaconid higher than that separating the paraconid from the other cusps. The trigonid measures: a-p =0.91, tr =0.75. Discussion.—The White Springs tooth is similar in morphology to those from Brooksville 2, but is smaller and more transversely compressed. It is also smaller transversely than measurements reported by Korth (1992) for Herpetotherium youngi. In size it is similar to the Rosebud Herpetotherium sp. (Morton & Green 1976). Since White Springs is the youngest Florida site, this could be indicative of the general trend in Herpetotherium of size reduction for younger samples (Fig. 5). CONCLUSIONS Eberle and Storer (1995) demonstrated that a single spe- cies of Herpetotherium, H. valens, from the Chadronian Calf Creek Fauna, exhibits a wide range of upper molar stylar cusp variation as well as a wide range of dental size. Early Arikareean Herpetotherium is also highly variable in dental morphology and size and this variation is continuously distributed in large samples. While size ranges are high in comparison to many mammals they are not outside the ranges of other marsupials such as Didelphis (Hough 1961) and although modern didelphid species appear to have consistent stylar cusp morphol- ogy that is suitable for Linnaean diagnosis, individual variation in the position of stylar cusps on the molars (anteriorly or more posteriorly located) is known in marmosine opossums (Case 1991). Recognition that Herpetotherium might have a considerable amount of dental variation is not new. In her much criticized review (Green & Morton 1976; Korth 1994) of Oligocene and Miocene didelphids, Hough (1961) discussed the large degree of size variation in modern didelphids due to sexual dimorphism and indi- vidual variation. She suggested this as the probable rea- Table 4. Dental measurements of Herpetotherium sp. from the “brown siltstone”, Gering Formation, and undifferentiated lower Arikaree Group (=”Monroe Creek Fm”), Wildcat Ridge, Nebraska. UNSM14947 UNSM11685 UNSM14951 m1 a-p 1.80 1.70 1.70 tr 0.94 0.98 1.04 m2 a-p 1.96 1.86 2.00 tr 1.10 1.15 1.18 m3 a-p 2.00 1.76 1.98 tr 1.10 1.14 2.11 UNSM11652 UNSM11584 UNSM11576 UNSM81321 M1 a-p 1.86 - 1.86 - tr 1.56 - 1.56 - M2 a-p 1.94 - - - tr 1.96 - 1.80 - M3 a-p 1.68 1.82 - 1.82 tr 2.10 2.12 - 2.12 M4 a-p 1.12 0.98 - - tr 1.06 2.16 - - 351 son for the taxonomic confusion of “Peratherium.” This agreed with Scott (1941) who had earlier suggested that all large “Peratherium” from the Oligocene were as- signable to “Peratherium” fugax and the large degree of size separation was due to sexual and individual varia- tion. However, later in her paper, Hough dismissed this proposal by Scott as an “impossibility” (Hough 1961: 220) simply based on the large numbers of specimens. In their review of Great Plains didelphids, Green and Morton (1976) again noted the high degree of variation found in South Dakota samples and proposed that all Herpetotherium after the Chadronian might be assign- able to Herpetotherium fugax. In Korth’s (1994) review of Herpetotherium, his taxonomic classification of species was based in part on size. Comparison of size relationships in Herpetotherium (Fig. 5) shows that while a case can be made for a morphocline in average size reduction, there is consid- erable overlap in observed ranges. Arikareean Herpetotherium, with the exception of H. merriami and H. cf. merriami from I-75, is similar in size to the mar- supials from the Whitneyan Cedar Ridge LF (Setoguchi 1978), which Korth referred to H. fugax as well as other reported Orellan H. fugax. While the degree of size separation between H. valens and H. fugax may be suitable for species distinction, the separation in dental size for most Arikareean Herpetotherium and H. fugax is not as pronounced. The early Arikareean samples show that they may be a distinct group from the older samples. The standard errors are distinct for the early Arikareean samples in comparison to the Orellan and Whitneyan populations. Later Arikareean samples re- ferred to H. youngi even show a slight increase in av- erage size in comparison to early Arikareean samples (Fig. 5). Korth (1994) also followed the trend in moving away from stratospecies, or species defined on the ba- sis of superposed stratigraphic occurrence, by organiz- ing Herpetotherium species on the basis of stylar cusp morphology, but at the time there