Front Cover

Group Title: Bulletin of the Florida State Museum, biological sciences, vol. 27, no. 3
Title: The squamate reptiles of the Inglis IA fauna (Irvingtonian Citrus County, Florida)
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00000976/00001
 Material Information
Title: The squamate reptiles of the Inglis IA fauna (Irvingtonian Citrus County, Florida)
Series Title: Bulletin of the Florida State Museum
Physical Description: 85 p. : ill. ; 23 cm.
Language: English
Creator: Meylan, Peter A ( Peter Andre )
Publisher: University of Florida
Place of Publication: Gainesville
Publication Date: 1982
Subject: Squamata, Fossil   ( lcsh )
Paleontology -- Florida -- Citrus County   ( lcsh )
Paleontology -- Pleistocene   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Bibliography: p. 72-76.
General Note: Cover title.
General Note: Originally presented as the author's thesis (M.S.--University of Florida)
Statement of Responsibility: by Peter A. Meylan.
 Record Information
Bibliographic ID: UF00000976
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 000438847
oclc - 08797074
notis - ACJ9018
 Related Items
Other version: Alternate version (PALMM)
PALMM Version

Table of Contents
    Front Cover
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
Full Text
of the
FLORIDA STATE MUSEUM Biological Sciences
Peter A. Meylan

Numbers of the BULLETIN OF THE FLORIDA STATE MUSEUM, Biological Sciences, are published at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in any one calendar year.
Oliver L. Austin, Jr., Editor Rhoda J. Bryant, Managing Editor
Consultants for this issue:
J. Alan Holman Robert M. Sullivan S. David Webb
Communications concerning purchase or exchange of the publications and all manuscripts should be addressed to: Managing Editor, Bulletin; Florida State Museum; University of Florida; Gainesville, Fl. 32611, U.S.A.
Copyright by the Florida State Museum of the University of Florida
This public document was promulgated at an annual cost of $5000.00 or $5.00 per copy. It makes available to libraries, scholars, and all interested persons the results of researches in the natural sciences, emphasizing the circum-Caribbean region.
Publication date: 30 June 1982
Price: $5.00

Peter A. Meylan1
Abstract: The early Pleistocene Inglis IA site has yielded the largest and most diverse sample of fossil squamates in eastern North America. About 4000 specimens, including 250 skull elements, represent 26 species of snakes, 4 species of lizards, and 1 amphisbae-nian. The herpetofauna is essentially modern; 21 of 26 snake species, 2 of 4 lizard species, and the amphisbaenian survive in Florida today. Three Inglis IA snakes now extinct are Diadophis elinorae, Xenodontinae (cf. Dryinoides), andRegina intermedia (n. sp.). The extant species Opheodrys vernalis and Heterodon nasicus are present in the Inglis IA fauna and have apparently been replaced in Florida by Opheodrys aestivus and Heterodon sirnus. The lizards include Ophisaurusventralis, Sceloporusundulatus, Gerrhono-tus, and an extinct new species, Eumeces carri. The amphisbaenian is Rhineura cf. R. floridana.
The ecological requirements of modern counterparts of the fauna suggest that Inglis IA represents a community from a region of high pine with xeric hammock interspersed.. This open high pine country was part of a savanna that extended around the Gulf of Mexico during the late Cenozoic. The savanna apparently served as a corridor for mammals going to and from South America, but it did not act in the same capacity for the squamate fauna, which remained essentially autochthonous. The composition of the squamate fauna suggests that this corridor maintained a connection to the west that had previously existed farther north.
SUMARIO: El sitio denominado Inglis IA correspondiente al temprano Pleistoceno ha dejado la mas grande y mas diversificada muestra de fosiles de reptiles en el Este de Norte America. Cerca de 4000 muestras, incluyendo 250 elementos de craneo, represen-tan 26 especies de serpientes, 4 especies de lagartijas y 1 especie de anfisbaenido. La herpetofauna es esencialmente moderna: 21 de las 26 especies de serpientes, 2 de las 4 especies de lagartijas y el unico anfisbaenido son los sobrevivientes actualmente pre-sentes en Florida. Las tres especies de serpientes extintas pertenecientes al sitio Inglis IA son Diadophis elinorae, Xenodontinae (cf. Dryinoides), yRegina intermedia (n. sp.). Opheodrys vernalis y Heterodon nasicus son especies existentes que estan presentes en la fauna del Inglis IA pero que aparentemente, ,en Florida, han sido reemplazadas por Opheodrys aestivus y Heterodon simus. Las especies de lagartijas incluyen Ophisaurus ventralis, Sceloporus undulatus, Gerrhonotus sp., y una nueva especie extinta, Eumeces carri.
Los requerimientos ecologicos de la presente fauna suguiere que sus antepasados en el Inglis IA representaron una comunidad de region de pino alto intercalada con formaci-ones xerofiticas de "hammock." Esta area abierta con pinos altos constituyo parte de savana que se extendio alrededor del Golfo de Mexico durante el tardio Cenozoico. Esta savana aparentemente sirvio como corredor para el paso de mamiferos haciay desde Sur America, pero no actuo con la misma ef icacia para los reptiles, los cuales permanecieron esencialmente como autoctonos. La composicion de la fauna de reptiles sugiere que este corredor, que previamente habia existido mas hacia el norte, mantuvo conexion con el Oeste.
'The author is a Laboratory Technologist in Herpetblogy, Florida State Museum, and a graduate student in the Department of Zoology, both at the University of Florida, Gainesville 32611. This paper was originally submitted in partial fulfillment for the degree of Master of Science at the University of Florida.
MEYLAN, P. A. 1981. The Squamate Reptiles of the Inglis IA Fauna (Irvingtonian: Citrus County, Florida). Bull. Florida State Mus., Biol. Sci. 27 (3): 111-

INTRODUCTION ............................................................ 3
ACKNOWLEDGEMENTS.................................................... 5
METHODS AND ABBREVIATIONS......................................... 5
SYSTEMATIC PALEONTOLOGY............................................ 6
Reptilia: Squamata.......................................................... 6
Suborder Lacertilia......................................................... 6
Family Anguidae......................................................... 7
Family Iguanidae......................................................... 15
Family Scincidae......................................................... 18
Suborder Amphisbaenia..................................................... 19
Family Rhineuridae....................................................... 19
Suborder Serpentes.........................................................20
Family Colubridae........................................................20
Subfamily Colubrinae...................................................22
Subfamily Lampropeltinae..............................................29
Subfamily Natricinae...................................................39
Subfamily Xenodontinae ................................................45
Subfamily Incertae Sedis................................................55
Family Elapidae..........................................................55
Family Viperidae......................................................... 56
PALEOECOLOGY ...........................................................61
Diagenetic Factors..........................................................63
Collection Bias and Sample Size .............................................63
Faunal Requirements.......................................................64
Ecological Interpretation....................................................64
LITERATURE CITED .......................................................72
APPENDIX 1 LIZARD SKELETONS EXAMINED...........................77
APPENDIX 2 SNAKE SKELETONS EXAMINED............................78
ho, 3

Studies of fossil squamates in North America have focused largely on two general geographic regions, the High Plains and Florida. Work by Estes, Holman, and others has documented the rich fossil history for squamates in the High Plains states. Although Florida's fossil squamates received much attention during the 1950's and 1960's (Brattstrom 1953; Auffenberg 1955,1956,1963; Holman 1958,1959a, 1959b, 1962; Gut and Ray 1963; Estes 1963), most of the material excavated in the last 15 years has not been described. Unreported fossil squamates are present in at least 8 localities: 1 Oligocene, 6 Miocene, and 1 early Pleistocene (Inglis IA). Of these, Inglis IA (a sinkhole in Citrus County, Florida) has the largest squamate fauna. Biostratigraphic correlations place it in the early Pleistocene (earliest Irvingtonian) (Webb 1974).
When Auffenberg (1963) made his major survey of fossil snakes in Florida, no early Pleistocene sites were known. He recognized the potential value of an early Pleistocene site to the study of Florida's paleoherpetology, and his work showed that the modernization of Florida's snake fauna occurred between the middle Pliocene and middle Pleistocene. The Inglis fauna documents these final steps in the evolution of Florida's snakes and lizards.
Inglis IA is a rich fauna, with over 4000 elements that represent 31 squamate species. This is the largest sample of Squamata known from Florida. The fauna's large size makes it suitable for paleoecological interpretation and allows determination of the zoogeographical affinities of southeastern squamates in the earliest Pleistocene. Furthermore, the Inglis IA fauna contributes to our knowledge of the evolutionary history of 31 taxa.
Inglis IA was discovered by Jean Klein and Robert Martin in 1967, when excavations for the Cross Florida Barge Canal exposed it. After several small collections were made in intervening years, the site was collected in its entirety in 1974. The following description is based primarily on Klein's (1971) account.
The site was exposed on the north bank of the now defunct Cross Florida Barge Canal in Section 8, R.16 E., T.17 S., Citrus County, Florida. The deposit occurred in a sinkhole in the Inglis member of the Ocala Group (or more simply called Ocala limestones) (late Eocene). The fauna accumulated in locally derived clastic sediments, which occupied a volume about 10 m by 20 m and 3 to 4 m deep. In January 1974, approximately 300 cubic meters of fossiliferous sands were removed. The sands were washed through screens at the site and the recovered fossils taken to the Florida State Museum for study.
Klein (1971) recognized six stratigraphic units within the sinkhole.

The lowest, termed the basal conglomerate, included a large quantity of bone fragments, many of which were waterworn and polished. He attributed the wear to the action of waters of a spring boil (see Paleo-ecology). The next four units varied from 1 to 2 m thick. They filled the bulk of the sinkhole and contained the majority of the fossils. Klein said the sediments were deposited during periods of high and low water tables. During high water low energy deposition within the sinkhole produced the deposition of clays, and during a lowered water table sand rapidly accumulated. The entire sequence was capped by a cemented silica sandstone.
Problems in determining the age of Florida's sinkhole faunas are discussed by Webb (1974). The absence of physical stratigraphic control is the major drawback. Because fossils accumulate as isolated deposits in sinkholes and caves, deposits of differing ages can occupy adjacent sinkholes in the same formation. Thus, superpositional data are not available and correlations of their contained faunas must be based wholly on biostratigraphic methods.
In his review of the chronological framework for the Florida Pleistocene, Webb (1974) cited 11 important biostratigraphic indicator lineages, seven of which assist in determining the age of the Inglis IA fauna.
The record of the cotton rat (Sigmodon) is perhaps the most valuable. Sigmodon curtisi, the most common cricetine rodent in the fauna, is known from late Blancan or early Irvingtonian sites in the American West, including the Curtis Ranch Fauna of Arizona, the Kentuck Fauna of Kansas, and the Vallecito sequence in California (Martin 1979).
Smilodon gracilis, a sabercat present in the Inglis fauna, first appears in the late Blancan or early Irvingtonian of North America. By the late Irvingtonian it is replaced by Smilodon fatalis (Webb 1974).
Members of the armadillo lineage Kraglievichia to Chlamytherium increase in size from the Blancan through the Rancholabrean (Robertson 1976). The Inglis IA chlamythere is intermediate in size between late Blancan Kraglievichia and Irvingtonian and Rancholabrean Chlamytherium septentrionalis.
The presence of Hemiauchenia rather than Palaeolama and oiPla-tygonus bicalcaratus rather than Platygonus cumberlandensis in the Inglis fauna suggests a late Blancan or early Irvingtonian age (Webb 1974). The absence of Bison is further evidence that Inglis IA is pre-Rancholabrean.
The freshwater turtlePseudemys is the only reptile that helped date Inglis I A. This genus was examined in detail by Klein (1971), who

determined that the Inglis IA Pseudemys was a morphological inter-grade between the Blancan Pseudemys platymarginata and Irvingtonian Pseudemys scripta petrolei. Fossil squamates are too poorly known to be of any help in determining the age of Inglis IA.
Based on the lineages discussed above, Webb(1974) placed the Inglis IA fauna in the earliest Irvingtonian. In particular the fauna from Inglis IA closely resembles that from Curtis Ranch in Arizona, which has been dated at 1.9 million years old.
The motivation to study fossil squamates and the guidance to see the study through have been provided by Walter Auffenberg. David Webb and Ron Wolff have also advised me, especially in the preparation of this thesis. To them I extend my thanks. Bruce J. MacFadden, Robert Sullivan, J. Alan Holman, and Rhoda J. Bryant also provided comments that improved this contribution.
A number of people helped with specific aspects of this study and deserve recognition and thanks. Tom Van Devender, Richard Zweifel (AMNH), and George Zug (USNM) provided comparative material. Pat Srygley and Howard Kochman helped me with the use of packaged computer programs. Computing was done at the Northeast Regional Data Center of the State University System of Florida, located on the campus of the University of F lorida in Gainesville. Ron Wolff and Donna Born Drake provided equ ip-ment and advice for the preparation of the included photographs. Esta Belcher prepared Figures 1, 3, 6, 9, 13, 14, and 15. Angela O'Brien and Kelly Howard cheerfully typed the final draft. Special thanks go to all of the people who have helped to assemble the Florida State Museum Herpetological Skeleton Collection, without which this study would have taken much longer.
My wife, Anne, provided endless encouragement and instruction in good English. She diverted time from her own studies to type the first draft of this thesis.
Identification. Most of the material reported here was identified by direct comparison or univariate statistics (student's Mest). Diagnostic characters reported in the literature were used when possible. Additional distinguishing features were determined from a large sample of comparative osteological material of lizards (Appendix 1) and snakes (Appendix 2).
Multivariate statistical programs were found to be appropriate for identifying isolated vertebrae. Discriminate analysis was used to distinguish the vertebrae of snake species that show extreme similarity in vertebral form.
Ratios of pairs of linear measurements from single vertebrae were used as the discriminating variables in the analyses. The use of ratios in multivariate statistical analysis has been questioned by Atchley et al. (1976) and requires justification. Ratios are used to remove the effect of individual vertebra size, and their employment has become standard in studies of snake vertebrae (Johnson 1955; Auffenberg 1963). However, there are some statistical consequences of combining two variables to produce a third stemming largely from changes in coefficients of variation (CV) of the denominator variable (Atchley et al. 1976). In the present study such effects were minimized by using variables with approximately equal coefficients of variation.
Heyer (1978) justified the use of ratios in a study of leptodactylid frogs by comparing results from analyses using ratios and linear measurements. In the present study, test comparisons of analyses using ratios and linear measurements produced similar results for two analyses, and better results using ratios in a third (Table 1).

Table 1. Results of discriminant analysis using measurements and ratios as variables. (Poor discrimination is indicated by a higher percentage of reclassification.)
Analysis Percentage of Reclassification
With Ratios With Measurements
Elaphe (2 species) 10.6 9.6
"Racers" (3 species) 8.8 7.4
Vipers (4 species) 10.1 24.3
Two types of analyses were used to maximize the information obtained from the discriminant analysis. The Statistical Analysis System (Barr et al. 1976) was used to identify the fossils because it includes a simple test data statement that automatically classifies unknowns. It also produced less reclassification than a similar stepwise BMDP analysis (Dixon and Brown 1979). The BMDP analysis was used to provide information on the relative importance of each of the variables used in the analysis.
Terminology. Terminology of snake bones follows: vertebrae, Auffenberg (1963); skulls, Bullock and Tanner (1966); compound and basiparasphenoid, Estes et al. (1970); basiparasphenoid, Underwood (1967); and compound and palatine, Marx and Rabb (1972). Lizard bone terminology follows: vertebrae, Auffenberg (1963); and skull, Oel-rich (1956) and McDowell and Bogert (1954). Measurements of snake vertebrae follow the methods of Auffenberg (1963).
Determination of Minimum Number of Individuals (MNI). For lizards, MNI can be determined using the same technique used for mammals, i.e. the most common right, left, or midline element is counted. Estimation accuracy can be increased by considering the size range of elements. The same technique can be used for snakes and amphisbaenians if skull material is common. If skull material is rare, which i&typically the case, an estimate of MNI can still be made if the maximum difference in centrum length (or any other measure) from a single column of a given species is known. By subtracting the centrum length of the smallest known fossil of a species from the largest, a range in size for the fossil vertebrae is determined. The MNI is found by dividing the range in size of the fossils by the maximum range in the column of that same species(and rounding up to the next whole number). Some precaution must be taken, because large gaps in the range of centrum lengths of the fossils can produce spurious results. No such gaps are present in data gathered from Inglis material. This method probably greatly underestimates the number of individuals in large samples.
The following abbreviations are used in the text: AMNH, American Museum of Natural History; UF, University of Florida/Florida State Museum; USNM, National Museum of Natural History; CL, centrum length; CTW, cotyle width; NAW, neural arch width; NH, neural spine height; NLU, neural spine length at upper edge; NSB, neural spine length at base; POPR, postzygaphophyseal to prezygapophyseal length; PRPR, width across the prezygapophyses; ZW, zygosphene width; R, right; L, left: MNI, minimum number of individuals; N, sample size; OR, observed range; X, mean; SD, standard deviation.
Order Squamata Oppell 1811 Suborder Lacertilia Gunther 1867
Four species of lizards, representing three families, are present in

the Inglis IA fauna. Comparisons of the fossils were made only to the lizard families now found in the New World (Anelytropsidae, Angui-dae, Gekkonidae, Helodermatidae, Iguanidae, Scincidae, Teiidae, Xantusidae, and Xenosauridae). Statements concerning lizard morphology for which no citation is provided are based on examination of the comparative material listed in Appendix 1.
Family Anguidae Cope 1864 Subfamily Gerrhonotinae Cope 1900
cf. Gerrhonotus Weigmann 1828 Figure 1 A
Referred Material. UF 26409,1 vertebra.
Description. The single vertebra has a subtriangular centrum 3.35 mm long and 2.1 mm wide. The centrum is gently rounded ven-trally with no haemal keel or subcentral ridges. The neural spine is low and rises from the neural arch at an angle of 25. It projects posteriorly beyond the edge of the neural arch. Rib articulations are not strongly differentiated into parapophyses and diapophyses.
Comparisons. The absence of subcentral ridges indicates that the referred vertebra is not that of a teiid or iguanid. The centrum is longer and more triangular than that of any gekkonids examined. It is more rounded ventrally than in Ophisaurus or Heloderma. The absence of the haemal keel in the fossil suggests that it is not a scincid or diploglossine anguid. The simple rib articulations eliminate the possibility thatXenosaurus is represented. The fossil is very similar to Recent Gerrhonotus wmlticarinatus, but due to a lack of material is not assigned to species.
distribution. The genus Gerrhonotus is found throughout western North America from northern Mexico to southern British Columbia. This genus may be represented by material from upper Cretaceous, upper Paleocene, and lower Eocene of the Midwest (Meszoely 1970). Later records include the Mio-Pliocene of Nebraska (Meszoely 1970), lower Pliocene of Kansas (Wilson 1968), and upper Pliocene (Blancan) of Texas (Rogers 1976).
Remarks. Verification of Gerrhonotus in the early Pleistocene of Florida is required before any zoogeographical conclusions can be drawn. Its presence in Florida could be a result of the great radiation of anguids that apparently occurred in the Tertiary of North America.

I-1 '-1
Figure 1. Limbed lizards and an amphisbaenian from Inglis IA: Gerrhtmolus (A) vertebra, ventral view, XT. iii,i<,c>tru cf. R. tloridana (B) vertebra, dorsal view, X10. Sr.eloporux umhdatus (C) dcntary, lingual view, X4; (D) maxilla, lingual view, X5: (F) frontal, dorsal view, X5;(F) vertebra, ventral view, X10.?w;etwK(n.sp.)Jeiitary, (I101,OTVI'R),X10;(G) lingual and (II) labial views; (I)vertebra, ventral view. X 10; (J) maxilla. (PARATYPE). linj-ual vi.iv, X10 lead, scale 2 mm).

Genus Ophisaurus Daudin 1803 Ophisaurus ventralis (linnaeus) 1766
Figure 2
Referred Material. UF 26410, 10 R and 11 l maxillae; UF 26411,12 R and 11l dentaries; UF 26412,3 R and 1l pterygoids; UF 26413, 3 R and 3 l compounds; UF 26414,1 R and 1 l frontals; UF 26415, 2 parietals; UF 26416, 3 occiputs; UF 26417, 1 l exoccipital; UF 26418,2 basisphenoids; UF 26419,1 quadrate; UF 26420,1 ilium; and 926 vertebrae including UF 26421, 24 cervicals; UF 26422, 526 presacrals; UF 26423,22 sacrals; UF 26424,48 nonautotomic caudals; and UF 26425, 306 autotomic caudals.
Description. The maxillae have a broad facial flange that contributes to the posterior border of the nares. Premaxillary processes are bifurcate; posterior processes are single. There are 13-17 teeth (X = 15.0 + 1.04) on 12 complete maxillae. The teeth of both the maxillae and dentaries are closely spaced and unicuspid; they have well developed striations on the lingual surface. Thirteen complete dentaries possess 16-21 teeth (X = 17.7 1.44); all have a surangular notch that extends anteriorly well beyond the posterior end of the tooth row. A single patch of 30-37 teeth is present on the pterygoids.
T he compounds consist of an angular and supra-angular anteriorly, and a splenial on the lingual surface. An articular is fused to the supra-angular. There is a posteroventrally directed retroarticular process. A reduced mandibular foramen is located well anterior to the articular condyle.
The frontals are paired and long, narrow, and pointed anteriorly. The parietals are longer than wide and lack anterolateral processes. The supratemporal processes are robust and nearly meet at their bases.
The occiputs are elongate with the prootic bones extending anteriorly. Crista prootica are poorly developed. The disassociated exoccipital is identical to those of the intact occiputs.
The two fossil basisphenoids have short, robust basipterygoid processes which extend anterolateral^. The processes originate from the ventral surface of the basisphenoid. The quadrate is triangular in lateral view, and taller than it is thick. The ilium is reduced, consisting of a short, curved shaft with a circular distal expansion.
The thoracic vertebrae are short and wide and have flat centra. The neural spines are broad and low. The sacral vertebrae have flat centra and irregular distal extremities on the transverse processes. The haemal arches of the caudal vertebrae are fused without a trace of suture. Autotomy septa are present in all but the largest (most proximal) caudal vertebrae.

Figure 2. o^m'hum^h.^ rmirnii-.-: from Inglis IA: (A) dentary, lingual view, X3; (B) maxilla, lingual view, X3; (C) pterygoid, ventral view, X3; (D) compound, lingual view, XSMlil parit-tal. durssl view, X3;(F) frontal, dorsal view, X3; (G) occiput, dorsal view, X:i; (Hi hasisphenoid, fro n to-dorsal view, X4; (I) quadrate, lateral view, X4; (J) ilium, lateral view, X4 (each scale 5 mm).

Comparisons. The fossil dentaries and maxillae can be distinguished from North American iguanids (except Phrynosoma), helo-dermatids, tends, Anniella, and xantusids (exceptKlauberina) by the closely spaced unicuspid teeth. They differ iromKlauberina and gek-konids in having well developed striations on the lingual surface of the teeth. These striations and a more symmetrical tooth shape distinguish the fossil dentaries and maxillae from Gerrhonotus. The dentaries differ from Phrynosoma and all scincids in having the surangular notch extended anteriorly beneath the tooth row. Bifurcate premaxil-lary processes distinguish the fossil maxillae from Phrynosoma. The fossil maxillae differ from those of scincids in having a single posterior process.
The fossil dentaries and maxillae have significantly fewer teeth than modern samples of Ophisaurus attenuatus and 0. compressus (t-te&t, P< .01) (Fig. 3). They have more teeth than modern 0. ventra-lis, but the difference is not significant.
The pterygoids of 0. ventralis, including the fossils reported here, are apparently unique among North American lizards in having a single patch of 15 or more teeth. 0. attenuatus has 10-20 teeth arranged
Figure 3. Number of teeth in the dentaries and maxillae of three Recent North American Ophisaurus and the Inglis IA fossils. Mean, 95% confidence limits, and range are indicated.

in one or two rows. Gerrhonotus and 0. compressus have 10 or fewer pterygoid teeth.
The fossil compounds differ from those of all North American lizards except Ophisaurus in having a small mandibular foramen located well anterior of the articular condyle. The mandibular foramen is reduced in Gerrhonotus but it is also more posteriorly located. There are no apparent differences in the compounds of the modern Ophisaurus species examined.
Among extant North American lizards, only the Scincidae, Xantu-sidae, Helodermatidae, some diploglossine anguids, and Ophisaurus have divided frontals. The divided fossil frontals are as long and narrow as those of Eumeces and Ophisaurus. As in these two genera, the frontals are more than three times longer than the width at their widest point. The frontals of Eumeces are not pointed anteriorly as they are in the fossils and in Ophisaurus.
The frontal of Ophisaurus attentuatus makes up a larger part of the dorsal edge of the orbit than in 0. compressus or 0. ventralis (Wilson 1968). In the Inglis fossil, the proximity of the prefrontal and postfron-tal sutures indicates that the frontals made up only a small part of their respective orbits, so the material could represent either 0. ventralis or 0. compressus.
The referred parietals are longer and more narrow than those of all North American lizards examined, except Ophisaurus. The parietals could not be identified to species, but are referred to 0. ventralis on the basis of associated material.
A characteristic feature of the occiput of all anguimorph lizards is the extended nature of the prootic bone, which extends anterior to the semicircular canal (McDowell and Bogert 1954). This condition, as well as the near absence of the crista prootica, distinguishes the fossil occiputs from those of all extant North American lizards, except Ophisaurus and Anniella. Anniella differs from the fossils in having reduced spheno-occipital tubercles on the basioccipital. There is no consistent interspecific variation in the occiputs of modern North American Ophisaurus. The fossils are referred to Ophisaurus ventralis on the basis of associated material.
The fossil basisphenoids were only compared to material of species thought to occur in the Inglis fauna. They differ from Eumeces, in which the lateral wall of the abducens canal is continuous with the anterior edge of the basipterygoid process. They differ from Scelopo-rus, which has relatively longer, thinner basipterygoid processes. In Gerrhonotus these processes are directed anteriorly rather than anterolateral^.
The quadrates of most North American lizards are D-shaped in

lateral view. Only in scincids, helodermatids, and Ophisaurus are they triangular, as they are in the Inglis material. The quadrates of helodermatids are unlike the fossils in having a distinctive T-shape from above. Those of Eumeces are thicker than those of Ophisaurus. The quadrates of Ophisaurus are not diagnostic at the species level.
The reduced size of the ilium of Ophisaurus is distinctive. The fossil differs from 0. compressus, which has a triangular rather than round distal expansion. The expansion is round in 0. attenuatus and 0. ventralis.
The body vertebrae of Ophisaurus differ from those of all other North American lizards in being short and wide and having a flat centrum and a broad, low neural spine (Etheridge 1961). The absence of articulating surfaces on the transverse processes of the sacral vertebrae indicates that these vertebrae belong to the limbless genus Ophisaurus. Only Anniella is similar, but it has lower neural spines. The caudal vertebrae differ from all limbed forms in having haemal arches fused to the centrum without a trace of suture. Again, Anniella is similar, but it has a much lower neural spine (Etheridge 1961).
In the present study the body vertebrae of modern Ophisaurus species were found to show more variation than Etheridge (1961) reported. He separated the three well known modern species in North America by differences in the length to width ratio of the centrum (CL/NAW) and/or by differences in the angle of the posterior border of the neural spine. Two Mexican species are known from a total of three specimens, none of which is prepared as a skeleton. 0. compressus has distinctly narrower vertebrae than either 0. ventralis or 0. attenuatus,
1.0 I.I 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 4. Vertebrae shape (centrum length/neural arch width) for three Recent North American Ophisaurus and Inglis IA fossils. Mean, standard deviation, two standard errors of the mean, and range are indicated.

but there is little difference between the latter two (Fig. 4). A similar figure for separating 0. compressus and 0. ventralis in Auffenberg (1955) includes a misleading error the scale of 0.1 to 1.0 should read 1.1 to 2.0. Etheridge (1961) used differences in the angle of the posterior edge of the neural spine to separate 0. attenuatus (45-65) from O. ventralis (65-84). In the present study, this angle was found to be 65-85 in 0. attenuatus and 70-90 in O. ventralis. The difference between Etheridge's data and those reported here may be explained in part by the geographical origin of the samples. Etheridge's O. attenuatus were from Kansas and Texas and represented only O. attenuatus attenuatus (Etheridge, pers. comm.). The sample used in the present study included both O. a. attenuatus from Texas and Kansas and O. a. longicaudus from Florida.
Although no clear difference exists between the vertebrae of O. attenuatus, O. ventralis, and the fossils, Figure 4 shows that the mean and the 95% confidence limits for the fossil body vertebrae are more similar to O. ventralis than to O. attenuatus.
The sacral and caudal vertebrae aiOphisaurus are not diagnostic at the species level except in one case. The caudals of O. compressus are unique in lacking fracture planes. The presence of these planes in the fossil caudals indicates that O. compressus is not represented.
Not all of the material referred to Ophisaurus ventralis is diagnostic. However, statistical variation observed in the fossils (see Table 2) is
Table 2. Diagnostic osteological characters for United States Ophisaurus species
and Inglis IA fossils.
Fossil O. attenuatus O. compressus O. ventralis
Maxillary Teeth (mean one S. D.) 15.0 1.04 16.8 1.14 21.0 1.58 14.4 0.91
Dentary Teeth (mean + one S. D.) 17.7 + 1.44 19.3 + 1.87 24.9 + 1.46 16.7 1.10
Pterygoid Teeth 30-37 (patch) 10-20 (in 1 or 2 rows) 10 (in 1 row) 10-25 (patch)
CL/NAW (body vertebrae) (mean + one S. D.) 1.33 + .123 1.51 .151 1.74 .097 1.37 .121
Frontals in Orbit Small Large Small Small
Distal Expansion of Ilium Round Round Triangular Round
Fracture plane in Caudal Vertebrae Present Present Absent Present

not greater than that in any one of the modern species examined, suggesting that only one species is present in the Inglis fauna. Analysis of seven characters in combination (Table 2) indicates thatO. ventralis is the species present in the Inglis fauna.
Distribution. Ophisaurus ventralis is found in the southeastern coastal plain and Piedmont from Louisiana to North Carolina. It is known as a fossil from the late Miocene of Nebraska (Holman 1975), the early Pliocene of Kansas (Wilson 1968), middle Pliocene of Florida (Auffenberg 1955), late Irvingtonian of Florida (Holman 1959a,b), and Rancholabrean of Florida (Auffenberg 1955; Holman 1958) and Missouri (Holman 1965). Ophisaurus attenuatus has been reported from late Plio-Pleistocene localities in Kansas and Oklahoma (Etheridge 1961). Holman (1970) described Ophisaurus canadensis from the upper Miocene of Saskatchewan and suggested that this species might be ancestral to 0. ventralis and 0. attenuatus. The Buda Local Fauna (Miocene: Arikareean) and Love Bone Bed (Miocene: Clarendonian) of Florida contain unstudied Ophisaurus material.
REMARKS. The Ophisaurus material from Inglis is the most extensive reported for the genus in the New World. It provides previously unavailable morphological data on Ophisaurus ventralis in the early Pleistocene. Minor differences between the Inglis O. ventralis and a modern sample include higher average numbers of dentary and maxillary teeth, and substantially higher numbers of pterygoid teeth in the fossils. No changes were detected in the cranium or dermal roofing bones. An important consistency is seen in the condition of the ilium. It is evident that 0. ventralis had reached its present limbless condition by the earliest Pleistocene and has not subsequently reduced the pelvic girdle. Speciation in North American Ophisaurus has been attributed to isolation of various stocks during the Pleistocene (McCon-key 1954). The fossil record described above shows that this is not the case for 0. ventralis and 0. attenuatus, which are both present before the Pleistocene.
Family Iguanidae Gray, 1827 Genus Sceloporus Weigmann 1828 Sceloporus undulatus (Latrielle) 1802
Figure 1 C-F
Referred Material. UF 26426,3 frontals; UF 26427,33 R and 36 L dentaries; UF 26428,11R and 10 L maxillae; UF 26432,2 humeri; UF 26433,1 femur; and 53 vertebrae including: UF 26429, 42 cervi-cals and thoracics; UF 26430, 2 sacrals; and UF 26431, 9 caudals.

Description. The frontals are single, unsculptured, narrow between the orbits, and have poorly developed crista cranii.
The tooth-bearing elements have tricuspate teeth posteriorly. The cusps are not strongly developed. The dentaries are long and slender with Meckle's groove open lingually. They have an average of 3.57 teeth per mm. Maxillae have an average of 3.14 teeth per mm.
The vertebrae have trapezoidal centra with moderately developed haemal keels and weakly developed subcentral ridges. Both sacral vertebrae are second members of disarticulated sacral pairs. They have posterolaterally directed tubercles on the posterior margins of the transverse processes.
Comparisons. Among extant North American lizards, only igua-nids, teiids, gekkonids, xenosaurids, and Gerrhonotus have single frontals (Camp 1923). Unlike the fossils, crista cranii are well developed in the frontals of both gekkonids and xenosaurids. The fossil frontals are relatively shorter than those of the teiid generaAmeiva and Cnemido-phorus. They are narrower between the orbits than the frontals of the iguanid Anolis carolinensis, but are very similar to Sceloporus woodi andS. undulatus. The dentigerous elements possess typically iguanid tricuspate teeth. The three cusps are less developed than those of Leiocephalus, and the central one is not as pronounced as in Anolis carolinensis. The teeth and the long slender shape of the dentary suggest a small Sceloporus. The referred vertebrae have centra similar in shape to most iguanid lizards. Subcentral ridges are less well
UNDULATUS 1-m"mm^mmmllm '
n =8
n = 2
WOODI ^^^^^
3.0 3.2 3.4 3.6 3.8 4.0 4.2
Figure 5. Tooth density in the dentaries and maxillae of RecentSceto^orws undulatus, Recent Sceloporus woodi, and fossil Sceloporus from Inglis IA. Mean, 95% confidence limits, and range are indicated.

developed than in the teiids Ameiva and Cnemidophorus, but they and the haemal keels are better developed than in any scincids, gekkonids, or anguids examined. The vertebrae of many small iguanids are similar, and the reference of these fossils to Sceloporus is based on the associated skull elements. All of the elements referred to Sceloporus undulatus closely resemble S. undulatus and S. woodi. These two species can be distinguished on the basis of tooth density (Fig. 5). Sceloporus woodi has a significantly larger number of teeth on the dentary and maxilla than does S. undulatus or the fossils (t test, P < 0.05). The fossils do not differ in tooth density from S. undulatus. Therefore the frontals and postcranial material, along with the den-tigerous elements, are assigned to S. undulatus.
Distribution. Sceloporus undulatus is found throughout the southern two-thirds of the United States from extreme southeastern Nevada to New Jersey and Florida. It is known as a fossil from the Blancan of Texas (Rogers 1976), Irvingtonian of Maryland (Holman 1977a) and Texas (Holman 1969a), and Rancholabrean of Arizona (Van Devender et al. 1977), Florida (Brattstrom 1953), Georgia (Holman 1967), Missouri (Holman 1974), and Texas (Holman 1968b, 1969a; Gehlbach and Holman 1974). The oldest known sceloporine is from the upper Miocene of Saskatchewan (Holman 1970).
Remarks. In their discussion of the evolution of the genus Sceloporus, Larsen and Tanner (1975) reported S. undulatus to be a member of the virgatus group, the most complex and recently derived of three species groups. They suggested that the initial division of this group occurred when its progenitor was isolated in various refugia during the first glacial advance, and that with the first interglacial virgatus group stock migrated east and west to produce occidentalis, undulatus, and woodi. The appearance of S. undulatus by the Pleistocene indicates that some division of thevirgatus group had occurred by the end of the Pliocene.
Jackson (1973) found Sceloporus woodi, a Florida endemic, to be more similar to southwestern members of the virgatus group than to Sceloporus undulatus undulatus. He cited the possible late arrival (Sangamonian = late Irvingtonian) of S. undulatus as additional evidence that it was a southwestern form which was isolated in Florida during the Pleistocene and gave rise toS. woodi. This new evidence of S. undulatus in Florida by the earliest Pleistocene renews the possibility that it is ancestral to Sceloporus woodi.

Family Scincidae Gray 1825 Genus Eumeces Weigmann 1834 Eumeces carri new species
Figure 1 G-J
Diagnosis. Eumeces carri is similar to several small Eumeces species (e.g. E. inexpectatus, E. fasciatus). However, it differs from these and all Recent species examined in features of the teeth and the dentary. The dorsal margin of the dentary in all available Eumeces (14 species) is straight. The anterior dilation of the fossil dentary (Fig. 1G) could be a result of wear, but the presence of a labial ridge, which apparently occurs in no other skink, suggests that the anterior end of the dentary is modified. The teeth are also modified, being rounded or even blunt and unstriated rather than striated and weakly cusped as in most skinks. The high degree of lingual tooth erosion seen in the fossils could be a result of depositional wear, but the identical pattern of wear on both toothed elements suggests a predepositional feature. Such erosion may indicate very rapid tooth replacement.
Holotype. UF 26435 (Fig. 1 G, H), a left dentary from Inglis IA, Irvingtonian of Citrus County, Florida.
Paratype. UF 26432, (Fig. 1J) a right maxilla.
Referred Material. UF 26434,11 thoracic vertebrae; 2 pairs of fused sacral vertebrae (tentative).
Description. The dentary is nearly complete and contains space for 17 teeth. The 10 preserved teeth are blunt, unstriated, and eroded lingually to about two-thirds their height. Meckle's groove is open along the entire length of the bone. In labial view the dentary narrows and then expands anteriorly. A well developed ridge occupies the middle of the labial surface at the same level as this expansion.
The maxilla is nearly complete. The most anterior portion of the tooth row and the facial wing are missing. The maxilla has 15 or 16 tooth positions. The teeth are eroded lingually to about two-thirds of their height.
The referred vertebrae are small (< 2 mm CL) and about twice as long as wide. Centra have nearly parallel sides and are rounded ven-trally. Haemal keels are broad and weakly developed. Subcentral ridges are absent. Neural spines rise from the neural arch at an angle of 15-20. The sacral vertebrae are about as wide as long and have long transverse processes. E ach pair is fused at the centrum and at the ends of the transverse processes.
Comparisons. The thoracic vertebrae compare best to those of the scincid genus Eumeces. The haemal keels are weaker than in iguanids, but stronger than in Gerrhonotus or Ophisaurus. The fossil vertebrae

differ from those of the teiids, Cnemidophorus and Ameiva, which have well developed subcentral ridges. They differ from gekkonid vertebrae, which tend to be relatively shorter and have neural spines that rise more vertically from the neural arch. Sacral vertebrae are fused only in scincids, xantusids, gekkonids, and some teiids (not fused in Cnemidophorus). Because no other material was found that could be referred to the Teiidae, Gekkonidae, or Xantusidae, the sacral vertebrae are tentatively referred to this skink.
The lingual erosion of the teeth in the maxilla and dentary suggests non-anguimorph, vertical tooth replacement. Among lizards with vertical tooth replacement, iguanids and teiids have some multicuspid teeth; Meckle's groove is closed in gekkonids and xantusids. The only remaining non-anguimorph New World families are the Scincidae and the Anelytropsidae. It is unlikely that the latter, a small mono-typic form known only from central Mexico, is represented.
Remarks. The Inglis material represents a new and specialized species of skink. The study of its relationships will require additional comparative material. At least two individuals of this skink are represented in the fauna.
Suborder Amphisbaenia Gray 1825 Family Rhineuridae Vanzolini 1951
Genus Rhineura Cope 1861 cf. Rhineura floridana (Baird) 1858
Figure 1 B
Referred Material. UF 26438, 32 thoracic and 1 caudal vertebra.
Description. The thoracic vertebrae (Fig. IB) are dorsoventrally compressed and as long as or slightly longer than wide. The large pre-and postzygapophyses are adjacent to the neural arch and connected by an interzygapophyseal ridge. Zygosphenes and zygantra are absent. Neural spines project posteriorly when enlarged, but typically are reduced to a ridge. On either side of the neural spine is a series of raised striations. The vertebral centra are rounded below and have no haemal keel or subcentral ridges. The paradiapophysis is single. The single caudal vertebra is as above, but, in addition, has anterolateral^ directed transverse processes and ventrally fused hemapophyses.
Comparisons. The vertebrae of rhineurids do not differ markedly between species. They are apparently unique among amphisbaenians in bearing numerous longitudinal striations on the neural arch (Ber-man 1976).

Distribution. Rhineura floridana is at present restricted to the northern half of peninsular Florida. Rhineurid vertebrae from the late Irvingtonian (Holman 1962, 1959a) and Rancholabrean (Holman 1958) of Florida are referred to R. floridana. The Inglis material extends the fossil record for rhineurids in Florida to the beginning of the Pleistocene, but unstudied material in the Buda Local Fauna (Arikareean) indicates an early Miocene arrival of rhineurids in Florida.
remarks. The identification of fossil rhineurids in the literature has been based completely on skull characters. The absence of skull material from Inglis makes a positive identification of this rhineurid impossible, but because these fossils occur within the present range of Rhineura floridana and are identical to recent examples, the Inglis material is tentatively referred to this species.
Suborder Serpentes Linneaus 1758
Living members of the suborder Serpentes are divided into three infraorders: the Scolecophidia, the Henophidia, and the Caenophidia (Underwood 1967). Each of these has a diagnostic vertebral form (Holman 1979). The Scolecophidia are primitive burrowing snakes. Their vertebrae are depressed and lack neural spines. The cotyle is oval in shape and the haemal keel and subcentral ridges are absent or poorly developed. The henophidians include boas and pythons. The vertebrae of this group have neural spines that tend to be short and thick, rarely as long as the neural arch. Paradiapophyses tend to be undifferentiated, and accessory processes are poorly developed or absent. The advanced snakes, or Caenophidia, have vertebrae with thin neural spines that are typically longer than those of henophidians; they are often nearly as long as the neural arch. Paradiapophyses are divided into parapophyses and diapophyses. Accessory processes tend to be well developed.
On the basis of these characters, all of the snake vertebrae from Inglis IA are referable to the Caenophidia. Smith et al. (1977) applied the name Colubroidea to this group and recognized four families within it. Members of three of these families, Colubridae, Elapidae, and Viperidae, constitute the snakes of the Inglis fauna.
Family Colubridae Oppel 1811
This study was facilitated by dividing American colubrid snake vertebrae into five artificial groups. Four of these morphological groups approximate natural phylogenetic assemblages of species recognized by various authors. These are the Natricinae, Xenodontinae, Lampropeltinae, and Colubrinae.

The Natricinae are easily distinguished by the presence of laterally compressed hypapophyses throughout the column (Underwood 1967). A second group has dorsoventrally flattened (depressed) neural arches. It approximates the Xenodontinae (Ophiinae of Dunn 1928) and includes Heterodon, which is clearly a xenodontine, andFarancia, which was placed in this subfamily by Dunn (1928) and Neill (1964). Underwood (1967) placed the genus in the Lycodontinae. The extent of vaulting or depression of the neural canal is difficult to determine in very small snakes. They are assembled in a third, clearly unnatural group which includes Carphophis, Diadophis, Rhadinaea, Tantilla, Stilosoma, and Sonora. The remaining larger colubrid vertebrae are of two kinds (Fig. 6, Table 3). One group is short and wide, with poorly developed epizygapophyseal spines and short laterally directed accessory processes. The other group is longer and more narrow, with epizygapophyseal spines that are moderately to well developed and long anterolateral^/ directed accessory processes. The former group includes Pituophis, Lampropeltis, Elaphe, Arizona, Rhinoceilus, and Cemophora. This assemblage approaches Dowling and Duellman's (1974) Lampropeltiinae, a tribe of their subfamily Colubrinae. Smith et al. (1977) raised this tribe to subfamilial status on the basis that knowledge of snake relationships is too uncertain to recognize tribes. The group with long vertebrae remains in the Colubrinae and includes Coluber, Drymarchon, Masticophis, and Opheodrys.
Table 3. Vertebral characters for the separation of the Colubrinae and Lampro-peltinae (as recognized in this study).
Lampropeltinae Colubrinae
(8 species, 29 (4 species, 10
Vertebral columns, 223 columns, 79
characters vertebrae) vertebrae)
CL/NAW less than 1.27 greater than 1.37
(except in (except in Drymarchon)
E pizygapophy seal weak to absent moderately to well
Spines (96.9%) developed (83.5%)
Accessory shorter than longer than
Process prezygapophyseal prezygapophyseal
Length width (85.4%) width (64.6%)
Accessory laterally anterolateral^'
Process directed (65.1%) directed (87.4%)

n = IS
0.8 1.0 1.2 1.4 1.6 1.8
Figure 6. Centrum length/neural arch width (CL/NAW) for lampropeltine and colubrine snakes. Mean, 95% confidence limits, and range are indicated.
The present study does not address the validity of these natural groups. The subfamilial names Colubrinae, Lampropeltinae, Natrici-nae, and Xenodontinae are used only to assist in the organization and identification of the fossils.
Subfamily Colubrinae Oppel 1811
For the purpose of this paper, the Colubrinae is considered to include the racer-like snakes Coluber, Masticophis, Drymarchon, Opheodrys, and Salvadora. The vertebral centra of these snakes are nearly half again as long as they are wide, except in Drymarchon. The

accessory processes are long and straight. Epizygapophyseal spines are better developed in this group than in other subfamilies (less so in Opheodrys and Salvadora).
Genus Coluber Linnaeus 1758 Coluber constrictor Linnaeus 1758
Figure 7 A-D, Table 4
Referred Material. UF 26360,334 vertebrae; UF 26361,1 L pterygoid; UF 26362,2 R compounds; UF 26363,1L and 2 R dentaries.
Description. The referred vertebrae have long narrow centra that are nearly square across the zygapophyses (Table 4). Accessory processes are long, narrow and anterolaterally directed. Haemal keels are variable, but are typically narrow and straight with a slight lateral expansion just anterior to the condyle. Epizygapophyseal spines are well developed.
The fossil left pterygoid is moderately large. It is broken at the posterior end of the toothrow. The pterygoid flange has a poorly developed, laterally directed ectopterygoid process. The flange is slightly constricted between the ectopterygoid process and the base of the quadrate process. A ridge is present on the dorsal surface; it extends anteriorly beyond the ectopterygoid process.
In the fossil compound bones the labial and lingual flanges are subequal in height, the lingual flange being slightly taller than the labial flange.
All three fossil dentaries are broken posterior to the 12th tooth. Meckle's groove closes completely by the 8th or 9th tooth. The anterior end of these dentaries is very slightly curved.
Comparisons. The referred vertebrae are similar to those of Masticophis and, to a lesser extent, Drymarchon. In recent individuals of D. corais from Florida the anterior edge of the neural spine is bevelled. In Masticophis, Coluber, and some Mexican and Central American Drymarchon, the anterior edge of the neural spine is square. Thus the bevel, used as a standard diagnostic character for Drymar-; ehon, does not always hold. This problem, and the fact that neural spines from many of the fossils were broken, encouraged me to include Drymarchon in a discriminant analysis with Coluber andMasticophis.
The analysis developed is independent of neural spine characters. It is based on 26 vertebrae from four Coluber constrictor, 16 vertebrae from two Drymarchon corais and 26 vertebrae from four Masticophis /lagellum. Of the 68 vertebrae, 6 (8.8%) are reclassified by the analysis. The most important characters for the separation of these groups ^ZW/NAW, PRPR/POPR, PRPR/NAW, PRPR/ZW) are shown in table 4.

Figure 7. Fossil < :.hiher and Moulin,,,/,^ from Inglis IA: Coluber constrictor (A) pterygoid, dorsal view, X5;(B) pterygoid, ventral view, X5; (C) compound, labial view, X2; (D.I dentary, lingual v irw, X3. !\ I u>:l imiih i^jh.iiii! I a if I I'j I max il la, occlusal view, X3; (F) compound, labial view. \2 leadi i.calc = 5 mm).

Table 4. The four most important ratios for the discrimination of Coluber constrictor, Drymarchon corais, and Masticophis flagellum (mean one standard deviation).
Coluber constrictor
Recent (N = 27) .98 .04 1.04 .04 1.79 + .10 1.83 .11
Fossils (N = 334) .96 + .98 .99 + .05 1.74 .11 1.90 .15
Masticophis flagellum
Recent (N = 26) .98 + .05 .98 + .05 1.81 .05 1.84 .11
Fossils (N = 331) .95 .08 .97 .05 1.80 + .11 1.89 .13
Drymarchon corais
Recent (N = 14) .81 + .02 1.12 + .06 1.65 + .07 2.03 .06
Fossils (N = 29) .84 .03 1.03 + .06 1.72 + .06 2.04 .06
'Zygasphene Width/Neural Arch Width
'^Prezygapophysis to Prezygapophysis Width/Prezygapophysis to Postzygapophysis Length 3Prezygapophysis to Prezygapophysis Width/Neural Arch Width 'Prezygapophysis to Prezygapophysis Width/Zygasphene Width
The fossil pterygoid is broken posteriorly. The presence of an ectopterygoid process (Fig. 7) indicates that it is not that of a viperid (see Figs. 5, 29, 30; Brattstrom 1964). The process in the fossil is less developed than that of the larger North American natricines and Farancia (App. 3-4). It differs from Heterodon in being laterally, rather than anteriorly, directed (Fig. 12B). From the large lampropel-tines, the fossil differs in having the dorsal ridge extended anteriorly beyond the ectopterygoid process. It is most similar to Coluber, Masticophis and Drymarchon, differing from the latter two only in having the ectopterygoid flange rise from the tooth row at a high angle and in having the ectopterygoid process more anteriorly located.
Among larger North American snakes very few genera have sub-equal flanges on the compound bone (Marx and Rabb 1972). The lingual and labial flanges are subequal in the referred fossils, as they are in Masticophis and Coluber. In both Masticophis and Coluber the lingual flange is slightly the larger, and in both the extension of these two flanges increases with age. However, the lingual flange of C. constrictor is always more dorsally expanded than in individuals of M. flagellum of the same size. Based on this criterion, the fossils are referred to C. constrictor rather than Masticophis.
The fossil dentaries are most similar to Coluber and Masticophis, which have Meckle's groove closed completely between the 8th and 11th teeth (App. 3-6). They are assigned to the former because they do not show the more distinct curve at the anterior end of the dentary of Masticophis.
distribution. Coluber constrictor is found throughout the United States, except in the desert southwest. Its extensive fossil record must

be regarded with caution because of the difficulty involved in distinguishing it from Masticophis flagellum. The oldest referred material is from the early Pliocene of Nebraska. Blancan fossils are known from Idaho (Holman 1968a), Kansas (Brattstrom 1967), and Texas (Rogers 1976). Irvingtonian specimens are reported from Arizona (Brattstrom 1955a; Lindsay and Tessman 1974), Florida (Holman 1959a), Kansas (Brattstrom 1967), and Maryland (Holman 1977a). Rancholabrean records occur in at least 10 states from coast to coast (see Holman 1981).
Extinct, but possibly related torms include Par aoxybelis floridanus from the early Miocene of Florida (Auffenberg 1963), Paracoluber storeri from the late Miocene of Wyoming and Saskatchewan (Holman 1970), and Coluber plioagellus from the late Pliocene of Kansas (Wilson 1968).
Remarks. It is apparent from the fossil record that Coluber constrictor is widely distributed by the Irvingtonian. The referred skull material from Inglis shows that the skull of this species has not changed over the last two million years.
Genus Drymarchon FiTZlNGER 1843 Drymarchon corais (Boie) 1827
Referred Material. UF 26367,29 vertebrae.
Description. The fossil vertebrae differ from those of Coluber constrictor only in being relatively wider. Neural spines, if present, are squared off anteriorly.
Comparisons. Long accessory processes and well developed epizygapophyseal spines identify these as colubrine vertebrae. The vertebrae were identified as D. corais by the discriminant analysis described under Coluber constrictor. They differ from modern D. corais from Florida, in which the anterior edge of the neural spine is bevelled. Specimens from other parts of the range often have unbe-velled neural spines (UF 11467,11782,11784).
Distribution. At present!), corais is found from southern Texas to northern Argentina, with a disjunct subspecies in Florida, Alabama, and Georgia. It is known as a fossil from 10 middle or late Pleistocene and 6 late Pleistocene localities in Florida (see Holman 1981). Material from the Pleistocene of Texas that may represent this species has been lost (Holman 1969a).
Remarks. Florida's Recent D. corais is clearly a relict of a once more widely distributed species. The Inglis fossils referred toD. corais show greater similarity to Recent populations from Mexico and Central America than to those in Florida. Apparently differentiation of Florida's D. corais populations has occurred since the early Pleisto-

cene. This suggests that Drymarchon became isolated in Florida after Inglis time.
Genus Masticophis Baird and Girard 1853 Masticophis flagellum (Shaw) 1802
Figure 7 E,F, Table 4
Referred Material. UF 26364,331 vertebrae; UF 26265,2 R, 1 L maxillae; UF 26366,1 R, 2 L compounds.
Description. In most respects the Masticophis vertebrae fit the description of Coluber constrictor above, but they are slightly narrower across the zygapophyses and the centrum (Table 4).
The complete left maxilla is from an adult snake; the two right maxillae are from subadults and are missing their anterior halves. The complete maxilla has space for 15 teeth with no diastema. The prefrontal process is located at the 5th and 6th teeth; the ectopterygoid process is located at the 12th tooth. There are three tooth sockets posterior to the latter process. A dorsal constriction of the maxillary ramus is present at the level of the ectopterygoid process. The constriction of the maxillary ramus is more pronounced in the smaller specimens than in the complete adult maxilla.
The three compound bones are complete. Labial and lingual flanges are subequal in height; the lingual flange is slightly higher than the labial flange.
Comparisons. The fossil vertebrae were identified using discriminant analysis (see Coluber constrictor). The most important characters for distinguishing M. flagellum from Coluber constrictor and Drymarchon corais appear in Table 4. Lack of adequate comparative material prevented the inclusion ofMasticophis taeniatus in this analysis. However, the largest referred vertebrae are much larger than those of Recent M. taeniatus.
The number and placement of teeth and the position of the two maxillary processes distinguish the referred maxillae from most other colubrids (App. 3-1). In the lampropeltine genera these processes are closer to one another and are located more posteriorly than in the fossils. North American natricines have more teeth; the minimum observed is 19 in Seminatrix. Heterodon has a diastema in the tooth row. The maxillae of Farancia are more robust and have the two processes more widely separated. Forms similar to the fossils include Drymarchon, which has more teeth (ca 22) in adults and C. constrictor, which differs only in lacking the constriction of the maxillary ramus at the level of the ectopterygoid process. BothM. flagellum andM. taeniatus have the constriction described in the fossils. Four maxillae of the latter differ from the fossils in having more (17-18) teeth.

The three fossil compound bones are referable to Masticophis based on the characters discussed under Coluber constrictor.
Distribution. M. flagellum is found throughout the southern half of the United States and in northern Mexico. The fossil record of M. flagellum must be regarded with caution because of the difficulty involved in distinguishing this species from C. constrictor and other species of Masticophis. Vertebrae of this species have been reported from the Blancan of Arizona (Lindsay and Tessman 1974) and Texas (Rogers 1976), the Irvingtonian of Florida (Holman 1959a; Auffenberg 1963) and Arkansas (Dowling 1958), and the Rancholabrean of Arizona, Arkansas, California, Florida, New Mexico, Nevada, and Virginia (see Holman 1981).
Remarks. The Inglis material is the earliest reported from Florida and includes the oldest known cranial material. It substantiates the appearance of Masticophis flagellum by the earliest Pleistocene.
Genus Opheodrys Fitzinger 1843 Opheodrys vernalis (Harlan) 1827
Referred Material. UF 26368,29 vertebrae.
Description. These vertebrae have centra which are longer than wide (CL/NAW X = 1.54 + .145). They are nearly square across the zygapophyseal faces (POPR/PRPR X = 1.03 .061). The cotyle is oval and epizygapophyseal spines are present. The accessory processes are short and often laterally directed. The haemal keel and subcentral ridges are poorly developed.
Figure 8. Diagnostic ratios for the vertebrae of North American Opheodrys. Mean, 95% confidence limits, and range are indicated. POPR/PRPR = postzygapophyseal to prezygapophyseal length/width across the prezygapophyses, CL/NAW centrum length/neural arch width.

Comparisons. The lack of development of the accessory processes and subcentral features serves to distinguish Opheodrys from other North American colubrids with long narrow vertebrae.
The two extant species of Opheodrys can be distinguished by differences in their centrum and neural arch shapes (CL/NAW and POPR/PRPR) (Auffenberg 1963). The fossils differ significantly from 0. aestivus in both these ratios (Fig. 8). They do not differ from 0. vernalis and thus are assigned to this species.
Distribution. 0. vernalis is at present restricted to the northeastern portion of the United States from northeastern Kansas to Maine. Isolated populations occur at medium to high elevations in several western states, North Carolina, the eastern coastal plain of Texas, and northern Mexico. O. aestivus occupies a nearly complimentary range from Florida to Texas, north to the Ohio River Valley and southern New Jersey. 0. aestivus, and not 0. vernalis, is present in the vicinity of Inglis IA today. 0. vernalis is known as a fossil only from the Irvingtonian of Maryland (Holman 1977a). 0. aestivus is known from Rancholabrean localities in Florida (Auffenberg 1963) and Texas (Holman 1969a).
Remarks. The Inglis Opheodrys vernalis, together with the large number of Recent isolated populations, is evidence of a once greater range for this species. The Inglis material indicates that in the early Pleistocene this species occurred as far south as Florida. The complimentary range of modern Opheodrys aestivus suggests that it has replaced 0. vernalis over much of its former range. In light of the present fossil record, this would have had to occur in the middle or late Irvingtonian.
Subfamily Lampropeltinae Dowling 1975
The subfamilial name Lampropeltinae is the best available for the North American constricting colubrids. In the present study, Arizona, Elaphe, and Pituophis have been added, on the basis of vertebral morphology, to the genera considered by Dowling and Duellman (1974) to be members of this group. Possibly the similarity in vertebral form of these snakes is due to convergence.
Genus Cemophora Cope 1860 Cemophora coccinea (blumenbach) 1788
Appendix 3-5, Table 5
Referred Material. UF 26269,28 vertebrae.
Description. The vertebrae are moderately small (CL = 1.75-2.35 mm) and slightly longer than wide. The cotyles are as large or larger than the neural canal. Epizygapophyseal spines are absent.

Neural spines are almost as high as long, and are not undercut anteriorly. The haemal keels are narrow and not expanded posteriorly. The accessory processes are as long or longer than the prezygapophyseal width, and are directed laterally.
Comparisons. The vertebrae are from a moderately small 1am-propeltine snake. The relative size of the neural canal indicates that the fossils are not those of juveniles of a larger species. The fossils differ from Rhinoceilus, which has neural spines undercut anteriorly (App. 3-5). Arizona is similar, but has wider vertebrae (CL/NAW, Table 4). The vertebrae of Cemophora coccinea can be separated from those of Lampropeltis triangulum only with great difficulty. The best character for separation is the ratio of the height to the length of the neural spine (NLU/NH). The neural spine of modern L. triangulum is not as high as that of C. coccinea (Table 5). Vertebrae thought to represent one or the other of these forms are referred to C. coccinea if NLU/NH is < 2.6 and toL. triangulum if NLU/NH is> 3.0. Vertebrae falling in the region of overlap (2.6 < NLU/NH > 3.0) were not referred to species unless they were very narrow (CL/NAW > 1.40), as in modern C. coccinea.
Distribution. Recent populations of C. coccinea occur throughout the southeastern United States from southern New Jersey and central Missouri to Louisiana and Florida. An isolated population occurs in southern coastal Texas. This species has been reported from a single Rancholabrean locality in Florida (Auffenberg 1963). A form with similar vertebral morphology, Pseudocemophora antiqua, is known from the Hemingsfordian of Florida (Auffenberg 1963) and Wyoming (Holman 1976). It is similar to both Cemophora and Lampropeltis, but its relationship to these two genera remains unclear.
Table 5. Diagnostic ratios for the vertebrae of Cemophora coccinea, Lampropeltis triangulum, and Arizona elegans.
Cemophora coccinea
Recent 21 1.8-3.0 2.44 .42 23 1.05-1.55 1.33 .13
fossil 27 1.4-2.8 2.00 .33 27 1.05-1.55 1.18 .13
Lampropeltis triangulum
Recent 26 2.6-5.0 3.46 .77 19 1.10-1.35 1.24 .07
fossil 10 3.0-4.0 3.34 .46 10 1.0-1.29 1.19 .08
Arizona elegans
Recent 16 .87-1.08 .96 .11
'Neural Spine Length/Neural Spine Height 2Centrum Length/Neural Arch Width

Remarks.Williams and Wilson (1967) suggested that Cemophora ; is a specialized derivative of some member of theL. triangulum group,
though!/, triangulum has the lowest neural spine of any Lampropeltis. This apparent specialization for burrowing would have had to be reversed in order to arrive at the higher neural spines of C. coccinea. ; Thus it is more likely that Cemophora diverged from some earlier ; form of Lampropeltis. The Inglis fossils constitute the oldest record of Cemophora coccinea.
The observed range in centrum lengths of the referred vertebrae
indicates that at least two individuals of C. coccinea were preserved in
the Inglis IA fauna.
Referred Material. UF 26370, 71 vertebrae; UF 26371, 1 L pterygoid; UF 26372,1 L compound.
Description. The referred vertebrae lack hypapophyses. The centra are as wide as long (Table 6). Haemal keels are moderately to well developed and have posterior expansions. Subcentral ridges are moderately developed and straight. Accessory processes are not longer than the prezygapophyseal width. Neural spines are as tall as long. Epizygapophyseal spines are very weak or absent. Zygosphenes are usually flat or convex from above.
Only the anterior third of the large left pterygoid is preserved. The anterior margin of the ectopterygoid flange rises from the toothbear-ing ramus at an angle of 25. The ectopterygoid process is not well developed. The tooth row is straight. No ridge is evident on the dorsal surface of this fragment.
The referred compound is large and nearly complete. The lingual flange is highly arched. Its dorsal edge is twice as high as the dorsal edge of the labial flange.
Comparisons. The pterygoid fragment has a pterygoid flange and thus cannot represent a viperid; it is too large to be Micrurus. Unlike the pterygoids of the colubrine group and Lampropeltis getu-. lus, there is no ridge on the dorsal surface at the level of the ectopterygoid process. The weak development of this process indicates that the fossil is not a xenodontine or natricine. The extension of the pterygoid flange is more anterior in the fossil than in Elaphe obsoleta, less anterior than inPituophis melanoleucus and identical XoE. guttata (see App. 3-3).
The compound differs from colubrines in having only one flange developed. The lingual flange is unlike E. obsoleta, Pituophis melano-
Genus Elaphe Fitzinger 1833 Elaphe guttata (LlNNEAUS) 1766
Figure 9 A-C, Table 6

leucus, and Lampropeltis calligaster in being highly arched. This condition is present in Farancia abacura, Lampropeltis getulus, Elaphe guttata, and viperids. The two former have taller labial flanges than the fossil. Viperids have a less symmetrically arched labial flange (see Figs. 31-32, Brattstrom 1964).
The referred vertebrae are of the lampropeltine type. Lampropeltis getulus differs from the fossils in having better developed subcentral ridges (strong in 83% of L. getulus and 47% of E. guttata) that curve inward near the condyle (Brattstrom 1955a; Auffenberg 1963). Pituo-phis is unlike the fossils in having higher neural spines and zygo-sphenes that are usually concave from above (Auffenberg 1963). The fossils have better developed subcentral ridges thanArizona. They are larger than the vertebrae oiLampropeltis triangulum and Cemophora coccinea (1.6-2.45 mm CL inL. triangulum, 1.75-2.35 mm CL in C. coccinea, and 3.2-6.6 mm in the fossils). The fossil vertebrae are similar to several species of Elaphe.
The vertebrae of various Elaphe species have been separated on the basis of the height of the neural spine (Auffenberg 1963; Holman 1968a). E. vulpina has a lower neural spine than E. guttata or E. obsoleta. It is not present in the Inglis fauna. Discriminant analysis was used to assign the vertebrae to E. obsoleta or E. guttata. The analysis is based on 28 vertebrae from four individuals of Elaphe guttata and 38 vertebrae from five individuals of Elaphe obsoleta. Of the 66 original observations, 7 (10.6%) were reclassified. Values for the most important measurements and ratios are found in Table 6.
Distribution. Elaphe guttata is found throughout the southern half of the United States from New Mexico, Colorado, and northeastern Mexico to New Jersey and Florida. Fossils of this species are known from the Blancan of Texas (Rogers 1976) and Kansas (Brattstrom 1967), the Irvingtonian of Kansas (Brattstrom 1967), and the
Table 6. The four most important ratios for the separation oiElaphe obsoleta and Elaphe guttata by discriminant analysis.
Elaphe guttata
Recent (N = 32) 1.10 .11 1.74 + .11 .81 + .06 1.35 .14
fossils (N = 71) 1.08 .09 1.49 .19 .81 + .06 1.34 + .13
Elaphe obsoleta
Recent (N = 33) 1.01 + .06 1.71 .15 .81 + .07 1.25 + .09
fossils (N = 98) .98 + .06 1.41 .18 .82 + .06 1.19 + .09
'Centrum Length/Neural Arch Width Centrum Length/Cotyle Width 'Zygasphene Width/Neural Arch Width 'Centrum Length/Zygasphene Width

Rancholabrean of Kansas, Missouri, Texas, and Florida (see Holman 1981).
Four extinct species of Elaphe have been described from the Miocene and Pliocene of North America (Holman 1979). There is no evidence that any of these are closely related to Elaphe guttata.
Remarks. The Inglis IA fossils establish the presence of E. guttata in Florida by the earliest Pleistocene. The observed range in centrum lengths of the fossils shows that at least three individuals are present.
Elaphe obsoleta (Say) 1823 Figure 9 D-F, Table 6
Referred Material. UF 26373, 98 vertebrae; UF 26374, 1 L maxilla; UF 26375,1 R palatine; UF 26376,1 parasphenoid.
Description. The vertebrae fit the description of Elaphe guttata. The only qualitative difference is that the neural spines tend to be higher than long.
The maxilla is nearly complete (Fig. 9 D). There is a diagonal break at the 13th tooth; probably a 14th tooth is lost. There is no diastema. The maxilla is large, but not robust. The shaft is curved medially anterior to the 4th tooth, posterior to that point it is straight. The prefrontal process is at the level of the 7th and 8th teeth. The ectopterygoid process is at the level of the 12th and 13th teeth. Both processes are subrectangular.
The palatine is moderately large. It is worn and broken posterior to the medial process. The posterior edge of the lateral process and the anterior edge of the medial process are opposite. There are four tooth positions anterior to the lateral process. The maxillary nerve foramen passes through the lateral process. The posterior opening of this foramen lies below the dorsal surface of the lateral process.
The base of the parasphenoid is missing posterior to the middle of the pituitary fossa. The cultriform process is stout, widens anteriorly, and extends beyond the suborbital flanges, which are well developed. There is a frontal step at the midpoint of the cultriform process. The suborbital flange has a smooth margin.
Comparisons. The referred vertebrae have been identified by the discriminant analysis discussed under Elaphe guttata.
The fossil maxilla is that of a large colubrid (App. 3-1). It has fewer teeth than any natricine. The absence of a diastema indicates that it is not Heterodon (Fig. 12 A). Either one or both of the maxillary processes are more posteriorly located in the fossil than in Farancia, Coluber, Drymarchon, or Masticophis. Most of the maxillary ramus of the fossil is straight, unlike that in Lampropeltis getulus or Pituophis melanoleucus, which is gently curved. The maxilla of E. guttata andi?.

obsoleta are similar to the fossil. That of the former differs slightly from both the fossil andE. obsoleta in having a posteriorly projecting prefrontal process.
The palatine differs from that of most large North American snakes in its basic form (App. 3-2). The two processes are absent in all viperids (Brattstrom 1964, Fig. 10). InHeterodon the base of this lateral process extends to the anterior end of the bone. The medial process of all large I Nerodia species has a shorter base than the fossil. Drymarchon, Coluber, and Masticophis have medial and lateral processes that arise i-, from the same section of the toothed ramus. In the fossil the base of the lateral process lies completely anterior to the base of the medial process. The fossil differs from the remaining large lampropeltines, except E. obsoleta andE. vulpina, in the placement of the maxillary nerve foramen in the lateral process. In the fossil the posterior opening of the foramen is below the dorsal surface of this process. In Elaphe guttata, Pituophis melanoleucus, Lampropeltis getulus, and Lampro- peltis calligaster, this opening is in the dorsal surface of the lateral process. The final assignment of the fossil toE. obsoleta rather ihanE. vulpina is based on the presence of vertebrae of the former.
The fossil parasphenoid differs from those of viperids in being narrower and in lacking the ventral keel found in that group (see Figs. 25-26, Brattstrom 1964). Unlike the fossil, larger natricines have no frontal steps on the cultriform process. The fossil differs from Heterodon, in which orbital flanges are absent or restricted to the basal part of the bone. The cultriform process is narrower than in Farancia, but :. wider than mDrymarchon, Coluber, andMasticophis. The latter three I also have suborbital flanges reaching the anterior end of the cultri-i=' form process. The suborbital flanges in Lampropeltis getulus and \ Elaphe guttata are poorly developed, which results in a narrower I cultriform process than in the fossil. The fossil is very similar to I Pituophis melanoleucus and E. obsoleta, but it lacks the small notches present in the suborbital flanges on either side of the parasphenoid in f Pituophis, and thus is assigned to E. obsoleta (see App. 3-4). t Distribution. Elaphe obsoleta is found today throughout the I eastern half of the United States. It is known as a fossil from the I Blancan and Irvingtonian of Kansas (Brattstrom 1967) and Blancan of I Texas (Rogers 1976). It is also present in seven middle to late Pleisto-cene sites in Florida (see Holman 1981). Holman (1973) provided a tentative phylogeny for Elaphe in which he designated E. buisi Hol-: man from the middle Pliocene of Oklahoma and E. kansensis (Gil-i, more) from the lower Pliocene of Kansas as ancestors to E. obsoleta. \ Remarks. The Inglis material documents the presence of this species in Florida at the beginning of the Irvingtonian. E. obsoleta is

thus widely distributed by the early Pleistocene. At least two individuals are represented in the fauna.
Genus Lampropeltis Fitzinger 1843 Lampropeltis getulus (Linneaus) 1766
Referred Material. UF 26377, 58 vertebrae.
Description. The referred vertebrae are short and wide with moderate to low neural spines. The neural arches are slightly depressed. The subcentral ridges are very well developed (in most they are so strong that the ventral surface of the centrum appears to be excavated between the subcentral ridges and haemal keel). Subcentral ridges curve inward near the cotyle. Haemal keels are usually widened posteriorly.
Comparisons. The extreme development of the subcentral ridges of Lampropeltis getulus distinguishes it from most other large 1am-propeltines (Auffenberg 1963). Lampropeltis ealligaster is similar, but has haemal keels that remain uniformly narrow to the condyle (100% vs. 23% inL. getulus). It also has less well developed subcentral ridges (88% moderate to weak vs. 83% strong inL. getulus). Farancia also has similar vertebrae, but has more depressed neural arches and wider, blunter accessory processes (Auffenberg 1963).
Distribution. L. getulus ranges from coast to coast across the southern United States and northern Mexico. This species is known as a fossil from Blancan localities in Kansas (Brattstrom 1967) and Texas (Rogers 1976), Irvingtonian localities in Kansas (Brattstrom 1967), Texas (Holman 1969b), and Florida (Holman 1959a), and Rancholabrean localities in Arizona, California, Florida, Georgia, Kansas, New Mexico, Nevada, Tennessee, and Texas (see Holman 1981). Two extinct species of Lampropeltis have been described, but they are more closely related to the tricolored king snakes, includingL. triangulum (Holman 1970), and are discussed under that heading.
Remarks. Blanchard(1921) and Blaney (1977) suggested that the Recent subspecies, Lampropeltis getulus splendida, from Texas and Mexico best approximates the ancestor of now widely dispersed and highly variable Lampropeltis getulus. Blaney (1977) considered the peninsular Florida form, L. g. floridana, to be a direct and closely related derivative of L. g. splendida. Thus a splendida-like form was probably once widespread. The Inglis record now indicates that this had occurred by the earliest Pleistocene.
At least two individuals of L. getulus, one large and one of medium size, were preserved in the Inglis fauna.

Lampropeltis triangulum (Lacepede) 1788 Table 4
Referred Material. UF 26378,10 vertebrae.
Description. These vertebrae are slightly longer than wide, moderately small, and have cotyles as large or larger than the neural canal. Neural spines are low. Accessory processes are moderate to long and typically anterolateral^ directed. Haemal keels are usually narrow and not flared posteriorly. Subcentral ridges are not well developed.
Comparisons. The vertebrae represent a moderately small 1am-propeltine snake. The relative size of the neural canal indicates that they are not juveniles of large species. The only moderately small lampropeltine with a neural spine as low as that of the fossils is L. triangulum (see Comparisons under Cemophora coccinea).
Distributions. L. triangulum is presently distributed throughout the eastern three-quarters of the United States and ranges south to northern South America. The species is known from the Blancan of Kansas and Oklahoma (Brattstrom 1967) and Texas (Rogers 1976); the Irvingtonian of Kansas (Brattstrom 1967) and Maryland (Holman 1977a); and the Rancholabrean of Arkansas, Florida, Georgia, Kansas, Missouri, Pennsylvania, Tennessee, Texas, and Virginia (see Holman 1981). Two related forms are the extinct species Lampropeltis similis Holman, known from the late Miocene and early Pliocene of the Midwest (Holman 1979), and Lampropeltis intermedins, known from the late Pliocene and earliest Pleistocene of Mexico and Arizona (Brattstrom 1955a, b).
Remarks. Brattstrom (1955a) considered Lampropeltis intermedins to be ancestral to L. triangulum and the mountain kingsnakes Lampropeltis pyromelana and Lampropeltis zonata. Holman (1964) suggested that Lampropeltis similis is ancentral toL. intermedins. It is noted thatL. intermedins is limited to the southwest, the suggested center of origin for L. triangulum (B\anchardl921). Blanchard(1921) hypothesized a pre-Pleistocene radiation of L. triangulum, from the southwest into the southeast. The Inglis material supports this hypothesis. Blanchard also suggested a postglacial migration of the northern forms of L. triangulum, but the fossil record indicates that this dispersal was a pre-Pleistocene or, at least an early Pleistocene event.
All the referred vertebrae could have come from a single vertebral column.

Genus Pituophis Holbrook 1842 Pituophis melanoleucus (daudin) 1803
Referred Material. UF 26379,129 vertebrae.
Description. These vertebrae have high neural spines which do not overhang anteriorly and overhang only slightly posteriorly. The zygosphene is convex from anterior, and flat to concave from above. Subcentral ridges are moderate to weak.
Comparisons. Large lampropeltine snakes other than Elaphe can be distinguished from adult Pituophis by their lower neural spines alone (Auffenberg 1963). Elaphe also has relatively lower neural spines but some overlap occurs (see Tables 15 and 16; Auffenberg 1963). The fossils are referred to Pituophis melanoleucus on the presence of very high neural spines, zygosphenes that are always convex from anterior (concave or flat in 41% of Elaphe), and poorly developed subcentral ridges (weak in 78% of P. melanoleucus, strong to moderate in 67% of Elaphe).
Distribution. P. melanoleucus ranges from coast to coast, from New Jersey to Florida and British Columbia to northern Mexico, with a significant hiatus in the Mississippi Valley.
Fossils are known from the Blancan of Texas (Brattstrom 1967) and Kansas (Rogers 1976), the Irvingtonian of Florida (Holman 1959a) and Kansas (Brattstrom 1967), and the Rancholabrean of Arizona, Arkansas, California, Florida, Illinois, Kansas, Missouri, Nevada, New Mexico, Oklahoma, and Texas (see Holman 1981).
Remarks. The Inglis material indicates thatP. melanoleucus was already widespread by the beginning of the Irvingtonian. The observed range in centrum length indicates that at least three individuals are present in the Inglis fauna.
Genus Stilosoma Brown 1890 Stilosoma extenuatum Brown 1890
Referred Material. UF 26380, 5 vertebrae.
Description. The referred vertebrae are small. The oval cotyles are about equal in size to the neural canal. The centra are nearly as wide as long. Accessory processes are pointed and laterally directed. Haemal keels are well developed but do not widen posteriorly.
Comparisons. The size of these vertebrae and the relative size of the neural canals suggest a small but mature snake. All small snakes extant in the eastern United States have anterolateral^ directed accessory processes except Rhadinaea flavilata and Stilosoma extenuatum, in which they are laterally directed. R. flavilata differs from the fossils in having blunt accessory processes and haemal keels that

broaden posteriorly. The Pliocene form, Stilosoma vetustum Auffenberg, differs from the extant species and from the Inglis fossils in having a narrower, more ridge-like haemal keel.
Distribution. Stilosoma is endemic to northern peninsular Florida. Records of this genus as a fossil fall within its present range. S. extenuatum may be represented by vertebrae from two middle to late Pleistocene sites near Gainesville, Florida (Auffenberg 1963; Martin 1974). S. vetustum is known only from the middle Pliocene, Haile VI locality (Auffenberg 1963).
Remarks. The Inglis material is the first clearly referable to Stilosoma extenuatum. The material could represent a single individual.
Subfamily Natricinae Bonaparte 1838
Various caenophidians, including all natricines, viperids, and elap-ids, have hypapophyses present throughout the vertebral column. Ia the natricines they are wide in lateral view and somewhat laterally compressed. The hypapophyses are fin-like, rather than cylindrical as in viperids and elapids. Because all caenophidians have hypapophyses present in the anterior portion of the column, natricine vertebrae are easily confused with the anterior precaudals of many species. Although no simple character will distinguish them, anterior precaudals of non-natricine colubrids can often be identified by their wide neural canals, reduced accessory processes, posteriorly sloping neural spines, or distinctly non-natricine hypapophyseal shapes.
Genus Nerodia Baird and Girard 1853 Nerodia cyclopion (Dumeril, Bibron and Dumeril) 1854
Referred Material. UF 26381, 5 vertebrae.
Description. These are moderately large vertebrae. The centra are short and wide. Neural spines are high and overhang anteriorly and posteriorly to the same extent, giving the neural spine a nearly symmetrical outline in lateral view (see Fig. 42 in Auffenberg 1963). Hypapophyses are short and broad.
Comparisons. The vertebrae are assigned toNerodia on the basis of their size, short, broad hypapophyses, and short, wide centra(Bratt-strom 1967). Holman (1962,1968a, 1970,1971) and others separated Nerodia into three artificial groups based on the length-height relationship of the neural spine. Nerodia sipedon, N. fasciata, and the extinct forms N. hibbardi, N. hillmani, and Neonatrix elongata have neural spines that are longer than high. Nerodia erythrogaster has neural spines that are as long as high. Nerodia cyclopion, N. taxispi-lota, and JV. rhombifera have neural spines that are higher than long.

The referred vertebrae belong to this last group. They differ fromN. taxispilota andiV. rhombifera, which have neural spines that overhang more posteriorly than anteriorly. Only N. cyclopion shows the near symmetry seen in lateral view of these fossils. Thus, this sample is assigned to N. cyclopion.
Distribution. Nerodia cyclopion is present today along the coastal plain from eastern Texas to southern South Carolina and up the M ississippi Valley to southern Illinois. N. cyclopion is known as a fossil from three Rancholabrean localities in Florida (Auffenberg 1963). Eight vertebrae from the late Blancan Beck Ranch locality in central Texas represent N. cyclopion orN. rhombifera (Rogers 1976).
Remarks. N. cyclopion is not common as a fossil. The Inglis material is the first evidence of its presence in the Irvingtonian. It and the material from the Beck Ranch (Rogers 1976) indicate that one or two high-spined forms of Nerodia had evolved by the Irvingtonian.
Nerodia erythrogaster (Forster) 1771
Referred Material. UF 26382,1 vertebra.
Description. This short, wide natricine vertebra has a neural spine which is as high as it is long. Accessory processes are longer than the prezgapophyseal width and are anterolateral^ directed.
Comparisons. The fossil resembles Nerodia erythrogaster in the size of the neural spine (seeNerodia cyclopion). It can be distinguished fromN.fasciata andAf. sipedon by its longer, more anteriorly directed accessory processes.
Distribution. N. erythrogaster is present throughout the south from Texas and Mexico to southern Illinois and the Delmarva Peninsula. As a fossil it is known from the late Blancan of Texas (Rogers 1976), the Irvingtonian of Texas (Holman 1969b), and the Rancholabrean of Texas (Holman 1969a) and Florida (Auffenberg 1963).
Remarks. As in theNerodia cyclopion group discussed above, N. erythrogaster appears to be distinct and widely distributed by the beginning of the Pleistocene. A single individual is represented in the fauna.
Genus Regina Baird and Girard 1853 Regina intermedia new species
Figure 10 A-C, J, K
Diagnosis. An early Pleistocene natricine snake closely resembling Regina rigida. It differs from all modern members of the genus in features of the dentary. The teeth are stout and blunt like those of R. alleni and R. rigida, but they are not hinged as in those species (Rojas and Godley 1979). The teeth are well ankylosed as in R. septemvittata

meylan: inglis ia squamates
x4:(D)fronta i vertebra, x4; (c) frontal, (h) (letn;u-v!ii()l()TYrE),x4:l,n
and H. grahami. These two species differ from II i.ntermedi-a in having thinner, sharper teeth.
Holotype. UK 26383; the middle third of a left dentary from the early Irvingtonian Inglis IA Local Fauna, Citrus County, Florida.
Referred Material. UF 26384, 7 vertebrae.
Description. The holotype is broken anteriorly and posteriorly. It is straight and shallow. Meckle's groove is open throughout its length. Thirteen tooth positions are present; eight retain well anky-!.losed teeth. Five unbroken teeth arc short, stout, and blunt.

The referred vertebrae have neural spines that are slightly flattened on the dorsal edge (= "ground off" of Auffenberg 1963). The neural spines are long and low and undercut anteriorly and posteriorly. The vertebrae have small hypapophyses and short blunt accessory processes.
Comparisons. The presence of an open Meckle's groove so far anterior in the dentary (App. 3-6), and short, stout, blunt teeth are not common among North American snakes. These characters do occur in the crayfish-eating snakes of the genus Regina. The type differs from all living members of the genus in having blunt and firmly ankylosed teeth.
Neural spines with flattened dorsal edges have been reported for the vertebrae of two natricine species, Storeria dekayi and Regina alleni (Auffenberg 1963). They also occur in adult Regina rigida (Fig. 10 D-F). The referred vertebrae are too large (2.75-3.50 mm CL) to beS. dekayi (1.13-2.13 mm CL). They differ from is?, alleni in having shorter, more blunt accessory processes and lower neural spines (Fig. 10 G-I). No apparent difference is evident between the vertebrae of R. rigida and R. intermedia.
Distribution. The genus Regina occurs throughout most of the midwest and eastern United States. The fossil record for the genus includes R. alleni from the Rancholabrean of Florida (Auffenberg 1963) and-R. grahami from the Irvingtonian of Kansas (Holman 1972).
Remarks. Wilson (1968) described an extinct natricine, Natrix hillmani, from the late Miocene of Kansas. He assigned it to Natrix only because he questioned the validity of the genus Regina, but he implied that its affinities lie with Regina. Thus, it appears that the genus Regina was established by the late Miocene and had begun diversification by the Pleistocene, with two species known in the Irvingtonian and three by the Rancholabrean.
Regina intermedia receives its name from its intermediate position between primitive and advanced Regina. The primitive forms have thin unhinged teeth and eat only freshly molted crayfish (Branson and Baker 1974). The advanced forms have evolved stout, blunt, hinged teeth and, at least in the case of R. alleni, distinctive behavior which allows them to capture and eat hardshelled crayfish (Franz 1977).
A scenario for the phylogeny of Regina is presented in Figure 11. The presence of R. intermedia in the Inglis IA fauna indicates that the split between the primitive and the advanced forms of the genus must have occurred by the end of the Pliocene as suggested by Rossman (1963b). Regina alleni is depicted as arising from the -proto-rigida (R. intermedia)-Regina rigida line following Rossman.

This material possibly represents a single individual of R. intermedia.
Genus Thamnophis Fitzinger 1843 Thamnophis cf. T. sauritus (Linnaeus) 1766
Table 7
Referred Material. UF 26385,18 vertebrae.
Description. These are natricine vertebrae with short, posteriorly directed hypapophyses. The centra are 1.44 to 1.77 times as long as wide. Neural arches are nearly square across the zygapophyseal faces (PRPR/POPR= .92-1.10).
Several authors (Brattstrom 1955a, 1967; Holman 1977b) have separated Thamnophis from Nerodia (Natrix) on the basis of the relatively longer vertebral form of Thamnophis, but it must be pointed
Recent Regina Regina
septemvittata granami
/ /
Regina Regina rigida alleni I I
I /
I /
I /
I X Regina alleni (Winter Beach
I 1 and Ichetucknee Faunas)
Irvingtonian \ J
I /
\ X Regina grahami //
^ 1 (Kanopolis Local / ^ \ Fauna) /
\\ /
\ /
\ X Regina intermedia (Inglis IA Fauna)
Blancan ^ j
\ / \
Regina hillmani (Wakeeny Fauna)
Figure 11. a phylogenetic scenario for the genus Regina.

Table 7. The four most important ratios in the discrimination of Thamnophis sirtalis, Thamnophis proximus, and Thamnophis sauritus.
Thamnophis proximus Recent (N = 22) 1.84 + .11 1.19 .04 1.55 .10 1.17 .06
Thamnophis sirtalis Recent (N = 27) 1.94 .10 1.21 .06 1.61 .10 1.16 + .06
Thamnophis sauritus
Recent (N = 25) 2.00 + .11 1.19 .03 1.69 + .11 1.16 .09 fossils (N = 17)_1.93 + .17 1.18 + .06 1.64 + .12 1.15 .11
Postzygapophysis to Prezygapophysis Length/Neural Arch Width ^Postzygapophysis to Prezygapophysis Length/Centrum Length "Centrum Length/Neural Arch Width
'Prezygapophysis to Prezygapophysis Length/Centrum Length
out that this practice may produce some error. A highly aquatic garter snake like Thamnophis melanogaster could be present in the later Cenozoic of North America, yet remain undetected because it has the short, wide vertebrae typical of Nerodia.
Comparisons. The vertebrae referred to Thamnophis were identified to species using discriminant analysis. Discriminant functions were created using 25 vertebrae from four T. sauritus, 27 vertebrae from four T. sirtalis and 22 vertebrae from three T. proximus. The program reclassified 15 of 74 (20.2%) of these vertebrae. The most important characters for the separation of these species are given in Table 7.
Distribution. T. sauritus occurs today throughout the United States east of the Mississippi River. The closely related form T. proximus occurs in the Great Plains states and the Mississippi Valley as far north as Wisconsin. Because of the difficulty in distinguishing species of Thamnophis, much material has been referred to Thamnophis sp. The genus may be present as early as late Miocene (Holman 1977b; Webb et al. 1981). There are numerous late Pliocene and Pleistocene records for Thamnophis sp., but the only material referred to T. sauritus is from the Rancholabrean of Tennessee (Guilday et al. 1978).
Remarks. Rossman (1963a) suggested that T. sauritus evolved from T. proximus or its prototype. He felt that the longer tail and reduced supralabial scale count in T. sauritus indicate specialization. Speciation was thought to have occurred during the Pleistocene, when the ribbon snake stock was isolated in Floridian and Mexican refugia. If Rossman's suggestion that?1, sauritus is derived from T. proximus is correct, then the Inglis material clearly indicates a pre-Pleistocene separation, but if asauritus-like form, at least in vertebral characters, is ancestral to T. proximus, this separation could have occurred during the Pleistocene.

GENUS Virginia Baird and GlRARD 1853
Referred Material. UF 26386,3 vertebrae. Description. The vertebrae are short (< 1.75mm) in centrum length. Hypapophyses are small. Neural canals are about equal in width to the cotyle. Neural spines are twice as long as they are high, and are not flattened dorsally.
Comparisons. The vertebrae represent a very small but mature natricine snake. Seminatrix has a neural spine which is as tall as long. The fossils (CL/NAW = 1.22-1.50) are wider thanStoreria (CL/NAW -1.67-1.83) and Tropidoelonion (CL/NAW = 1.62-2.00). The width of ;, the fossils falls into the range of modern Virginia (CL/NAW 1.20-;-1.66). Like the fossils, the neural spines of Virginia are long and low and not dorsally flattened. As no diagnostic characters for the two |extant species of Virginia were found, no attempt was made to assign [ the fossils to species.
Distribution. Virginia is presently found throughout the eastern United States from Iowa and Texas to New Jersey and North Florida. There are records of Virginia from the late Pleistocene of Texas (Holman 1963) and Virginia (Guilday 1962).
Remarks. The Inglis material extends the record of this genus into the early Pleistocene. Thus Virginia, like the natricines discussed above, had differentiated by the earliest Pleistocene.
Subfamily Xenodontinae Bonaparte 1845 Genus Diadophis Baird and Girard 1853 Diadophis elinorae Auffenberg 1963
Referred Material. UF 26387, 3 vertebrae.
Description. The vertebrae are very small, but represent a mature snake. Accessory processes extend anteriorly beyond the prezygapophyseal facets, but are not longer than the width of the facet. They are hooked anteriorly and are not markedly pointed. The neural spine is nearly as tall as the cotyle and overhangs posteriorly. The zygosphene is crenate from above and flat or nearly flat from anterior. The haemal keel is narrow and the subcentral ridges are weakly to moderately developed.
Comparisons. The shape and position of the accessory processes distinguish Diadophis from Stilosoma, Tantilla, Sonora, and Rhadi-naea (App. 3-5). The shape of the zygosphene and the taller neural spine with a posterior overhang separates it from Carphophis (Auffenberg 1963). The Inglis material compare more closely to Auffen-berg's Diadophis elinorae (Hemiphillian, Florida) than to Recent D. punctatus. As described forD. elinorae, the haemal keel of the fossil is narrower and the subcentral ridges better developed than in D. punc-

tatus. Neural spines are taller than those inD. punctatus and overhang only slightly in one of the three vertebrae.
Distribution. Recent Diadophis punctatus is found throughout most of North America. It is absent from the northern Rocky Mountains, the northern Great Plains, and coastal Mexico. Fossils of D. punctatus are recorded from the Rancholabrean of Florida (Holman 1959a; Auffenberg 1963), Georgia (Holman 1967), Maryland (Holman 1977a), and Texas (Holman 1969a) and from the Irvingtonian of Kansas (Brattstrom 1967).
Remarks. Haemal keels and subcentral ridges tend to be better developed in terrestrial rather than fossorial snakes. Extreme reduction of these features occurs in burrowers (scolecophidians^eterodow) (Holman 1979). Thus the slight difference betweenD. elinorae andD. punctatus may indicate a minor shift in the ecology of ringneck snakes from the Irvingtonian to present times. The slightly higher neural spine and the greater development of subcentral features suggest that D. elinorae was less fossorial and perhaps less secretive than modern D. punctatus.
All three vertebrae could be from a single individual.
cf. Dryinoides Auffenberg 1958 Figure 12 F-H
Referred Material UF 26388,26 vertebrae; UF 26389,1R, 1L maxillae (tentative).
Description The vertebrae are dorsoventrally compressed. Haemal keels and subcentral ridges are moderately to well developed. Subcentral ridges extend three-quarters or more of the length of the centrum. In some vertebrae the subcentral ridges are so well developed that the area between them and the haemal keel appears to be excavated. Neural spines are low.
The right maxilla is unbroken but lacks teeth. It has sockets for eight teeth. No diastema is present. The sockets for the two posterior-most teeth are slightly enlarged. The entire bone is short and straight. The prefrontal process is at the level of the 3rd and 4th teeth. The ectopterygoid process is at the level of the 6th and 7th teeth. The left maxilla is broken across the prefrontal process. It is slightly larger than the right maxilla, and has an additional tooth socket posterior to the prefrontal process. It is also very straight. Teeth are present in the last and the third-to-last sockets. The last tooth is broken but is larger at its base than the single tooth anterior to it.
Comparisons. The dorsoventrally flattened or depressed neural arches of the vertebrae are typical of the xenodontine genus Heterodon, but in Heterodon the subcentral ridges, if present, do not extend

mkvi.ak: tk;i .is ia squam ati :s

more than halfway down the centrum. T he development of subcentral ridges allowed Auffenberg (1958) to distinguish the vertebrae of a new genus, Dryinoides, from Heterodon. The referred vertebrae compare well with Auffenberg's (1958) description of the middle thoracic vertebrae of Dryinoides oxyrachis from the Miocene of Montana.
The fossil maxillae have fewer teeth than any colubrid examined, except Arrhyton and Lystrophis. These two genera have a maxillary diastema that is lacking in the fossils. The fossil maxillae are similar to those of H. nasicus and H. simus in length and proximity of the two maxillary processes, but they are much straighter and lack the diastema present in all Heterodon. The type and only known specimen of Dryinoides oxyrachis lacks maxillae, so comparison is impossible. Auffenberg(1958) found Arrhyton, Lystrophis, andHeterodon, as well as Conophis, to be similar toD. oxyrachis in certain skull and vertebral characters.
Remarks. It is possible that the maxillaries and the vertebrae represent two different snakes, but considered separately, each suggests a snake similar to Heterodon, Arrhyton, or Lystrophis. Just such a snake is Dryinoides oxyrachis Auffenberg. The vertebrae are the best evidence for assigning the material to Dryinoides.
The Inglis fossils probably represent a new species. The skull of the type of D. oxyrachis is longer than that of Heterodon. The short, complete maxilla described above suggests that the Inglis fossil had a short skull, similar to Heterodon.
If the maxillae are correctly associated with the referred vertebrae, they suggest that the Inglis Dryinoides may be the relict of an ancestor toHeterodon as indicated in Figure 14. To derive aHeterodon maxilla from the Inglis cf.Dryinoides (Fig. 12 F) requires only that a diastema and a gentle curve be added. The species of Heterodon vary only in the degree of curvature and the number of teeth anterior to the diastema. The vertebrae are also easily imagined as a model for Heterodon, requiring only reduction of subcentral ridges and an additional depression of the neural arch.
genus Farancia gray 1832 Farancia abacura (holbrook) 1836
Figure 13 A, C
Referred Material. UF 26390,1 L maxilla; UF 26391, 5 thoracic vertebrae.
Description. The single maxilla is complete. It has space for 17 teeth and no diastema. The prefrontal process is simple and located anteriorly at the level of the 6th and 7th teeth. The ectopterygoid process is nearly terminal, located at the 15th and 16th teeth. The

meylan: inglis ia squamates
entire bone is heavy and only very slightly curved.
The vertebrae are nearly square across the centrum and have depressed neural arches. They have well developed haemal keels, moderate to strong subcentral ridges, and blunt, laterally directed accessory processes.
Comparisons. Among North American snakes only Farancia has such astout maxilla with the prefrontal process so anterior in location (Fig. 13 A, B, App. 3-1). The anterior location of this process undoubtedly adds structural strength to the skull of this burrowing snake. The maxilla of F. abacura differs from F. enjlhrngrunvma in having a simple rather than recurved or hooked prefrontal process (Fig. 13 B).
In the past, isolated vertebrae of the two species of Farancia have been considered difficult, if not impossible, to separate (Auffenberg 1963; Holman 1978,1981). To distinguish these species, adiscriminant analysis was developed using49 vertebrae, 31 from four individuals of F. abacura and 18 from three individuals o( F. e,ryl.hrogramma. The analysis reclassified three of the F. abacura vertebrae, indicating reliable discrimination. The analysis assigned all of the fossil vertebrae to F. abacura.
Distribution. F. abacura is at present limited to the Gulf coastal plain of the United States, the Mississippi Valley, and the Atlantic coastal plain as far north as Maryland. The fossil record oiFa.rancta. is so far restricted to the Rancholabrean of Florida (see Holman 1981),
Figure 13.-Fossil and Recent./'Vmrn,,:,,. Fossil rmrltl iihnr,irii (A) maxilla occlusal view, X3; (C) vertebra, dorsal view, X3. Recent F,,r

The closely related snake Paleofarancia brevispinosa was described from a single vertebra from Lithia Springs, Florida, a site that includes material thought to be middle Pliocene in age.
Remarks. The Inglis material establishes the presence of Farancia abacura in the early Pleistocene of Florida. If Paleofarancia is ancestral to Farancia, it suggests modernization of the lineage by the late Pliocene.
The material probably represents a single individual. Apparently this individual had serious vertebral pathologies in the cervical region (Fig. 13 C).
Genus Heterodon Latreille 1802 Heterodon nasicus Baird and GlRARD 1852
Figure 12 D, Table 8
Referred Material. UF 26392, 57 vertebrae; UF 26393, 1 L pterygoid.
Description. The vertebrae are depressed and have flattened haemal keels. The subcentral ridges are weak.
The single left pterygoid has a moderately developed ectopterygoid process on the pterygoid flange. The anterior border of the flange between the ectopterygoid process and the tooth-bearing ramus is straight (Fig. 12 D).
Comparisons. The specific identification of the vertebrae was carried out using discriminant analysis. The analysis is based on 21 vertebrae from three if. nasicus, 28 vertebrae from fouriJ. platyrhi-nos and 30 vertebrae from four H. simus. The program reclassified two (2.5%) of these 79 vertebrae. The four most important variables in
Table 8. The four most important ratios for the discrimination of Heterodon nasicus, Heterodon platyrhinos, and Heterodon simus (mean one standard deviation).
Heterodon nasicus
Recent (N = 22) 1.59 + .13 1.68 + .20 1.40 .12 2.35 .18
fossils (N = 57) 1.62 .17 1.70 .19 1.40 .11 2.37 .23
Heterodon platyrhinos
Recent (N = 27) 1.72 + .16 1.57 .17 1.41 .08 2.20 + .18
fossils (N = 128) 1.75 .17 1.62 + .15 1.39 .11 2.23 .16
Heterodon simus
Recent (N = 30) 1.33 .08 1.56 .15 1.49 .14 2.30 .12
Postzygapophysis to Prezygapophysis Length/Neural Arch Width Centrum Length/Neural Arch Width 'Prezgapophysis to Prezgapophysis Width/Centrum Length 'Prezygapophysis to Prezygapophysis Width/Zygasphene Width

this analysis are POPR/NAW, CL/ZW, PRPR/CL, and PRPR/ZW (Table 8).
The pterygoid is referred toHeterodon because of the anterior orientation of the ectopterygoid process. It is similar to H. nasicus and H. simus in the moderate development of this process. H. simus has a curved anterior border to the pterygoid flange, thus the assignment to H. nasicus.
Distribution. At present H. nasicus ranges throughout the Great Plains from northern Mexico to southern Canada. The closely related form H. simus occupies the southeastern coastal plain from Mississippi to North Carolina.
The hognose snakes have a long fossil record. H. nasicus is known from one Blancan and two Irvingtonian sites in Kansas (Brattstrom 1967). It is also known from the Irvingtonian of Texas and from Rancholabrean sites in Kansas, New Mexico, Oklahoma, and Texas (see Holman 1981). H. simus is known as a fossil only from the Rancho-. labrean of Florida (Holman 1981). Likely ancestors of H. nasicus include Heterodon plionasicus Peters from the Blancan of Texas (Rogers 1976) and Kansas (Peters 1953; Brattstrom 1967) and Paleoheter-odon tiheni from the upper Miocene of Nebraska and South Dakota and the lower Pliocene of Kansas (Holman 1964,1977b).
Remarks. The fossil record of H. nasicus and its ancestors is surprisingly complete. It suggests the evolution oiH. nasicus and//. plionasicus fromPlioheterodon tiheni (Fig. 14). As a final step in this sequence, the evolution of Heterodon simus from H. nasicus is indicated by the present study. This supports Piatt's (1969) rebuttal of Edgren's (1952) conclusion that//, nasicus evolved from//, simus or the two evolved in parallel fashion from H. platyrhinos.
The fossil record suggests that two lines of Heterodon have been separate since the late Miocene (Fig. 14). Two forms oiHeterodon are present in the recently discovered Love Bone Bed of Clarendonian age in Florida (Webb et al. 1981). Although the material has not been thoroughly studied, it is apparent that both the H. nasicus and H. platyrhinos lines are represented. At the time the Inglis fauna was trapped, gene flow between western populations and the southeast was apparently unrestricted. It seems likely that the isolation of eastern populations, probably caused by increased interglacial sea levels, resulted in the differentiation of H. simus.
Heterodon platyrhinos Latreille 1802
Figure 12 A-C, E, Table 8
Referred Material. UF 26344,128 vertebrae; UF 26395,2 L dentaries; UF 26396,2 R maxillae; UF 26397,1 L and 1 R pterygoids;

Heterodon platyrhinos
Heterodon nosicus
ct.Dryinoides | INGLIS IA
REXROAD LOC3 Heterodon
cf. Dryinoides MIOCENE
"WE ct Heterodon
cf. Heterodon I love BONE bed
Poleoheterodon tiheni
Dryinoides I
oxyrachis MADISON valley
Figure 14. A phylogenetic scenario for the North American xenodontine genera Dryinoides and Heterodon.
and UF 26398,1 basiparasphenoid.
Description. The referred vertebrae fit the description oiHeter-odon nasicus. Measurements and ratios important in the identification of the vertebrae of this species are given in Table 8.
One maxilla is complete, the other consists of only the central portion. Each possesses a diastema at the level of the ectopterygoid pro-

cess. The complete specimen is a gently curved bone with space for two enlarged teeth posterior to a diastema, and 11 smaller teeth anterior to the diastema. In neither specimen do the tips of the prefrontal and ectopterygoid processes approach each other. In the dentaries, Meck-le's groove is open widely to the level of the 4th tooth; from that point to the symphysis it is reduced to a narrow slit. These bones are moderately long and narrow, tapering gradually to a point anteriorly. The compound notch on the labial surface approaches the level of the 9th tooth. The dental foramen lies at the level of the 7th tooth.
The pterygoids exhibit an anteriorly directed, round ectopterygoid process for the articulation of the posterior end of the ectopterygoid.
The basiparasphenoid has a short cultriform process. The basisphe-noid and the pituitary fossa are distinctly wider than long, and suborbital flanges are well developed, but do not extend onto the cultriform process.
Comparisons. The four most important characters in the discrimination of vertebrae of the extant species of Heterodon appear in Table 8.
Among North American snakes, only the maxillae of Heterodon fit the description of the fossils given above.//, simus and//, nasicus have fewer maxillary teeth and shorter, more curved maxillae with the prefrontal and ectopterygoid processes closely approaching each other.
Among snakes examined in this study, only the dentaries of Heterodon and a few natricines were found to have Meckle's groove present to the symphysis, but reduced to a slit at some point posterior to it (App. 3-6). This reduction in natricines varies with the species and the individual, but generally occurs posterior to the level at which it occurs in Heterodon. The dentaries of H. simus and H. nasicus are shorter, deeper and less tapered than H. platyrhinos. Among North American colubrid snakes Heterodon is apparently unique in having the ectopterygoid process of the pterygoid anteriorly directed. The shape of this process is variable, but it always consists of an anteriorly produced portion of the pterygoid flange. This process is best developed in//, platyrhinos (see Fig. 6 in Weaver 1965).
The shortened Heterodon skull includes a short, wide basisphenoid with a very short cultriform process. Heterodon is extreme among the examined colubrids in this regard (App. 3-4). The suborbital flanges in the referred fossils are more developed than those in any//, simus or H. nasicus examined. They are similar to the condition seen in adult//. platyrhinos.
Auffenberg (1963) described a related form, H. brevis, from the middle Pliocene of Florida. Its shorter neural spine distinguishes it

fromH. platyrhinos. The ratio of centrum length to basal neural spine length (CL/NSB) for the type of H. brevis is 2.07. This ratio for 25 Inglis fossils ranges from 1.35 to 1.85, similar to the range of 1.32 to 1.82 for if. platyrhinos given by Auffenberg (1963).
Distribution. H. platyrhinos is found throughout the United States east of the Great Plains. It is present as a fossil in the Blancan of Kansas and Nebraska (Brattstrom 1967) and Texas (Rogers 1976), the Irvingtonian of Florida, Kansas, and Maryland, and the Rancholabrean of Florida, Georgia, Kansas, Missouri, Tennessee, and Texas (see Holman 1981).
The related and possibly ancestral tormH. brevis is known only from the middle Pliocene of Florida (Auffenberg 1963).
Remarks. In view of the fossil record, the presence of H. platyrhinos in the early Pleistocene of Florida is not unexpected. However, it is of interest for two reasons. F irst, it suggests the replacement ofH. brevis byH. platyrhinos in Florida between middle Pliocene and early Pleistocene. Second, the referred skull material shows that H. platyrhinos had reached its present degree of specialization by the earliest Pleistocene.
Genus Rhadinaea Cope 1863 Rhadinaea cf. R. flavilata (Cope) 1871
Referred Material. UF 26399,3 vertebrae.
Description. The referred vertebrae are small (CL < 2.0 mm). Accessory processes are laterally directed, short, straight, and blunt. Haemal keels extend well below the centrum, and are very slightly expanded posteriorly.
Comparisons. The vertebrae have shorter, blunter accessory processes than Stilosoma, Tantilla, or Sonora (App. 3-5). The processes are not curved in the fossils as in Diadophis punctatus and Carphophis amoenus. The ventral extension of the haemal keel in the fossils is identical to that in modern R. flavilata.
Distribution. The 45 modern members of this genus are found from Argentina to the southeastern United States. A single species, Rhadinaea flavilata, is known from the southeastern United States. Fossils referred toR. flavilata were previously reported from the late Irvingtonian (Holman 1959a) and Rancholabrean (Auffenberg 1963) of Florida.
Remarks. Myers (1974) considered the flavilata group of the genus Rhadinaea unique in having no semblance of geographic unity. The two Recent members, Rhadinaea flavilata from the southeastern United States and Rhadinaea laureata from western Mexico, are peripheral relicts of a once widespread species that Myers suggested

had its origin in Mexico. He speculated thati?. flavilata became isolated in Florida in the late Pleistocene and spread to its present range in postglacial times, but Inglis material indicates thatRhadinaea had already arrived in Florida by the earliest Pleistocene.
Subfamily--Incertae Sedis
Genus Tantilla Baird and Girard 1853
Referred Material. UF 26400,6 vertebrae.
Description. The referred vertebrae are small (< 1.75 mm CL), with cotyles equal in size to the neural canals. Neural spines are long and low. Subcentral ridges are poorly developed or absent. Accessory processes are straight, anterolateral^ directed, and as long as or longer than the prezygapophyseal facet width.
Comparisons. The size of the cotyle relative to the neural arch identifies the fossils as those of a very small adult snake. The accessory processes are like those of Tantilla relicta, which are longer and straighter than in any small North American snakes examined (App. 3-5). The vertebrae oiSonora are similar to the fossils, but have shorter and more laterally directed accessory processes and stronger subcentral ridges. The fossils compare well with T. relicta, but because of the similarity in vertebrae of a number of North American Tantilla species no specific identification was attempted.
Distribution. The present distribution of Tantilla includes the southern United States from coast to coast, and Central and South America to northern Argentina. The only reported fossils of this genus are from the late Pleistocene of Florida, New Mexico, and Texas (see Holman, 1981).
Remarks. The Inglis material is the oldest fossil record of this genus. It confirms Telford's (1966) suggestion that Tantilla had arrived in Florida by the early Pleistocene and thus was present during interglacial periods. Telford (1966) suggested that during these interglacials various populations of Tantilla became isolated, and then differentiated into the various forms found in Florida today.
Family Elapidae Boie 1827 Genus Micrurus Wagler 1824 Micrurus cf. Micrurus fulvius (LinnAeus) 1766 Referred Material. UF 26401,6 vertebrae. Description. These are small elongate vertebrae with long, low neural spines. Hypapophyses are long, sharp, and posteriorly projected. Accessory processes are short, laterally directed and subtrian-gular.

Comparisons. Among North American snakes, only the natricines, viperids, and elapids have hypapophyses on all precaudal vertebrae. No viperids have vertebrae with such low neural spines or hypapophyses as sharp and posteriorly projected as in this material from Inglis. Two small natricine genera, Virginia and Storeria, are similar. The former differs from Micrurus in having longer, sharper, and more anteriorly projecting accessory processes. The latter has dorsally flattened neural spines (Auffenberg 1963) and broader hypapophyses.
distribution. M. fulvius occurs from northeastern Mexico and central Texas to southern North Carolina and Florida. It is known as a fossil from the later Irvingtonian of Florida (Holman 1959a) and the Rancholabrean of Florida and Texas (see Holman 1981). Material referred to Micrurus sp. is known from the upper Miocene of Nebraska (Holman 1977a) and Pliocene of Florida (Auffenberg 1963).
Remarks. The specific assignment to cf. M, fulvius is based on geographical considerations. The lack of comparative material and incomplete knowledge of this large genus has deterred closer examination of the Inglis material at the species level.
The great diversity of Micrurus in South America suggests that the center of origin of the genus may be in that continent, but the oldest known fossils are from the M iocene of Nebraska. Whatever the origin, the Inglis material, along with that from Haile VI referred to Micrurus sp. by Auffenberg (1963), documents the arrival of coral snakes in Florida by the Pleistocene.
The material could represent a single individual.
Family Viperidae Oppel 1811
Viperid vertebrae are short and wide and have short accessory processes. Throughout the column they have hypapophyses that are thick, but narrow in lateral view. They are usually long, and tend to point ventrally in the anterior portion of the column. In the rear portion of the column they point more posteriorly.
The viperid skull is highly modified to incorporate a venom delivery system. This has resulted in a number of skull elements that are diagnostic at the family level.
Genus Crotalus Linnaeus 1758 Crotalus adamanteus (Beauvois) 1799
Figure 15 A-d, Table 9
Referred Material. UF 26402, 979 vertebrae; UF 26403, 5 R and 4 L maxillae; UF 26404,1R and 1L dentaries; UF 26405,1R and 1 L palatines; and UF 26406, 3 R and 3 L compounds.

Description. The vertebrae all have long and narrow hypapo-physes that are laterally thickened. Cotyles are oval. Accessory processes are much shorter than prezygapophyseal widths.
The maxillae are short, massive, and modified to hold a fang. From the basal portion of each maxilla rises a vertical flange (= dorsal process of Holman 1959b) which forms the medial wall of the facial pit cavity (= loreal fossa). At the top of this flange is a laterally directed process (= dorsal process) that extends to about the midline of the maxilla. On the anterior border of the flange is a short lateral extension in the frontal plane. There is a small tubercle in the middle of this extension. The anterior expansion and a similar posterior one form the lateral walls of the facial pit cavity.
The dentaries are short and deep with Meckle's groove open to the symphysis. Three complete dentaries average 9.3 teeth.
The palatines are small, thin sheets of bone that bear teeth on their ventral edges. Both fossil palatines are triangular but are broken posteriorly.
The compounds have expanded lingual and reduced labial flanges. The labial flange is dorsally directed.
Comparisons. The maxillae are typically viperid (Brattstrom 1964). They differ from Agkistrodon in having anterior and posterior margins to the facial pit cavity. In North American Agkistrodon only the medial wall is evident in the maxilla. The fossils differ from Crotalus atrox and Crotalus horridus in having the basal portion of the bone dorsally expanded at the distal edge. The maxilla of Sistrurus can be distinguished from the fossils by a feature visible in the frontal plane. The labial margin of the basal portion tends to be directed dorsomedially rather than dorsally. The maxillae are most similar to Crotalus adamanteus.
Marx and Rabb(1972) include several Crotalus in a group of snakes with fewer than 9.5 dentary teeth. Crotalus adamanteus is the only large snake included in their list. Although they report a dentary of C. atrox that has seven teeth, most dentaries examined in this study have 10 teeth. The fossils are identical to C. adamanteus in the configuration of Meckle's groove and in the tooth shape.
The thin, high palatines are typical of crotaline snakes (Brattstrom 1964). They are triangular with a longer trailing edge. They do not have a step in the anterior margin, as in C. horridus, Sistrurus, and Agkistrodon. The compounds are indistinguishable from C. adamanteus and C. atrox. In Crotalus horridus and Agkistrodon the reduced labial flange is more laterally directed.
The fossil vertebrae can be distinguished from Sistrurus miliarius by their wider cotyles and less vaulted neural arches (seeS. miliarius).

Separation from other North American viperid snakes is more difficult. The vertebrae of large Crotalus and Agkistrodon are very similar. Holman (1963) distinguishes Agkistrodon from Crotalus by the presence in the former of distinct pits on either side of the cotyle. Agkistrodon is reported to have one large fossa in each pit; Crotalus has one or more small fossae. Examination of two complete vertebral columns of C. adamanteus reveals intracolumnar variation in pitting and in the presence of fossae. Pits are present in the anterior one-third and posterior one-third of both columns. They are strongly developed in the vertebrae just anterior to the vent. Two small fossae are present in the pit throughout the column in one specimen but single larger fossae are present throughout the other. Holman's characters would correctly identify a majority of midbody vertebrae, but discriminant analysis was used here in an effort to identify all of the precaudal vertebrae. The fossils were identified using discriminant functions developed from 30 vertebrae from four A. piscivorous, 31 vertebrae from four C. adamanteus, 31 vertebrae from seven C. atrox, and 16 vertebrae from two C. horridus. Of the original 108 observations, 11 (10.1%) were reclassified; 10% of C. adamanteus were reclassified as A piscivorous, and 10% were reclassified as C. atrox. Twelve ratios were used to separate these species. The most important are ZW/NAW, PRPR/POPR, CL/CW, and PRPR/CTW (Table 9).
In a sample of 100 fossil viperid vertebrae, 60% were referred to Crotalus adamanteus by this analysis, 30% were referred to C. atrox, and 10% to Agkistrodon piscivorous. The small number referred to Agkistrodon is within the error known to occur in this analysis. Thus the presence of A. piscivorous remains uncertain. The absence of any
Table 9. The four most important ratios for the discrimination of Crotalus adamanteus, Crotalus atrox, Crotalus horridus, and Agkistrodon piscivorous (mean + one standard deviation).
Agkistrodon piscivorous
Recent (N = 32) .88 .07 1.41 .12 1.25 .17 2.92 + .14
Crotalus atrox
Recent (N = 31) .80 .06 1.36 + .12 1.27 + .12 2.80 .33
Crotalus horridus
Recent (N = 17) .85 .07 1.28 + .14 1.44 .24 2.68 + .14
Crotalus adamanteus
Recent (N = 31) .82 + .06 1.43 .14 1.24 + .17 2.70 + .31
fossils (N= 100) .79 .06 1.46 + .11 1.20 + .13 2.75 .27
'Zygasphene Width/Neural Arch Width
2Prezygapophysis to Prezygapophysis Width/Postzygapophysis to Prezygapophysis Length
^Centrum Length/Neural Arch Width
4Prezygapophysis to Prezygapophysis Width/Cotyle Width

skull material ofAgkistrodon and the paleoecology of the Inglis site are further evidence for the absence, or at least extreme rarity, of this species in the fauna.
The 30% referred to Crotalus atrox cannot wholly be attributed to error. It suggests that vertebrae with a morphology more similar to C. atrox than to C. adamanteus are present in the early Pleistocene of Florida. C. adamanteus is also clearly represented among the vertebrae. Although it is possible that C. atrox and C. adamanteus are sympatric in Florida in the earliest Pleistocene, it seems more likely that adamanteus evolved from C. atrox and that the vertebrae of early Pleistocene C. adam,anteus had not differentiated to the extent seen in the modern members of the species.
The skulls of Crotalus adamanteus and C. atrox are very similar. Of the four elements representing Crotalus found preserved in the Inglis fauna, only the maxillae and dentaries show more similarity to C. adamanteus than to C. atrox. This is further evidence of a close relationship between these forms.
Distribution. Crotalus adamanteus is found today in the eastern coastal plain of the United States from the Mississippi River to North Carolina. Klauber(1972) listed the complete fossil record of the genus. He includes 12 middle to late Pleistocene and 10 late Pleistocene to Recent localities for Crotalus adamanteus in Florida. Ten late Pleistocene to Recent records are listed for New Mexico, Nevada, and Texas for Crotalus atrox.
Remarks. At one time Crotalus atrox was considered a subspecies of C. adamanteus. At present their relationship is considered to be one of direct common ancestry (F ig. 3:4 in Klauber 1972). The similarity of some of the Inglis vertebrae to C. atrox supports this relationship. It seems likely that the two were once a single species which was split by a Mississippi Embayment. The Inglis IA record is the oldest for the species.
Differences in the size of the maxillae indicate the presence of at least seven individuals.
Genus Sistrurus Garmen 1883 Sistrurus miliarius (LlNNAEUS) 1766
Figure 15 E, F
Referred Material. UF 26407, 384 vertebrae; UF 26408, 2 L dentaries.
Description. These small vertebrae have relatively small, round cotyles. The hypapophyses are long, straight, and posteroventrally directed. The neural arches are highly vaulted and not distinctly wider than long. Neural spines are about as high as they are long.

The dentaries are small. Meckle's groove is open to the symphysis. The dental foramen lies immediately adjacent to the anterior end of the compound notch.
Comparisons. Long, straight, posteroventrally directed hypapo-physes are typical of the Viperidae. The fossil vertebrae can be distinguished from young Crotalus, Agkistrodon, and Sistrurus catenatus by their smaller, round cotyles.
Few snakes have dentaries with Meckle's groove open to the symphysis (App. 3-6). This condition occurs inHeterodon and some natri-cines, but the dentaries of these species differ from the fossils in having the groove reduced to a slit by the 4th tooth. In pit vipers it is open to the symphysis, although it may become ventrally oriented and not laterally visible. Among the crotalines examined, only Sistrurus miliarius is similar to the fossils in having the dental foramen immediately adjacent to the compound notch (Fig. 15 F). Its exact position is variable, but it is typically much closer than in any other pit vipers examined.
Distribution. Sistrurus miliarius is found today in the southeastern United States from eastern Texas and Oklahoma to southern North Carolina and Florida. The only records of the species as a fossil are from the middle to late Pleistocene of Florida (see Holman 1981). A related form, Sistrurus catenatus, is known from the Blancan of Kansas (Brattstrom 1967) and Texas (Rogers 1976) and the Irving-tonian of Kansas (Holman 1972).
Remarks. The Inglis material represents the oldest record of Sistrurus miliarius. Klauber (1972) suggests that the most primitive living member of the genus, Sistrurus ravus, gave rise to both S. miliarius and S. catenatus in the Miocene or Pliocene. The Inglis material verifies at least a Pliocene arrival, presumably from the west, of S. miliarius. Material reported by Rogers (1976) shows that S. catenatus lived in Texas during the Blancan.
Geological evidence (fide Klein 1971) indicates that the Inglis IA sample accumulated in a sinkhole trap. Webb (pers. comm.) suggested that additional material may have been added by local slope wash. Excavations revealed that in cross-section the sinkhole had a hemispherical roof with a 3-meter breach in the center. Such a sinkhole would have been an ideal trap. Indeed the sediments were thickest directly below the opening and thinned out toward the edges. The clastic sediments were locally derived and appear to have been washed or blown in through the roof.
'Literally the study of burials (for a good example see Behrensmeyer 1975).

The excellent preservation of the fossils is also an indication of a natural trap. Many fragile specimens, including pterygoids, dentaries, and maxillae of both lizards and snakes, have been well preserved. These elements would presumably bear evidence of any transportation. Furthermore, the occiputs of several Ophisaurus and short sections of snake vertebral columns remain articulated, suggesting limited maceration and transport. Klein (1971) suggested that windblown sands may have filled the sinkhole at a rapid rate. The covering of newly trapped animals by these sands would account for the crisp detailed preservation of much of the recovered material (Behrens-meyer 1975).
Further evidence of a natural trap is drawn from the large Ophisaurus sample. The 11 left and 10 right maxillae, 11 left and 12 right dentaries (1 each per individual), 926 body vertebrae (50-55 per individual), 22 sacral vertebrae (2 per individual), and 48 nonautotomic caudal vertebrae (5 per individual) all provide an MNI of 9 to 12. This equality of counts suggests that each Ophisaurus was preserved completely. Complete preservation would be best accounted for by entrapment.
Few of the fossils are waterworn. Only 11 snake vertebrae are too waterworn to be identified. Klein (1971) gave several reasons why he thought abrasion by water had occurred within the sinkhole after deposition. He considered that the shape of the deposit was incompatible with the concave cross-section of a stream, and noted the absence of crossbedding, graded bedding, or lenticular bedding that is usually associated with stream deposits. During the 1974 excavation some crossbedding was discovered (Webb, pers. comm.), which indicates that water-lain sediments were present in portions of the deposit not seen by Klein. Klein also noted the lack of stream-dwelling forms among the mammals preserved in the fauna. Although one stream-and three marsh- or pond-dwelling snake species are represented, they are clearly in the minority.
Further evidence that I nglis was a sinkhole trap comes from a study of the frogs (Meylan, MS) which reveals a scarcity of hylids. Their rarity in the Inglis fauna may reflect their ability to escape a natural trap by climbing the vertical walls.
Most numerous among the Inglis anurans is a new species of Bufo that is related to modern Bufo terrestris. At least 168 individuals are represented. A comparison of the population structure of the fossil toads to Recent populations reveals a surplus of adults. Several explanations are possible, but the most plausible seems to be that the surplus adults were trapped during breeding migrations.

Animals would be expected to accumulate in a natural trap in proportion to their density, though differences in trapability could affect the observed results. Larger snakes are less likely to be trapped, being better able to climb out, but any bias toward small snakes in the fauna might be offset by their ability to leave the sinkhole via small crevices or holes. The effects of such possibilities cannot be measured, so it is assumed that the squamates were preserved in the fauna in proportion to their abundance in the vicinity of the sinkhole.
Diagenetic Factors
Between the time a fauna is deposited and excavated a number of factors can modify it. Destruction of bone can result from a variety of processes (Behrensmeyer 1975). Most of the Inglis fauna was preserved in fine sands and sandy clays (Klein 1971). Bones in a sandy matrix are rarely subject to compaction and fracturing, and have little chance of being perturbed, provided that the sands are not too coarse (Behrensmeyer 1975). At Inglis there was no evidence of slumping, differential settling, or reworking of the sediments (Klein 1971). Thus, diagentic factors appear to have had little effect on the fauna.
Collection Bias and Sample Size
The 100 tons of sediment excavated from the Inglis IA site were screened through both coarse and fine screens. Unfortunately the finest screen used was ordinary window screen (16 meshes per inch). Trials with modern skeletons of Carphophis, Diadophis, Rhadinaea, Stilosoma, and Tantilla showed that some vertebrae and most of the longer skull elements (maxillae, pterygoids, compounds, etc.) of these species pass through window screen. Others become trapped in the screen. Fossils trapped in such a fashion would most likely be broken and lost during washing. This may account for the absence of the small scincids Scincella laterale, Neoseps reynoldsi, and Eumeces egregius in the Inglis fauna. Neoseps occurs only in isolated areas of fine sands which may not have been present near Inglis I A, but Eumeces egregius and Scincella laterale are common in high pine and xeric hammock (Carr 1940) and should have been present in the Inglis fauna. It is possible that Scincella had not yet arrived in Florida by the early Pleistocene. The diversification of Eumeces egregius into four subspecies in peninsular Florida is attributed to isolation during Pleistocene interglacials (Mount 1965).
It is fortunate that vertebrae of some of the very small snakes were recovered. However, because of the probable bias created by the screening procedure, they do not weigh heavily in the following paleo-ecological interpretation of the fauna.

The tremendous quantity of matrix recovered and washed from Inglis IA insures that the faunal sample is large enough to provide a valid picture of all but the smallest vertebrates of the surrounding communities. In a discussion of sampling and sample sizes for paleo-ecological interpretations using fossil mammals, Wolff (1975) considered it necessary to have a sample of 5,000-10,000 kg of bulk sediment, based on the density of fossils in several localities in California. He concluded that for mammalian taxa, the relative abundance of the most common forms are roughly approximated when at least 500 identifiable specimens have been recovered. This represents a minimum of 25 specimens and three individuals per taxon of mammals at the California sites. In the present study, more than 3,500 elements, with an average of 125 specimens per species, represent squamate reptiles. This certainly qualifies as an adequate sample for paleoeco-logical interpretation.
Faunal Requirements
The use of fossil faunas to identify ancient environments can only be successful when the habitat requirements of the faunal members are known. These requirements must be inferred for more than half of the 98 Pleistocene mammal genera from North America because they are now extinct. North American squamate taxa have experienced far fewer extinctions since the Pleistocene (1 genus, Dryinoides), and because they have survived, often to the species level, they are more useful than mammals for paleoecological studies.
Of 31 squamate species identified from Inglis, four are extinct. The remainder, with two exceptions, live in Florida today. The specific Florida habitats occupied by these species have been recorded by Carr (1940) and others (Table 10). Recorded habitat preferences and the relative abundance of the taxa in the Inglis IA fauna are used below in an attempt to classify the community from which the Inglis fauna was trapped.
An often unstated assumption in paleoecological studies is that the habits of a species do not change through time, but just as the habitats occupied by species may change through space (e.g. Rana areolata), they can also change through time (Van Devender et al. 1976). Still the Inglis sample is large enough so that many changes would have had to occur to confound the observed results.
Ecological Interpretation
It has been suggested that the Inglis IA fauna sampled a savanna that extended around the Gulf of Mexico during the late Cenozoic and expanded markedly during the early Pleistocene (Webb 1978). Evidence from the squamates, presented below, supports this thesis, and

Table 10. Preferred habitats of Inglis IA squamates. A number in a given column indicates that the species is found in that community.
The number is the MNI for that species in the Inglis fauna. Data are largely from Carr (1940), but habits of Diadophis and Rhadinaea are taken from Myers (1965 and 1967), data for Tantilla is taken from Telford (1966), and data for Cemophora is taken from 60 individuals trapped by National Fish and Wildlife Service in various habitats in Wakulla County, Florida.
Longleaf Xeric Marshes,
Pine Hammock Ponds
(= High (= Upland Mesophytic (Near
Species MNI Pine) Hammock) Hammock Flatwoods Water) Rivers
Nerodia erythrogaster 1 1
Nerodia cyclopion 1 1
Virginia sp. 1 1 1 1
Regina sp. 1 1
Thamnophis sauritus 1 1
Heterodon platyrhinos 2 2 2
Diadophis sp. 1 1 1 1
Farancia abacura 1 1
Rhadinaea flavilata 1 1 1
Coluber constrictor 4 4 4 4 4 4
Masticophis flagellum CO 3 CO
Drymarchon corais 2 2 2
Elaphe obsoleta 2 2
Elaphe guttata CO 3 3 3
Lampropeltis getulus 2 2 2
Pituophis melanoleucus CO CO
Lampropeltis triangulum 1 1 1 1
Cemophora coccinea 1 1 1 1
Stilosoma extenuatum 1 1 1
Tantilla sp. 1 1 1 1
Crotalus adamanteus 7 7 7 7 7
Sistrurus miliarius 3 3 3 3
Sceloporus undulatus 36 36 36
Ophisaurus ventralis 12 12 12 12
Rhineura floridana 1 1 1 1
80 78 16 40 9 1

provides further information as to the nature of this savannah in the earliest Pleistocene of Florida.
In an attempt to analyze the communities represented by the Inglis fauna, the occurrence of each species (or living equivalent) was recorded for six general habitat types (Table 10). Data are almost entirely from Carr (1940), but other sources have been used for taxa not conclusively covered by him. The minimum number of individuals (MNI) for all of the taxa occurring in a single habitat was added. The resulting scores appear in the last line of Table 10. The two xeric habitats, longleaf pine and xeric hammock, receive the highest scores (80 and 78 respectively). Flatwoods are next (40), although 6 of the taxa and 31 of the individuals so assigned could have come from the two more xeric habitats. Another species, Sistrurus miliarius, is found in longleaf pine and xeric hammock. Carr's reference to this snake occurring in flatwoods probably reflects its greater abundance in that community. Seven of the eight species recorded from mesophytic hammock are also recorded for xeric hammock. Data for the eighth, Diadophis elinorae, has been inferred from Diadophis punctatus. It is possible that the minor changes in morphology discussed in the account of Diadophis elinorae coincided with a change in habitat preference of ringneck snakes.
The seven taxa that usually occur near water are represented by a minimum of 11 individuals. Three of them may also be found in xeric hammock and/or longleaf pine. The other four species are each represented by a minimum of one individual. Nerodia cyclopion and Tham-nophis sauritus are both active snakes that often leave water. I have found both along the borders of isolated, seasonally dry ponds in north-central Florida. Farancia abacura and Recent Regina species also travel overland, especially during heavy rain. The single stream dweller, Nerodia erythrogaster, is also known to travel far from water (Conant 1975). The few occurrences of these normally aquatic species in the Inglis fauna could thus be attributable to chance entrapment during terrestrial movements. This view is in accord with the apparent subsurface flow of water that deposited the Inglis site (see p. 4).
The composition of the squamate fauna overwhelmingly indicates longleaf pine and/or xeric hammock as the principle community sampled. Longleaf pine, when mature, is open and savanna-like; xeric hammock is not, having a more closed canopy. True grasslands are not present in Florida, thus the only habitat more open is wet treeless marsh maintained by seasonal flooding. The fauna clearly shows that this latter type of habitat did not make a significant contribution to the Inglis fauna.

Klein (1971) cited the abundance of Capromeryx (pronghorns), Odo-coileus (deer), Platygonus (peccary), Hemiauchenia (llama), and Lepus (jackrabbit) in the Inglis fauna as evidence of open habitat, but he noted that these grazers, as well as the carnivores in the fauna, probably required cover for resting spots. The very large sample of Geomys indicates dune-like sandy terrain well above the water table.
The combined evidence of the herpetofauna, other vertebrates, and the sedimentary context suggests a mixed habitat of mature longleaf pine with xeric hammock interspersed. Xeric hammock would be expected in the depressions characteristic of a karst topography. That one or more of these depressions contained water at least seasonally, is suggested by the presence of Nerodia, Regina, Thamnophis, Alligator, Pseudemys, and Rana catesbeiana.
With few exceptions, the Inglis squamates have ancestors in the fossil record of North America. Thus the fauna is autochthonous and shows no South American or West Indian influence. The Inglis IA fauna shows continuity with the widespread mid- to late Tertiary xeric communities of North America. Post-Inglis changes in Florida's squamates appear to result largely from the variation in climate and sea level resulting from glaciation.
The mid-Tertiary fossil record of temperate North America includes probable ancestors for 84% of the Inglis squamates (Table 11). Twenty-four Inglis genera and 11 species are known from earlier localities. Two genera, Cemophora and Farancia, are known from morphologically similar forms. Thus, nearly all the species known from Inglis could have evolved in North America.
Only five genera present in Inglis IA do not have known ancestors in the North American fossil record. Three of these, Drymarchon, Rhadinaea, and Tantilla, appear to have a Central American origin. A detailed study of Rhadinaea (Myers 1974) reveals that only 12 of 45 known species are present in South America. The primitive and probably ancestral godmani group has its center of dispersal and probable place of origin in Nuclear Central America. Similarly, only 8 of 47 known Tantilla species and 3 of 8 subspecies of Drymarchon corais are found in South America. It seems likely that studies of these genera would reveal Central American origins.
Opheodrys vernalis and Virginia sp. also lack ancestors in the fossil record. Virginia is only found in North America and undoubtedly evolved on this continent. The genus Opheodrys includes four Asian species, which suggests the possibility that it evolved in the Old World (Schmidt and Necker 1936).

Table 11. A listing- of the probable closest relative for members of the Inglis IA fauna known from the previous fossil record (citations given in species accounts).
Inglis IA
Previous Record
Coluber constrictor Dryni a reh on cor a is Mont icopli is flayellurn Opheodrys rernalis Ccmophora coccinea Elaphe guttata Elaphe obsoleta La m pro pelt is getul us Lampropeltis triangulum Pituoph is melanoleucus Stilosoma extenuatum Diadophis elinorae cf. Dry irio ides Farancia abacura cf. Rhadinaea flarilata Tantilla sp. Heterodon nasicus Heterodon platyrh inos Nerodia cyclopion Nerod ia erythrogaxter Regina intermedia Virginia sp. Thamnopkii sa;uritus Crotalus adamanteus Sistriirus miliarius Micrurus fulvius Rhine ur a. flo ridan a Seeloporus undulatus Ophisaurus centralis
Eumeces c.arri Gerrhonotus sp.
Coluber constrictor (Blancan, Texas) NONE
Masticophis flagellum (Blancan, Texas) NONE
Pseudocemophora antiqua (Arikareean, Florida) Elaphe guttata (Blancan, Texas) Elaphe obsoleta (Blancan, Texas) Lampropeltis getulus (Blancan, Kansas) Lampropeltis triangulum (Blancan, Kansas) Pituophis melanoleucus (Blancan, Texas) Stilosoma uetustum (Pliocene, Florida) Diadophis elinorae (Pliocene, Florida) Dryinoides oxyrachis (Miocene, Montana) Paleofarancia. brecispinosa (Pliocene, Florida) NONE NONE
Heterodon plionasicus (Blancan, Kansas)
Heterodon breris (Pliocene, Florida)
Nerodia, eyclopicm or N. rhombifera (Blancan, Texas)
Nerodia erythrogaster (Blancan, Texas)
Regina hillmani (Lower Pliocene, Kansas)
Thamnophis sp. (Hemphillian, Florida) Crotalus atrox (Blancan, Texas) Sistrurus catenatus (Blancan, Texas) Micrurus sp. (Miocene, Nebraska; Pliocene, Florida) Rhineura marslandensis (Miocene, Kansas) Sceloporus undutatus (Blancan, Texas) Ophisaurus uentralis (Miocene, Nebraska; Pliocene, Kansas) Eumeces sp. (Miocene, Florida) Gerrhonotus sp. (Blancan, Texas)
The only substantial evidence of a West Indian influence in the squamate history of Florida is seen in the Miocene Thomas Farm Fauna. The presence of Pseudoepicrates (Auffenberg 1963), Leioce-phalus, a possible iguanine, and a gecko (Estes 1963) documents this influence in the early Miocene. No West Indian squamates except Anolis are present in the fossil record after that time. The Inglis fauna is the next oldest of sufficient size to be examined for the presence of a West Indian influence in Florida. It documents a complete disappearance of the West Indian faunal elements by the early Pleistocene.
Anolis was not found in the Inglis IA fauna, though it should have been if it occurred in Florida at that time. It is common in longleaf pine

and in xeric hammock, thus its absence cannot be attributed to ecological factors. Perhaps Anolis carolinensis had not yet arrived in Florida by the earliest Pleistocene. It is known from Rancholabrean deposits in Florida (Auffenberg 1956).
The modernization of Florida's squamates can easily be seen by comparing the Thomas Farm and Inglis IA faunas. The Thomas Farm fauna includes four boid snakes and two colubrid snakes. Inglis IA includes 23 colubrids, 2 viperids, 1 elapid, and no boids. This change in the squamate fauna is similar to that reported for western North America where domination by colubrids and disappearance of boids (except two species on the west coast) occurs by the end of the Blancan. The Thomas Farm lizards are dominated by iguanids, with at least four species present. An anguid, a gecko, a skink, and Cnemidophorus are also present (Estes 1963). Unlike the Thomas Farm fauna, the Inglis lizard fauna is reduced in comparison to the snakes. This could reflect collecting bias in the Thomas Farm, but more likely indicates a change from nearly equal diversity of lizards and snakes (8:6) in the Miocene to a greater diversity of snakes by Inglis time (4:26), which still is evident today (14:36).
All the subfamilies (family, in the case of Rhineura) represented in Inglis IA are known in North America by the end of the Miocene (E stes 1963; Holman 1970,1979; Berman 1976). Thus, the change in Florida's squamate fauna between the early Miocene (Thomas Farm) and early Pleistocene (Inglis IA) is largely due to dispersal and evolution of an established North American fauna rather than immigration and mixing of North and South American faunas, as is apparent in the mammals from Inglis IA (Webb 1978). The participation of colubrids in the faunal interchanges that began seriously in the late Blancan was apparently one-way south. There is no evidence of South American snakes arriving by Inglis time along with the mammals.
The Inglis fauna documents an important 1 ate Tertiary link between the savannas of Florida and those of the American West. Savannas were widespread in North America during the Miocene (Webb 1977). The snakesAniliodes, Calamagras,Pseudocemophora, andOgmophis, the lizards Ophisaurus, Eumeces, and Cnemidophorus (or a similar form), and the rhineurid amphisbaenians are found in deposits of Miocene age in both the High Plains and in Florida. Thus it is assumed that ecological continuity between the east and west existed in the late Tertiary and was an important factor in the arrival of many forms in the southeastern United States (Blair 1958). Such an east-west corridor would help explain an emerging pattern of survival in the Southeast for some forms that were widespread in the Miocene of the High

Plains. Dryinoides, the rhineurid amphisbaenians, the tortoise genus Stylemys, and the lizard Ophisaurus ventralis are present in the mid-Tertiary of the western United States and have their last surviving members in the Southeast. The rhineurid amphisbaenians are present from the Eocene to the Miocene in the middle and far West (Berman 1976). A single extant species, Rhineura floridana, exists today in peninsular Florida. Stylemys is known from the Eocene to Miocene of the western United States, but one of its most specialized members, the related genusFloridemys, is known from the Miocene of Florida. Ophisaurus ventralis is known from the Miocene of Nebraska and Pliocene of Kansas but is now limited to the Southeast. Dryinoides, from the Miocene of Montana (Auffenberg 1958), is now apparently present in the earliest Pleistocene of Florida.
These and other squamates were members of mid-Tertiary xeric communities that were widespread in North America. They probably survived longer in Florida because the peninsula acted as a refuge during periods of climatic deterioration. Further investigation of the fossil record may reveal other species that fit this pattern. Perhaps such endemics as Stilosoma and Neoseps are relicts of widespread mid-Tertiary faunas.
The present distribution of 15 xeric-adapted North American squamates may reflect the past distribution of xeric communities. Their ranges (taken from Conant 1975) are similar in that they all avoid the area of Tertiary Mississippi E mbayments (Clark and Stearn 1968). Crotaphytus collaris, Heterodon nasicus, Sonora episcopa, Tantilla gracilis, and Crotalus atrox all approach the Southeast from the West, but are now restricted north of the area of most Mississippi Embayments. Cnemidophorus sexlineatus, Sceloporus undulatus, Cemophora coccinea, Lampropeltis calligaster, and Micrurus fulvius are found east and west of the M ississippi River but are absent in much of the embayment area. Eumeces anthracinus, Masticophis flagellum, and Pituophis melanoleucus are completely separated by the Mississippi Basin.
In the early Pleistocene, when the climate was perhaps more arid and sea levels lower, this embayment may have been breached by savanna biotas. But during the late Tertiary the main avenue for east-west connection of xeric faunas was probably across higher latitudes.
A mid-continent corridor, xeric in nature, is thus indicated. It apparently provided an east-west dispersal route for late Tertiary faunas. The presence in Inglis IA of Heterodon nasicus and an undifferentiated Drymarchon corais, suggests that gene flow was at least periodically maintained with western populations until the end of the

Pliocene. The separation of at least three xeric-adapted genera(Ophi-saurus, Sistrurus, and Crotalus) into eastern and western entities by the early Pleistocene may have occurred during isolation when this corridor was closed.
Few species present in the modern squamate fauna of xeric Florida habitats are missing from Inglis IA. The absence of small scincids may be due to screening techniques and is discussed under Paleoecology. Ophisaurus attenuatus, Cnemidophorus sexlineatus, and Opheodrys aestivus, which would be expected in xeric settings, are apparently post-Inglis arrivals. Sceloporus woodi, Ophisaurus compressus, and Heterodon simus all have close relatives in the Inglis fauna, which suggests that speciation of these forms occurred by post-Inglis evolution in peninsular Florida.
Climatic fluctuations during the Pleistocene undoubtedly produced important changes in Florida's squamate fauna. Lowered sea levels during each glacial interval produced broad circum-Gulf corridors of savanna habitat (Auffenberg and Milstead 1965) and probably provided an avenue for the arrival of Ophisaurus attenuatus, Cnemidophorus sexlineatus, and Opheodrys aestivus. During interglacials, higher sea levels would have separated parts of peninsular Florida from the mainland providing the isolation necessary for speciation to occur in Sceloporus undulatus (to S. woodi), Heterodon nasicus (to H. simus), and Ophisaurus ventralis (toO. compressus). Inglis IA represents one of the first glacial intervals and therefore, in a broad sense, the beginning of the Pleistocene mode of herpetofaunal evolution in Florida.

Atchley, W. R., C. T. Gaskins, and D. Anderson. 1976. Statistical properties of ratios 1.
Empirical results. Syst. Zool. 25(2): 137-148. Auffenberg, W. 1955. Glass lizards (Ophisaurus) in the Pleistocene and Pliocene of
Florida. Herpetologica 11(2): 133-136. __ 1956. Additional records of Pleistocene lizards from Florida. Quart. J.
Florida Acad. Sci. 19(2-3): 157-167. __ 1958. A new genus of colubrid snake from the Upper Miocene of North
America. Am. Mus. Novit. 1874: 1-16.
__ 1963. The fossil snakes of Florida. Tulane Stud. Zool. 10(3): 131-216.
_, and W. W. Milstead. 1965. Reptiles in the Quaternary of North America. Pp.
557-568 in The Quarternary of the United States, VII Congress. Int. Assoc. Quar-
ternary Research, Princeton Univ. Press. Barr, A. J., J. H. Goodnight, J. P. Sail, and J. T. Helwig. 1976. A user's guide to SAS 76.
SAS Institute Inc., Raleigh. 329 p. Behrensmeyer, A. K. 1975. The taphonomy and paleoecology of Pleistocene vertebrate
assemblages east of Lake Rudolf, Kenya. Bull. Mus. Comp. Zool. 146(11): 473-578. Berman, D. S. 1976. A new amphisbaenian (Reptilia: Amphisbaenia) from the
Oligocene-Miocene John Day Formation, Oregon. J. Paleon. 50(1): 165-174. Blair, W. F. 1958. Distributional patterns of vertebrates in the southern United States in
relation to past and present environments. Pp. 433-468 in Zoogeography. C. L.
Hubbs (ed.). Amer. Assoc. Adv. Sci. Publ. No. 51. Blanchard, F. W. 1921. A revision of the kingsnakes: Genus Lampropeltis. Bull. U.S.
Natl. Mus. No. 114.
Blaney, R. M. 1977. Systematics of the common kingsnake, Lampropeltis getulus (Linnaeus). Tulane Stud. Zool.-Bot. 19(3-4): 47-103.
Branson, B. A., and E. C. Baker. 1974. An ecological study of the QueensnakeRegina septemvittata (Say) in Kentucky. Tulane Stud. Zool.-Bot. 18(4): 153-171.
Brattstrom, B. H. 1953. Records of Pleistocene reptiles and amphibians from Florida. Quart. J. Florida Acad. Sci. 16(4): 243-248.
__1955a. Pliocene and Pleistocene amphibians and reptiles from southeastern
Arizona. J. Paleon. 29(1): 150-154.
__ 1955b. Records of some Pliocene and Pleistocene reptiles and amphibians
from Mexico. Bull. So. California Acad. Sci. 54(1): 1-4.
-- 1964. Evolution of the pit vipers. Trans. San Diego Soc. Nat. Hist. 13(11):
__ 1967. A succession of Pliocene and Pleistocene snake faunas from the High
Plains of the United States. Copeia 1967(1): 188-202. Bullock, R. E., and W. W. Tanner. 1966. A comparative osteological study of two species
of Colubridae (Pituophis and Thamnophis). Brigham Young Univ. Sci. Bull. 8(3):
Camp, C. L. 1923. Classification of the lizards. Bull. Am. Mus. Nat. Hist. 48: 289-491. Carr, A. F. 1940. A contribution to the herpetology of Florida. Univ. Presses of Florida, Gainesville, 118 p.
Clark, T. H and C. W. Stearn. 1968. Geological evolution of North America. Ronald
Press Co., New York, 569 p. Conant, R. 1975. A field guide to reptiles and amphibians of eastern and central North
America. Houghton-Mifflin Co., Boston. 429 p. Dixon, W. J., and M. B. Brown. 1979. BMDP biomedical computer programs, P-series.
Univ. California Press, Berkeley, 880 p.

Dowling, H. G. 1958. Pleistocene snakes of the Ozark Plateau. Am. Mus. Novit. 1882:1-9. -_, and W. E. Duellman. 1974. Systematic herpetology: A synopsis of families
and higher categories. HISS Publ. Herpetol. 7:1-240. Dunn, E. R. 1928. A tentative key and arrangement of the American genera of Colubri-
dae. Bull. Antivenin Inst. Amer. 2(1): 18-24. Edgren, R. A. 1952. A synopsis of the snakes of the genus Heterodon, with the diagnosis
of a new race of Heterodon nasicus Baird and Girard. Nat. Hist. Misc., Chicago
Acad. Sci., 112:1-4.
Estes, R. 1963. Early Miocene salamanders and lizards from Florida. Quart. J. Florida
Acad. Sci. 26(3): 235-256. _, T. H. Frazzetta, and E. E. Williams. 1970. Studies on the fossil snake
Dinilysia patagonica Woodward: Part 1. Cranial morphology. Bull. Mus. Comp.
Zoo'l. 140(2): 25-73.
Etheridge, R. 1961. Late Cenozoic glass lizards (Ophisaurus) from the southern Great
Plains. Herpetologica 17(3): 179-186. Franz, R. 1977. Observations on the food, feeding behavior and parasites of the striped
swamp snake (Regina alleni). Herpetologica 33(1): 91-94. Gehlbach, F. R., and J. A. Holman. 1974. Paleoecology of amphibians and reptiles from
Pratt Cave, Guadaloupe Mountains National Park, Texas. Southwestern Nat. 19(2):
Guilday, J. E. 1962. The Pleistocene local fauna of the Natural Chimneys, Augusta County, Virginia. Ann. Carnegie Mus. 36(9): 87-122.
__H. W. Hamilton, E. Anderson, and P. W. Parmalee. 1978. The Baker Bluff
Cave Deposit, Tennessee and the late Pleistocene faunal gradient. Bull. Carnegie Mus. Nat. Hist. 11:1-68.
Gut, H. J. and C. E. Ray. 1963. The Pleistocene vertebrate fauna of Reddick, Florida. Quart. J. Florida Acad. Sci. 26(4): 315-328.
Heyer, W. R. 1978. Systematics of the/sew,s group of the frog genus Leptodactylus (Amphibia, Leptodactylidae). Nat. Hist. Mus. Los Angeles Co. Sci. Bull. 29:1-85.
Holman, J. A. 1958. The Pleistocene herpetofauna of Saber-tooth Cave, Citrus Co., Florida. Copeia 1959(4): 276-280.
__ 1959a. Amphibians and reptiles from the Pleistocene (Illinoian) of Willis-ton, Florida. Copeia 1959(2): 96-102.
__1959b. A Pleistocene herpetofauna near Orange Lake, Florida. Herpetologica 15(3): 121-125.
__ 1962. Additional records of Florida Pleistocene amphibians and reptiles.
Herpetologica 18(2): 115-119. __1963. Late Pleistocene amphibians and reptiles of the Clear Creek and Ben
Franklin Local Faunas of Texas. Southern Methodist Univ. J. Grad. Res. Cen. 31(3):
__ 1964. Fossil snakes from the Valentine Formation of Nebraska. Copeia
1964(4): 631-637.
__ 1965. A late Pleistocene herpetofauna from Missouri. Trans. Illinois Acad.
Sci. 58(3): 190-194.
__1967. A Pleistocene herpetofauna from Ladds, Georgia. Bull. Georgia Acad.
Sci. 1867(3): 154-166.
__ 1968a. Upper Pliocene snakes from Idaho. Copeia 1968(1): 152-158.
__ 1968b. A Pleistocene herpetofauna from Kendall Co., Texas. Quart. J.
Florida Acad. Sci. 31(3): 165-172. __ 1969a. The Pleistocene amphibians and reptiles of Texas. Michigan State
Univ. Publ. Mus. 4(5): 163-192.

__1969b. Herpetofauna of the Slaton local fauna of Texas. Southwestern Nat.
14(2): 203-212.
__ 1970. Hepetofauna of the Wood Mountain Formation (Upper-Miocene) of
Saskatchewan. Canadian J. Earth Sci. 7(5): 1317-1325. __1971. Herpetofauna of the Sandahl Local Fauna (Pleistocene: Illinoian) of
Kansas. Univ. Michigan Contrib. Mus. Paleon. 23(22): 349-355. __ 1972. Herpetofauna of the Kanopolis Local Fauna (Pleistocene: Yarmouth)
of Kansas. Michigan Academician 5(1): 87-98. _1973. A new Pliocene snake, germs Elaphe, from Oklahoma. Copeia 1973(3):
__1974. A late Pleistocene herpetofauna from southwestern Missouri. J. Herp.
8(4): 343-346.
__1975. Herpetofauna of the Wakeeney Local Fauna (Lower Pliocene: Claren-
donian) of Trego Co., Kansas. Univ. Michigan Pap. Paleon. 12: 49-66.
__1976. Snakes of the Split Rock Formation (Middle Miocene), Central Wyoming. Herpetologica 32(4): 419-426.
__ 1977a. The Pleistocene (Kansan) herpetofauna of Cumberland Cave, Maryland. Ann. Carnegie Mus. 46(11): 157-172.
__ 1977b. Upper Miocene snakes (Reptilia, Serpentes) from southeastern
Nebraska. J. Herp. 11(3): 323-335.
__ 1978. The late Pleistocene herpetofauna of Devil's Den Sinkhole, Levy Co.,
Florida. Herpetologica 34(2): 228-237.
__ 1979. A review of North American Tertiary snakes. Mus. Michigan State
Univ. Publ. Mus. Paleon. Series 1(6): 203-260.
__1981. A review of North American Pleistocene snakes. Michigan State
Univ. Publ. Mus. Paleo. Series 1(7): 263-306.
Jackson, J. F. 1973. Distribution and population phenetics of the Florida scrub lizard, Sceloporus woodi. Copeia 1973(4): 746-761.
Johnson, R. G. 1955. The adaptive and phylogenetic significance of vertebral form in snakes. Evolution 9: 367-388.
Klauber, L. M. 1972. Rattlesnakes. 2nd ed. Univ. California Press, Berkeley and Los Angeles, 1533 p.
Klein, J. G. 1971. The ferungulates of the Inglis IA Local Fauna, early Pleistocene of
Florida. M.S. Thesis, Univ. Florida, Gainesville, 115 p. Larsen, J. R., and W. W. Tanner. 1975. Evolution of the sceloporine lizards (Iguanidae).
Great Basin Nat. 35(1): 1-20. Lindsay, E. H., and N. T. Tessman. 1974. Cenozoic vertebrate localities and faunas in
Arizona. J. Arizona Acad. Sci. 9(1): 3-24. Martin, R. A. 1974. Fossil vertebrates from the Haile XIVA fauna, Alachua County,
Florida. Pp. 100-113 in Pleistocene Mammals of Florida. S. D. Webb (ed.). Univ.
Presses of Florida, Gainesville. __ 1979. Fossil history of the rodent genus Sigmodon. Evolutionary Monographs, Univ. Chicago. 2: 1-36. Marx, H., and G. B. Rabb. 1972. Phyletic analysis of fifty characters of advanced snakes.
Fieldiana: Zool. 63: 1-321. McConkey, E. H. 1954. A systematic study of the North American lizards of the genus
Ophisaurus. Am. Midi. Nat. 51(1): 133-171. McDowell, S. B., and C. M. Bogert. 1954. The systematic position of Lanthonotus and the
affinities of the Anguinomorphan lizards. Bull. Am. Mus. Nat. Hist. 105(1): 1-142. Meszoely, C. A. M. 1970. North American fossil anguid lizards. Bull. Mus. Comp. Zool.
139(2): 87-149.

Mount, R. H. 1965. Variation and systematics of the scincoid lizard, Eumeces egregius
(Baird). Bull. Florida State Mus., Biol. Sci. 9(5): 184-213. Myers, C. W. 1965. Biology of the ringneck snake, Diadophis punctatus, in Florida. Bull.
Florida State Mus., Biol. Sci. 10(2): 43-90. __. 1967. The pine woods snake, Rhaddnaea flavilata (Cope). Bull. Florida State
Mus., Biol. Sci. 11(2): 47-97. -- 1974. The systematics of Rhadinaea (Colubridae), a genus of New World
snakes. Bull. Am. Mus. Nat. Hist. 153(1): 1-262. Neill, W. T. 1964. Taxonomy, natural history, and zoogeography of the rainbow snake,
Farancia erthrogramma. Am. Midi. Nat. 71(2): 257-295. Oelrich, T. M. 1956. The anatomy of the head of Ctenomurapectinata (Iguanidae). Univ.
Michigan Misc., Publ. Mus. Zool. 94:1-122. Peters, J. A. 1953. A fossil snake of the genus Heterodon from the Pliocene of Kansas. J.
Paleon. 27(3): 328-331. Piatt, D. R. 1969. Natural History of the hognose snakes Heterodon platyrhinos and
Heterodon nasicus. Univ. Kansas Publ. Mus. Nat. Hist. 18(4): 253-420. Robertson, J. S. 1976. Latest Pliocene mammals from Haile XVA, Alachua County,
Florida. Bull. Florida State Mus., Biol. Sci. 20(3): 111-186. Rogers, K. l. 1976. Herpetofauna of the Beck Ranch Local Fauna (Upper Pliocene:
Blancan) of Texas. Michigan State Univ. Publ. Mus. Paleon. Series 1(5): 167-200. Rojas, N. N., and J. S. Godley. 1979. Tooth morphology in crayfish-eating snakes, genus
Regina. Abstract, Joint Ann. Mtg. Herpetologists League and Soc. Study Amphibians Reptiles, 12-16 August 1979. Rossman, D. A. 1963a. The colubrid genus Thamnophis: A revision of the sauritus
group. Bull. Florida State Mus., Biol. Sci. 7(3): 99-178. --1963b. Relationships and taxonomic status of the North American natricine
snake generaLiodytes, Regina and Clonophis. Louisiana State Univ. Occ. Pap. Mus.
Zool. 29:1-29.
Schmidt, K. P., and W. L. Necker. 1936. The scientific name of the American smooth
green snake. Herpetologica 1(2): 63-64. Smith, H. M., R. B. Smith, and H. L. Sawin. 1977. A summary of snake classification
(Reptilia, Serpentes). J. Herp. 1(2): 115-121. Telford, S. R. 1966. Variation among the southeastern crowned snakes, genus Tantilla.
Bull. Florida State Mus., Biol. Sci. 10(7): 261-304. Underwood, G. 1967. A contribution to the classification of snakes. British Mus. Nat.
Hist. Publ. 653:1-179. Van Devender, T. R., K. B. Moodie, and A. H. Harris. 1976. The desert tortoise(Gopherus
agassizi) in the Pleistocene of the northern Chihuahuan Desert. Herpetologica
32(3): 298-304.
Van Devender, T. R., A. M. Phillips III, and J. I. Mead. 1977. Late Pleistocene reptiles and small mammals from the lower Grand Canyon of Arizona. Southwestern Nat. 22(1): 49-66.
Weaver, W. G. 1965. The cranial anatomy of the hog-nosed snakes (Heterodon). Bull. Florida State Mus., Biol. Sci. 9(7): 275-304.
Webb, S. D. 1974. Pleistocene mammals of Florida. Univ. Presses of Florida, Gainesville. 270 p.
__ 1977. A history of savanna vertebrates in the New World. Part 1: North
America. Ann. Rev. Ecol. Syst. 8: 355-380. __ 1978. A history of savanna vertebrates in the New World. Part II: South
America and the great interchange. Ann. Rev. Ecol. Syst. 9: 393-426. _, B. J. MacFadden, and J. A. Baskin. 1981. Geology and paleontology of the
Love Bone Bed from the late Miocene of Florida. Amer. J. Sci. 281: 513-544.

Williams, K. L., and L. D. Wilson. 1967. A review of the colubrid snake genus Cemophora
Cope. Tulane Stud. Zool. 13(4): 103-124. Wilson, R. L. 1968. Systematics and faunal analysis of a Lower Pliocene vertebrate
assemblage from Trego Co., Kansas. Univ. Michigan Contrib. Mus. Paleon. 22(7):
Wolff, R. G. 1975. Sampling and sample size in ecological analysis of fossil mammals. Paleobiology 1(2): 195-204.

APPENDIX 1 Lizard Skeletons Examined
Anniella pulchra (1) Anguidae
Celestus contains (1) C. crusculus (2)
Gerrhonotus multicarinatus (3) Ophisaurus apodus (2), 0. attenuatus (7) 0. compressus (5) O. ventralis (14)
Aristelliger praesignis (2) Coleonyx variegatus (2) Hemidactylus garnoti (2) if. turcicus (2) Gekko gecko (3)
Phyllodac.tylus tuberculosus (2) Sphaerodactylus cinereus (1) Thecadactylus rapicaudus (2)
Heloderma horridum (2) //. suspectum (2)
Amblyrynchus subcriMatus (2) Anolis bimaculatus (1) A. carolinensis (1) A distichus (1) A equestris (2) A garmani (1) A scriptus (1) Basiliscus vittatus (2) Callisaurus dracinoides (2) Corythophanes cristatus (1) Crotaphytus callaris (2) Ctenosaura pectinata (1) Cyclura carinata (1) C. cornuta (1) Dipsosaurus dorsalis (1) Holbrookia maculata (2) Iguana iguana (1) Lewcephalus carinatus (3) Phrynosoma cornutum (2) f. solare (2) Sceloporus clarki (3) 5. cyanogenys (1)
S. jarrovi (1) & magister (1) S. occidentalis (2) S. poinsetti (1) S. scalaris (1) S. undulatus (7) S. woodi (8) f/to, stansburiana (2)
Eumeces brevilineatus (3) Z?. callicephallus (1) i?. copei (1) -E. egregius (3) Z?. fasciatus (2)
(/tteerti (2) A1, inexpectatus (4) Z?. laticeps (4) Z?. multivirgatus (2) i?. obsoletus (3)
parvulus (1) Z?. septentrionalis (3) Z?. skiltonianus (2)
tetragra,mmus (1) Leilopisma cherrei (1) Mabuya. mabouya (2) Neoseps reynoldsi (2) Scincella laterale (2)
Ameiva am,eiva (2)
A griswaldi (2)
A. quadrolineata (3)
Cnemidophorus sexlineatus (6)
C. sonorae (2)
C. Tigris (2)
Klauberina riversiana (2) Lepidophyma flavimaculatum (2) Xantusia vigilis (2)
Xenosaurus platyveps (4)

APPENDIX 2 Snake Skeletons Examined
Alsophis cantherigerus (2) Antillophis parvifrons (1) Arizona elegans (4) Carphophis amoenus (3) Cemophora coccinea (6) Coluber constrictor (6) Diadophis punctatus (7) Drymarchon corais (6) Drymobius margartiferus (1) Elaphe guttata (6) E. obsoleta (8)
vulpina (4) Farancia abacura (6) jP. erythrogramma (3) Heterodon nasiais (3) J. platyrhinos (8) JY. simus (6) Hypsirynchus ferox (1) Lampropeltis calligaster (5) L. getulus (9) L. triangulum (5) Leptodeira annulata (1) L. septentrionalis (1) Lystrophis dorbigny (1) Masticophis flagellum (6) A/, taeniatus (2) Nerodia eyclopion (6) JV. erythrogaster (3) AT. fasciata (6) JV. rhombifera (2) AT. sipedon (2) JV. taxispilota (4) Opheodrys aestivus (4) 0. vernalis (4) Oxybelis fulgidus (1) Phyllorynchus decurtatus (2) Pituophis melanoleucus (7) Pseudoeryx plicatilis (1) Regina alleni (4) i2. grahami (2)
rigida (3) JS. septemvittata (2) Rhadinaea decorata (1) R. flavilata (4) Rhinoceilus lecontei (3) Salvadora hexalepis (3) Scaphiodontophis annulatus (1) Seminatrix pygea (5) Si6ow nebulata (2)
Spilotes pullatus (2) Stilosoma extenuatum (3) Storeria dekayi (4) S. occipitomaculata (2) Tantilla. atriceps (1) T. planiceps (1) T. reJicta (5) T. semicincta (1) r. taeniata (1) Thamnophis cyrtopsis (2) 71. marcianus (2) T. melanogaster (2) r. proximus (4) T. sauritus (4) T. sirtalis (4) Trimorphodon tau (2) Tropidoclonion lineatum (1)
Micrurus a/finis (1) M. fulvius (8)
Agkistrodon contortrix (3)
A piscivorous (6)
Bothrops asper (2)
J3. schlegelii (1)
Crotalus adamanteus (10)
C. afrw (7)
C. horridus (5)
C. molossus (2)
C. viridis (2)
Sistrurus catenatus (4)
S. miliarius (7)

3-2. Palatines of selected Recent North American snakes: (A) Elaphe guttata. (B) Neriitlitifiiticiata. ((') Elt: )>!,< nb.iuhln. (I)) Si-nut ia r,/,-!,,,,;,,,,. (V,) Law prapvll in

meylan: incus ia squamates
3-3. pterykoids view. Coluber rouxtrirt,,,- (i) ventral view. (.1) dorsal view. Masti.....hi. fin,,-It,- 1k1
ventral view, (l) dorsal view. Si rati in time iota (m) ventral view. Fh mucin abacura (n) ventral view. all x3 (each scale = 5 mm).

bulletin florida state museum

meylan: inglis ia squamates
orsfl.-.-lf.iUm-iil.Xoi-lhAm [OltlmtiimnijlurU
can snakes: (AH 'ti rphtiphis aiiim-in. t. (D)Sow.ra cpismpa. iE) Stiltistin >ts.(U)Cfmiiphnm cwcinea.il) fil<>><

Number of teeth and condition of Meckle's groove in the dentary of selected North American snakes. A zero in the right hand column indicates that Meckle's groove is open to the symphysis.
Meckle's Groove
Species N Number of Teeth Reduced to Slit at Tooth Number Closed at Tooth Number
Arizona elegans 2 14-15 7 (1 case), not reduced (1 case) 6
Cemophora coccinea 1 10 not reduced 4
Coluber constrictor 7 19-22 not reduced 8-9
Drymarchon corais 5 17-19 7 (1 case), not reduced (4 cases) 5-8
Elaphe guttata 7 19-22 7-10 (6 cases), not reduced (3 cases) 4-6
Elaphe obsoleta 7 21-23 10-12 3-6
Elaphe vulpina 2 18-19 7 3-4
Farancia abacura 6 20-23 not reduced 4-6
Heterodon platyrhinos 7 16-18 3-4 0
Heterodon simus 3 14-15 3-4 3-0
Lampropeltis calligaster 2 14-15 6 4
Lampropeltis getulus 4 13-17 7 (2 cases), not reduced (2 cases) 3-5
Lampropeltis triangulum 2 16-17 not reduced 4-5
Masticophis flagellum 5 19-23 9 (1 case), not reduced (4 cases) 6-9
Nerodia cyclopion 11 19-23 not reduced 7-8
Nerodia erythrogaster 1 28 not reduced 6
Nerodia fasciata 9 26-29 not reduced 6-9
Nerodia rhombifera 1 25 not reduced 8
Nerodia taxispilota 4 22-26 not reduced 6-9
Opheodrys aestivus 3 22-24 12 (1 case), not reduced (2 cases) 6-8
Pituophis melanoleucus 6 17-19 7-9 3-4
Regina alleni 5 27-30 11-12 (4 cases), not reduced (1 case) 3-8
Regina grahami 1 34 18 0
Regina rigida 2 29 10-11 0-5
Regina septemvittata 4 23-27 5-9 0-4

Rhimoceilus lecontei 6 13-16
Seminatrix pygea 6 20
Thamnophis eques 1 31
Tha-mnophis marcianus 1 29
Thamnophis melanogaster 2 26-27
Thamnophis sauritus 4 33-39
Thamnophis sirtalis 5 26-30
Trimorphodon tau 2 16-17
Agkistrodon mokeson 5 15-17
Agkistrodon piscivorous 7 16-19
Crotalus adamanteus 8 8-10
Crotalus atrox 4 9-10
Sistrurus miliarius 2 9-10
6-7 (2 cases), not reduced (4 cases)
not reduced
not reduced
not reduced
not reduced
not reduced
not reduced
3 (2 cases), not reduced (3 eases) 2-4 (6 cases), not reduced (1 case) 1-5 (6 cases), not reduced (2 cases) 1-2
not reduced

Contributions to the BULLETIN OF THE FLORIDA STATE MUSEUM, BIOLOGICAL SCIENCES SERIES, may be in any field of biology. Manuscripts dealing with natural history of systematic problems involving the southeastern United States or the New World tropics are solicited especially. Manuscripts should be of medium length circa 35 to 200 pages (10,500-16,000 words). Examination for suitability is made by an Editorial Board.
The BULLETIN is distributed worldwide through institutional standing orders and exchanges. It is considered the responsibility of the author to distribute his paper to all interested individuals. To aid in this the author(s) receivers) 50 copies free, and may purchase additional separates at cost, if ordered when galley proof is returned. The author is also responsible for any charges incurred for alterations made by him on galley or page proofs. The Museum will send an invoice to the author for this amount upon completion of publication.
Contributors should consult recent numbers of the BULLETIN for preferred style and format. Highly recommended as a guide is the CBE Style Manual, 3rd Edition, 1972 (Amer. Inst. Biol. Sci.). More specific instructions are available upon request.
MSS must be submitted in triplicate (NO ONIONSKIN, XEROX COPIES PLEASE) and satisfy the following minimal requirements: They must be typewritten, double-spaced throughout (INCLUDING tables, figure captions, and Literature Cited), with triple-spacing around all headings, on one side of numbered sheets of standard (8-1/2 x 11 in.) bond paper, with 1 -1/2 in. margin on left and 1 in. margins on other three sides. Each table must be typed on a separate sheet of paper; figure captions should be typed run-on. All illustrations are referred to as figures. They must comply with the following standards: Photographs should be sharp, with good contrast, and printed on glossy paper. Drawings must be made with dense black waterproof ink on quality paper or illustration board and have a cover sheet. All lettering must be medium weight, sans-serif type (e.g. Futura Medium, News Gothic) in cutout, dry transfer, or lettering guide letters. Make allowance so that after reduction no lowercase letter will be less than 1 mm high (2 mm is preferred) nor any capital letter greater than 5 mm high. The maximum size for illustrations is 8-5/8 x 14 in. (twice typepage size); illustrations should not be less than typepage width (4-5/16 in.). Designate the top of each illustration and identify on the back with soft pencil by author's name, MS title, and figure number.
All manuscripts not submitted in BULLETIN format will be returned to the author for retyping.
ManuxeripLt and all editorial matters should, he addressed to: Managing Editor of the BULLETIN Florida State Museum University of Florida Gainesville, FL 32611

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs