early irvingtonian (latest pliocene) rodents from inglis 1c ... - eDocs

Journal of Vertebrate Paleontology 21(1):153–171, March 2001 q 2001 by the Society of Vertebrate Paleontology

EARLY IRVINGTONIAN (LATEST PLIOCENE) RODENTS FROM INGLIS 1C, CITRUS COUNTY, FLORIDA DENNIS R. RUEZ, JR.* Florida Museum of Natural History, University of Florida, Gainesville, Florida 32611 ABSTRACT—A new early Irvingtonian (latest Pliocene) rodent fauna is reported here from Inglis 1C, Citrus County, Florida. This sinkhole deposit near the west coast of the state contains 33 species of mammals, including 11 rodents, of which six are now extinct. A new species of Peromyscus, P. sarmocophinus, sp. nov., is described from this locality and the slightly older Inglis 1A. Inglis 1C contains the oldest records of three rodents (Peromyscus polionotus, Reithrodontomys humulis, and Atopomys texensis) and the youngest occurrences of two others (Reithrodontomys wetmorei and Ondatra idahoensis). Additionally, the paleogeographic range of two taxa, R. wetmorei and Baiomys sp., is extended east of the Mississippi River. Deposition at Inglis 1C was nearly contemporaneous with, the better-known, early Irvingtonian Inglis 1A fauna. The two localities share Sciurus sp., Glaucomys sp., Orthogeomys propinetis, Sigmodon curtisi, P. sarmocophinus, sp. nov., R. wetmorei and O. idahoensis. A slightly younger age at Inglis 1C is implied by the presence of P. polionotus, R. humulis, and Atopomys texensis. This fauna serves as an older complement to Florida’s other well studied Irvingtonian localities, Leisey Shell Pits, Hillsborough County, and Coleman 2A, Sumter County.

INTRODUCTION Inglis 1A (Citrus County, Florida; Fig. 1) was the only significant early Irvingtonian site in the eastern United States until 1982 when the Leisey Shell Pits (Hillsborough County, Florida) were found to be richly fossiliferous (Hulbert and Morgan, 1989; Hulbert et al., 1995). Leisey, at about 1.4 Ma in age, differs in being younger than Inglis 1A (Fig. 2), and in having a relatively much smaller rodent component. Inglis 1A, therefore, retains its prominent position with an abundance of small herpetological, avian, and mammalian fossils, but now shares that position with the fauna from Inglis 1C. This study introduces the rodent fauna of Inglis 1C (Citrus County, Florida; Florida Museum of Natural History locality CI019), which helps to fill the faunal gap between Inglis 1A and the Leisey Shell Pits (Figs. 1, 2; Table 1). The recovery of many thousands of vertebrate fossils from Leisey 1A provided the basis for the publication of an excellent survey of Florida Blancan and Irvingtonian faunas (Hulbert et al., 1995). The description of mammals from Inglis 1C complements these works by expanding on the sparse rodent component from Leisey. The only other significant Irvingtonian rodent fauna described from Florida is the latest Irvingtonian Coleman 2A (Martin, 1974a), which at ;400 Ka is substantially younger than Inglis 1C. Inglis 1C produced over 5,000 identified mammalian specimens. Of the 33 species of mammals (Table 1), the 11 rodent taxa are described here. The rodent assemblage includes six extinct species, including a new species of Peromyscus, P. sarmocophinus, described from this locality and Inglis 1A. The Inglis 1C fauna also extends the geographic ranges of three taxa and the temporal ranges of five others. Future studies of the other Inglis sinkholes may yield a unique perspective within Florida. The temporal variation between sinkholes may be such that incremental changes in the taxa can be witnessed within a single geographic area. While the physical distance between the sinkholes has no relationship to their age, perhaps some chronologic differences can be distinguished through evolutionary progression of morphological characters. This sequence could * Present adress: Department of Geological Sciences, The University of Texas at Austin, Austin, Texas 78712.

possibly show the gradation often missing from lineages of fossil species pieced together from various localities, as has necessarily been done in Florida due to the near absence of stratified fossil deposits. While cutting a canal across Florida in an attempt to bypass the long shipboard journey around the Keys, the U. S. Army Corps of Engineers in 1942 intersected sand-filled fissures within the middle Eocene Inglis Formation of the Ocala Group. In 1967, Jean Klein and Robert Martin discovered vertebrate fossils in one of the fillings on the north bank, 2.5 km southwest of Inglis, Citrus County, Florida. Four years later Klein (1971) discussed the carnivorans and ungulates from the sinkhole denoted as Inglis 1A. Intense collecting at the site continued until 1974, when the hole was excavated to a logistically difficult depth of five meters below current sea level. Subsequent research on the original locality focused on birds (Steadman, 1980; Carr, 1981; Emslie, 1996) and squamates (Meylan, 1982). The numerous works on various mammalian components are summarized in Morgan and Hulbert (1995). Inglis 1A still lacks a comprehensive faunal study. The microfauna dominates the Inglis 1A assemblage, both in number of identifiable specimens and number of species. Ondatra idahoensis (Repenning, 1987) and Neofiber from this site occur as Florida’s oldest arvicoline rodents. Inglis 1A is the type locality of three mammalian and two squamate species. Sylvilagus webbi occurs just before the evolutionary explosion of modern rabbits (White, 1991). It lies in an ecologically interesting position—ancestral to the marsh and swamp rabbits yet descended from Sylvilagus hibbardi, a form found only in arid localities. The late Pliocene and early Pleistocene pocket gopher Orthogeomys propinetis is best represented at Inglis 1A (Wilkins, 1984). A porcupine known only from Inglis 1A, Erethizon kleini, is unique in resembling the modern Mexican and South American genus Coendou in size (Frazier, 1981). Meylan (1982) described two new species of squamates, Eumeces carri (a skink) and Regina intermedia (a water snake), in the only herpetological study of the site. A more inclusive summary of Inglis 1A is included in Morgan and Hulbert (1995). The discovery of Inglis 1A proved to be of critical importance to Florida and North American paleontology because of its unique biochronologic position. The age of the site shortly following the connection of North and South America via the

153

154

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 21, NO. 1, 2001

FIGURE 2. Time scale of Florida Pliocene and Pleistocene localities. Geomagnetic polarity time scale (GPTS) dates are taken from Wei (1995); subdivisions of North American land mammal ages (NALMA) follows that used by Morgan and Hulbert (1995) for Florida faunas. FIGURE 1. Location of Florida Pliocene and Pleistocene localities mentioned in text. 1, Haile, Alachua County; 2, Inglis, Citrus County; 3, Coleman, Sumter County; 4, Leisey Shell Pit, Hillsborough County; 5, De Soto Shell Pit, De Soto County. Many sites contain multiple localities (1A, 1B, 2, etc.) that are not necessarily similar in age.

Panamanian land bridge, coupled with the simultaneous evolutionary explosion of native taxa, make it critical with regards to the late Pliocene–early Pleistocene faunal history of Florida. While there are numerous references to the fauna, field work at the canal was not resumed until late March 1996. The excavations resulted in collections from five additional sites (Inglis 1C–1G) exposed by the construction of the canal. In addition to abundant avian remains, a large assemblage of mammalian and herpetological specimens was recovered. At present, seven sinkholes from the canal have been excavated: two early discoveries, which are at least partially evaluated, and five recent discoveries currently under study. Although none of the newly found sites have the diversity of fossils recovered from Inglis 1A, Inglis 1C has produced thousands of mammalian specimens. The geology of Inglis 1C was discussed by Emslie (1998), who used the presence of Lepus sp. as partial evidence for an age similar to Inglis 1A. No Lepus remains are identified from any of the new sites. Inglis 1C was likely deposited in a similar fashion to most of Inglis 1A (wind blown sands; Meylan, 1982), but the total time interval represented may be less. While Inglis 1A can be divided into discrete stratigraphic units, the 1C accumulation shows no such differentiation and may represent a shorter interval of deposition. Pale yellow quartz sand dominates the clastic fraction at Inglis 1C, with small amounts of rose quartz (;5 per 1,000 grains) and smoky quartz (;1 per 1,000) occurring. The quartz grains are roughly equidimensional and sub-rounded to subangular. Small foraminifers eroded from the surrounding Inglis Formation account for about two percent by volume of the matrix. Also weathered from the limestone walls are tests of the large foraminifer Lepidocyclina ocalina, gastropod external molds, and Galeocerdo alabamensis (extinct tiger shark) teeth.

METHODS AND MATERIALS All sediments were washed at the site through 0.64 cm and 0.16 cm screens, and samples were seived at 0.8 mm. Comparisons were made with specimens in the Vertebrate Paleontology and Mammalogy collections of the Florida Museum of Natural History (FLMNH) and the Vertebrate Paleontology Laboratory of the Texas Memorial Museum. All Inglis 1C fossils are curated in the Vertebrate Paleontology collection of the FLMNH. Measurements of teeth were made through a Wilde monocular microscope with an optical micrometer. All illustrations were drawn with a camera lucida attachment on a Wilde binocular microscope. In all illustrations the occlusal surfaces are oriented so that anterior is toward the top of the page. The use of North American Land Mammal Ages and their subdivisions follows Morgan and Hulbert (1995). Tests of statistical differences were performed at the 90% confidence level. All measurements are given in mm. Abbreviations—TMM, Vertebrate Paleontology Laboratory, Texas Memorial Museum, The University of Texas at Austin; UF, Vertebrate Paleontology, Florida Museum of Natural History, University of Florida, Gainesville. SYSTEMATIC PALEONTOLOGY RODENTIA Bowdich, 1821 SCIURIDAE Fischer de Waldheim, 1817 SCIURINAE Fischer de Waldheim, 1817 SCIURUS Linnaeus, 1758 SCIURUS sp. Referred Specimens—UF 198381–198382, LI1; UF 196291, 196778, LP4; UF 196231, 198455, LM1; UF 196779, 198451, LM2; UF 198383–198385, RI1; UF 196232, 196292, 197912–197913, 198633, RM1; UF 194918–194919, 196780, 197922, RM2; UF 197914, L mandible with m1–2; UF 186061, 194913, L edentulous mandible; UF 198379–198380, Li1; UF 194917, Lp4; UF 196234–196325, 198452, Lm1; UF 196236– 196237, 197911, 198453, Lm2; UF 196233, 196781–196782, 197910, Lm3; UF 198454, Rp4; UF 196238, 197923–197924, Rm1; UF 196239–196240, 196293, Rm2; UF 197925, Rm3;

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA UF 198265, RdP4; UF 197909, Rdp3; UF 194916, L humerus; UF 194914, L humerus, distal end; UF 194915, R humerus, distal half; UF 198627, R humerus, missing proximal epiphysis; UF 198264, L innominate; UF 197921, L femur, proximal half; UF 194945, 198628, L femur, distal half; UF 198629, L tibia; UF 198630, L tibia, proximal epiphysis; UF 198266–198267, L astragalus; UF 198200, L calcaneum; UF 190792, R calcaneum; UF 198634, thoracic vertebra; UF 197920, UF 198631– 198632, lumbar vertebra. Diagnosis—Cheekteeth brachydont; P4 triangular and M1–2 quadrate in occlusal outline; trigonid ‘‘U’’ shaped on P4–M2; mesostyle present on M1–2; p4 lacks protolophid; anterior cingulum joins metalophid midway between the protoconid and metaconid and extends to the metaconid on m1–2; m3 only slightly larger than to m1 and m2. Incisors and postcrania are assigned to Sciurus sp. due to the indication in the cheekteeth that only a single species of tree squirrel is present in the fauna. Description—Dental terminology follows Black (1963) for both Inglis 1C sciurids. The upper incisors are identical to those of modern S. carolinensis. They are recognized by their large size, very narrow transverse width, and extreme antero-posterior thickness. Although the poorly preserved nature of the P4 precludes a complete description, it can be discerned that the protoloph was poorly developed, and the valley between the protoconule and anterior cingulum is much narrower than in extant Sciurus carolinensis. There does not appear to be a mesostyle, but this may be due to the worn condition of the teeth. The well–developed parastyle, paracone, mesostyle, and metacone form a straight line labially on the M1. The mesostyle is half the height of the other cusps and lies closest to the paracone. In extant S. carolinensis the mesostyle is less developed, slightly elongate antero-posteriorly, and more centrally located between the paracone and metacone. Within the Inglis 1C specimens the anterior and posterior cingula form a pronounced ridge, except for a slight depression just before connection to the protocone. The protoconule and metaconule are more prominent in the fossil specimens than in extant S. carolinensis. There is one M1 (UF 196292) that fits the S. carolinensis description of the mesostyle and cingula and matches the dimensions, but the protoconule and metaconule are present. It is only tentatively referred to the same species as the other Inglis 1C teeth. The M2, like the M1, is quite similar to the same tooth in modern S. carolinensis. The Inglis 1C specimens differ from the extant form in having a prominent protoconule, metaconule, and a generally better-developed parastyle, although development of the parastyle is variable. The protoconule and metaconule disappear with moderate wear. Compared with extant S. carolinensis the inferior dental foramen of the fossil dentary is elongate and situated closer to the condyloid process. Additionally, recent specimens have a deeper pterygoid fossa that often curves under the posterior incisor cavity. The lower incisors are flattened latero-medially. They may be distinguished from the uppers by the greater arc of curvature and a twist that allows the teeth to meet parallel though the dentaries are angled. The metaconid is the most prominent cusp on the p4. Adjacent to that cusp is a mesostylid separated completely from the entoconid. The development and complexity of the posterolophid varies in both fossil and recent Sciurus. The metalophid also varies, from well developed to absent, in the Inglis 1C collection. In the extant gray squirrel the anterior cingulum is much better developed to the extent that it forms a cusp-like feature that may be as large as the protoconid. The Inglis 1C squirrels possess a labial cingulum below the mesoconid. It originates near the center of the hypoconid and extends to the base of the protoconid.

155

The cusps (hypoconid, mesoconid, and protoconid) forming a labial line on the m1 are rounded, equally spaced, and connected by a high, thin ectolophid. A fourth ‘‘cusp’’ is formed from the labial extension of the anterior cingulum, which is rounded and does not connect to the protoconid in unworn specimens. Posterior to the very small mesostylid is a large lingual depresseion into the talonid basin; the notch is considerably larger and deeper than that in recent S. carolinensis. The posterolophid and metalophid are both high and slender. The anterior cingulum joins the metalophid midway between the protoconid and metaconid and extends to the metaconid. The m2 closely resembles the m1, except that the m2 has a larger anterior than posterior half; in the m1 the posterior portion is larger. The anterior cingulum can vary as a slender ridge or a slightly elongate ‘‘cusp.’’ In UF 196240 there is a gap in the anterior cingulum at the midline of the tooth. The m3s are indistinguishable from those in modern gray squirrels. All available postcranial materials are nearly identical to that of S. carolinensis, although the limbs of the fossil species are slightly more gracile. Discussion—The Inglis 1C fossils most closely match extant and fossil Sciurus carolinensis in morphology and size. Fossils of S. carolinensis are recorded from hundreds of fossil localities throughout the eastern United States, closely tracing the current distribution. The gray squirrel represents one of the most commonly reported rodent fossils in the literature in spite of its reduced preservational potential as an arboreal animal. Moderate size, high population densities, and extensive range make them a frequent find, especially in Rancholabrean faunas. The size of the Inglis 1C teeth is equivalent to that found in three species of North American Sciurus, S. carolinensis, S. griseus, and S. aberti. The triangular outline of the P4s, the well developed protoconule of the P4s, and the high metaloph of the M1s and M2 in the Inglis 1C collection indicate a closer affinity to S. carolinensis. Minor differences, however, especially the better-developed cusps and cingula, preclude definite referral of the Inglis 1C species to S. carolinensis, although the two taxa are obviously closely related if not conspecific. PTEROMYINAE Simpson, 1945 GLAUCOMYS Thomas, 1908 GLAUCOMYS sp. Referred Specimens—UF 196247, L mandible, with i1; UF 196248, R edentulous mandible; UF 196246, Lm1. Diagnosis—Small sciurid; gracile mandible with a relatively smooth masseteric scar; tear-drop to oval shaped inferior dental foramen; trigonid slightly higher than talonid on m1. Description—The most similar mandibles to Glaucomys (Fig. 3) are those of Sciurus and Tamias; the latter two being much larger than jaws of the flying squirrel. Tamias also has a longer diastema, a lateral protrusion on the lower extent of the masseteric scar, and the shallower pterygoid fossa. Sciurus has the largest mandible of the three sciurids discussed here and may further be distinguished by its robustness, relative depth of the dentary under the alveoli, and circular inferior dental foramen. The lower incisors are very similar to those described above for Sciurus, but they differ in their smaller size and in having a lateral bend, rather than a twist, to move the incisors parallel in front of the mandibular symphysis. The m1 is elongated antero-posteriorly, with a narrower anterior end. It resembles Glaucomys volans and G. sabrinus in having a deep talonid basin that is surrounded by high ridges. The anterior cingulum is lacking, completely opening the trigonid basin. The mesostylid is absent, leaving a larger lingual gap. Small breaks occur in the posterolophid just before attachment to the entoconid and hypoconid. Finally, the ectolophid is

156


TABLE 1. Non-marine mammals of Inglis 1C, Citrus County, Florida and other Florida early Irvingtonian faunas. The fauna lists are derived from Morgan (1991), Morgan and Hulbert (1995), and Morgan and White (1995) and from personal observation. Inglis 1A Xenarthra Dasypodidae Dasypus bellus Pachyarmatherium leiseyi

Inglis 1C De Soto Leisey

1 2

1 2

1 2

1 1

Pampatheriidae Holmesina floridanus

1

1

1

1

Glyptodontidae Glyptotherium arizonae

1

2

1

1

Mylodontidae Paramylodon harlani

1

1

2

1

Megalonychidae Megalonyx leptostomus Megalonyx wheatleyi

1 2

2 2

1 2

2 1

Megatheriidae Eremotherium eomigrans Nothrotheriops texanus

1 2

2 2

1 2

1 1

Insectivora Soricidae Cryptotis parva1 Blarina carolinensis1

1 1

1 1

2 1

2 1

Talpidae Scalopus aquaticus

1

1

2

2

1

2

2

2

1 1 1 1 1 1 2 2

2 2 2 2 2 2 1 1

2 2 2 2 2 2 2 2

2 2 2 2 2 2 2 2

Chiroptera Phyllostomidae Desmodus archaeodaptes Vespertilionidae Pipisterllus subflavus Antrozous sp. Eptesicus sp. Lasiurus sp. Plecotus sp. Myotis sp. Myotis cf. M. austroriparius Indet. large Vespertilionidae Carnivora Canidae Urucyon sp.2 Canis armbrusteri Canis edwardii

TABLE 1. (Continued) Inglis 1A


Lagomorpha Leporidae Sylvilagus webbi Sylvilagus floridanus Lepus sp.

1 1 1

1 2 2

1 2 2

2 1 16

Rodentia Sciuridae Sciurus sp. Glaucomys sp.

14 1

1 1

14 2

2 2

2

2

2

1

2 1

2 1

2 1

1 2

2 1

2 2

2 2

1 2

1 2

2 2

2 2

2 1

1 1 2 2 1 ? 1 2 1 2 1 2 2 1 2 2 2

2 1 2 2 1 1 2 2 1 1 2 1 1 1 2 2 2

2 1 2 1 ?5 2 2 2 2 2 1 2 1 1 2 2 2

2 2 1 2 2 2 2 1 2 2 2 2 2 2 1 1 1

2 1 2

1 2 2

2 2 2

2 1 1

1 1 2

1 2 1

1 1 2

1 2 1

Castoridae Castoroides leiseyorum Geomyidae Geomys pinetis Orthogeomys propinetis Erethizontidae Erethizon dorsatum Erethizon kleini Hydrochaeridae Hydrochaeris holmesi Neochoerus sp. Muridae Neotoma sp. Sigmodon curtisi Sigmodon libitinus Sigmodon minor Peromyscus sarmocophinus Peromyscus polionotus Peromyscus small spp. Podomys, sp. nov. Reithrodontomys wetmorei Reithrodontomys humulis Reithrodontomys sp. Baiomys sp. Atopomys texensis Ondatra idahoensis Ondatra annectens Pedomys, sp. nov. Synaptomys sp. Artiodactyla Tayassuidae Platygonus bicalcaratus Platygonus vetus Mylohyus fossilis

1 2 1

1 2 2

2 2 1

1 1 1

Ursidae Arcotodus pristinus

1

2

2

1

Camelidae Hemiauchenia macrocephala Hemiauchenia, sp. nov. Palaeolama mirifica

Procyonidae Procyon, sp. nov.

1

2

1

1

Antilocapridae Capromeryx arizonensis

1

2

1

2

Cervidae Odocoileus virginianus

1

1

1

1

Bovidae Indeterminate Bovidae

1

2

2

2

2 1

1 2

2 1

1 2

1

1

1

1

2

2

1

1

Mustelidae Mustela frenata Trigonictis macrodon Spilogale putorius Satherium piscinarium Lutra canadensis

2 1 13 2 2

2 2 1 2 1

2 1 2 1 2

1 1 1 2 1

Felidae Smilodon gracilis Homotherium sp. Lynx rufus Lynx sp. Miracinonyx inexpectatus Felis sp.

1 1 2 1 1 1

1 1 2 2 2 2

2 2 2 2 2 2

1 1 1 2 1 2

Hyaenidae Chasmaporthetes ossifragus

1

2

1

2

Perissodactyla Tapiridae Tapirus haysii Tapirus, sp. nov. Equidae Equus sp. Proboscidea Gomphotheriidae Cuvieronius tropicus

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA

157

TABLE 1. (Continued) Inglis 1A


Mammutidae Mammut americanum

1

2

1

1

Elephantidae Mammuthus hayi

2

2

2

1

Indeterminate Proboscidea

2

1

2

2

1

Cryptotis cf. C. parva and Blarina cf. B. carolinensis of Morgan and White (1995); 2includes Urocyon n. sp. and U. cinereoargenteus; 3listed as Spilogale sp. by Morgan and Hulbert (1995); 4referred to Sciurus carolinensis by Morgan and White (1995); 5Peromyscus large sp. of Morgan and White (1995) may represent additional specimens of Peromyscus sarmocophinus, but the material is inadequate to be certain; 6Lepus cf. L. townsendii of Morgan and White (1995).

very poorly developed, leaving the mesoconid alone, nestled between the hypoconid and protoconid. Length 5 1.70 mm; width 5 1.43 mm. Discussion—The p4–m3 alveolar lengths for the mandibles are 7.10 and 7.24 mm. These values are near the maximum reported for Glaucomys volans, but near the minimum for Glaucomys sabrinus (Martin, 1974a). The paucity of material precludes specific designation at this time. The Glaucomys of Inglis 1C may represent a new species and might also be conspecific with the Inglis 1A species. GEOMYIDAE Bonaparte, 1845 GEOMYINAE Bonaparte, 1845 ORTHOGEOMYS Merriam, 1895 ORTHOGEOMYS PROPINETIS (Wilkins, 1984), comb. nov. Geomys pinetis Wilkins, 1984 Referred Specimens—UF 197892–197893, fused L and R frontals; UF 198417, fused L and R edentulous premaxillae; UF 191858, 197894–197895, fused L and R edentulous maxillae; UF 197898, L edentulous maxilla; UF 196768, 197899, 198341–198348, LI1; UF 198424–198425, LM1or2; UF 198352, LM3; UF 196769, 197901, 198353–198356, RI1; UF 198262–198263, 198357, 198419, RP4; UF 198434–198437, RM1or2; UF 196770, 198350–198351, RM3; UF 197902– 197903, 197926–197927, 198261, M1or2; UF 196774, L mandible with i1, p4; UF 194924, L mandible with i1; UF 190809, L mandible with p4; UF 186062–186063, 191859, 197900, 197938–197940, 198349, L edentulous mandible; UF 198358, Li1; UF 196772, 197931–197934, Lp4; UF 198426, Lm1; UF 197908, 198427–198429, Lm2; UF 196771, 197905–197907, 198430, Lm3; UF 194925, R mandible with i1, p4; UF 197935, R mandible with i1; UF 197896–197897, 197936–197937, R edentulous mandible; UF 196773, 197904, 197929–197930, 198420, Rp4; UF 198431–198432, Rm1; UF 198421–198423, Rm2; UF 198418, Rm3; UF 198657, 3 cheekteeth in pellet; UF 191860, axis; UF 198656, R humerus, missing proximal epiphysis; UF 198655, R ulna; UF 198438, L innominate. Chronologic and Geographic Range—Late Blancan (late Pliocene) to late early Irvingtonian (early Pleistocene) of Florida. Diagnosis—Moderate sized geomyid; dolichocephalic skull; shallow retromolar fossa in mandible; unisulcate upper incisors; enamel present on posterior surface of the P4 metaloph; M3 noticeably bicolumnar and longer than wide; m1–2 lacking enamel on the anterior surface. Description—Dolichocephalic specialization is apparent on the frontals, which lack visible temporal ridges. The Inglis 1C premaxillae are fused in adults. Geomys pinetis tends to show a slight ridge at the connection, although this feature varies in

FIGURE 3. Glaucomys sp., Inglis 1C, left mandible with i1, labial view, UF 196247. Scale bar equals 4 mm.

modern specimens. Of the 16 upper incisors recovered, all have a deep medial groove and three have a faint lingual groove (Fig. 4A). The P4 consists of two transversely elongate ovals, the protoloph and metaloph united centrally by a short stem (Fig. 4B). In O. propinetis the metaloph is wider; in modern Geomys the protoloph is wider. Dentine tracts extend the full length of the labial and lingual surfaces of both the protoloph and metaloph. Enamel partially covers the posterior edge of all teeth. In the Inglis 1A Orthogeomys the posterior enamel is restricted to the lingual half to two-thirds of the metaloph, except for one specimen that has only a trace of enamel on the labial side. At Inglis 1C three P4s have enamel that extends over one-half of the lingual surfaces and one-fourth of the labial. The other P4 (UF 198262) has only a thin (one-fourth transverse length) sliver of enamel on the lingual side of the occlusal surface that would expand with increasing ontogenetic age. UF 198419 is an immature specimen, barely showing wear. A small, secondarily developed enamel ‘‘column’’ extends anterolabially from the anterior surface of this tooth, 0.5 mm below the occlusal surface. Its tip is just above the beginning of the lateral dentine tract of the metaloph and well above the dentine tract of the protoloph. With wear this third ‘‘column’’ would coalesce with the protoloph. I am unaware of this feature on any other pocket gopher tooth, fossil or recent. The M1 and M2 are difficult to differentiate, although there appears to be a tendency to slightly increase the overall tooth curvature posteriorly in the toothrow. Both teeth are columnar with enamel on the anterior and posterior surfaces, but not on the lateral edges. The primitive bicolumnar pattern can be seen in the M3 (Fig. 4D). Enamel covers the anterior surface as in the other cheekteeth, with dentine tracts labially and lingually.

FIGURE 4. Inglis 1C Orthogeomys propinetis. Occlusal views. A, RI1, UF 196769; B, Rp4, UF 194925; C, RP4, UF 198263; D, RM3, UF 198351. Scale bar equals 1 mm.

158


A prominant reentrant on the labial side separates the protoloph from metaloph. The lingual reentrant is less well-developed. The mandible is robust and similar to that of modern geomyines except for smaller size. The anterior extent of the masseteric scar lies just ventral to the reentrant of the p4. The position of the mental foramen varies from a point halfway between the masseteric scar and the i1 alveolus, to one-fourth the distance from the scar, as in Geomys pinetis. The ventralmost portion of the scar bulges laterally. The depth of the retromolar fossa is greater in Geomys than in O. propinetis. The elongate dental foramen is halfway between the m3 alveolus and the condyloid process. The large lower incisors have a wide, flattened outer surface. The lower premolar, like the upper, is distinctly bicolumnar, with a triangular metalophid and an oval hypolophid (Fig. 4B). The lophs are connected just labial to the midline. Both the metalophid and hypolophid have labial and lingual dentine tracts. Reentrants are V-shaped. A fourth enamel plate makes up the posterior edge of the tooth. The m1, m2, and m3 resemble the M1 and M2, but only the posterior surface has enamel. The plate is centered or slightly extended lingually. The anteroposterior curve of the teeth increases posteriorly in the toothrow. The postcrania are similar to modern geomyines and are assigned to O. propinetis because it is the only geomyid taxon indicated by the dentition. Discussion—Referral of the Inglis 1C specimens to Orthogeomys propinetis represents a substantial temporal and geographic range extension for the genus. Such affinities were suggested in the earliest reference to the species (Martin, 1974b). Referred to Geomys in the type description (Wilkins, 1984), it is the morphology of the upper incisor that first indicated referral to Orthogeomys may be more accurate. Russell (1968) characterized geomyines according to the number of grooves in the upper incisors. Geomys, Zygogeomys, and Pliogeomys were characterized on the basis of both a lingual and medial groove; Pappogeomys and Orthogeomys have only a medial groove; and Pliosaccomys and Thomomys have no grooves. Upper incisors are not known for Dikkomys or Plesiothomomys. An earlier work (Merriam, 1895) described all geomyines as having a lingual groove, but placed no taxonomic weight on the character. Additionally, a later revision of Thomomys listed the lingual groove as present, but occasionally indistinct (Bailey, 1915). In a work focusing solely on the morphology of upper incisor grooves, Akersten (1973) reported at least a faint lingual groove in 99% of Thomomys, 75% of Pappogeomys, and 34% of Orthogeomys. Further, he warned against using groove patterns to interpret phylogenetic relationships, though he did suggest an evolutionary trend: primitively smooth, formation of lingual groove, addition of medial groove, subsequent reduction of lingual groove. This sequence matches well with the proposed phylogeny of Russell (1968). To this I would add that the formation of the lingual groove likely occurred before the Entoptychinae–Geomyinae split in the Oligocene. Upper incisors from entoptychines in the Arikareean Brooksville 1B Local Fauna (Hernando County, Florida) show a faint lingual groove that is even more expressed than that of fossil and modern Thomomys (pers. obs.). The development of the lingual groove in fossil and recent Geomys pinetis is always obvious macroscopically. This species even has a rare third groove, as previously recorded by Akersten (1973) and Russell (1968). Inglis 1A O. propinetis has the lingual groove in seven of the 61 upper incisors (11%), with the majority of the teeth showing no development of that groove, and even lacking the ridge shown in Thomomys. In the development of the lingual groove, this species is nearest to Orthogeomys. The large medial groove of that genus, however,

tends to sit slightly lingual to the midline of the tooth, rather than labially, as in Geomys. In the list of diagnostic dental traits presented by Russell (1968), the propinetis samples match every character of Orthogeomys and is clearly distinct from Geomys. Orthogeomys always has an enamel plate on the posterior wall of P4, more developed on the lingual side; the M3 is noticeably bicolumnar, and longer than wide due to elongation of the metaloph; and the upper incisors are unisulcate. Geomys differs by lacking enamel on the posterior wall of P4, having an unicolumnar M3 no longer than wide, and having distinctly bisulcate upper incisors. Orthogeomys propinetis additionally fits the description of subgenus Heterogeomys as defined by Russell (1968). It is to this subgenus that the only fossils of Orthogeomys have been referred (Russell, 1968). Wilkins (1984:176) briefly discussed the relationships of O. propinetis: ‘‘The apparent position of G. propinetis in the Geomys clade is intermediate between the less advanced G. adamsi . . . and the more advanced G. quinni . . . and modern G. pinetis. G. propinetis most closely resembles G. jacobi . . . and G. tobinensis . . . These relationships are indicated by the degree of development of the retromolar fossa, deployment of enamel on cheekteeth (especially on the posterior face of P4), and general size.’’ Varying degrees of development of the retromolar fossa can be found among independent geomyid lineages. This character may be a useful aid in differentiating species but of little validity in determining higher level relationships. Geomys adamsi (Hibbard, 1967) and Geomys quinni (McGrew, 1944) have no enamel on the posterior loph of the P4. Geomys tobinensis has a trace of enamel on the posterior loph of 5% of the P4s, but this is coupled with enamel on the anterior face of the lower molars (Paulson, 1961). The specimens may represent a separate lineage. Upper teeth are not known for Geomys jacobi (Hibbard, 1967). The five million year lineage in Geomys leading to the modern G. pinetis has consistently lacked enamel on the posterior wall of the P4. The evolutionary trends previously noted between O. propinetis and G. pinetis (Wilkins, 1984; Morgan and White, 1995) are based on the occurrence of the former species at Inglis 1A and Haile 16A. The geomyid from the Leisey Shell Pits, is indistinguishible from the modern G. pinetis, even though it likely closely follows the deposition of Haile 16A. This tightly constrains the transition, and suggests displacement by G. pinetis rather than rapid morphological evolution. MURIDAE Illiger, 1811 SIGMODONTINAE Wagner, 1843 SIGMODON Say and Ord, 1825 SIGMODON CURTISI Gidley, 1922 Referred Specimens—UF 198621, L premaxilla with I1; UF 198622, 198626, R premaxilla with I1; UF 195772, 195139, 196697, 197867–197869, 198456, L maxilla with M1–3; UF 195776, 196249, L maxilla with M1–2; UF 196250, 196698, 197870, L maxilla with M1; UF 196699, L maxilla with M2; UF 191887, 195140, L edentulous maxilla; UF 195163– 195165, 195779–195780, 196251–196252, 196700, 196776, 198457, LM1; UF 195781–195782, 196153–196254, 196701– 196702, 197871, LM2; UF 195783–195785, 196703–196704, 197872, 195812, LM3; UF 195142, 197943, R maxilla with M1–3; UF 195141, 195773–195775, 196255, R maxilla with M1–2; UF 195777, 196256–196257, R maxilla with M1; UF 195778, 196695–196696, 198458, R edentulous maxilla; UF 195162, 195786–195790, 196258, 196705–196707, 197873, RM1; UF 195791–195793, 196708–196709, 197874–197875, 198459–198460, RM2; UF 195166, 195794–195797, 196259, 196710–196711, 198461, RM3; UF 186071, 195133, 195134, L mandible with i1–m3; UF 196260–196261, 197876, L man-

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA dible with i1–m2; UF 198462, L mandible with i1; UF 196262– 196265, L mandible with m1–3; UF 186072, 196266, L mandible with m1–2; UF 196267–196268, L mandible with m1; UF 196269, 196712–196713, 198463, Lm1; UF 195167, 195801– 195804, 196714, 197877, 198331, 198464–198465, Lm2; UF 195798, 196715, 198466, Lm3; UF 191886, 195799, 196270– 196271, R mandible with i1–m3; UF 195811, 196276, R mandible i1-m2; UF 195135, R mandible with i1, m2–3; UF 186073–186074, 195138, 196277, R mandible with m1–3; UF 195137, 195800, 196272, 197878, 198467, R mandible with m1–2; UF 196716, 197880, R mandible with m1; UF 195136, 197879, R mandible with m2; UF 195805–195806, 195810, 196273–196275, 197881, Rm1; UF 195168, 195807–195808, 197882–197885, 198468, Rm2; UF 195809, 196717, 196777, 197886, 198469, Rm3; UF 195177, L humerus, missing proximal epiphysis; UF 198623, sacrum, S1–S3; UF 198624, sacrum, S1; UF 195180, L femur; UF 195178, R tibia, proximal half, missing epiphysis; UF 195179, R tibia, missing proximal epiphysis. Chronologic and Geographic Range—Late Blancan (late Pliocene) of Arizona, early Irvingtonian, (late Pliocene) of Colorado and Florida, and late early Irvingtonian (early Pleistocene) of Kansas and Texas. Diagnosis—Assigned to Sigmodon based on the ‘‘S’’ shape occlusal pattern on m3 and presence of 3 roots on m1; smaller than Sigmodon hudspethensis, S. lindsayi, and some modern forms; more hypsodont than S. medius, similar in crown height to S. hudspethensis, and more brachydont than all other species of Sigmodon; anteroconid generally symmetrical; anterior cingula on m2 and m3 well-developed; dental pattern similar to S. leucotis, except that S. curtisi has a deep first lingual reentrant angle (LRA1) on the m2 and a slight LRA1 on the m3 whereas those folds are absent in S. leucotis; reentrants not nearly as wide as in S. hudspethensis. Description—Dental terminology follows Martin and Prince (1989). The upper dentition resembles that of the extant Sigmodon hispidus, but is more brachydont. Within Sigmodon morphological variation of the upper molars has not been determined to be of taxonomic utility (Martin, 1979). The assignment of these teeth to S. curtisi is based on the presence of only a single species of cotton rat as indicated by the lower molars. The anteroconid of the m1 is symmetrical, with the anteriormost tip flattened in some specimens (Fig. 5A). Lingual reentrants are narrower anteroposteriorly than labial reentrants. The second lingual reentrant angle (LRA2) is deflected anteriorly rather than protruding into the protoconid. The dentine isthmuses that join the protoconid both anteriorly and posteriorly are pulled far off the midline. A three-rooted form predominates, but two and four roots also occur. Only the labial portion of the anterior edge of the m2 touches the m1 (Fig. 5A). The metaconid swings widely lingually and turns posteriorly. The very thick enclosing enamel reaches its extreme at the postero-lingual termination. Labially, the BRA2 is separated anteriorly by the anterior cingulum that only appears with wear. With an increase in ontogenetic age, the BRA2 is separated into a discrete pit. This enclosed pit can only be seen in individuals of certain ontogenetic age. The dentine isthmus connecting the protoconid with the entoconid is consistently extremely wide, nearly as thick as in the protoconid itself. This contrasts with the complete constriction of the entoconidhypoconid junction. The hypoconid extends anteriorly nearly as far as the entoconid and has only very thin enamel on its anterior edge. Posteriorly the hypoconid sweeps broadly down to the posterolophid, which does not extend lingually as far as either the metaconid or the entoconid. The postero-lingual edge of the m2 is withdrawn under the posterolophid, allowing the anterior cingulum of the m3 to rest underneath.

159

The third lower molar (Fig. 5A) is the largest, and simplest, tooth in the mandible. Left specimens show the diagnostic ‘‘S’’ shape of the genus. The metaconid is straight across its anterior edge and only slightly bulbous lingually. The pit from the erosion of the BRA2 is usually more developed in the m3 than the m1. There is no constriction between the protoconid and the entoconid. Non-dental material referred to S. curtisi is based on similarity to postcranial elements to extant Sigmodon. Discussion—Species of Sigmodon are well represented at nearly every late Pliocene and Pleistocene microvertebrate locality in the southern United States. It is the second most abundant rodent at Inglis 1C with 173 specimens. This abundance coupled with a rapid evolution ensures continued study for its biostratigraphic utility (Martin, 1979). Five species of Sigmodon occur as fossils in Florida: Sigmodon minor (including records of S. medius) in the late Blancan and early Irvingtonian; S. curtisi and S. libitinus from the early Irvingtonian; S. bakeri from the middle Irvingtonian through early Rancholabrean; and S. hispidus found from early Rancholabrean to today. Sigmodon fossils from Inglis 1C are not identical to previously described S. curtisi. The Inglis 1C specimens have a much better developed pit (derived from the enclosure of the BRA2) on the antero-external corner of the m2–3. In the S. curtisi sample from Inglis 1A this character is often reduced to a thickening of the enamel, and in a few specimens it is entirely absent. In this feature the Inglis 1C fossils may be more primitive, showing closer affinities to S. hudspethensis. This character, however, varies considerably within the cotton rat lineages. A quantitative comparison between samples of S. curtisi and S. hudspethensis is presented in Table 2. PEROMYSCUS Gloger, 1841 PEROMYSCUS SARMOCOPHINUS, sp. nov. (Fig. 6) Type Specimens—Holotype, UF 195148, L mandible with i1–m2, collected by S. Emslie and others on 1 April 1996 (Fig. 6A, B). Paratype, UF 195823, R maxilla with M1–3 (Fig. 6C). Type Locality—Inglis 1C, 2 km SSW of Inglis, SW¼, Sec. 10, T17S, R16E, Yankeetown 7.59 Quadrangle; Citrus County, Florida (FLMNH locality CI019). Chronologic and Geographic Range—Early Irvingtonian (late Pliocene) of Florida. Referred Specimens—UF 195821, 197825, L maxilla with M1–2; UF 196727, L maxilla with M1; UF 197826, L maxilla with M2–3; UF 195822, L maxilla with M2; UF 191988, L maxilla with M3; UF 196219–196223, 196718, 196728, 197827–197829, 198374–198375, 198377, 198470–198473, L edentulous maxilla; UF 195173–195175, 195830–195836, 196719, 196721, 196729–196734, 197830–197833, 198474– 198475, LM1; UF 195837–195839, 196278–196279, 196735– 196737, 197834, 197835, LM2; UF 195160, R maxilla with M1–3; UF 197836, R maxilla with M1–2; UF 195824, 198376, 198476, R maxilla with M1; UF 196738, 198477, R maxilla with M2; UF 195826–195829, 196723–196726, 197837, 198652, R edentulous maxilla; UF 195176, 195840–195849, 196739–196742, 197838–197839, 198478–198479, 198681, RM1; UF 195850, 196201, 197840, 198480–198481, RM2; 196280, L mandible with i1–m2; UF 186065, 195143–195146, 197841–1978142, L mandible with i1; UF 196217–196218, L mandible with m1–3; UF 195149, L mandible with m1–2; UF 196722, L mandible with m2–3; UF 195150, 196281, 197843– 197846, L mandible with m3; UF 186064, 191890, 191987, 195147, 196202–196204, 196282–196283, 197847, 197848– 197851, 197919, 198275, 198482, L edentulous mandible; UF 196205–196207, 196720, 196743–196744, 197852–197853,

160


FIGURE 5. Sigmodontine rodents from Inglis 1C. Occlusal views. A, Sigmodon curtisi, Rm1–3 in mandible, UF 195138. B, Peromyscus cf. P. polionotus, Rm1, UF 198645. C, Reithrodontomys humulis, Rm2, UF 198649; D, Baiomys sp., LM1, UF 198637. Anterior is toward the top of the page. Scale bar equals 1 mm.

198483, Lm1; UF 195169, 196208–196211, 196745, 197854, Lm2; UF 198651, Lm3; UF 195155, 196284, R mandible with i1, m3; UF 186070, 191985, 195152–195154, 196224, 196225, 196285–196286, 197855–197857, R mandible with i1; UF 197858, R mandible with m2–3; UF 186068–186069, 191888, 196287, 197859, R mandible with m3; UF 186066–186067, 191889, 195151, 196226–196228, 196288–196290, 197860– 197864, R edentulous mandible; UF 195170–195172, 196212– TABLE 2. Measurements of Sigmodon m1. Methodology for measuring Sigmodon m1 length and width follows Martin (1979) and Czaplewski (1987). The enamel height (ht) is measured on the lingual side of the tooth from the tip of the metaconid to the lower extent of the enamel, perpendicular to the straight portion of the linea sinuousa below the anteroconid, metaconid, and entoconid. All measurements are in millimeters. Localities: Inglis 1A, Citrus County, Florida (FLMNH CI001); Fyllan Cave, Travis County, Texas (TMM 40682); Madden Arroyo, Hudspeth County, Texas (TMM 40240); Red Light, Hudspeth County, Texas (TMM 40857). Sigmodon hudspethensis

Sigmodon curtisi Inglis 1C Number

39

Inglis 1A 36

Fyllan Cave 9

Madden Arroyo Red Light 4

2

m1 length range

2.31 2.06–2.45

2.32 2.17–2.51

2.46 2.4–2.6

2.18 2.1–2.2

2.25 2.0–2.5

m1 width range

1.59 1.41–1.79

1.62 1.49–1.79

1.59 1.4–1.8

1.45 1.4–1.6

1.3 1.2–1.4

Enamel ht range

0.84 0.54–1.29

0.92 0.65–1.14

1.00 0.8–1.4

0.66 0.6–0.8

0.96 0.7–1.2

Number of roots % with 2 6 % with 3 66 % with 4 28

4 78 18

— — —

0 75 25

0 100 0

FIGURE 6. Type specimens of Peromyscus sarmocophinus, sp. nov. from Inglis 1C. Labial (A) and occlusal view (B) of holotype, L mandible with m1–2, UF 195148. C, paratype, partial R maxilla with M1– 3, occlusal view of molars, UF 195823. Anterior is toward the left for A and toward the top of the page for B and C. Scale bar equals 4 mm for A and 1 mm for B and C.

196216, 196746–196747, 197865, Rm1; UF 195825, 196748, 197866, 198484, 198675, Rm2. Additional Locality—Inglis 1A, Citrus County, Florida (CI019): UF 200197, R maxilla with M1–3; UF 200198, RM1; UF 200092, L mandible with i1; UF 200093, L edentulous mandible; UF 200087–200089, R mandible with i1; UF 200090– 200091, 200094, R edentulous mandible; UF 200196, Rm1. Diagnosis—Referable to Peromyscus (sensu lato) due to the short coronoid process of the mandible, marked reduction of the M3, alternation of cusps, brachydont molars, narrow flexi and flexids, absence of a distinct labial cingulum, and conical metacone, paracone, entoconid, and metaconid; assigned to subgenus Peromyscus based on the large number of accessory tubercles between the principal cusps, and with a loop extending to the outer edge of the m1 and m2. Mandibular-alveolar length (mean 5 4.44 mm) shorter than Podomys floridanus (4.85 mm) and longer than that of the largest Peromyscus, P. gossypinus (4.07 mm); ramus-alveolar length (9.48 mm) slightly exceeds that found in P. floridanus (9.10 mm) and much larger than other peromyscines. Accessory

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA TABLE 3. Peromyscine accessory cusp frequency expressed as a percentage. Abbreviations: MS, mesostyle or mesostylid; ML, mesoloph or mesolophid; ES, enterostyle or ectostylid; EL, enteroloph or ectolophid. Sample size is shown in parentheses. MS

ML

ES

EL

P. sarmocophinus1 M1 (52) M2 (20) m1 (27) m2 (19)

Tooth

100 100 100 100

100 100 100 100

29 55 56 21

0 0 15 11

O. nuttalli2 (36) M1 M2 m1 m2

100 100 100 100

100 100 95 95

31 25 95 90

0 0 90 75

P. gossypinus3 (70) M1 100 M2 100 m1 99 m2 93

100 100 37 4

24 26 83 60

0 0 49 16

1

Inglis 1C; 2recent specimens from Kentucky, North Carolina, and Virginia (Hooper, 1957:fig. 18); 3recent specimens from Florida (Bader, 1959:table 1).

cusp frequency unique among peromyscines (Table 3); first and second upper molars identical to each other in frequency of accessory cusps and lophs except the enterostyle; first and second lower molars likewise show identical cusp frequencies save the extostylid; mesostyle, mesostylid, mesoloph, and mesolophid always strongly expressed; ectolophid occurs very rarely and poorly developed when present; enterostyle typically very weak or absent; m1 shows an ectostylid in 15 of 27 referred specimens, but in only 4 of the 21 available m2s—opposite of the trend shown in the enterostyle within the upper molars; mesolophid arises from the median murid, while mesostylid often has an independent connection to the entoconid; ectostylid (when present) shows an antero-posterior elongation; mesostyle central to, but never connected to, mesocone and paracone. Etymology—The specific epithet, sarmocophinus, is derived from sarma (Greek) meaning chasm in the earth and from cophinus (Latin) meaning coffin. It refers to the sinkhole deposits that contain all currently known specimens of this species, as well as the majority of fossil mammals found in Florida. Description—The descriptions here use the terminology of Reig (1977). One additional term is introduced to describe the bulbous labial extension of the posterolophid, posterolophulid. The use of ‘‘-lophulid’’ follows Reig’s (1977) definition as an occasional ridge that extends only for a short distance. ‘‘Postero-’’ was chosen due to its point of origin in the posterolophid. The outline of M1 is trapezoidal (Fig. 6C). The center of the anterocone is centered slightly labial of the midline of the tooth. The anteromedian flexus makes only a very shallow, but broad, indentation. A very weak protostyle sits diagonally on some specimens. The parastyle connects strongly to the anterolabial conule. When present (15 of 52 specimens), the enterostyle occurs weakly, as a fold in the relatively broad hypoflexus. The mesostyle sits at the extreme edge of the tooth connected by a prominent mesoloph to the median mure. The mesostyle/mesoloph only touch the metacone or paracone in extremely worn individuals. The postero-lingual corner of the hypocone cuts an angle similar to that in the anterior enamel of the protocone. The posteroflexus is nearly absent, and when present occurs only as a slight indentation on the postero-labial corner. This results in a point on the posterior edge of the tooth. All specimens have three distinct roots with 50% showing a fourth rootlet. When present it lies nearly adjoining the anterior root, between the paracone and anterolabial conule. The development

161

of the fourth root varies greatly from a barely perceptible bump to a length equivalent to that of the posterior root. Its development is not associated with any other qualitative or quantitative character in the Inglis 1C population. UF 195845 has a fourth rootlet originating directly under the paracone and extending the length of the posterior root. It is fused to the anterior root along its entire extent. The anterior cingulum on the M2 is slightly concave (Fig. 6C). The protoflexus truncates the antero-lingual corner of the protocone. A slender anteroloph connects to a lophate parastyle that closes the paraflexus. The remainder of the tooth is identical to the M1, except for the frequency of occurrence of the enterostyle—in 11 of 20 M2s (55%). The M3 is triangular in outline and much reduced compared to M1 and M2 (Fig. 6C). The anterior cingulum is slightly concave, as in M2. Parastyle is only slightly developed, closing the paraflexus completely in worn specimens. The large paracone cuts across the paraflexus at its posterior turn, connecting to the anterior mure. This creates an enamel lake in the middle of the tooth. A thick mesoloph extends posteriorly to the tooth margin. The mesostyle connects to the metacone, closing the metaflexus. The opening of the mesoflexus is as wide as the paracone. The mental foramen on the dentary is large and circular (Fig. 6A). It sits directly in line with the upper limit of the masseteric crest, one-fourth of the distance to the incisor alveolus. The masseteric crest is very rugose and extends anteriorly to immediately below the anterior edge of the anterior root of the m1. An enlarged capsular process on the condyloid can be seen when the dorsal aspect of the mandible is viewed. The anterior opening of the oval-shaped dental foramen lies half the distance between the posterior edge of the alveoli and the condyloid process. The coronoid process extends only one-fourth the distance from the base of the sigmoid notch that the condyloid does. The m1 anteromedian flexid varies greatly, ranging from not expressed at all to deeply incised in the anteroconid with either a wide or narrow notch (Fig. 6B). The small anteroconid lies just labial to the midline. The shape of the metalophid may be associated with the development of the anteromedian flexid; the deeper protrusion of the latter is usually accompanied by a wider opening of the former. The protostylid stretches from the anterolabial conulid posteriorly via the anterolabial cingulum to the bottom of the protoconid at the margin. The result is a closed protoflexid in worn teeth. A very distinct mesolophid and mesostylid are found in all specimens with each character connecting independently to the median murid and entoconid respectively. This results in an extremely wide entolophid, occasionally with a small enamel lake formed from the labial portion of the entoflexid. The ectostylid is present in just over half the specimens and is much more conspicuous than the corresponding lingual accessory style on the M1 (enterostyle). The ectolophid rarely extends completely to the ectostylid, but rather consists of a slender ridge extending out from the median murid. On the postero-lingual corner the posterostylid grades into the base of the entoconid, closing the posteroflexid. That fold is further obscured in its antero-labial extent by the rapid wear of the hypoconid. This additionally widens the entolophid/ mesoconid connection. In specimens with moderate wear, as in the holotype, the posterior edge of the hypocone does not represent the edge of the tooth. It slants postero-lingually to the posterolophid which has a slight, bulbous, labial projection, here termed the posterolophulid. The m2 occlusal outline is rectangular, slightly longer than wide (Fig. 6B). The anterior cingulum is rounded convexly anterior to the protolophulid, concave near the metaconid, and comes to a point at the protostylid. The mesostylid sits at the tooth margin, bisecting the opening between the protolophulid

162


and entoconid. Unlike the same character on the m1, the mesostylid usually only touches the mesolophid, which may link to either the median murid or the entolophid. The ectostylid appears in only four of 19 specimens and the ectolophid in only 11% (2 of 19). In none of the teeth do these two features connect. The description of the entolophid, hypoconid, posterolophid, and posterolophulid follow that of the m1. The occlusal outline of the m3 changes from near rectangular in slightly worn teeth to a rounded triangle in those from older individuals. The anterior cingulum is straight, only occasionally rounded at the protolophulid. Closure of the protoflexid is completed by the protostylid anteriorly and the anterolabial cingulum labially. These features lie very low on the tooth, not appearing on the occlusal surface until late in life. A small ectostylid appears in some teeth, while a mesostylid is found in all. It is difficult to tell if a mesolophid is present. If so, it has fused to the entoconid/entolophid and is not obvious as in all other teeth of this species. The posterolophid connects to the posterior end of the entoconid, forming a small, circular enamel lake from the posteroflexid. Discussion—A quick count of Peromyscus (sensu lato) in Carleton (1989) yields 70 species. Nowak (1991) includes 60 species. Kurteń and Anderson (1980) list 24 species as occurring in the fossil record, 13 of which are extinct. In spite of this apparent diversity, only three species were recorded in the literature from Florida sites as recently as 1992 (Hulbert, 1992). Citations of these mice were limited temporally from the middle Pleistocene forward. Subsequently, Morgan and White (1995) noted Podomys, sp. nov. from the early Pleistocene (late–early Irvingtonian) Leisey Shell Pit, Hillsborough County, Florida. Although only a single isolated m2 was described, this record greatly extended the geological range of peromyscines in Florida from 400 Ka to 1.5 Ma. The lack of study of Irvingtonian peromyscines is certainly not due to a shortage of material (with the exception of Leisey). Though localities representing the early Irvingtonian are few, sites where the matrix has been screen washed invariably turn up the small rodents. Especially notable in Florida for their abundance of peromyscines are Inglis 1A, Haile 16A, and Inglis 1C. Hooper (1957) fell short of his ultimate goal of including every living species of Peromyscus in his analysis of dental variation, but finished with an excellent odontological summary of nearly 2,000 individuals. The primary focus of his study was evaluation of the frequency of accessory cusp occurrence and shape. Subsequently, Bader (1959) added the descriptions of three more species, bringing the total to 20. Those not covered are found either only in Central America or on small, isolated islands in the Gulf of California (Nowak, 1991). These synthetic works allow for quick comparison of fossil to recent Peromyscus species. In spite of his work, Hooper (1957:52) doubted ‘‘that the true relationships of the species of Peromyscus can be determined if one relies principally on those structures which convert a simple pattern into a complex one, namely the accessory styles and lophs.’’ Further, the limitations in pure taxonomic usage of dental patterns was lamented: ‘‘In at least two species, population-to-population differences in the lophs, styles and dental patterns exceed those that contrast some full species’’ (Hooper, 1957:53). Many of the discrepencies that concerned Hooper were resolved with additional systematic revisions (Carleton 1989). Wolfe and Layne (1968) echoed the warning about the variation in the frequency of accessory cusps. In spite of these assertions, the consistency of the same characters has been argued as useful for species distinction (Schmidly, 1973). The absence of other data neccessitates the use of tooth morphology. In the case of Peromyscus sarmocophinus, the numerous absolute frequency occurrences (100% and 0%) of accessory

cusps (Table 3) lend credence to a discrete biological entity with minimal variation. The interrelation of Peromyscus dental measurements was analyzed by Van Valen (1962), and the third molars were determined to be independent from the other teeth. The coefficients of variation for third molars greatly exceed that of the others, further reducing any utility of the M3 and m3 in species descriptions. The first two molars in each row were found to function as a unit, and even M1–m1 and M2–m2 couplets were found to correlate well. Martin (1967, 1968) later developed mandibular measurements to help discriminate some species. With this advancement, even edentulous mandibles in mixed assemblages of late Pleistocene age could occasionally be assigned to species. Comparison with the accessory cusps frequency charts of Hooper (1957) and Bader (1959) reveals the closest similarity of Peromyscus sarmocophinus to Ochrotomys nuttalli and Peromyscus gossypinus. The M1s of the three species are identical, while the M2 of P. sarmocophinus differs by having a more common occurrence of enterostyle, although the discrepancy is small enough to be easily explained by population variance. The lower teeth of the three peromyscines, however, differ greatly. The mesostylid always occurs in all three species, but the mesolophid of P. gossypinus is only present in 37% of the m1s and 4% of the m2s. The latter character is present in all specimens of O. nuttalli and P. sarmocophinus. The Inglis 1C sample of Peromyscus can further be differentiated from P. gossypinus by the latter’s higher frequency and development of ectolophids on the m1, and the mesolophid characteristically originating in the entoconid rather than the median mure. Ochrotomys nuttalli differs in having very brachydont molars and a consistent appearance of both the ectostylid and ectolophid. The m1 of P. sarmocophinus has an ectostylid occurrence of about 50%, but this feature on the m2 and the ectolophid of both m1 and m2 are essentially absent. The sample of Peromyscus nudipes (considered a variant of P. mexicanus by Huckaby [1980] and Carleton [1989]) measured by Hooper (1957) show accessory cusp and loph frequency identical to that of O. nuttalli, but may be differentiated by its more severely convoluted enamel, especially obvious on the anteroconid. In size, Peromyscus sarmocophinus fits between P. gossypinus and Podomys floridanus. Table 4 lists mandibular dimensions for all recent Peromyscus found in the eastern United States and some fossil forms from the Irvingtonian of Florida. The average mandibular-alveolar length of P. sarmocophinus exceeds the range of P. gossypinus and falls in the lower extent of P. floridanus. Peromyscus sarmocophinus barely exceeds the average ramus-alveolar length of P. floridanus. Paleontological specimens of the subgenus Peromyscus are almost exclusively found in the Rancholabrean (Kurteń and Anderson, 1980; Tomida, 1987; Hulbert, 1992); the abundant records of Irvingtonian Peromyscus (senso lato) are generally referable to subgenus Haplomylomys. Only P. hagermanensis from Idaho (Hibbard, 1962; Zakrzewski, 1969), Arizona (Tomida, 1987), and California (Albright, 1999) and P. nosher from Washington (Gustafson, 1978) are species in the subgenus Peromyscus from deposits older than late Irvingtonian. Albright (1999) recently named two new species of late Blancan Peromyscus, (P. maximus and P. complexus) which are likely referable to the subgenus Peromyscus based on their extreme complexity. The description of P. hagermanensis specimens from southern Arizona (Tomida, 1987) is used for comparison to P. sarmocophinus due to the scarcity of the topotypic material. Only significant differences are mentioned. The mesoloph in P. hagermanensis is present in all specimens, but does not always reach the mesostyle (it reaches in only 8 of 12). The mesostyle is absent in one of twelve specimens, and is poorly developed


163

TABLE 4. Mandibular measurements of Peromyscus (senso lato). Species above the dashed line are extant; those below are from the early Irvingtonian of Florida. Abbreviations: N, number of measured specimens; OR, observed range; SD, standard deviation of the sample. Measurements are in millimeters. Mandibular–alveolar length Podomys floridanus1 Peromyscus gossypinus1 Peromyscus gossypinus2 Ochrotomys nuttalli1 Peromyscus leucopus1 Peromyscus leucopus2 Peromyscus leucopus3 Peromyscus maniculatus1 Peromyscus polionotus1

Ramus–alveolar length

N

Mean

OR

SD

N

Mean

OR

SD

34 49 20 41 26 19 101 27 50

4.85 4.05 4.1 3.95 3.85 3.7 3.7 3.6 3.3

4.4–5.35 3.75–4.35 3.8–4.3 3.75–4.35 3.5–4.3 3.5–4.0 3.1–4.4 3.4–3.85 3.1–3.7

0.22 0.15 0.12 0.11 0.15 0.16 — 0.13 0.15

34 49 20 41 26 20 — 27 50

9.1 8.4 7.9 7.3 7.5 7.4 — 7.6 6.7

8.35–10.00 7.5–9.1 7.1–8.3 6.6–7.8 6.9–8.1 7.0–7.8 — 7.1–8.0 6.2–7.3

0.4 0.38 0.3 0.27 0.31 0.24 — 0.28 0.27

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Peromyscus sarmocophinus4 Peromyscus sarmocophinus5 Peromyscus, sp. nov.6 Peromyscus, sp. nov.5 Podomys, sp. nov.6

57 7 36 41 23

4.44 4.42 3.86 3.74 4.68

4.00–4.76 4.23–4.71 3.59–4.12 3.54–3.98 4.43–4.95

0.16 0.17 0.14 0.11 0.15

26 — 11 16 4

9.48 — 7.29 7.42 9.36

8.64–10.07 — 6.79–7.74 6.57–7.42 9–9.52

0.36 — 0.31 0.62 0.24

Sources: 1from Martin (1967:figs. 2–3); 2from Martin (1968:table 2); 3from Tamsitt (1957:table 1); 4Inglis 1C; 5Inglis 1A; 6Haile 16A.

in the other eleven. This contrasts with the extremely well developed mesoloph and mesostyle in all P. sarmocophinus. Peromyscus hagermanensis has a mesolophid that reaches the mesostylid in only 6 of 13 specimens. When it does occur, it originates on the entolophid. The mesolophid is well developed in P. sarmocophinus, and typically originates from the median mure, while the mesostylid makes a connection to the entoconid or, rarely, the entolophid. Tomida (1987) related P. hagermanensis in general morphology to extant P. maniculatus. Quantitatively (Table 5), P. sarmocophinus is significantly larger, especially in the M3 (a caveat concerning the diagnostic use of M3 has already been discussed). Though Tomida (1987) separated P. nosher from P. hagermanensis by virtue of the more strongly bilobed anteroconid and small, cuspoid parastyle of the former, these characters may not be of taxonomic value (Albright, 1999). Because they are very similar in other morphological landmarks and in size, these taxa could prove to be conspecific. Peromyscus hagermanensis was named from a single specimen, and P. nosher was described from only four. With the high number of rodent fossils recovered in screen washing techniques and the immense variation shown in modern populations, caution should be used when diagnosis of a new species is based on small collections. Peromyscus sarmocophinus differs from both of the species described by Albright (1999) by being intermediate in size.

Qualitatively, Peromyscus maximus differs greatly in complexity; the protocone splits and attaches to both the labial and lingual lobes of the anterocone. Another distinct character is the connection of the metacone to the posterostyle rather than by the metaloph to the hypocone. Peromyscus complexus can quickly be differentiated from all other species by its five projections extending off the M2 paracone. Peromyscus hagermanensis shows a dental pattern comparable with that of a hypothetical ancestor of P. sarmocophinus. It is not difficult to imagine an evolutionary sequence from the Blancan Hagerman mice to those in the Irvingtonian of Florida, with molar complexity increasing with time. The unique set of characters in P. sarmocophinus makes it doubtful that this species has any extant descendents. Peromyscus gossypinus would represent the most likely possibility. That progression, however, would require an unusual trend: significant increase in ectolophid and ectostylid frequency coupled with a massive decrease (from 100% to 4% in m2) in the frequency of the mesolophid. Based solely on accessory cusp frequency, Ochrotomys nuttalli also represents a possible descendent. Ochrotomys differs in its more brachydont molars, compressed and thicker enamel folds, and the high placement of an oval mental foramen on the mandible instead of a circular foramen located lower in the jaw. The similarities between O. nuttalli and P. gossypinus are likely due to convergence. They share identical macro-habitat (south-

TABLE 5. Comparison of Peromyscus sarmocophinus with P. hagermanensis and P. nosher. Peromyscus sarmocophinus includes all referred specimens from Inglis 1A and Inglis 1C. Peromyscus hagermanensis is a composite of specimens from San Timoteo Formation (Albright, 1997), Duncan (Tomida, 1987), and 111 Ranch (Tomida, 1987). Specimens of P. nosher are from White Bluffs (Gustafson, 1978). Measurements are in millimeters. Abbreviations: L, length; W, width; N, number; OR, observed range. Peromyscus sarmocophinus M1 M2 M3 m1 m2 m3

L W L W L W L W L W L W

Peromyscus hagermensis

Peromyscus nosher

N

Mean

OR

N

Mean

OR

N

Mean

OR

53 54 23 23 4 4 25 24 17 17 19 18

1.87 1.18 1.43 1.13 0.92 1.00 1.74 1.11 1.46 1.12 1.13 0.96

1.65–2.09 1.09–1.35 1.31–1.52 0.94–1.29 0.88–0.96 0.96–1.03 1.58–1.92 0.94–1.25 1.35–1.57 1.00–1.21 0.90–1.28 0.84–1.06

16 19 18 18 7 7 16 18 21 22 8 8

1.71 1.20 1.26 1.08 0.79 0.84 1.59 1.02 1.31 1.02 1.06 0.82

1.62–1.85 1.00–1.22 1.18–1.50 1.00–1.28 0.70–0.97 0.76–1.04 1.46–1.70 0.94–1.19 1.16–1.52 0.82–1.12 0.94–1.14 0.76–0.86

3 3 — — — — 1 1 2 2 1 1

1.67 1.1 — — — — 1.4 1.0 1.25 0.95 1.1 0.8

1.6–1.8 1.1 — — — — 1.4 1.0 1.2–1.3 0.9–1.0 1.1 0.8

164


eastern United States), micro-habitat (very moist with dense vegetation), and diet (seeds and insects) (Linzey and Packard, 1977; Wolfe and Linzey, 1977). Their reciprocal density relationship, as determined by capture and release studies (Pearson, 1953; McCarley, 1958), further suggests an equivalent ecological niche. PEROMYSCUS POLIONOTUS (Wagner, 1843) Referred Specimens—UF 197825, L maxilla with M1–2; UF 196749, LM1; UF 198647, 198653–198654, LM2; UF 191985, R mandible with i1; UF 198643, 198645, Rm1. Chronologic and Geographic Range—Early Irvingtonian (late Pliocene) and Rancholabrean (late Pleistocene) of Florida, and recent of the southeastern United States. Diagnosis—Referable to Peromyscus (sensu lato) due to the short coronoid process of the mandible, alternation of cusps, brachydont molars, narrow flexi and flexids, conical metacone, paracone, entoconid and metaconid, and absence of a distinct labial cingulum; assigned to subgenus Peromyscus based on the principal cusps consisting of a loop extending to the outer edge of the tooth and dentine spaces of molars confluent. Description—The center of the M1 anterocone lies just labial of the midline of the tooth and lacks an anteromedian flexus. Relatively thick dentine isthmuses connect each major loph. The first upper molar is the most complex tooth. A mesostyle and mesoloph are apparent in both specimens, while an enterostyle occurs only in UF 196749. Lengths: 1.41 and 1.49 mm; widths: 0.88 and 0.94 mm. The anterior cingulum of the M2 is straight. The mesostyle appears in all specimens, while the presence of the mesoloph is more tenuous. A weakly developed mesoloph occurs on three of the four specimens, but no trace of one can be found on UF 198647. Average length: 1.15 mm; width: 0.83 mm. Length of the m1–3 alveolus (3.42 mm) is average for P. polionotus, barely within the observed range for P. maniculatus, and smaller than all other members of the genus. The lower teeth both appear to have been digested, slightly impeding the description. A poorly developed anterolabial cingulum projects from the m1 anteroconid (Fig. 5B). A very weak ectostylid is present in both, and one (UF 198645) has a small mesolophid extending anteriorly from the entolophid. The two m1s are probably conspecific with the other specimens, based both on size and the resemblance to recent P. polionotus. Lengths: 1.37 and 1.43 mm; widths: 0.81 and 0.83 mm. Discussion—The allocation of the Inglis 1C specimens to Peromyscus polionotus, the cotton and beach mice, extends its chronologic range into the late Pliocene. Previously recognized specimens are confined to sites of late Rancholabrean age. This placement contrasts sharply with earlier suggestions of more recent origination from its possible ancestor, Peromyscus maniculatus. The relation of beach mice to P. maniculatus was recognized early in the study of the group by Osgood (1909). The current geographic distributions of the two species do not overlap, with the woodland form P. m. nubiterrae of Alabama nearest to the modern range of beach mice. Bowen (1968) cited the tendency for P. m. bairdii to spread into deforested areas as a modern example of the evolution of P. polionotus from a grassland form. Peromyscus polionotus is most similar morphologically to modern P. m. pallescens from Texas, so Blair (1950) proposed a continuous population spanning the narrow strip around the Gulf of Mexico, and that the final geographic split was not until the Mississippi Delta built up at the end of the Pleistocene. The presence of P. polionotus in Sangamonian localities of Florida (Holman, 1959; Gut and Ray, 1963) makes this scenario unlikely. Alternatively, the continuous population of Blair (1950) could have become isolated from the parental

stock during the Sangamon interglacial, with the two species allopatric since that time. REITHRODONTOMYS Giglioli, 1874 REITHRODONTOMYS WETMOREI Hibbard, 1952 Referred Specimens—UF 195813–195814, L edentulous maxilla; UF 198637, LM1; UF 198639, LM2; UF 195159, 195815, R maxilla with M1–2; UF 195816–195817, R edentulous maxilla; UF 198641, 198646, RM1; UF 191984, L mandible with i1–m1; UF 197887–197888, L mandible with i1; UF 195157, L mandible with m1–2; UF 191982, 195818, 195820, 196294, 198489, L edentulous mandible; UF 198642, Lm1; UF198638, Lm2; UF 196295, R mandible with i1–m2; UF 196296, 191983, R mandible with i1; UF 197890, R mandible m1; UF 195819, R mandible with m2; UF 195158, R edentulous mandible; UF 198644, Rm1. Chronologic and Geographic Range—Early Blancan (late early Pliocene) of Kansas and early Irvingtonian (late Pliocene) of Florida. Diagnosis—Small peromyscine, yet large for the genus; limits of the masseteric scar not continuous with the anterior edge of the ascending ramus nor the bottom of the mandible as it is in R. humulis; masseteric scar less rugose than other Reithrodontomys; capsular process is extremely enlarged and has a fold above it; lacking a labial cingulum on m1–2; m1 with a short, narrow, and shallow metaflexid; ‘‘S’’ shaped occlusal pattern on m3. Description—The anterior edge of the M1 is greatly sloped from the alveolar to occlusal level, exposing a large surface in front of the anterocone (Fig. 7C). The transversely elongated anterocone has a rounded lingual margin and a pointed labial margin that connects to the parastyle. The anterior margin is nearly straight, slightly convex in UF 195815, and with a very faint anteromedian flexus in UF 195159. There is a very small lingual cingulum. A long anterior mure connects to the protocone. The paraloph extends straight across the tooth (labially) to the paracone. The paracone is greatly elongate laterally, and the metacone is slightly so. The median mure is long and thick. The protoflexus, paraflexus, hypoflexus, and metaflexus are all wide and deep, thanks largely to the elongate mures. A very small enterostyle is present, rising off the cingulum. The posterostyle and posteroloph are absent. It is difficult to examine the roots since the teeth available are still in the maxilla, but UF 195815 appears to have a tiny rootlet under the paraflexus in addition to the 3 main roots. There is a high degree of variation within the M2 (Fig. 7C). The occlusal outline ranges from square (UF 195815) to oval (UF 198639, 195159), all with a nearly straight anterior margin. A weakly developed protostyle connects to the anterior arm of the protocone and continues to the labial edge of the tooth via a distinct anteroloph and parastyle in UF 195159. In UF 198639 the protostyle is deflected anteriorly and is even fainter. UF 195815 does not appear to possess a protostyle, but this character may simply not be visible due to the more advanced wear of this specimen. The paraflexus is narrow and deep. Both the hypoflexus and metaflexus are wide and deep except in UF 195815, which appears considerably more compact antero-posteriorly. UF 195815 also lacks the elongate enterostyle present in the other specimens. In fact, this tooth probably would not be allocated to R. wetmorei had it not occurred in a maxilla with a diagnostic M1. The masseteric scar on the dentary extends anteriorly to the m1 alveolus (Fig. 7A). Only the ‘‘.’’ shape at the anterior edge is developed, and it lies at the same level of the mental foramen. The capsular process bulges only slightly. There is a deep pterygoid fossa, separated from the inferior dental foramen by a pronounced ridge.


165

TABLE 6. Reithrodontomys wetmorei from Inglis 1C. Abbreviations: L, length; W, width; OR, observed range; N, number; SD, standard deviation. Mean

OR

M1–3 alveolar length M1 L W M2 L W

3.38 1.27 0.86 0.95 0.82

— 1.26–1.28 0.82–0.90 0.88–1.02 0.80–0.86

N 1 2 2 3 3

SD — — — — —

m1–3 alveolar length m1 L W H m2 L W

2.99 1.25 0.78 0.50 1.02 0.79

2.87–3.08 1.24–1.27 0.72–0.81 0.44–0.63 0.96–1.19 0.69–0.84

13 6 6 4 6 6

0.08 0.01 0.04 0.09 0.10 0.06

locality, Fox Canyon, Kansas (Hibbard, 1952). The range of R. wetmorei is extended geographically from Kansas to Florida and chronologically from the early Blancan to the early Irvingtonian. These geographic and temporal differences are sufficient to immediately question the validity of this taxonomic assignment, but the qualitative characters mirror the early Blancan form exactly. Measurements of the Inglis 1C specimens (Table 6) do not differ significantly from the Kansas material, but the type material consists of only two mandibles. REITHRODONTOMYS HUMULIS (Audubon and Bachman, 1841)

FIGURE 7. Reithrodontomys wetmorei from Inglis 1C. A, labial view of R mandible, UF 195158; B, occlusal view of Rm1–2 in mandible, UF 19629; C, occlusal view of RM1–2, in maxilla, UF 195815. Anterior is to the right for A and toward the top of the page for B and C. Scale bar equals 2 mm for A and 0.5 mm for B and C.

The m1 occlusal outline is oval (Fig. 7B). The anteroconid is small, rounded, and connects to a well-developed anterolabial cingulum that extends almost to the protoconid. The metaflexid is short, narrow, and shallow. The metaconid and entoconid are transversely elongate and less rounded than the labial cusps. In UF 196295 the metaconid is isolated and does not connect to the anterior murid. The entoconid (mislabeled metaconid in the original species description) protrudes lingually into the mouth. A greatly extended median murid separates the very wide and deep hypoflexid and entoflexid. The result is an isolation of the posterior cusps (hypoconid and entoconid). An ectostylid appears on the poorly developed labial cingulum. Only the two primary roots are present with no indication of rootlets. A distinct protostylid on the m2 connects to the metalophid, forming a straight anterior margin abutting the m1 (Fig. 7B). The narrow and deep protoflexid extends postero-labially to the tooth margin. The tooth is identical to the m1 except for the lack of an anteroconid and the further lingual protrusion of the entoconid. The m3 is not represented in the Inglis 1C collection, but does occur in Inglis 1A; the description of this tooth is taken from those specimens. The protostylid connects to the anterior arm of the protoconid. The anterior mure is slightly concave. The hypoflexid and entoflexid both make a deep cut halfway across the tooth. Each of them is partially closed: the hypoflexid by the ectostylid, and the entoflexid by a well-developed entolophulid. The occlusal pattern is ‘‘S’’ shaped. Discussion—The occurrence of Reithrodontomys wetmorei at Inglis 1C represents the first appearance outside the type

Referred Specimens—UF 198649, Rm2. Chronologic and Geographic Range—Early Irvingtonian (late Pliocene) through recent of Florida, Rancholabrean (late Pleistocene) of California, and recent throughout the southeastern United States. Diagnosis—Smallest species in the genus; distinctive due to the very pronounced labial cingulum; concave anterolingual margin and convex anterolabial margin. Description—This tooth is indistinguishable from that of extant R. humulis. The metaconid is not yet connected to the protoconid due to the slight wear (Fig. 5C). Thick enamel curves from the metaconid, through the protostylid, and posteriorly along the anterolabial cingulum. A slightly developed ectostylid is indicated on the labial cingulum. The thin and elongate protoconid and entoconid are connected by a slender isthmus of dentine that is wider than that joining the other cusps. The lunate hypoconid lies nearly on the midline. The posterolophid reaches as far lingually as does the entoconid (only because of the tooth’s stage of wear) and expands transversely into the posterostylid. Discussion—This specimen may be differentiated from R. wetmorei by the very large labial cingulum. Additionally, the anterior margin of this tooth and all recent R. humulis examined is lightly concave lingually and convex labially, while the R. wetmorei m2s have a straight edge. Rancholabrean R. humulis are significantly smaller than the Inglis 1C specimens. UF 198649 measures 1.11 by 0.83 mm, while Rancholabrean specimens from Florida (Ruez, 2000; Table 2) average 0.92 by 0.78 mm. Study of the large sample of Reithrodontomys Haile 16A, Alachua County, Florida will help quantify the variation, but this is beyond the scope of this study. The presence of both R. wetmorei and R. humulis at Inglis 1C warrants a brief summary of the differences between the two species. Reithrodontomys wetmorei may be differentiated from R. humulis by the latter having a thick labial cingulum, wide metaflexid on the m1, and ‘‘C’’ pattern on the m3. On mandibles of R. wetmorei, the limits of the masseteric scar are not continuous with the anterior edge of the ascending ramus nor the bottom of the mandible as it is in R. humulis. The scar

166


on R. wetmorei is much less rugose. The extant species also has the scar higher on the mandible, above the mental foramen. Additionally the capsular process is extremely enlarged and has a fold above it in R. humulis. REITHRODONTOMYS sp. Referred Specimens—UF 198368–198373, 198636, LI1; UF 198359–198367, 198635, 198640, RI1. Remarks—Upper incisors of Reithrodontomys are very distinctive due to their small size and deeply grooved lateral surface. The presence of more than one type of harvest mouse at Inglis 1C precludes specific assignment. BAIOMYS True, 1894 BAIOMYS sp. Referred Specimens—UF 197891, right edentulous mandible; UF 198637, LM1; UF 198641, RM1. Diagnosis—Smallest and most brachydont of New World mice; coronoid process enlarged, broad, and strongly recurved; capsular process enlarged; complete lack of labial cingulum; simple occlusal pattern, lacking accessory cusps. Description—The M1 occlusal outline is trapezoidal, with the small, rounded anterocone greatly shifted labially (Fig. 5D). The protocone, paracone, hypocone, and metacone are all softly rounded, as in all Baiomys. Likewise, the paraflexus, protoflexus, metaflexus, and hypoflexus are wide and deep folds. A posterior expansion of the metacone completes the edge of the tooth. The teeth are very brachydont (enamel height at paracone50.46 and 0.43 mm) in spite of the nearly unworn nature of the teeth. Their widths (0.81 and 0.83 mm) are similar to Reithrodontomys although their lengths are much less (1.16 and 1.16 mm). In addition to the three main roots, UF 198637 has a small bump located centrally. The masseteric scar on the dentary is poorly developed, with the anterior point under the alveolus of the anterior root of m1 and lateral and above the mental foramen. The capsular process is enlarged. The pterygoid fossa is separated from the inferior dental foramen by a pronounced ridge that extends to the m3 alveolus. The m1–3 alveolar length is 2.78 mm. Discussion—The presence of Baiomys at Inglis 1C is the first fossil record of the genus east of Texas. The Inglis 1C specimens are referable to Baiomys as their size is less than any other murid rodent; only recent Reithrodontomys humulis and late Blancan R. pratincola (Hibbard, 1941a) are close. Pygmy mice also differ from harvest mice in being more brachydont, lacking a labial cingulum, having a more enlarged capsular process, and possessing a relatively larger coronoid. Baiomys may be discerned from the small Peromyscus, P. polionotus, by the simpler occlusal pattern and much smaller size. Unfortunately, almost every described species of fossil Baiomys is known only from lower dentition. It is impossible to tell if the Inglis 1C specimens belong to a named species. The mandible alone is not sufficient to permit specific identification. Based on size and the general simplicity of the lower teeth B. rexroadi (Hibbard, 1941b, c, 1950) and B. kolbi (Hibbard, 1952) from the early Blancan of Kansas and an undescribed species from Hagerman (Zakrzewski, 1969) are most similar to Inglis 1C Baiomys. ARVICOLINAE Gray, 1821 ATOPOMYS Patton, 1965 ATOPOMYS TEXENSIS Patton, 1965 ATOPOMYS SALVELINUS Zakrzewski, 1975 Referred Specimens—UF 200227, L maxilla with M1; UF 200228, LM1; UF 200229, LM2; UF 200230–200231, RM1; UF 200232, Lm1. Chronologic and Geographic Range—Early Irvingtonian

FIGURE 8. Arvicolines of Inglis 1C. Occlusal views. A–C, Atopomys texensis: A, RM1, UF 200230; B, Lm1, UF 200232; C, LM2, UF 200229. D, Ondatra idahoensis, RM3, UF 200233. Anterior is toward the top of the page. Scale bar equals 1 mm for A, B, and C and 2 mm for D.

(late Pliocene) of Florida, late early Irvingtonian (early Pleistocene) of Florida and Texas, and early late Irvingtonian (middle Pleistocene) of Maryland and West Virginia. Diagnosis—Small, hypsodont, rooted arvicoline; well developed dentine tracts; cement absent in reentrant angles; relatively thick enamel; loose enamel folds, lacking sharp angles; triangles widely confluent; m1 consisting of an anteroconid complex (ACC), three triangles (T1, T2, and T3), and posterior loop; crenulations on the (ACC) of the m1 variable in development, but always more complex in the ontogenetically younger individuals. Description—Terminology of arvicoline occlusal patterns follows van der Meulen (1978). The anterior loop of the M1 has a symmetrical leading edge and sits in front of four alternating triangles (Fig. 8A). The T1 is widely confluent with the T2, as is the T3 with the T4. In each case the dentine isthmus is twice the width of the T2–T3 connection. The posterior edge of the tooth curves uniformly, with the apex just labial of the midline. The posterior and anterior roots are double the length of the centrally located lingual root. There is no cementum in the enamel folds. The labial dentine tracts are extremely long and thin. On the lingual side only the dentine tract on second buccal salient angle (BSA2) is well developed. The M2 anterior loop is slightly bulbous lingually (Fig. 8C). The three triangles alternate with a roughly equivalent confluence between each. The posterior edge of the tooth ends in a

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA rounded right angle. The first, second, and third buccal salient angles (BSA1, BSA2, and BSA3) are arranged linearly, rather than with the BSA2 greatly extended labially. This is due to UF 200229 being only slightly worn. There are only two roots. Only BSA1 does not have an extensively developed dentine tract. The m1 anteroconid complex is very simple for an arvicoline (Fig. 8B), having only a slight indication of third buccal reentrant angle (BRA3). A wide opening connects the ACC to the three alternating triangles. T1 and T2 are widely confluent, almost twice the width of the dentine separating T2 and T3. The posterior loop terminates with a straight edge. LSA1 has a slightly developed dentine tract, while all other salient angles have extensive ones. There are two roots. Discussion—Atopomys texensis was previously recorded only from Fyllan Cave and Kitchen Door, Texas (Patton, 1965; Winkler and Grady, 1990). Specimens from Trout Cave, West Virginia and Cumberland Cave, Maryland later served as the basis for the description of an additional species, A. salvelinus. At that time the distinctions between the species seemed clear. Subsequent finds referred to A. salvelinus, especially that in Hamilton Cave, West Virginia, resulted in a review of the dental variation by Winkler and Grady (1990). Their description of new specimens blurred the distinction between the two species, yet they deferred synonymizing in favor of waiting for additional collections of A. texensis. Atopomys salvelinus was originally differentiated from A. texensis by m1 having better developed dentine tracts, simpler ACC, shallower depth of the LRA2, and deeper BRA2 (Zakrzewski, 1975). With the material available at the time, the first two characters clearly separated the species. The supposedly different penetration of BRA2, however, is not seen on the holotypes of either of the two species; they have identical measurements. The average depth of BRA2 of the samples of both species are not significantly different either (Winkler and Grady, 1990). Differences in LSA2 development and dentine tract height noted by Zakrzewski (1975) disappeared with the addition of the Hamilton Cave specimens (Winkler and Grady, 1990). The only characters Winkler and Grady considered reliable were the length of the m1 and complexity of the ACC. The m1s of Hamilton Cave Atopomys averaged 6% smaller than the Texas form (1.67 vs. 1.78 mm), a statistically significant, though very small, difference. Based on the amount of overlap in range (Fyllan Cave51.65–2.00 mm; Hamilton Cave51.45– 1.90 mm), and the high variance, I suggest that this character is not reliable. No other measurements presented by Winkler and Grady (1990) show a significant difference. In fact, the Hamilton Cave teeth wholly include the range of the Texas specimens for every measurement except the length and width of m1 and the width of M1. The complexity of the m1 may give the only valid difference. Atopomys from Fyllan Cave has crenulated enamel on the ACC in 80% of the specimens, while Hamilton Cave Atopomys is smooth in 91%. In both samples the feature exhibits a complete gradation in degree of complexity. When viewed laterally, the crenulations on the ACC extend only slightly below the occlusal surface. The folds disappear with wear. In many arvicolines crenulations are present in juveniles, but are lost as the tooth wears (e.g., Zakrzewski, 1969). Assuming the validity of both species, the occurrence of A. texensis in the late early Irvingtonian of Texas and A. salvelinus in the early late Irvingtonian of Maryland and West Virginia made a middle Irvingtonian species transition somewhat plausible. This temporal placement of Fyllan Cave is contrary to dates that have placed it in the middle or late Irvingtonian based on selected arvicolines (Taylor, 1982; Repenning, 1987). Examination of the entire mammalian assemblage (especially the presence of Ondatra hiatidens, Sigmodon curtisi, and Sylvilagus hibbardi) supports the earlier date. The introduction of Florida

167

TABLE 7. Florida Atopomys specimens. Measurements are in millimeters. N Length Inglis 1C De Soto Shell Pit Haile 16A

M1 M2 m1 M2 m1 m2 m1 m2

4 1 1 1 1 2 1 1

1.74 1.42 1.74 1.51 1.89 1.47 1.62 1.35

Range

Width

Range

1.68–1.78 — — — — 1.46–1.48 — —

1.13 0.85 0.93 1.21 1.09 1.06 1.06 1.1

1.08–1.11 — — — — 1.05–1.06 — —

specimens complicates this potential transition. The Haile 16A specimen was referred to A. salvelinus because of its small size and simple ACC (Winkler and Grady, 1990). Although Lundelius et al. (1987) tentatively assigned a late Irvingtonian age to the locality, most authors have referred it to the late early Irvingtonian (Martin, 1987; Morgan and Hulbert, 1995). This makes Haile 16A very similar in age to Fyllan Cave, with the former being only slightly younger based on the presence of Sigmodon libitinus, although S. libitinus may be restricted to Florida, making this point moot. This severely confines the (possible) transition within Atopomys, while the subsequent period shows no trend. The De Soto Shell Pit and Inglis 1C specimens share a combination of the features present in the two potential species. They have the simple uncrenulated ACC found in the Hamilton Cave collection, while the large size of the m1 and shallowness of the BRA2 imply affinities with the Fyllan Cave fauna. The Florida teeth are more than just a mixture of the characters; they are extreme in their expression. The ACC is simpler than in the West Virginia specimens, and the length of the m1 is greater than the maximum size observed in the Texas fossils. Measurements of the Florida Atopomys specimens are presented in Table 7. Even without the Florida specimens, the similarities and variability of characters suggest synonymy of the two species of Atopomys. The variation exhibited in this group is even exceeded by intraspecific characters in extant populations of arvicolines (Nelson and Semken, 1970; Martin, 1987; Harris, 1988). I propose that the single species of Atopomys changed from a crenulated form seen in the central United States to the simple ACC in the eastern US, the trend following a geographic, not chronologic, cline. This transition is not without precedent; an extant arvicoline, Microtus pennsylvanicus exhibits such a change in crenulation. Semken (1966) noted an increase not simply in complexity, but in the addition of entire new triangles on the m1 along a westward transect of the United States. The existing records of Atopomys also indicates a latitudinal variation, with larger the southern localities exhibiting larger size. ONDATRA Link, 1795 ONDATRA IDAHOENSIS Wilson, 1933 Referred Specimens—UF 200233, RM3. Chronologic and Geographic Range—Late Blancan (late Pliocene) of Arizona, Idaho, Kansas, Nebraska, and Wyoming and early Irvingtonian (late Pliocene) of Arizona and Florida. Diagnosis—Smallest species of Ondatra, with the least amount of cementum in reentrant angles and least developed dentine tracts; distinguished from Pliopotamys by possessing cementum in reentrant angles and having better developed dentine tracts. Description—Enamel pattern mirrors that found in modern Ondatra zibethicus, consisting of an anterior loop, four alternating triangles, and a broad posterior complex (Fig. 7D). Each

168


opposite pair of triangles is equivalently confluent and slightly more open than in the extant species. Approximately 0.5 mm below the occlusal surface there are small protrusions on the lateral side of the salient reentrant angles, the posterior cap, and the anterior loop. Slight wear would erode these ridges, as they appear very high on the tooth. Only a trace of cementum fills the reentrants. There are two prominent roots. Dentine tracts are moderately well developed. Discussion—The muskrat lineage is characterized by an increase in size, development of dentine tracts, and increase in amount of cementum (Hollister, 1911; Zakrzewski, 1969) from the ancestral Pliopotomys through extant Ondatra zibethicus. Quantification of these traits on the m1 permitted the development of a chronocline of western and mid-continent deposits (Semken, 1966; Nelson and Semken, 1970). Comparisons of the measurements to teeth from Inglis 1A are difficult because the Inglis 1C specimen is very young and has not yet worn down to the size shown in the adult. The Inglis 1C Ondatra measures 2.69 by 1.59 mm. UF 69011 from Inglis 1A has 70% the enamel height of the Inglis 1C specimen and measures 2.93 by 1.55 mm. Two complete maxillae are also preserved, but both are from very old individuals. The four M3s average 3.20 by 1.75 mm. In their extreme age, the teeth tend to wear partially on the side, creating a wider occlusal surface. It is difficult to use the length of this immature tooth as a chronological indicator, especially at this time in the ondatrine lineage. During the approximate time of Inglis 1C, Ondatra exhibits a short reversal in the trend of increasing size. Based on the presence of cementum in the reentrants, the Inglis 1C specimens presumably occurred later in the general sequence than does Pliopotamys. DISCUSSION Chronology Determining the age of karstic fillings means starting with very little information usually used to assess the geological age of a fossil locality. The only immediate conclusion is that the sediments are younger than the surrounding Eocene limestone. Two approaches were attempted here to refine the age of the fauna: paleomagnetics and mammalian biochronology. Magnetic stratigraphy typically involves matching the pattern of reversals in a section to the geomagnetic polarity time scale (Fig. 2). Sinkholes rarely have adequate stratigraphic sequences for such evaluation, and even when they do, their geologic rapid filling typically precludes deposition spanning several polarity reversals. Therefore, rather than typical magnetic stratigraphic interpretations, only a normal or reversed signal can be hoped for. Nevertheless, when used with other chronologic data, paleomagnetism can further refine geological age estimates. The mammalian biochronology of Plio–Pleistocene sites in Florida was recently revised (Morgan and Hulbert, 1995), providing an excellent framework in which to evaluate Inglis 1C. Thermal and alternating field demagnetization (with a Schonstedt TSD-1 thermal demagnetizer and GSD-1 AF demagnetizer respectively) of the samples produced random declinations and inclinations and in most cases failed to decrease the intensity. This indicated an unstable magnetization and no polarity interpretation was possible. Inglis 1C is located at sea level, and part of the excavation site currently lies under water at high tide. This repeated exposure of the sands to moving water may have randomized any initial signal. Biochronologically useful mammals found at Inglis 1C are presented in Figure 9. The following biochronologic assessements are based on taxon ranges compiled by Morgan and Hulbert (1995). A maximum possible age can be assigned to Inglis 1C based on the first occurrence of Dasypus and Holmesina in the late Blancan. The minimum age of late early Irvingtonian

is based on the last occurences of Orthogeomys propinetis and Sylvilagus webbi during this period. Inglis 1C best fits the description of early Irvingtonian faunas in containing the characteristic taxa Sigmodon curtisi and Ondatra idahoensis. The Haile 7C fauna represents the only latest Blancan assemblage in Florida. This age is based on an undescribed species of Chelydra and the intermediate size of Holmesina floridanus as compared with late Blancan and early Irvingtonian specimens. No Chelydra is known from Inglis 1C, and Holmesina floridanus osteoderms from the site are equivalent in dimensions to those from the early Irvingtonian Inglis 1A. Inglis 1C is best placed between the early Irvingtonian localities Inglis 1A and De Soto Shell Pit. Inglis 1A and Inglis 1C share Peromyscus sarmocophinus, Reithrodontomys wetmorei, and Ondatra idahoensis, taxa not identified from De Soto Shell Pit. Inglis 1C and De Soto Shell Pit both contain Atopomys texensis, which does not occur at Inglis 1A. Present at both Inglis 1A and De Soto Shell Pit, but not at Inglis 1C are the gracile Hemiauchenia, Capromeryx arizonensis, Mammut americanum, Trigonictis macrodon, and Chasmaporthetes ossifragus. Assignment of an early Irvingtonian age also serves as a compromise between the late Blancan, as suggested by the presence of Platygonus bicalcaratus, and the late early Irvingtonian, as suggested by Reithrodontomys humulis and Palaeolama mirifica. Integration of sea level data with biochronology has refined the dates of Florida’s Pliocene and Pleistocene deposits (Morgan and Hulbert, 1995; Emslie, 1998). The De Soto Shell Pit lies 10 m above modern sea-level, yet contains abundant marine vertebrates. This indicates deposition during an interglacial period with high sea level. Conversely, Inglis 1A lacks most aquatic forms and lies, in part, below sea level, suggesting the sinkhole was filled during a glacial period. Inglis 1C also formed during a glacial advance, but is clearly not contemporaneous with Inglis 1A. Morgan and Hulbert (1995) suggest Inglis 1A formed between 2.01 and 1.87 Ma and De Soto Shell Pit between 1.89 and 1.78 Ma. This creates an age bracket for Inglis 1C of 2.01 to 1.78 Ma, with a most likely age of around 1.9 Ma. Taphonomy Vertebrate remains can accumulate in sinkholes through a variety of pathways: from carcasses brought in by water, wind, or animals, or from live animals physically trapped. The relative paucity of large carnivorans and bats implies that the sinkhole did not serve as a den or roost. Water abrasion is nearly absent and evidence of predation is limited to a single tooth puncture on a deer phalanx and partial digestion of a few rodent teeth. The majority of fossils were likely transported only short distances. Larger animals may have fallen or climbed into the hole, only to starve. The fossils are typically in excellent condition, although associated elements do not occur. A few bones have a soft, chalky surface, caused by post-depositional dissolution. Paleoecology The large fauna from the slightly older Inglis 1A indicates a more arid and open ecosystem than is present today in the same area. Many constituents of Inglis 1A have distinct western U. S. affinities, such as the jackrabbit (Lepus sp.) and pronghorn (Capromeryx arizonensis). The slightly more recent Inglis 1C depicts a wetter, more densely forested system. The animals with western associations are mostly absent, with only Baiomys representing those forms in Inglis 1C. Large grazers, which overwhelmingly dominate Inglis 1A in abundance, are equally matched with browsers at the younger Inglis 1C locality. Taxonomically, Inglis 1C lacks the grazers Capromeryx, an undescribed gracile Hemiauchenia, and a bovid, while adding the


169

FIGURE 9. Selected mammalian biochronologic ranges. Abbreviations: Irv., Irvingtonian; Rancholab., Rancholabrean. Modified in part from Morgan and Hulbert (1995).

browsing Palaeolama mirifica. Additionally, the greater abundance of squirrels and insectivores, as well as the presence of Lutra canadensis, lend support to a habitat similar to that of modern central Florida. The ecologicaly interpretation is similar to that reached by study of the avian fossils, which indicates a comibination of palm, dense-scrub, and prairie habitats (Emslie, 1998). CONCLUSIONS The unique faunas of Florida are not easily compared to the numerous localities west of the Mississippi River. Fortunately, the productivity of these sites attracts continued scrutiny. While the prolific study of large mammals has greatly advanced our knowledge, the micromammals of the state have not enjoyed the same progress. Screen washing at the newly discovered early Irvingtonian Inglis 1C locality produced thousands of small mammal fossils, including the rodents reported here. This is only the second study examining an abundant small mammal community from the Irvingtonian of Florida. The other locality, Coleman 2A, differs in being much younger (0.4 Ma compared to 1.9 Ma). Other eastern United States small mammal localities rarely represent time periods earlier than the Rancholabrean. Study of the diverse mammal assemblage at Inglis 1C helps to fill the faunal picture for the southeastern United States. Six of the Inglis 1C rodents occur outside of their previously recognized temporal and geographic range. Reithrodontomys humulis is identified from Inglis 1C and the late early Irving-

tonian Haile 16A locality, which are the two earliest records of the species. Prior to this study, Reithrodontomys wetmorei was reported only from the early Blancan of Kansas. The substantial extension into the Irvingtonian of Florida is supported by hundreds of teeth from both Inglis 1C and the slightly older Inglis 1A. The Inglis 1C fauna records two range extensions. Reithrodontomys wetmorei and Baiomys sp. are reported as fossils east of the Mississippi River for the first time. The site marks another record of the rare arvicoline Atopomys, slightly predating previous records from Haile 16A and De Soto Shell Pit. Review of the morphology here suggests that complexity in the anteroconid complex of the m1 may vary along a geographic cline, instead of having taxonomic implications. A new species of mouse, Peromyscus sarmocophinus, is described from Inglis 1A and 1C. Mandibular dimensions and accessory cusp frequency separate the new form, and suggest that it may not have any living descendents. The latest Blancan/early Irvingtonian pocket gopher in Florida is shown to belong within the genus Orthogeomys. Further, the suggestion that it was the evolutionary precursor to Geomys pinetis is demonstrated to be doubtful. The change in pocket gopher taxa within Florida sites is likely due to ecological replacement. Pliocene squirrels from Florida are described in detail for the first time. In the case of both Sciurus and Glaucomys, the morphological similiarities to extant sciurids in Florida is readily discerned. It is inconclusive whether they are conspecific with recent squirrels. If not, the Inglis 1C fossil forms should be

170


considered possible immediate ancestors of Sciurus carolinensis and Glaucomys volans. ACKNOWLEDGMENTS This study comprised a portion of my master’s thesis and I would therefore like to thank my committee members, Bruce MacFadden, David Webb, Neil Opdyke, and Jon Martin, for their patience. This locality could not have been studied without the work of Steve and Suzan Hutchens, who brought attention to the site, and Steve Emslie who directed the excavations. I am greatly indebted to those who allowed access to specimens in their care (FLMNH vertebrate paleontology range—Bruce MacFadden, David Webb, Marc Frank, and Russell McCarty; mammalogy range—Charles Woods, Candace McCaffrey, and Laurie Wilkins; TMM Vertebrate Paleontology Laboratory— Ernie Lundelius and Laura Froelich). This manuscript was reviewed by the UT Paleoseminar. LITERATURE CITED Akersten, W. A. 1973. Upper incisor grooves in the Geomyinae. Journal of Mammalogy 54:349–355. Albright, L. B., III. 1999. Biostratigraphy and vertebrate paleontology of the San Timoteo Badlands, Southern California. University of California Publication, Geological Sciences 144:1–121. Bader, R. S. 1959. Dental patterns in Peromyscus of Florida. Journal of Mammalogy 40:600–602. Bailey, V. 1915. Revision of the pocket gophers of the genus Thomomys. North American Fauna 39:1–138. Black, C. C. 1963. A review of the North American Tertiary Sciuridae. Bulletin of the Museum of Comparative Zoology 130:109–248. Blair, W. F. 1950. Ecological factors in speciation of Peromyscus. Evolution 4:253–275. Bowen, W. W. 1968. Variation and evolution of Gulf Coast populations of beach mice (Peromyscus polionotus). Bulletin of the Florida State Museum, Biological Sciences 12:1–91. Carleton, M. D. 1989. Systematics and Evolution; pp. 7–142 in F. L. Kirkland, Jr. and J. N. Layne (eds.), Advances in the Study of Peromyscus (Rodentia). Texas Tech University Press, Lubbock, Texas. Carr, G. S. 1981. An early Pleistocene avifauna from Inglis, Florida. Ph.D. dissertation, University of Florida, Gainesville, Florida, 174 pp. Czaplewski, N. J. 1987. Sigmodont rodents (Mammalia: Muroidea: Sigmodontinae) from the Pliocene (Early Blancan) Verde Formation, Arizona. Journal of Vertebrate Paleontology 7:183–199. Emslie, S. D. 1996. A fossil scrub-jay supports a recent systematic decision. The Condor 98:675–680. 1998. Avian community, climate, and sea-level changes in the Plio-Pleistocene of the Florida Peninsula. Ornithological Monographs 50:1–113. Frazier, M. K. 1981. A revision of the fossil Erethizontidae of North America. Bulletin of the Florida State Museum, Biological Sciences 27:1–76. Gustafson, E. P. 1978. The vertebrate faunas of the Pliocene Ringold Formation, south-central Washington. Bulletin of the Museum of Natural History, University of Oregon 23:1–62. Gut, H. J., and C. E. Ray. 1963. The Pleistocene vertebrate fauna of Reddick, Florida. Quarterly Journal of the Florida Academy of Sciences 26:315–328. Harris, A. H. 1988. Late Pleistocene and Holocene Microtus (Pitymys) (Rodentia: Cricetidae) in New Mexico. Journal of Vertebrate Paleontology 8:307–313. Hibbard, C. W. 1941a. The Borchers fauna, a new Pleistocene interglacial fauna from Meade County, Kansas. Bulletin of the Geological Survey of Kansas 38:197–220. 1941b. Mammals of the Rexroad fauna from the Upper Pliocene of southwestern Kansas. Transactions of the Kansas Academy of Science 44:265–313. 1941c. New mammals from the Rexroad fauna, Upper Pliocene of Kansas. American Midland Naturalist 26:337–368. 1950. Mammals of the Rexroad Formation from Fox Canyon,

Kansas. Contributions from the Museum of Paleontology, University of Michigan 8:113–192. 1952. A contribution to the Rexroad fauna. Transactions of the Kansas Academy of Science 55:196–208. 1962. Two new rodents from the early Pleistocene of Idaho. Journal of Mammalogy 43:482–485. 1967. New rodents from the late Cenozoic of Kansas. Papers of the Michigan Academy of Science, Arts, and Letters 52:115– 131. Hollister, N. 1911. A systematic synopsis of the muskrats. North American Fauna 32:1–47. Holman, J. A. 1959. Birds and mammals from the Pleistocene of Williston, Florida. Bulletin of the Florida State Museum 5:1–25. Hooper, E. T. 1957. Dental patterns in mice of the genus Peromyscus. Miscellaneous Publications of the Museum of Zoology, University of Michigan 99:1–59. Huckaby, D. G. 1980. Species limits in the Peromyscus mexicanus group (Mammalia: Rodentia: Muroidea). Contributions in Science, Los Angeles County Museum of Natural History 326:1–24. Hulbert, R. C., Jr. 1992. A checklist of the fossil vertebrates of Florida. Papers in Florida Paleontology 2:1–35. 1997. A new late Pliocene porcupine (Rodentia, Erethizontidae) from Florida. Journal of Vertebrate Paleontology 17:623–626. , and G. S. Morgan. 1989. Stratigraphy, paleoecology, and vertebrate fauna of the Leisey Shell Pit Local Fauna, early Pleistocene (Irvingtonian) of southwestern Florida. Papers in Florida Paleontology Number 2:1–19. , , and S. D. Webb (eds.). 1995. Paleontology and geology of the Leisey Shell Pits, early Pleistocene of Florida. Bulletin of the Florida Museum of Natural History 37:1–660. Klein, J. G. 1971. The ferungulates of the Inglis IA local fauna, early Pleistocene of Florida. MS thesis, University of Florida, Gainesville, 115 pp. Kurteń, B., and E. Anderson. 1980. Pleistocene Mammals of North America. Columbia University Press, New York, 442 pp. Linzey, D. W., and R. L. Packard. 1977. Ochrotomys nuttalli. Mammalian Species 75:1–6. Lundelius, E. L., Jr., T. Downs, E. H. Lindsay, H. A. Semken, Jr., R. J. Zakrzewski, C. S. Churcher, C. R. Harington, G. E. Schultz, and S. D. Webb. 1987. The North American Quaternary sequence; pp. 211–235 in M. O. Woodburne (ed.), Cenozoic Mammals of North America: Geochronology and Biostratigraphy. University of California Press, Berkeley. Martin, R. A. 1967. A comparison of two mandibular dimensions in Peromyscus, with regard to identification of Pleistocene Peromyscus from Florida. Tulane Studies in Zoology 14:75–79. 1968. Further study of the Friesenhahn Cave Peromyscus. The Southwestern Naturalist 13:253–266. 1974a. Fossil mammals from the Coleman IIA fauna, Sumter County; pp. 35–99 in S. D. Webb (ed.), Pleistocene Mammals of Florida. University of Florida Presses, Gainesville. 1974b. Fossil vertebrates from the Haile XIVA fauna, Alachua County; pp. 100–113 in S. D. Webb (ed.), Pleistocene mammals of Florida. University of Florida Presses, Gainesville. 1979. Fossil history of the rodent genus Sigmodon. Evolutionary Monographs 2:1–36. 1987. Notes on the classification and evolution of some North American fossil Microtus (Mammalia; Rodentia). Journal of Vertebrate Paleontology 7:270–283. , and R. H. Prince. 1989. A new species of Early Pleistocene cotton rat from the Anza-Borrego Desert of Southern California. Bulletin of the Southern California Academy of Sciences 88:80– 87. McCarley, W. H. 1958. Ecology, behavior and population dynamics of Peromyscus nuttalli in eastern Texas. Texas Journal of Science 10: 147–171. McGrew, P. O. 1944. An early Pleistocene (Blancan) fauna from Nebraska. Field Museum of Natural History, Geological Series 9:33– 66. Merriam, C. H. 1895. Monographic revision of the pocket gophers family Geomyidae (exclusive of the species of Thomomys). North American Fauna 8:1–259. Meylan, P. A. 1982. The squamate reptiles of the Inglis 1A fauna (Irvingtonian: Citrus Co., Florida. Bulletin of the Florida State Museum, Biological Sciences 27(3):1–85.

RUEZ—IRVINGTONIAN RODENTS FROM INGLIS 1C, FLORIDA 1995. Pleistocene amphibians and reptiles from the Leisey Shell Pit, Hillsborough County, Florida; pp. 273–297 in R. C. Hulbert, Jr., G. S. Morgan, and S. D. Webb (eds.), Bulletin of the Florida Museum of Natural History 37. Morgan, G. S., and R. C. Hulbert, Jr. 1995. Overview of the geology and vertebrate biochronology of the Leisey Shell Pit Local Fauna, Hillsborough County, Florida; pp. 1–92 in R. C. Hulbert, Jr., G. S. Morgan, and S. D. Webb (eds.), Bulletin of the Florida Museum of Natural History 37. , and J. A. White. 1995. Small mammals (Insectivora, Lagomorpha, and Rodentia) from the early Pleistocene (early Irvingtonian) Leisey Shell Pit Local Fauna, Hillsborough County, Florida; pp. 397–461 in R. C. Hulbert, Jr., G. S. Morgan, and S. D. Webb (eds.), Bulletin of the Florida Museum of Natural History 37. Nelson, R. S., and H. S. Semken. 1970. Paleoecological and stratigraphic significance of the muskrat in Pleistocene deposits. Geological Society of America Bulletin 81:3733–3738. Nowak, R. M. 1991. Walker’s Mammals of the World, 5th ed. Johns Hopkins University Press, Baltimore, 1629 pp. Osgood, W. H. 1909. Revision of the mice of the American genus Peromyscus. North American Fauna 28:1–285. Patton, T. H. 1965. A new genus of fossil microtine from Texas Journal of Mammalogy 46:466–471. Paulson, G. R. 1961. The mammals of the Cudahy Fauna. Papers of the Michigan Academy of Science, Arts, and Letters 46:127–153. Pearson, P. G. 1953. A field study of Peromyscus populations in Gulf Hammock, Florida. Ecology 34:199–207. Reig, O. A. 1977. A proposed unified nomenclature for the enamelled components of the molar teeth of the Cricetidae (Rodentia). Journal of Zoology, London 181:227–241. Repenning, C. A. 1987. Biochronology of the microtine rodents of the United States; pp. 236–268 in M. O. Woodburne (ed.), Cenozoic Mammals of North America, Geochronology and Biostratigraphy. University of California Press, Berkeley. Ruez, D. R. 2000. Late Pleistocene (Rancholabrean) records of the harvest mouse (Reithrodontomys) in Florida. Florida Scientist 63:182– 190. Russell, R. J. 1968. Evolution and classification of the pocket gophers of the subfamily Geomyinae. University of Kansas Publications, Museum of Natural History 16:473–579. Schmidly, D. J. 1973. Geographic variation and taxonomy of Peromyscus boylii from Mexico and southern United States. Journal of Mammalogy 54:111–150.

171

Semken, H. A., Jr. 1966. Stratigraphy and paleontology of the McPherson Equus Beds (Sandahl local fauna) McPherson County, Kansas. Contributions from the Museum of Paleontology, University of Michigan 20:121–178. Steadman, D. W. 1980. A review of the osteology and paleontology of turkeys (Aves: Meleagridinae). Natural History Museum of Los Angeles County, Contributions to Science 330:131–207. Tamsitt, J. R. 1957. Peromyscus from the late Pleistocene of Texas. Texas Journal of Science 9:355–363. Taylor, A. J. 1982. The mammalian fauna from the mid-Irvingtonian Fyllan Cave Local Fauna, Travis County, Texas. MS thesis, University of Texas, Austin, 105 pp. Tomida, Y. 1987. Small Mammal Fossils and Correlation of Continental Deposits, Safford and Duncan Basins, Arizona, USA. National Science Museum, Tokyo, 141 pp. van der Meulen, A. J. 1978. Microtus and Pitymys (Arvicolidae) from Cumberland Cave Maryland, with a comparison of some New and Old World species. Annals of Carnegie Museum 47:101–145. Van Valen, L. 1962. Growth fields in the dentition of Peromyscus. Evolution 16:272–277. Wei, W. 1995. Revised age calibration points for the geomagnetic polarity time scale. Geophysical Research Letters 22:957–960. White, J. A. 1991. A new Sylvilagus (Mammalia: Lagomorpha) from the Blancan (Pliocene) and Irvingtonian (Pleistocene) of Florida. Journal of Vertebrate Paleontology 11:243–246. Wilkins, K. T. 1984. Evolutionary trends in Florida Pleistocene pocket gophers (genus Geomys), with description of a new species. Journal of Vertebrate Paleontology 3:166–181. Winkler, A. J., and F. Grady. 1990. The middle Pleistocene rodent Atopomys (Cricetidae: Arvicolinae) from the eastern and south-central United States. Journal of Vertebrate Paleontology 10:484–490. Wolfe, J. L., and J. N. Layne. 1968. Variation in dental structures of the Florida mouse, Peromyscus floridanus. American Museum Novitates 2351:1–7. , and A. V. Linzey. 1977. Peromyscus gossypinus. Mammalian Species 70:1–5. Zakrzewski, R. J. 1969. The rodents from the Hagerman local fauna, upper Pliocene of Idaho. Contributions from the Museum of Paleontology, University of Michigan 23:1–36. 1975. The late Pleistocene arvicoline rodent Atopomys. Annals of Carnegie Museum 45:255–261. Received 22 February 2000; accepted 15 August 2000.