were few reported large samples. The material from Calf Creek (Eberle & Storer 1996) and the material described here show that coexistent Herpetotherium populations have highly vari- able stylar cusp morphology. In early Arikareean samples, such as the largest sample described here (Dw-121, Nebraska), the upper molars show variation in central stylar cusp configuration that range from having a single central stylar cusp to teeth possessing two clearly dis- tinct central stylar cusps (Fig. 4). In part, this relation- ship appears to be a function of size of the molar, with the smaller teeth possessing the more fused central cusps. Large samples also allow consideration of worn and unworn molar states which can give a more accu- rate picture of stylar cusp size relationships. In the Brooksville 2 and DW-121 marsupials, the unworn sty- lar cusp B is often as large as the central stylar cusp or cusps (Fig. 5). Korth (1994) and Eberle and Storer (1995) both used the relative size of B in relation to the central cusps as a diagnostic feature. Comparison of the Arikareean unworn teeth to unworn teeth of older spe- cies is needed to determine the applicability of this char- acter. Because of the small sample size, it remains to be seen if Herpetotherium youngi is a valid species diag- nosed on the basis of possessing a single fused central cusp and reduced stylar cusp B. At present, there is no evidence that suggests that H. youngi has variable sty- lar cusp morphology. Larger samples of middle to late Arikareean Herpetotherium are needed to conclusively determine this. H. youngi is therefore tentatively re- tained as a species with a single pointed stylar cusp and a reduced stylar cusp B in comparison to older Herpetotherium species. Several taxonomic alternatives present themselves in describing the Nebraska and Florida Arikareean Herpetotherium samples. Resurrect Herpetotherium spindleri as an early Arikareean species that differs only in slight average size, but has considerable size range overlap (Fig. 5), from Herpetotherium fugax (as sug- gested by Setoguchi 1978). Herpetotherium spindleri would not be separable on the basis of size from later Arikareean H. youngi but does have a variable stylar cusp condition in contrast to a single central stylar cusp and a large cusp B. The alternative (as suggested in part by Green and Morton, 1976, discussed above) is to refer Herpetotherium of similar size and variable stylar cusp morphology after the Chadronian to H. fugax, rec- ognizing that younger well-sampled Arikareean popula- tions have a smaller average size than Orellan and Whitneyan H. fugax. This solution to the problem, fa- vored here due to the demonstrated high degree of vari- ability between and within Herpetotherium populations, retains H. youngi as a later Arikareean species without variable cusps. H. merriami is enigmatic, it may be a larger northwestern Pacific Arikareean species, or there is the possibility that H. merriami, from the lower sec- tion of Logan Butte (Stock and Furlong, 1922), is late Whitneyan in age (Hunt, pers. comm. 2004) and agrees in size and morphology to the relatively large Whitneyan Herpetotherium species in Florida from I-75. This would essentially solve the problem of H. merriami in that it HAYES: Arikareean Herpetotherium From Nebraska and Florida 352 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb would fall close to, and within (for some measurements), the size range of H. fugax in the Whitneyan and Orellan and it could be referred to this species. ACKNOWLEDGEMENTS For their thoughts and discussion of this project, my ap- preciation goes to William Korth, David Webb, Gary Morgan, Michael Voorhies, Bruce Bailey, and Robert Hunt. Access to collections at the Florida Museum of Natural History was granted by D. Webb, R. Hulbert, and B. MacFadden. Access to the Nebraska State Mu- seum collection was granted by B. Bailey, M. Voorhies, and R. Hunt. The following deserve thanks for recov- ery and screen washing of specimens: S. Emslie, R. McCarty, H. Mead, G. Morgan, R. Portell, A. Poyer, A. Pratt, E. Simons, E. Taylor, and S. Tucker. My appre- ciation goes to the Nebraska Highway Department for funding collection and curation of the Dw-121 locality. This research was supported in part by funding from the Nebraska State Museum of Natural History, Vertebrate Paleontology Department. LITERATURE CITED Albright, L. B. 1998. The Arikareean Land Mammal Age in Texas and Florida: southern extension of Great Plains faunas and Gulf Coastal Plain endemism. Pp. 160-183 in D. O. Terry, H. E. LaGarry, & R. M. Hunt, eds. Deposi- tional Environments, Lithostratigraphy, and Biostratig- raphy of the White River and Arikaree Groups (Late Eocene to Early Miocene, North America). Geological Society of America Special Paper 325. Bailey, B. E. 1992. A new Early Arikareean microfauna from northwestern Nebraska. Nebraska Academy of Science, 102nd Annual Meeting, 1992:65. Bailey, B. E. 1999. New Arikareean/Hemingfordian micromammal faunas from western Nebraska and their biostratigraphic significance. Journal of Vertebrate Pale- ontology, 19:30A-31A. Bailey, B. E. 2004. Biostratigraphy and biochronolgy of early Arikareean through late Hemingfordian small mammal faunas from the Nebraska panhandle and adjacent ar- eas. Paludicola, 4:81-113. Case, J. A. 1991. Constancy in stylar cusp patterns in extant marsupials: implications for late Cretaceous marsupial systematics. Journal of Vertebrate Paleontology, 11(3):20A-21A. Cope, E. D. 1873. Third notice of extinct Vertebrata from the Tertiary of the Great Plains. Palaeontology Bulletin, 16:1- 8. Crochet, J.-Y. 1977. Les Didelphidae (Marsupicarnivora, Marsupialia) holarctiques tertiares. Comptes-rendus Academie Science, Paris, 24(Series D):357-360. Eberle, J., & J. E. Storer. 1995. Herpetotherium valens (Lambe), a didelphid marsupial from the Calf Creek local fauna. Journal of Vertebrate Paleontology, 15(4):785-794. Emry, R. J., Bjork, R.P., & L. S. Russell. 1987. The Chadronian, Orellan, and Whitneyan North American land mammal ages. Pp. 118-152 in M. O. Woodburne, ed. Cenozoic Mammals of North America: Geochronolgy and Bios- tratigraphy. University of California Press, Berkeley. Frailey, D. 1978. An early Miocene (Arikareean) fauna from north central Florida (the SB-1A Local Fauna). Occa- sional Papers of the Museum of Natural History, Univer- sity Kansas, 75:1-20. Frailey, D. 1979. The large mammals of the Buda Local Fauna (Arikareean: Alachua County, Florida). Bulletin of the Florida State Museum, 24(2):123-173. Green, M., & J. E. Martin. 1976. Peratherium (Marsupialia: Didelphidae) from the Oligocene and Miocene of South Dakota. Pp. 155-168 in C. S. Churcher, ed. Athlon: Es- says on Paleontology in Honour of Loris Shano Russell. Royal Ontario Museum, Ontario. Hayes, F. G. 2000. The Brooksville 2 Local Fauna (Arikareean, latest Oligocene): Hernando County, Florida. Bulletin of the Florida Museum of Natural History, 43(1):1-47. Hayes, F. G. 2004. Paleomagnetics and biostratigraphy of the Pine Ridge Arikaree Group (late Oligocene-early Mi- ocene), Nebraska. Dissertation, University of Nebraska- Lincoln, 211 p. Hough, J. R. 1961. Review of the didelphid marsupials. Journal of Paleontology, 35(1):218-228. Jones, D. S., & R. Portell. 1988. Fossil invertebrates from the Brooks Sink, Bradford County, Florida. Pp. 41-52 in F. L. Pirkle and J. G. Reynolds, eds. Southeastern Geological Society, 1988 Annual Field Trip Guidebook 29. Jones, D. S., P. A. Mueller, D. A. Hodell, & L. A. Stanley. 1993. 87Sr/86Sr geochronology of Oligocene and Miocene ma- rine strata in Florida. Pp. 15-26 in V. A. Zullo, W. B. Har- ris, T. M. Scott, & R. W. Portell, eds. The Neogene of Florida and Adjacent Regions. Proceedings, 3rd Bald Head Island Conference on Coastal Plains Geology: Florida Geological Survey Special Publication 37. Korth, W. W. 1992. Fossil small mammals from the Harrison Formation (late Arikareean: earliest Miocene), Cherry County, Nebraska. Annals of the Carnegie Museum, 61:69-131. Korth, W. W. 1994. Middle Tertiary marsupials (Mammalia) from North America. Journal of Paleontology, 68(2):376- 397 Macdonald, J. R. 1963. The Miocene faunas from the Wounded Knee area of western South Dakota. Bulletin of the Ameri- can Museum of Natural History, 125(3):139-238. Macdonald, J. R. 1970. Review of the Miocene Wounded Knee faunas of southwestern South Dakota. Bulletin of the Los Angeles County Museum of Natural History, 8:1-82. Macdonald, L. J. 1972. Monroe Creek (early Miocene) micro- fossils from the Wounded Knee area, South Dakota. 353 South Dakota Geological Survey Report of Investiga- tions 105:1-43. Marshall, L. G., J. A. Case, & M. O. Woodburne. 1990. Phylo- genetic relationships of the families of marsupials. Pp. 433-505 in, H. H. Genoways, ed. Current Mammalolgy, Volume 2. Plenum Press, New York. Martin, J. E., & M. Green. 1984. Insectivora, Sciuridae, and Cricetidae from the early Miocene Rosebud Formation in South Dakota. Carnegie Museum of Natural History Special Publication 9:28-40. McGrew, P. O. 1937. New marsupials from the Tertiary of Ne- braska. Journal of Geology, 45(4):448-455. Morgan, G. S. 1989. Miocene vertebrate faunas from the Suwannee River Basin of north Florida and south Geor- gia. Pp. 26-53 in G. S. Morgan, ed. Miocene Paleontol- ogy and Stratigraphy of the Suwannee River Basin of North Florida and South Georgia. Guidebook 30, South- eastern Geological Society, Tallahassee. Morgan, G. S. 1993. Mammalian biochronology and marine- nonmarine correlations in the Neogene of Florida. Pp. 55-66 in V. A. Zullo, W. B Harris, T. M. Scott, & R. W. Portell, eds. The Neogene of Florida and Adjacent Re- gions. Proceedings, 3rd Bald Head Island Conference on Coastal Plains Geology, Florida Geological Survey Spe- cial Publication 37. Morgan, G. S., & A. E. Pratt. 1988. An early Miocene (late Hemingfordian) vertebrate fauna from Brooks Sink, Bradford County, Florida. Pp. 53-69 in F. L. Pirkle and J. G. Reynolds, eds. Southeastern Geological Society, 1988 Annual Field Trip Guidebook 29. Obradovich, J. D., G. A. Izett, & C. W Naeser. 1973. Radiomet- ric ages of volcanic ash and pumice beds in the Gering sandstone (earliest Miocene) of the Arikaree Group, southwestern Nebraska. Geological Society of America, Abstracts with Programs 5(6):499-500. Patton, T. H. 1969. An Oligocene land vertebrate fauna from Florida. Journal of Paleontology, 43(2):543-546. Rothecker, J., & J. E. Storer. 1996. The marsupials of the Lac Pelletier lower fauna, middle Eocene (Duchesnean) of Saskatchewan. Journal of Vertebrate Paleontology, 16(4):770-774. Scott, W. B., & G. L. Jepsen. 1941. The mammalian fauna of the White River Oligocene. Part V. Transactions of the Ameri- can Philosophical Society, 28: 747-980. Setoguchi, T. 1978. Paleontology and geology of the Badwater Creek area, central Wyoming. Part 16. The Cedar Ridge local fauna (late Oligocene). Bulletin of the Carnegie Museum of Natural History, 9:1-61. Simpson, G. G. 1928. American Eocene didelphids. American Museum Novitates, 307:1-7. Slaughter, B. H. 1978. Occurrences of didelphine marsupials from the Eocene and Miocene of the Texas Gulf Coastal Plain. Journal of Paleontology, 52(3):744-746. Stock, C., & E. L. Furlong. 1922. A marsupial from the John Day Oligocene of Logan Butte, eastern Oregon. Univer- sity of California Publication in Geological Sciences, 3:311- 317. Swinehart, J. B., V. L. Souders, H. M DeGraw, & R. F. Diffendal, Jr. 1985. Cenozoic paleogeography of western Nebraska. Pp. 209-229 in R. M. Flores & S. S. Kaplan, eds. Cenozoic Paleogeography of West-Central United States. Special Publication Rocky Mountain Section-S.E.P.M., Denver, Colorado. Swisher, C. C. 1982. Stratigraphy and biostratigraphy of the eastern portion of the Wildcat Ridge, western Nebraska. Master’s Thesis, University of Nebraska-Lincoln, 172 p. Tedford, R. H., T. Galusha, M. F. Skinner, B. E. Taylor, R. W. Fields, J. R. Macdonald, J. M. Rensberger, S. D. Webb, & D. P. Whistler. 1987. Faunal succession and biochronology of the Arikareean through Hemphillian interval (late Oligocene through earliest Pliocene epochs) in North America. Pp. 153-210 in M. O. Wooburne, ed. Cenozoic Mammals of North America, Geochronolgy and Biostratigraphy. University of California Press, Berke- ley, California. Tedford, R. H., L. B. Albright, III, A. D. Barnosky, I. Ferrusquia- Villafranca, R. M. Hunt, Jr., J. E. Storer, C. C. Swisher, III, M. R. Voorhies, S. D. Webb, and D. P. Whistler. 2004. Mammalian biochronolgy of the Arikareean through Hemphillian interval (late Oligocene through early Pliocene Epochs); pp 169-231 in M. O. Woodburne, ed. Late Cretaceous and Cenozoic Mammals of North America, Biostratigraphy and Geochronology. Colum- bia University Press, New York, New York. Wolff, R. G. 1987. Late Oligocene-Middle Miocene didelphid marsupials from Florida. Journal of Vertebrate Paleontol- ogy, 7(supplement to 3):14A. HAYES: Arikareean Herpetotherium From Nebraska and Florida 354 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb