Front Cover
 Back Cover

Group Title: Bulletin of the Florida Museum of Natural History
Title: Ungulates of the Toledo Bend local fauna (late Arikareean, early Miocene), Texas coastal plain
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00087412/00001
 Material Information
Title: Ungulates of the Toledo Bend local fauna (late Arikareean, early Miocene), Texas coastal plain
Series Title: Bulletin of the Florida Museum of Natural History ; volume 42, no. 1
Physical Description: 80 p. : ill., map ; 23 cm.
Language: English
Creator: Albright, L. Barry III ( Lynn Barry ), 1957-
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 1999
Copyright Date: 1999
Subject: Ungulates -- Texas   ( lcsh )
Ungulates, Fossil -- Texas   ( lcsh )
Paleontology -- Miocene   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references (p. 75-80).
Language: Abstract in English and Spanish.
Statement of Responsibility: L. Barry Albright III.
 Record Information
Bibliographic ID: UF00087412
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: oclc - 42819890
issn - 0071-6154 ;

Table of Contents
    Front Cover
        Front cover 1
        Front cover 2
        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
    Back Cover
        Page 80
Full Text


of the



L. Barry Albright III

Volume 42 No. 1 pp. 1-80 1999



published at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in
any one calendar year.


Communications concerning purchase or exchange of the publications and all manuscripts should be
addressed to: Managing Editor, Bulletin; Florida Museum of Natural History; University of Florida;
P. O. Box 117800, Gainesville FL 32611-7800; U.S.A.

This journal is printed on recycled paper.

ISSN: 0071-6154


Publication date: 8-12-99

Price: $ 7.50


L. Barry Albright III'


The Toledo Bend Local Fauna includes the most diverse assemblage of mammals yet reported
from the earliest Miocene Gulf Coastal Plain. Faunal affinities with the Buda Local Fauna of northern
Florida, as well as with faunas of the northern Great Plains and Oregon, suggest a "medial" to late
Arikareean age. Of 26 mammalian taxa represented, the 17 ungulates are discussed in this report; the
carnivores, small mammals, and lower vertebrates are discussed elsewhere. All ungulate taxa are
browsers rather than grazers, and many have not previously been recorded from the Gulf Coastal Plain.
Riparian forms such as anthracotheres and tapirs account for a large percentage of mammalian remains
recovered, although horses, rhinoceroses, and a small, new species of protoceratid, Prosynthetoceras
orthrionanus sp. nov., are also common. Other ungulates include a small chalicothere, a new small
rhinoceros, Gulfoceras westfalli gen. et sp. nov., and a giant entelodont. The abundance of
protoceratids, tapirs, anthracotheres, and rhinoceroses, the absence of oreodonts, and the rarity of
camelids at Toledo Bend contrasts with Arikareean faunas in Florida where protoceratids and
anthracotheres are absent, tapirs and rhinos are extremely rare, oreodonts are present, and camelids are
relatively common.


La fauna local de Toledo Bend incluye el ensemble mis divers de mamlferos jamAs reportado
para el Mioceno mAs temprano de la Planicie Costera del Golfo, en Texas. Las afinidades faunisticas
con la fauna local de Buda, del norte de la Florida y con la de las Grandes Planicies del Norte y de
Oregon, sugieren una edad del Arikareano medio a tardio. De los 26 tax6nes de mamfferos
representados, los 17 ungulados se discuten en el present report, en otros trabajos se discuten los
carnivores, pequefios mamfferos, y vertebrados primitives. Todos los ungulados son ramoneadores mas
que pasteadores y muchos no han sido previamente descritos para la Planicie Costera del Golfo. Un
gran porcentaje de los restos de mamfferos recuperados corresponde a formas ribereflas, tales como
anthracotheros y tapires, aunque tambidn son comunes los caballos, rinocerontes y una nueva especie
de protoceratido pequeflo, Prosynthetoceras orthrionanus sp. nov. Otros ungulados incluyen un
pequeflo chalicothero, un nuevo rinoceronte, Gulfoceras westfalli gen. et sp. nov., y un entelodonte
gigante. La abundancia de protoceratidos, tapires, anthracotheros y rinocerontes; asi como la ausencia
de oreodontos y la rareza de camelidos en la curva de Toledo contrastan con las faunas Arikareanas en
Florida donde los protoceratidos y anthracotheros estAn ausentes, los rinocerontes y tapires son
extremadamente raros, los oreodontes estAn presents y los camelidos son relativamente comunes.

I Currently, Curator of Paleontology, Museum of Northern Arizona, 3101 N. Fort Valley Rd., Flagstaff AZ 86001; formerly, Postdoctoral
Research Associate, Vertebrate Paleontology, Florida Museum of Natural History, University of Florida, Gainesville FL 32611.

ALBRIGHT, L. BARRY III. 1999. Ungulates of the Toledo Bend Local Fauna (Late Arikareean,
Early Miocene), Texas Coastal Plain. Bull. Florida Mus. Nat. Hist. 42(1):1-80.



Introduction .......................................................................................................................................... 2
Abbreviations ............................ ................... ............................................................. 5
A know ledgm ents............................................................................................ ............................ 5
Geographic and Geological Setting................................................................... ....... ............ 6
Sedimentological and Taphonomic Context .................................................... .................... 6
System atic Paleontology ................................................................................... ........................... 9
A ge and C correlation ........................................................................................ ............................. 71
Sum m ary .............................................................................................................. ........ ............... 73
L literature C ited ............................................................................................... ............. ............... 75


The most diverse late Arikareean assemblage yet recorded from the Gulf
Coastal Plain, the Toledo Bend Local Fauna, was recovered during summer months
of 1989-90 from a submerged site in a canal that connects the Toledo Bend
Reservoir with the Sabine River in Newton County, Texas. A detailed description
of the site is provided in previous papers that focused on the lower vertebrate
(Albright, 1994) and non-ungulate mammalian taxa (Albright, 1996), and on a
comparison with other early Miocene faunas of the Gulf Coastal Plain and their
relationship with faunas of the northern Great Plains as a function of dispersal and
endemism (Albright, 1998a). This report, the last of a series on the Toledo Bend
Local Fauna, provides the systematic account of all ungulates recovered with the
exception of the tapir found there. The latter taxon, determined to represent a new
genus, is described and discussed in Albright (1998b). Ungulates in addition to the
tapir include two horses, three rhinoceroses, a chalicothere, an entelodont, three
peccaries, an anthracothere, a camelid, two protoceratids, a hypertragulid, and an
unidentified small artiodactyl. All are low-crowned browsing forms. Table 1 is an
updated faunal list of all vertebrate taxa currently comprising the Toledo Bend
Local Fauna.
Prior to study of the Toledo Bend Local Fauna, nearly all knowledge
regarding the Arikareean Land Mammal "Age" for the Gulf Coastal Plain was based
on a handful of assemblages in Florida such as the Buda, SB-1A, Franklin
Phosphate Pit No. 2, Brookesville, White Springs, and Cow House Slough local
faunas (see Albright, 1998a, and references within). Current study of the newly
discovered Brookesville 2 Local Fauna by Florida Museum of Natural History
personnel will add considerably to our understanding of this interval in this region
(G. Hayes, pers. comm.). Localities in Texas, such as the Cedar Run Local Fauna,
also have yielded mammals indicative of the Arikareean, yet the quality and


Table 1. Faunal list of all early Miocene vertebrates from the Toledo Bend Local Fauna. Compiled and
updated from Albright (1991, 1994, 1996, and 1998) and Manning, (1990).

Lepisosteus sp.
Atractosteus spatula
Ictalurus punctatus
Micropterus sp.
?Centropomus sp.
Batrachosauroides sp.
Dermatemys sp.
Emydid, small sp.
Emydid, large sp.
Hesperotestudo sp.
Peltosaurus sp.
Anilioides nebraskensis
Erycinaeid boid
Alligator olseni
Gavialosuchus sp.
Insectivora sp. indet.
Protospermophilus sp.
Neatocastor hesperus
Eomyid or cricetid indet.

Mammalia (cont.)
Proheteromys toledoensis
Proheteromys sabinensis
Texomys sp.
?Palaeogale sp.
Daphoenodon notionastes
?Miohippus sp.
Anchippus texanus
Moropus sp.
Nexuotapirus marslandensis
Gulfoceras westfalli n. gen. et sp.
Diceratherium annectens
Diceratherium armatum
?Dinohyus sp.
"Cynorca" social
?Floridachoerus olseni
?Hesperhys sp.
Arretotherium acridens
Nothokemas sp.
Prosynthetoceras orthrionanus n. sp.
Prosynthetoceras texanus
Nanotragulus sp.
Small artiodactyl indet.

amount of material, as well as the diversity of taxa, is substantially less when
compared with sites in Florida. The Toledo Bend Local Fauna is significant,
therefore, because it finally provides a view into the Arikareean of the western
region of the Gulf Coastal Plain that is based on an abundance of high quality
fossils representing a wide diversity of species. As a consequence, new views
regarding this distinct biogeographic province during the earliest Miocene,
unrecognized when seen solely through the Florida "window," are now
Of additional importance is the interval of time represented by the Toledo
Bend Local Fauna and by the correlative assemblages in Florida noted above.
Recent magnetostratigraphic studies in northwestern Nebraska have determined that
strata of the Arikaree Group (in which the defining faunas of the Arikareean Land
Mammal "Age" originated) spanning the interval from about 27 to 22 Ma are
apparently missing (MacFadden and Hunt, 1998). Importantly, it is the later part of
this interval, the early part of the late Arikareean, thought to be sampled by these


Gulf Coastal Plain assemblages (Albright, 1998a). It may be, therefore, that the
"endemic" early Miocene Gulf Coast fauna is, to some degree, an artifact of
sampling a temporal interval not previously represented in the North American
The Toledo Bend Local Fauna differs from other early Miocene Coastal Plain
faunas in a number of aspects. First is the unusual abundance of aquatic and
riparian taxa such as fish, turtles, alligators, tapirs, anthracotheres, and
protoceratids. Second, many of its members, such as the tapir and anthracothere,
were previously known only from Arikareean faunas of the northern Great Plains.
Third, many of its members are not shared with comparably aged faunas in Florida
such as those noted above. Finally, grazing ungulates are absent.
These factors can be attributed, in part, to the preservation of a
paleoenvironment that apparently is rarely preserved in the early Miocene of North
America-that of a subtropical to tropical coastal plain forest with a predominant
fluvial component. In turn, the preservation of this environment was largely due to
sedimentological influences imposed by a paleo-Mississippi River system that was
migrating from east Texas to Louisiana during the late Oligocene-early Miocene
(Galloway et al., 1986, 1991). Northern Florida, however, which undoubtedly
shared similar climatic conditions during this interval, was not subjected to similar
sedimentological influences, as is reflected in preservation there that is essentially
limited to karst fissure infilling rather than strictly fluvial processes. In fact, the
extent to which the Florida Platform was subaerially exposed during this interval is
still poorly understood (see, e.g., Huddlestun, 1993, and Scott, 1997). This
preservational bias may account for some of the faunal differences between the two
areas, but ecological factors and distance from, or proximity to, the northern Great
Plains may have contributed as well. Furthermore, vaguely defined temporal
inequivalencies that may exist between the Toledo Bend site and those in Florida
are difficult to resolve because of the lack of accurate geochronological control
other than that provided by the mammals themselves. There are no
radioisotopically dateable volcanic beds in appropriately aged strata of the Texas
Coastal Plain or Florida, and magnetostratigraphy is also of limited use, as few, if
any, of these sites occur in an amenable stratigraphic context; sites are either
underwater or they exist as widely distributed isolated outcrops or karst fissure-fill
deposits. Despite these limitations, the Toledo Bend assemblage still provides
important new information toward an understanding of the vertebrate fauna that
inhabited the Gulf Coastal Plain during the earliest Miocene, as well as toward the
paleoecology and biogeography of the region. Further details of the paleoecology
are provided in Albright (1991, 1994).



Fossils from the Toledo Bend site are curated in the vertebrate paleontological collection of the
Louisiana State University Museum of Geoscience, Baton Rouge, abbreviated throughout as LSUMG-
V. Other abbreviations used are as follows: AC, Amherst College Museum, Amherst, Massachusetts;
AMNH, American Museum of Natural History, New York; ANSP, Academy of Natural Sciences,
Philadelphia; CM, Carnegie Museum, Pittsburgh; F:AM, Frick:American Mammals collection at the
AMNH; UF/FGS, Florida Geological Survey, now housed with the University of Florida (UF)
collection at the Florida Museum of Natural History (FLMNH), Gainesville; MCZ, Museum of
Comparative Zoology, Harvard University, Cambridge; SDSM, South Dakota School of Mines, Rapid
City; TMM-TAMU, Texas Agricultural and Mechanical University collection at Texas Memorial
Museum, Austin; UCMP, University of California Museum of Paleontology, Berkeley; UNSM,
University of Nebraska State Museum, Lincoln; USNM, United States National Museum of Natural
History, Smithsonian Institution; YPM, Yale Peabody Museum, New Haven. Capital letters "P" or "M"
followed by a number refer to upper premolars and molars, respectively; lower case letters "p" or "m"
followed by a number refer to lower premolars and molars, respectively. AP and TR refer to
anteroposterior and transverse measurements, respectively. FAD denotes First Appearance Datum. All
measurements are in millimeters. Subdivisions of North American Land Mammal Ages as used in this
report (e. g., early late Arikareean) follow Tedford et al. (1987).


This report is based on a detailed study of the Toledo Bend site completed as an M.S. thesis at
Louisiana State University. J. Schiebout is thanked for her support as my major professor and for
providing a Museum of Geoscience Research Grant that helped fund field work, travel, and other
expenses. E. Manning first introduced me to the site, and to the amateurs who collected there
previously, and made many of the initial identifications. L. Jacobs, R. Tedford, and M. Woodbume
critically reviewed early manuscripts of the thesis and provided numerous helpful comments. Others
who reviewed specific sections and provided helpful comments include M. Coombs, J. Hazel, R. Hunt,
B. MacFadden, E. Manning, C. McCabe, J. Martin, and D. Prothero. R. Hunt and R. Tedford are
additionally thanked for their careful review of the manuscript which resulted in this contribution, and
for sharing with me over several years their vast knowledge of the Arikareean and Hemingfordian
mammal ages. Similar discussions with D. Webb and G. Morgan were always enlightening, as well, and
greatly appreciated.
The following either allowed access to fossil collections or provided helpful discussions on
various aspects of my research or both: M. Winans and J. A. Wilson, University of Texas; P. Bjork, J.
Martin, R. Macdonald, and J. Whitmore Gillette, SDSM (JWG is currently at the Museum of Northern
Arizona); R. Hunt, M. Voorhies, and M. Stout, University of Nebraska; L. Martin, University of Kansas;
D. Webb, B. MacFadden, G. Hayes, and G. Morgan, Florida Museum of Natural History, University of
Florida (Morgan is now at the New Mexico Museum of Natural History and Hayes is currently at the
University of Nebraska); R. Tedford, B. Evander, and J. Alexander, AMNH; M. Coombs, L. Thomas,
and D. Wright, Amherst College (Wright is now at the University of Washington); L. Flynn and C.
Schaff, Harvard MCZ; and M. A. Turner, Yale Peabody Museum. Others who helped with
identifications include J. Baskin, R. Evander, C. D. Frailey, H. Galiano, and R. Hulbert. Frailey is
additionally thanked for allowing me to examine his unpublished manuscript on anthracotheres from
Gratitude is also extended to numerous avocational collectors. J. Hudson, of Baton Rouge, first
brought the site to academic attention. K. Griffin from Lafayette, LA, and J. Stewart, from Leesville,
LA, generously donated and made available many important specimens, including the tapir maxilla and
the anthracothere ramus. R. Westfall, from Rosepine, LA, also donated and made available many
interesting and important specimens, including the chalicothere teeth, some peccary teeth, and the well
preserved dwarf rhino tooth. I additionally want to thank Mr. Westfall for help he provided during one


field trip when we were plagued by truck problems. B. Fite is thanked for making available for study
the only known virtually complete ramus of a small chalicothere. D. Cring, from the Department of
Sociology and Anthropology, USL, Lafayette, also donated some specimens he had collected, and
provided helpful discussions about the site early in my investigation. D. Lescinsky, L. Albright, S.
McLaughlin, T. Delage, and E. Manning helped in the underwater exploration of the site and all found
many important specimens.
Helpful discussions on the geology of the site were provided by J. Rogers, USGS retired, and P.
Heinrich, then a Ph.D. candidate in the LSU Department of Geology. Mr. Rogers generously provided
detailed, unpublished field and electric log data which accurately placed the site in stratigraphic context,
as well as helpful discussions regarding the geology and stratigraphy of the western Louisiana-eastern
Texas region. Mr. Rogers also found and donated a fragment of an entelodont tooth. Mr. Heinrich also
visited the site and was helpful in interpreting its sedimentological aspects.
S. Murray and F. Demiers of the LSU Coastal Studies Institute provided much of the scuba
equipment used in this investigation. Most photographs were taken and processed by LSU Coastal
Studies Institute photographer K. Lyle. Others were taken by Erica Simons and Terry Lott of the
Florida Museum of Natural History. Figure 1 was prepared by M. L. Eggert and C. Duplechin of the
LSU Cartographic Services Division.
Additional funds for field work and travel to various museums were provided by a Society of
Vertebrate Paleontology Bryan Patterson Award, an American Museum of Natural History Theodore
Roosevelt Award, a Geological Society of America Student Research Grant, and a LSU Museum of
Geoscience Research Grant made possible through the generosity of the Mcllhenny family.
Contributions for page charges were generously provided by the University of Florida Museum's
Department of Natural History, the Florida Museum's Pony Express fund, and the Museum's Office of
the Director. This report is University of Florida Contribution to Paleobiology Number 505.


Figure 1 shows the location of the Toledo Bend site in Newton County, Texas,
approximately 27 km northeast of Burkeville in the "tailrace" canal that joins the
Toledo Bend Reservoir to the Sabine River. The Sabine River marks the
Texas/Louisiana state boundary and the site is about 400 m northwest of that
boundary. As discussed at length in previous papers (Albright, 1991, 1994, 1996,
1998a, 1998b), the outcrop pattern of the Fleming and underlying Catahoula
formations is mapped differently across the state line resulting in placement of the
site in the upper part of the Catahoula Formation as mapped in Texas, but in the
Carnahan Bayou Member of the Fleming Formation as mapped in Louisiana.


Most fossils from the Toledo Bend site were collected from a lag deposit that
accumulated on the bottom of the canal since its excavation in 1969. The fossils
originated and eroded from a paleochannel conglomerate that crops out
immediately adjacent to the moder-day lag. Laminated floodplain siltstones lie
stratigraphically above and below this outcrop. Because nearly all lower vertebrate


Figure 1. Location of Toledo Bend site in Newton County, Texas.


taxa in the fauna are indicative of quiet, slow-flowing bodies of water (Albright,
1994), and because Galloway et. al. (1986) considered the east Texas region as
having very low relief during the early Miocene, the conglomeratic deposit was
likely during an episode of violent flood mixing. There is no indication of sediment
sorting throughout the deposit. Clay and siltstone rip-up clasts, quartz and chert
pebbles, and bones and teeth of various sizes occur randomly distributed. No
imbrication of pebbles is apparent, and there is no obvious indication of
paleocurrent direction based on the few specimens collected in situ. The only
graded bedding observed is at the top of the deposit where it abruptly grades into
the overlying siltstone presumably after the high energy event subsided.
Almost all bones show at least some degree of abrasion, although very few are
worn beyond recognition and most, although fragmented, show only light to
moderate wear and/or polishing. Exceptions include bone recently eroded from the
paleochannel deposit that fractured during the process. Well rounded bone cobbles
illustrate the most extreme cases of wear. Many specimens were likely exposed to
streamflow abrasion or subjected to numerous episodes of transport and reworking
prior to entrapment in the Toledo Bend deposit (see Behrensmeyer, 1982).
Supporting this, a highly worn entelodont astragalus was found in situ, whereas
another, virtually unworn specimen was found free in the modern lag deposit.
Fossils in the lag deposit are additionally subjected to abrasion caused by strong
currents produced during electricity generation (there is a hydroelectric plant at the
head of the canal). These currents not only facilitate erosion of fossils from their
matrix, but also chum the bottom sediments, thus exposing fossils previously
eroded and deposited into the lag.
No remains were found in an articulated or partially articulated state, nor were
any complete, unbroken bones found except isolated elements of the manus and pes
(rare vertebrae were recovered in worn, but relatively complete condition). The
frequency of occurrence of various elements recovered is largely a function of the
mechanical stability of those elements. Teeth, teeth fragments, and manus and pes
elements make up the greatest percentage of remains recovered. Of the isolated
teeth collected, none retain their complete root system and most consist of only the
crown. Mandibular fragments are uncommon. Skull material is limited to those
parts of relatively greatest mechanical strength including rare maxillary fragments,
isolated petrosals, and other bones of the ear region. Similarly, limb bone material
is limited primarily to fragments of the proximal and distal ends. Because the
scapula and pelvis are readily transported and abraded (Voorhies, 1969), recovered
fragments of these elements also include only their strongest parts, the glenoid
region of the former and the acetabular part of the latter.
Presently the Toledo Bend Local Fauna does not include a large enough series
of any one taxon to perform population studies. Stage of tooth wear indicates a
mixture of all age groups. From the diversity of the fauna and the condition of the
remains recovered, the fossils from Toledo Bend, although concentrated in the
deposit by a flood event, appear to represent an attritional mode of mortality. The


accumulation of remains in a channel deposit, combined with (1) many abraded
elements, (2) the absence of certain elements and concentration of others, and (3)
the lack of articulated skeletons and/or complete skulls, limb bones, etc., provides
evidence indicating that most, if not all, of the fossils were variably subjected to
episodes of reworking prior to their final emplacement in the Toledo Bend deposit
(see Behrensmeyer, 1982). Further details on the taphonomy of the Toledo Bend
site can be found in Albright (1991).


Class MAMMALIA Linnaeus, 1758
Family EQUIDAE Gray, 1821
?Miohippus sp.
Figure 2

Type Species.-Miohippus annectens Marsh, 1874
Referred Specimens.-LSUMG V-2509, incisor; V-2510, left M3; V-2251,
right upper cheek tooth; V-2252, partial left lower m3; V-2511, left astragalus; V-
2512, partial right cuboid; V-2513, distal metapodial III; V-2514, proximal phalanx
of medial digit; V-2515, two proximal phalanges of lateral digits; V-2768, proximal
phalanx of lateral digit.
Description.-V-2510 is thought to be an M3 based on the lack of interstitial
wear on the posterior surface (Fig. 2A). The labial surface of the paracone and
metacone is broad anteroposteriorly and ribbed. The parastyle and mesostyle, but
not the metastyle, are prominent. An anterior cingulum is continuous with the
parastyle. There is no lingual cingulum. The metaconule is submerged within the
metaloph. The metaloph has no crochet or plications, and it meets a lingually
directed spur originating from the ectoloph (the medicrista). The hypostyle shows
the "type 3" morphology of Prothero and Shubin (1989:144, fig. 10.1). Cement is
entirely absent. The tooth measures 13.7 mm AP by 15.5 mm TR. Another upper
cheek tooth, V-2251, differs from V-2510 in having a non-connected metaloph.
Although the partial m3 (V-2252) shows no trace of a cingulum, it may have been
worn away, as the tooth is highly water-worn. The astragalus measures 22.3 mm
proximo-distally by 17.5 mm transversely.
Discussion.-Tentative referral to Miohippus is based on dental similarity to a
small horse in the collections of the FLMNH referred to cf. Miohippus from the late
early Arikareean Cow House Slough Local Fauna, Hillsborough County, Florida
(Morgan, 1994). It is important to note, however, that the distinction between later
members of Miohippus and early members of Archaeohippus (extinction of


1 E

Figure 2. (A) ?Miohippus sp., stereo view of left M3, LSUMG V-2510; (B) Anchippus texanus, stereo
view of right ml-3, V-2257; (C) Labial view of same; (D) A. texanus, stereo view of left p4, V-2526.


Miohippus and first appearance of Archaeohippus overlaps in the Arikareean) based
solely on dentition remains tenuous, at best, and both groups are long overdue for a
major revision. Typically, distinction centers around the connection of the metaloph
to the ectoloph, i. e., not connected in Miohippus, connected in Archaeohippus
(Stirton, 1940; Prothero and Shubin, 1989:169). However, many specimens referred
to Miohippus show the connected condition, including F:AM specimens 116414,
116415, and 116416 from the "lower Protoceras channels" of Washington County,
South Dakota (labeled M. gidleyi), and AMNH specimens 12917d and 19914 from
the "lower Rosebud" near Porcupine Creek, South Dakota, whereas specimens of A.
blackbergi in the large sample from the early Hemingfordian Thomas Farm Local
Fauna, Florida, rarely show an unconnected metaloph. Additionally problematic are
several specimens from other Arikareean localities in Florida, other than Cow House
Slough, that show an intermediate, very weakly connected condition such as that
seen in LSUMG V-2510. On the other hand, V-2251, as noted above, retains the
disconnected condition more typical of Miohippus.
Upper cheek teeth of ?Miohippus sp. from Toledo Bend and Florida differ
from those of Archaeohippus blackbergi (also known from the early Hemingfordian
Garvin Gully Fauna, Texas) in larger size and absence of a crochet or plications on
the metaloph. ?Miohippus sp. also has an anteroposteriorly broader ectoloph than
does A. blackbergi and the cingulum labial to the metacone does not continue
posteriorly into as prominent a metastyle. Archaeohippus minimus (Douglass,
1899), A. ultimus (Cope, 1886), and A. mourning (Merriam, 1913) are larger than
A. blackbergi and they have additional plications on the metaloph. The lower cheek
tooth of ?Miohippus sp. is slightly lower crowned than in Archaeohippus
blackbergi, and the "V" shaped trigonid and talonid in the latter appear
anteroposteriorly broader and rounder in the former. Although not comparable due
to lack of material in the Toledo Bend hypodigm, R. Hunt (pers. comm., 1998) also
noted that lower premolars are conspicuously broader transversely than the molars
in Miohippus, whereas in Archaeohippus there is less of a size difference between
p4 and ml.

Subfamily EQUINAE Gray, 1821
Genus Anchippus Leidy, 1868
Anchippus texanus Leidy, 1868
Figures 2, 3

Anchippus texanus Leidy, 1868
Anchippus texanus Leidy, 1869
Parahippus texanus (Leidy). Gidley, 1907
Parahippus texanus (Leidy). Osborn (1918)
Parahippus texanus (Leidy). Stirton, 1940
Parahippus cf. P. texanus (Leidy). Forsten, 1975, in part


Table 2. Measurements of dentition ofAnchippus texanus. a LSUMG V-2620, right maxillary fragment
with P3-MI; b LSUMG V-2257, right ramal fragment with ml-3; C LSUMG V-2259, right ramal
fragment with dp4-ml; d LSUMG V-2530, right ramal fragment with ml-2; e LSUMG V-2531.

P2 17.5 17.7 p3 16.9 14.4
19.8 19.5 p4 17.9 17.0
P3 19.0a 17.3 21.2 ml 15.0b 12.3
P4 18.6a 17.8 22.0 14.1d 12.4
18.0 17.0 22.6 14.7c 10.0
MI 18.5a 17.2 21.5 m2 15.2b 12.0
16.3 21.6 14.3d 12.7
17.8 17.0 22.0 15.7 12.2
17.4 16.6 22.4 14.10 13.0
M2 18.1 17.6 23.4 m3 19.1b 11.1
M3 17.5 15.5 21.5 20.4 10.0
17.7 16.3 21.0

Holotype.-ANSP 11275, left M1 or M2 from "Hutchen's well," Washington
County, Texas.
Referred Specimens.-Toledo Bend: LSUMG-V-2516, three incisors; V-
2517, two right P2s; V-2518, right P2; V-2519, right P4; V-2520, worn right upper
cheek tooth; V-2248, right Ml; V-2521, right Ml?; V-2522, left M2; V-2523, right
M3; V-2524, left M3; V-2620, right maxillary fragment with P3-M1; V-2525, left
p3; V-2526, left p4; V-2531, right m2; V-2527, left m2; V-2528, left m3; V-2529,
partial right lower cheek tooth; V-2259, right ramal fragment with dp4-ml; V-2530,
right ramal fragment with ml-2; V-2257, right ramal fragment with ml-3; V-2532,
scapula fragments; V-2533, proximal left radius; V-2534, two distal radius
fragments; V-2535, two proximal ulna fragments; V-2536, right magnum; V-2537,
two distal tibia fragments; V-2538, seven astragali; V-2767, two astragali (juvenile);
V-2539, calcanea fragments; V-2540, left and right cuboid; V-2541, left and partial
right navicular; V-2542, partial ectocuneiform; V-2543, sesamoid fragment; V-
2769, two proximal lateral metapodials; V-2544, distal lateral metapodial; V-2545,
three proximal metatarsal III fragments; V-2546, two distal metapodial III
fragments; V-2547, distal metatarsal III; V-2548, three proximal phalanges; V-2258,
proximal phalanx; V-2549, four medial phalanges; V-2550, proximal lateral
phalanx. Cedar Run Local Fauna: TMM 40068-7, left M3.
Description.-Measurements of upper and lower cheek teeth are provided in
Table 2. The incisors from Toledo Bend have cupped crowns and the cheek teeth
show a primitive "parahippine" morphology. Upper cheek teeth (Fig. 3A-F) lack
cement and a lingual cingulum. The metaloph is non-plicated and its connection to
the ectoloph is at a primitive stage in that a lingually directed blade derived from the
ectoloph connects with a labially directed extension of the metaconule. The


Figure 3. Anchippus texanus. (A) Right P3-MI, LSUMG V-2620; (B) Stereo view of right P2, V-2518; (C) Stereo view of right Ml?, V-2521; (D) Stereo
view of right Ml?, V-2248; (E) Stereo view of left M2?, V-2522; (F) Stereo view of right M3, V-2523; (G) Proximal phalanges, V-2258 (top), V-2549

ro- 2
"_I I


ectoloph is ribbed with the rib on the paracone being generally more prominent than
that on the metacone. The parastyle and mesostyle are prominent. As seen in
unworn specimens, the anterior cingulum is shelf-like between the protocone and
protoconule, and continuous with the parastyle. The protocone and protoconule are
entirely separated to the extent that even in medium wear stage their dentine does
not merge. Dentine of the two cusps does not unite until the grinding surface of the
tooth is highly worn as in V-2519, 2520, and 2620. Prominence of the metaconule
varies, but it can always be distinguished from the hypocone by an anterior and
posterior constriction on the metaloph.
The crochet varies from non-existent (V-2523 and V-2620), to a slight trace
(V-2519 and V-2248), to relatively prominent (V-2520 and V-2524). In no case
does the crochet connect to the protoconule except in P2 (V-2518). The hypostyle
in V-2521, 2522, and 2620 resembles the "Type 3" hypostyle of Prothero and
Shubin (1989:144, fig. 10.1). In other cheek teeth, the hypostyle is triangular and
encloses a hypostylar fossette. The dentine of the hypostyle is not continuous with
that of the metaloph, each being bound by its own enamel until extreme stages of
wear. P4 and Ml are nearly the same size, and M2-3 are progressively smaller, but
not to the point where M3 would be considered reduced.
Lower cheek teeth (Fig. 2B, C, D) show a prominent anterior, posterior, and
labial cingulum and lack cement. The p3-4 are wider transversely than ml-3 with
p4 being the widest. V-2531, an m2, shows a separated metaconid and metastylid,
but not as separated as in lower molars of Parahippus leonensis. Crown height is
lower than in P. leonensis, resembling more closely that found in the type specimen
of P. pawniensis (AMNH 9085).
Appendicular skeleton (Fig. 3G): Proximal phalanges of digit III are distinctly
broader for their length than those of P. leonensis, but not as broad for their length
as in Anchitherium clarencei from Thomas Farm. The distal tibia fagments have
highly oblique astragalar facets, the medial of which is narrower and deeper than the
lateral. The nine astragali show a wide range of size variation.
Discussion.-Referral to Anchippus texanus is based on the similarity of the
teeth in size and morphology to the (admittedly poor) type specimen from
"Hutchen's Well" in Washington County, Texas (Leidy, 1868), and because of the
uncertain status of the genus Parahippus (see discussion in Voorhies, 1990:A170-
A171). Recovery of the type from a well precludes accurate stratigraphic
placement, although Forsten (1975) referred a tooth from another site in Washington
County, the Cedar Creek fossil locality, to this species. Wood and Wood (1937)
correlated the Cedar Creek assemblage (Cedar Run Local Fauna of Tedford et al.,
1987) to faunas from the Harrison Formation in the northern Great Plains. Although
illustrations of the type specimen (Leidy, 1869; Osborn, 1918) show an anterior
cingulum that continues lingually around the protocone, personal observation of a
cast of the specimen shows that this lingual continuation is exaggerated in the
illustrations and that the cingulum actually ends before rounding the lingual surface
of the protocone.


Similarity of tooth morphology suggests that Anchippus texanus, Parahippus
nebrascensis, and Parahippus wyomingensis may be closely related. The
resemblance between A. texanus and P. nebrascensis was noted by Peterson (1907),
Osborn (1918), Matthew (1913; see Osborn 1918:79, 84), and Schlaikjer (1937).
The type of P. nebrascensis from the "upper Harrison beds" of Sioux County,
Nebraska (Peterson, 1907), and AMNH 12924, a specimen from the "Upper
Rosebud" referred to P. texanus by Osbom (1918), differ from Toledo Bend
specimens in larger size, stronger crochets, less conical hypocones, and less
constricted separation between the protocone and protoconule. The relative degree
of hypsodonty is similar. Parahippus wyomingensis, from the "lower part of the
upper Harrison formation in the Goshen Hole Area, Wyoming" (Schlaikjer,
1937:255), is also larger, although one anomalously large lower premolar from
Toledo Bend (V-2519) closely resembles the p4 of the type specimen (MCZ 6390).
Teeth of the type specimen of P. tyleri Loomis (1908) differ from A. texanus in
being larger, more hypsodont, in having a less conical and more anteroposteriorly
crescentic hypocone and protocone, and a more crescentic metaloph resulting in the
hypocone being positioned posterior to the metacone rather than directly lingual to
it. Also, the paracone rib is reduced in P. tyleri. Although Loomis (1908:164) listed
the type (AC 1079) from "the upper part of the Lower Harrison beds, 8 miles
northeast of Agate, Sioux County, Nebraska," Cook and Cook (1933) listed it from
the Upper Harrison beds of Nebraska, and R. Tedford (written comm., 1990) also
suggested that the type locality was more likely in the Upper Harrison or
Runningwater formations. Parahippus tyleri is probably a latest Arikareean to early
Hemingfordian form and therefore not likely ancestral to P. nebrascensis as Loomis
The type specimen of P. pawniensis (AMNH 9085) differs from Toledo Bend
specimens in having a weakly plicated metaloph, a reduced M3, and a less distinct
posterior constriction between the protocone and protoconule. The prominent labial
cingulum on the lower cheek teeth from Toledo Bend is absent on the m3 of the P.
pawniensis type specimen. Gidley (1907) described the type from the Pawnee
Creek beds, Colorado, which were not clearly differentiated from the Martin
Canyon Formation of early Hemingfordian age (Tedford et al., 1987). The dental
morphology of AMNH 9085 suggests an older age, and specimens from the early
Hemingfordian Flint Hill Local Fauna, South Dakota, (e.g., UCMP 32360, 32622,
and 37285) and from late Arikareean deposits near the Lusk area, Wyoming (e.g.,
F:AM 109859), have also been referred to this species.
North American species of Anchitherium differ from Anchippus texanus in
larger size, lack of a ribbed paracone and metacone resulting in a strongly W-shaped
ectoloph, absence of a crochet, and a reduced M3. Also, in Anchitherium, the
metaloph is not always connected to the ectoloph (e.g., early Arikareean F:AM
specimens from "Harris Ranch just north of the Pine Ridge Reservation" and
AMNH 105167 from the John Day Formation, as well as UF 22137 [a cast of a
specimen] from the late Arikareean "Harrison beds near Van Tassell, Wyoming").


The M3 of the type specimen of Anchitherium navasotae Hay (1924) from the
Garvin Gully Fauna (TMM-TAMU 2385) is generally similar to some of the Toledo
Bend cheek teeth (e. g., V-2522), but differs in having a paracone that is higher than
the metacone (in the Toledo Bend teeth they are of equal height), in having a
relatively low metaloph and protoloph resulting in shallow pre- and post-fossettes in
contrast to deep ones in the Toledo Bend teeth, and in having a more oblique
ectoloph. Kalobatippus agatense is similar in size to Anchippus texanus, but differs
from the latter in the same manner as does Anchitherium. A review of Kalobatippus
and Anchitherium is currently underway by B. MacFadden (pers. comm., 1998).
Lower premolars referred to Anchitherium navasotae, also cataloged under
TMM-TAMU 2385, are indistinguishable from those of the Toledo Bend species,
sharing the slightly separated metaconid and metastylid and a continuous anterior,
labial, and posterior cingulum with no lingual cingulum. However, in Anchitherium
clarencei from Thomas Farm, both upper and lower cheek teeth, as well as post-
cranial elements such as astragali, are much larger than similar elements from
Toledo Bend.
Forsten (1975:20) referred two upper cheek teeth, one from Cedar Creek (i.e.,
Cedar Run Local Fauna) and one from Hidalgo Bluff, Washington County, Texas,
to Parahippus cf. P. texanus and additionally noted the absence of this species at
Thomas Farm. In her descriptions she noted that the protocone and protoconule
were of nearly equal size and that the teeth had a subtriangular outline caused by a
protocone "centrally situated and larger than the reduced hypocone." Personal
observation found this statement true only for the Hidalgo Bluff specimen, TMM
40067-63, but not for the Cedar Run tooth, TMM 40068-7. The Cedar Run tooth
referred to P. cf. P. texanus, a left M3, closely resembles LSUMG-V 2523, a right
M3 in the following features: size, lack of cement, relatively tall protoloph and
metaloph resulting in deep anterior and posterior fossettes, distinct ribbing on the
ectoloph with the paracone rib the more prominent, a simple crochet not connected
to the protoloph, a larger protocone than hypocone, a simple, connected metaloph
with no plications other than the crochet, and a protoloph with a distinct anterior and
posterior groove constricting the protocone from the protoconule. Differences are
minor including, in TMM 40068-7, a slightly more prominent crochet, the lack of an
anterior cingulum, a less bulbous parastyle, and a posteriorly notched hypostyle.
The type specimen of Anchippus texanus also has a posteriorly notched hypostyle.
In contrast to Forsten's (1975) findings, the Hidalgo Bluff tooth, TMM 40067-
63, appears to belong to a separate species, as it differs considerably from the Cedar
Run and Toledo Bend teeth. The Hidalgo Bluff tooth must certainly be the one to
which Forsten referred when she noted the subtriangular outline resulting from a
centrally located protocone. TMM 40067-63 additionally differs in smaller size,
presence of cement in the fossettes, a weak protocone-protoconule separation, a
crescentically-shaped rather than conical protocone, a strongly posteriorly-directed
metaloph, and greater hypsodonty. Overall, TMM 40067-63 appears most similar to
teeth of Parahippus leonensis, with the exception of a non-plicated metaloph, and it


is to this species that this tooth is here referred. Compared with P. leonensis, the
most common horse at Garvin Gully and Thomas Farm, the Toledo Bend teeth (and
TMM 40068-7) are slightly larger, relatively less hypsodont, lack cement, have a
less crescentic, less posteriorly directed protoloph and metaloph, a more conical
protocone, protoconule, and hypocone, and a simple metaloph.
Another specimen in Forsten's (1975) hypodigm of P. cf. P. texanus, TMM
40623-9 from Push Creek, Tyler County, Texas, shares few similarities with TMM
40068-7, TMM 40067-63, the Toledo Bend teeth, or the type ofAnchippus texanus.
TMM 40623-9 resembles most closely Anchitherium in having an almost W-shaped
ectoloph with no ribs and a plicated metaloph that does not connect to the ectoloph.
There is abundant cement and there is no posterior constriction between the
protocone and protoconule. Based on the above comparisons, it is here concluded,
in contrast to Forsten's findings, that TMM 40068-7, TMM 40067-63, and TMM
40623-9 each belong to separate species and that only TMM 40068-7 should be
referred to A. texanus.
Although lower cheek teeth from Toledo Bend vary in size (Table 2), this is
likely due to sexual dimorphism rather than multiple species. In contrast to earlier
studies by Simpson (1932), Quinn (1955), and particularly White (1942), who
suggested the presence of as many as seven different species of horses at Thomas
Farm, Hulbert (1984), in a detailed population dynamics study of P. leonensis from
Thomas Farm, documented sexual dimorphism and concluded that it was "a good
indication that [the] social structure [of P. leonensis] still resembled that of earlier,
browsing equids."

Subfamily SCHIZOTHERIINAE Holland and Peterson, 1914
Genus Moropus Marsh, 1877
Moropus sp.
Figure 4

Type Species.-Moropus distans Marsh, 1877
Referred Specimens.-LSUMG V-2703, partial right P2; V-2490, right ml or
m2; V-2489, right ramal fragment with m3; V-2408, distal left tibia; V-2409, distal
left tibia; V-2410, right astragalus; V-2260, right metatarsal IV; LSUMG V-2411,
proximal phalanx of digit IV; V-2261, proximal phalanx of digit III; V-2412,
proximal phalanx; V-2413, proximal phalanx; V-2414, medial phalanx.
Description.-In its broken state the partial P2 measures about 15.3 mm AP
by about 14 mm TR. There is a prominent, shelf-like anterior cingulum and a strong
posterior cingulum continuous with a lingual cingulum (Fig. 4B). The anterior and
lingual cingula only just meet at the anterolingual corner of the tooth. A conical,
unworn protocone is the only lingual cusp. Extending labially from the protocone is










a prominent protoloph and metaloph, the former more worn than the latter.
Although damaged labially, the paracone was at least as prominent as the protocone.
This tooth shows the wear pattern noted by Coombs (1978, 1979) of both Moropus
oregonensis and Tylocephalonyx skinneri where the protoloph and the metaloph
become worn before the protocone.
The ml measures 25.0 mm AP by 14.0 mm TR. It has a prominent anterior
and posterior cingulum and a small cingular segment at the labial entrance of the
valley between the trigonid and talonid. The talonid is larger than the trigonid.
Measuring about 11.0 mm AP by 12.5 mm TR, the trigonid resembles that of a right
lower molar from the Buda Local Fauna (Frailey, 1979) that measures 10.3 mm AP
by 13.0 mm TR. That from Buda, however, is relatively less hypsodont than the
Toledo Bend specimen and it has no cingula.
The relatively unworn state of the m2 (Fig. 4A) shows a separated metaconid
and metastylid. The trigonid and talonid are nearly the same size. Although
partially broken, the labial cingular segment is more prominent in m2 than in ml.
The tooth measures 29.2 mm AP and 15.4 mm TR.
The medial malleolus of the two distal tibia fragments is not nearly so
prominent as in Moropus elatus. Resembling M. elatus, however, is the broad and
shallow lateral astragalar facet (see Coombs, 1978:34-45). The medial facet is
narrower and deeper than the lateral, but not nearly to the extent seen in horses,
rhinos, and tapirs. V-2409 measures 44.4 mm TR and 33.2 mm AP; V-2408
measures 36.2 TR by 29.4 mm AP.
The astragalus measures 40.5 mm transversely by 36.4 mm dorsoventrally. It
resembles that of M. elatus, and differs from that of horses and rhinos in having a
broad sulcus between the medial and lateral trochleae, a highly reduced neck, and a
navicular facet that does not extend laterally to the distal termination of the lateral
trochlea (Fig. 4E).
The metatarsal (Mt) IV has separate dorsal and volar facets on the medial
surface for articulation with the Mt III (Fig. 4C). The volar facet, though somewhat
abraded, is smaller than the dorsal facet. The area just distal to the dorsal and volar
Mt III facets is rugosely textured for muscle attachment. The cuboid facet has a
roughly rectangular outline with its transverse diameter the greatest. There is no
distinct ectocuneiform facet. The cross-sectional outline of the shaft is roughly
quadrate. Immediately proximal to the dorsal surface of the distal articular surface
is a chevron-shaped sulcus. Proximal and medial phalanges do not unite to form a
duplex digit, apparently common in other species (Fig. 4F, G). Proximal phalanx
measurements are provided in Table 3. The medial phalanx from Toledo Bend
measures 24.5 mm in length, 16.5 mm in distal width, and 19.0 mm in proximal
Discussion.-Chalicothere remains from the Coastal Plain were previously
known only from the Buda Local Fauna (Patton, 1967c; Skinner, 1968; Patton and
Webb, 1970; Coombs, 1978; Frailey, 1979; and Coombs, 1989). Several additional


Table 3. Measurements of proximal phalanx for specimens ofMoropus sp. from Toledo Bend and
Buda local faunas. Frailey (1979).

Toledo Bend Buda*

V-2262 V-2412 V-2411 UF 24130

Length 32.3 34.6+ 41.8 45.7
Proximal width 25.4 25.8 24.6 26.5
Distal width 16.2 17.2 16.4 18.5
Depth at dorsal margin of
metapodial facet 22.0 20.3 20.2 21.3
Depth at volar termination of
distal articulating surface 16.9 18.8 16.5 16.8

specimens closely resembling those from Buda were recently recovered from a
undisclosed locality in Pasco County, Florida, but are retained in a private collection
unavailable for study. Additional taxa from this locality indicate age equivalence
with the Buda Local Fauna (pers. observe 1998). Another specimen recently
recovered from the panhandle region of Florida, and also retained in a private
collection, consists of a nearly complete right ramus with p3-m3, missing p2. This
specimen, a cast of which is housed at the FLMNH (UF 180233), represents a
smaller individual than those represented by material from Toledo Bend, which, in
turn are slightly smaller than the specimens from Buda. Sexual dimorphism, well
documented in the Chalicotheriidae (Coombs, 1975), may account for these
variations in size. Late Arikareean and Hemingfordian species from the High Plains
(e.g., Moropus elatus and M. hollandi; see Coombs, 1978, 1989) are considerably
Other small North American chalicotheres are known from the John Day
Formation, Oregon, and include the genoholotype Moropus distans Marsh (1877)
and M. oregonensis (Leidy, 1873). Coombs (1978, 1989) suggested that these two
species may be conspecific. Although Frailey (1979) concluded that the Buda
species was not M. oregonensis based on the quadrangular shape of M1 (UF 24131),
Coombs (pers. comm., 1996) found that UF 24131 is more likely a dP4 (and the
"M2" an Ml) noting that dP4 is typically more quadrate and less elongate than Ml.
Lack of matching elements precludes direct comparisons between the Oregon and
Toledo Bend specimens, but the P2 from Toledo Bend is the size expected for M.
oregonensis based on the P3-4 of the latter (see Coombs, 1978:46). All specimens
from the John Day Formation for which there is reliable stratigraphic data originate
from strata overlying a tuff dated by C. Swisher at 22.6 Ma (R. Hunt and E.


Table 4. Measurements and proportions of Mt IV for various specimens of Moropus. MXL = maximum
length; MXDW = maximum distal width; MNSW = minimum shaft width; L/DW = length/distal width;
L/MNSW = length/minimum shaft width. Frailey (1979:147); ** Coombs (1978:39).

Toledo Bend Buda* St. GCrand** M. hollandi M. elatus**
V-2260 UF 24129 F:AM (apx mg)

MXL 86.6 93.7 88.9 129.8 124-165
MXDW 25.0 26.2 28.0 38.5 35-51
MNSW 17.4 18.0 19.8 27.5 23-36
L/DW 3.46 3.57 3.20 3.4 3.2-3.8
L/MNSW 4.97 5.2 4.5 4.7 4.8-5.5

Stepleton to T. Fremd, pers. comm., 1996). Small chalicotheres also occur in the
Aquitanian St. Gdrand-le-Puy fauna of France, the upper boundary of which is
placed at about 20.52 Ma by Berggren et al. (1995).
The Mt IV from Toledo Bend differs from the Buda specimen in smaller size
and in having a triangular, rather than rounded, volar facet. The smaller size, less
robust build, and less rugosely textured area of muscle attachment of the Toledo
Bend specimen relative to the Buda specimen may be due to sexual dimorphism.
Proximal and medial phalanges from Toledo Bend are also smaller and less robust
than those from Buda (Table 3).
Neither the Toledo Bend nor the Buda Mt IV show a distinct facet for the
ectocuneiform. Coombs (1978) listed an ectocuneiform facet on the Mt IV as
diagnostic of the genus Moropus. The absence of this feature resembles the
condition in the large, dome-skulled chalicothere, Tylocephalonyx (Coombs, 1979).
However, Coombs (1978) also noted that in M. elatus only the proximal one-fourth
of the dorsal Mt III facet articulates with the Mt IV facet of the ectocuneiform.
Thus, the articulation may have been minimal in the Toledo Bend and Buda species,
and not obvious without the associated Mt III and ectocuneiform.
Although the Mt IV from Toledo Bend is substantially smaller than that of M.
elatus and M. hollandi, the length to distal width ratios and the length to minimum
shaft width ratios are similar (Table 4). A Mt IV from St. Grand shares similar
ratios, but differs from the Toledo Bend specimen in having a distinctly divided
dorsal facet on the medial side for articulation with both the ectocuneiform and the
Mt III. Also, the volar facet in the St. Grand specimen is relatively larger than in
the Toledo Bend or Buda specimens.


Family TAPIRIDAE Burnett, 1830
Genus Nexuotapirus Albright, 1998b
Nexuotapirus marslandensis (Albright, 1998b)

Nexuotapirus marslandensis (Schoch and Prins, in Schoch, 1984). Albright, 1998b

See Albright (1998b) for a description and discussion of the tapir from Toledo

Family RHINOCEROTIDAE Owen, 1845
Gulfoceras, gen. nov.

Type Species.-Gulfoceras westfalli.
Etymology.-See below.
Diagnosis.-As for species.

Gulfoceras westfalli gen. et sp. nov.
Figure 5

Holotype.-LSUMG V-2622, right M3.
Referred Specimens.-LSUMG V-2249, left M3; V-2621, right M3; V-2574,
left astragalus.
Etymology.-Gulfoceras, rhinoceros from the Gulf Coastal Plain; westfalli,
for Mr. Robert Westfall, an enthusiastic avocational collector who made available to
the LSU Museum of Geoscience the holotype, as well as many other important
Diagnosis.-Much smaller than Menoceras and Diceratherium. Smaller but
similar relative hypsodonty to Subhyracodon. Less prominent crochet than
Menoceras. M3 lacks continuous lingual cingulum and groove on protoloph to
mark protocone.
Description.-In contrast to the highly worn condition of V-2249 and V-2621,
the type specimen, V-2622, is a lightly worn specimen in excellent condition (Fig.
5A, B). The later measures 26.8 mm AP by 30.5 mm TR. The maximum, slightly
worn crown height is 27.6 mm. There is an anterior cingulum and a weak posterior
cingulum. The lingual cingulum is not continuous. There is no cristid but there is a
sharp ridge that begins at the apex of the hypocone and descends the anterior
surface. Once worn, this ridge becomes a simple crochet as seen in V-2249. V-
2249 measures 23.0 mm AP and, in its broken state, about 27.5 mm TR. The
transverse width would have been only slightly greater were it not broken. There is
a simple crochet and a simple cristid. Most enamel is worn off the tooth. V-2621 is
slightly larger than V-2249 and is also highly worn. It measures 25.3 mm AP, but
its broken state prevents a transverse measurement.



Figure 5; Gulfceras westfalli gen. et sp. nov., (A) Worn left M3, LSUMG V-2249; (B) Right M3, V-

The astragalus is similar in size to those of Anchippus texanus and
Nexuotapirus marslandensis. It resembles the tapir more closely than the horse in
the broader angle between the two trochleae and in a relatively shallow, rather than
deep, sustentacular facet. It further resembles the tapir in that the medial surface of
the medial trochlea is relatively flat, whereas in the horse the medial surface of the
medial trochlea has a distinct flange that follows the curvature of the trochlea. In
the rhino and horse, but not the tapir, the medial trochlea contacts or merges with
the navicular facet. The astragalus of the Toledo Bend chalicothere differs
substantially in having a shallower angle between trochleae, a transversely reduced
navicular facet, and a transversely expanded sustentacular facet.
Discussion.-Although material representing this tiny rhino is limited,
comparisons have found no currently described species to which these specimens
can be referred. Thus it seems reasonable to conclude that yet another Gulf Coastal
Plain endemic species has been discovered (see below). The diminutive Hyracodon
may have lasted into the early Arikareean (Prothero et al., 1989), but M3 of
Hyracodon differs from the Toledo Bend teeth in having an ectoloph that extends
posteriorly beyond the metaloph resulting in a vaguely "70" shaped occlusal outline.
The taxon with an M3 most closely resembling the Toledo Bend tooth is the
common Orellan rhino, Subhyracodon. General shape and hypsodonty are
essentially the same, although the M3 of Subhyracodon is larger in size, it lacks the
ridge that ascends the anterior surface of the hypocone, and it has a prominent
lingual cingulum that is continuous with the anterior cingulum.


Table 5. Upper cheek teeth measurements for various specimens of Diceratherium. a= approximate; =
average of Peterson's (1920) and Troxell's (1921) measurements. The first column includes Toledo Bend
specimens referred to D. annectens. The last column is the specimen referred to D. armatum.

Toledo D. annectens D. annectens D. "nanum" D. "nanum" D. "armatum"
Bend YPM 10001 AMNH 7324 AMNH 7325 AMNH 7343 ACM 1828

PI: L 18.9 20.4 19.0 20.0
W 18.3 17.2 17.0 16.0

P2: L 21.8 23.2 24.0 24.3
W 24.4 27.8 a28.0 30.5

P3: L 25.7 27.5 28.0 23.9 28.9
W 30.0 34.3 35.0 26.7 37.2

P4: L 27.7, 28.8 30.0 31.0
W 33.7, 37.1 39.0 40.4

Ml: L 35.0
W 41.0

M2: L 41.7 40.0 40.8
W 45.0 41.0 41.8

M3: L 39.5 33.0 34.4
W 41.8 38.0 42.2

Prothero and Manning (1987) described dwarf rhinoceroses from the
Barstovian of the Texas Coastal Plain and Prothero and Sereno (1982:24-26), in
discussing these dwarf forms, noted that the Miocene faunas of the Gulf Coast were
characterized by a "high degree of endemism" and an "unusually high diversity of
forms" apparently due to the mixing of endemic coastal forms with High Plains
immigrants. The Toledo Bend Local Fauna clearly indicates at least limited mixing
with faunas from the High Plains by the mutual occurrence of Arretotherium,
Nexuotapirus, Nanotragulus, "Cynorca," and Diceratherium in both regions (see
Albright, 1998a, and below).
The occurrence of "dwarf' rhinoceroses, chalicotheres, and amphicyonids in
the Gulf Coastal Plain, to their exclusion in the paleontologically well sampled
northern Great Plains, would seem to confirm the unique paleoenvironmental


Table 5. Extended

D. sp. D. sp. D. armatum* D. gregorii Toledo Bend
AMNH 7342 AMNH 7346 YPM 10003 AM 12933 V-2560

PI: L 19.7 29.0 21.0
W 17.0 25.5 20.0

P2: L 32.0 26.0
W 39.5 32.0

P3: L 37.0 31.0
W 46.0 44.0

P4: L 28.5 39.0 34.0 35.4
W 32.7 50.0 48.0 44.5

MI: L 48.0 38.0
W 53.0 45.0

M2: L 54.0 45.0
W 55.0 47.0

M3: L 46.0 38.0
W 50.0 44.0

setting of this region and its amenability to endemic evolution. However, as noted
earlier, the Toledo Bend Local Fauna may sample a temporal interval missing in
strata of the northern Great Plains thus biasing the apparent endemic nature of the
Gulf region during the Arikareean. In other words, are these small forms the
primitive members of lineages that would attain greater size upon dispersal
elsewhere? This is supported by the fact that the small Gulf Coast chalicothere and
amphicyonid probably predate the much larger and morphologically more derived
Moropus elatus and Daphoenodon superbus from late Arikareean faunas of the
Great Plains. Hence, "dwarf' as used above is not meant to imply a size decrease
from a larger ancestral member of the lineage.


Subfamily DICERATHERIINAE Dollo, 1885
Genus Diceratherium Marsh, 1875

In addition to Gulfoceras westfalli, discussed above, two larger rhinoceroses
also occur at Toledo Bend based primarily on the presence of two larger, yet
significantly different sized P4s and magna. The larger of the two is referred to
Diceratherium armatum; the smaller to D. annectens. It should be noted, however,
that comparison with specimens at the ACM, AMNH, MCZ, and YPM resulted in
the conclusion that there exists considerable confusion regarding the distinction
between Diceratherium armatum Marsh (1875), D. annectens (Marsh, 1873), and D.
gregorii Peterson (1920).
The type specimens of D. armatum (the genoholotype) and D. annectens were
originally described from the middle John Day Formation, Oregon. Peterson (1920)
noted that both were from the same horizon, and Tanner (1969) determined from
YPM records that they were collected at "Turtle Cove." Although Macdonald
(1963) reported that D. annectens had not been recovered in the Wounded Knee
faunas, South Dakota, Prothero et al. (1989) reported both males and females of D.
armatum and D. annectens from the early late Arikareean 77 Hill Quarry near Lusk,
Wyoming. The type specimen of D. gregorii is from the "Lower Rosebud beds,
near Rosebud Indian Agency, South Dakota" (Peterson, 1920:421).
Peterson (1920) considered only two of six John Day species of Diceratherium
to be valid (D. armatum and D. annectens), placing the others incertae sedis. He
synonymized D. nanum Marsh (1875) with D. annectens, and erected a new species,
D. gregorii, for material from the High Plains. Agreeing with Peterson's
conclusions, Troxell (1921) nonetheless erected two more species from the John
Day, D. lobatum and D. cuspidatum. Green (1958:588) considered D. gregorii a
valid taxon and reported additional specimens from "a nodular zone above the basal
ash" in South Dakota. The "basal ash" is the Rockyford Ash at the base of the early
Arikareean Sharps Formation (see Tedford et al., 1987, 1996, for further
discussion). Later, Macdonald (1963, 1970) reported D. gregorii from the Wounded
Knee Sharps Fauna. Green (1958) also concluded that D. armatum, with which he
synonymized D. lobatum, occurred in South Dakota. For reasons discussed below,
it is suggested here that D. cuspidatum be synonymized with D. annectens. The
validity of D. gregorii, in my opinion, remains questionable pending a thorough
review of the genus.
The largest Toledo Bend rhino material, including a large P4 and fragments
of the zygoma, post-glenoid process, and distal femur, are virtually identical to
material referred to D. armatum from the 77 Hill Quarry and to some specimens
from the John Day Formation (e. g., AMNH 7321). With the exception of the large
P4, the remainder of the upper cheek teeth are substantially smaller than those of the
type specimen ofD. armatum (YPM 10003). These teeth compare well with those
of the type specimen of D. annectens and the type specimen of D. nanum (= D.
annectens), YPM 10001 and AMNH 7324, respectively, and with other specimens


referred to D. annectens including YPM 11184 and YPM 12493. They also
compare well with AC 1828 and AC 4509 labeled D. armatum, although more
likely representing D. annectens based on size (see Table 5).
The late Arikareean-early Hemingfordian immigrant Menoceras Troxell
(1921), from early Miocene faunas of the Great Plains and Coastal Plain (Prothero
and Manning, 1987), has a number of derived characters absent in the Toledo Bend
rhinos. These include the prominent, highly plicated crochet and the weak to absent
lingual cingula in the upper molars and increased hypsodonty of the lower molars
(Prothero et al., 1986; Prothero et al., 1987; Prothero et al., 1989). Whether the
weak to absent upper molar cingulum is truly a derived feature is debatable.
Cingulum prominence is often variable and Wood (1964), in describing the type
specimen of Menoceras barbouri, noted complete lingual cingula on P4-M3. Most
post-cranial elements from the Toledo Bend species are larger than those of both
Menoceras arikarense (Barbour, 1906) and the larger M. barbouri (Wood, 1964)
(including the still larger M. marslandensis Tanner, 1969, which was synonymized
with M. barbouri by Prothero et al., 1989:328).
Note: In his discussion of the Rhinocerotidae, Prothero (1998:600) listed
Menoceras barbouri and Floridaceras white as occurring in the Toledo Bend Local
Fauna, evidently following the preliminary faunal list of Manning (1990). More
detailed study revealed that this material represents Diceratherium annectens and D.
armatum. Thus, the record of Menoceras and Floridaceras at Toledo Bend is in

Diceratherium annectens (Marsh, 1873)
Figures 6-9

Rhinoceros annectens Marsh, 1873
Diceratherium nanum Marsh, 1875
Diceratherium annectens (Marsh). Loomis, 1908
Diceratherium annectens (Marsh). Peterson, 1920
Diceratherium cuspidatum Troxell, 1921

Holotype.-YPM 10001, upper left premolars with associated upper incisor.
"lower to middle John Day Formation" (Peterson, 1920:417).
Referred Specimens.-LSUMG V-2250, left II; V-2526, left II; V-2527, left
P1; V-2528, left P2; V-2529, right P2; V-2530, right P3; V-2531, left P3 fragment;
V-2532, left P3 fragment; V-2533, left P4; V-2534, right P4; V-2535, right P4
fragment; V-2572, upper cheek tooth (dP4?) fragment; V-2536, right Ml or M2
(worn); V-2265, right M2; V-2266, left M3; V-2537, upper tooth fragments; V-
2538, right p2; V-2539, right p3; V-2763, right p3, figured; V-2540, right ml? in
jaw fragment; V-2541, talonid of right m2; V-2542, left m2 fragment; V-2543, left
m3; V-2544, talonid of right m3; V-2545, lower tooth fragments; V-2546,


50- O

45- [
O Toledo Bend D. annectens,
V-2533, 2534
S40- D. annectens type (YPM 10001)
C D. annectens (AMNH 7324)
35 ( D. "armatum" (AC 1828)
35 -
C ), (8) D. sp. (AMNH 7342)
] D. armatum type (YPM 10003)
25 30 35 D. gregoriitype (AMNH 12933)
AP Length ] Toledo Bend D. armatum,

Figure 6. Bivariate scatter plot of various Diceratherium P4s (mm).

mandibular symphysis fragment; V-2547, fragment of post-glenoid process; V-
2548, fragments of atlas vertebra; V-2549, two proximal right scapulae fragments;
V-2550, proximal right humerus; V-2551, distal left humerus; V-2552, distal right
humerus; V-2553, proximal right ulna; V-2554, proximal right Mc III; V-2555,
distal left Mc IV; V-2556, magnum; V-2557, left astragalus; V-2558, distal right
calcaneum; V-2559, right periotic.
Description and Discussion.-Table 5 and Figure 6 compare the size of upper
cheek teeth of various species of Diceratherium, including the Toledo Bend
specimens. Dental terminology follows that of Prothero et al. (1986:354, fig. 6).
Incisors from Toledo Bend compare closest with those of AMNH 7312 and 7346
from "Turtle Cove." The crown of the Ii is short anteroposteriorly and narrow
transversely (Fig. 7E).
Premolars from Toledo Bend also most closely resemble those from various
John Day specimens. The P1 (Fig. 8A) compares best with AMNH 7346 and 7324
(type of D. nanum). P1 of D. annectens is more anteroposteriorly compressed and
transversely broader than that of Menoceras, and it has a labial cingulum continuous
with the parastyle and metastyle not seen in Menoceras. There is also a lingual
cingular segment between the parastyle and protocone and between the protocone
and the hypocone. P2-3 compare well with AC 1828 and AC 4509. Although
labeled "Diceratherium armatum from the Lower Harrison near Agate, Nebraska,"


Figure 7. (A) Diceratherium armatum, stereo view of right P4, LSUMG V-2560; (B) D. annectens,
stereo view of left P4, V-2533; (C) D. annectens, stereo view of labial surface of right p3, V-2763; (D)
Same specimen, lingual surface; (E) D. annectens, left II, V-2250.




0 0

0 0



m U)




AC 1828 and 4509 are of size more suggestive of D. annectens (Table 5).
Premolars and molars in the two skulls match those from John Day and Toledo
Bend in size and in the weak to nonexistent crochets and cristae. P2s from Toledo
Bend are nearly square in occlusal outline with a cingulum that is continuous
around the tooth, although weak at the hypocone (Fig. 8B). There is also a "mure"
connecting the metaloph to the protoloph (see Green, 1958:592) and an incipient
crista. P3 (V-2530) has a continuous anterior, lingual, and posterior cingulum and
a small crochet and crista (Fig. 8C). The labial surface is too water worn to
determine if there was a labial cingulum. A fragment of the lingual part of another
P3 shows no crochet or cristae, but, like P2, it has a tiny "mure."
P4 (V-2534) has a strong labial cingulum that ascends the posterolabial corner
of the crown (Fig. 8D). The anterior, lingual, and posterior cingula are nearly
continuous, broken only at the posterolingual corner of the hypocone. The
transverse median valley is narrow and there is no crochet or crista. V-2533 is
similar with only a slight hint of a crochet. There is a prominent cusp at the
lingual entrance of the transverse median valley, immediately labial to the lingual
cingulum (Fig. 7B).
That V-2533 has a lingual tubercle, whereas V-2534 does not, leads one to
question the validity ofD. cuspidatum. This John Day species is so named because
the type specimen (YPM 12007) shows a prominent cusp from the floor of the
lingual part of the median valley in the molars. Furthermore, the premolars of D.
cuspidatum each show a "mure" joining the metaloph to the protoloph which is one
of the features that Green (1958) used to synonymize D. lobatum with D. armatum.
Diceratherium annectens and the Toledo Bend species also show this small "mure"
and D. cuspidatum is of similar size to D. annectens. For these reasons, D.
cuspidatum is here placed in synonymy with D. annectens. Except for the cusp of
V-2533, the two P4s from Toledo Bend closely resemble the type specimen of D.
nanum (AMNH 7342) from Turtle Cove.
Upper molars (Fig. 8E, F) also compare well with the type and referred
specimens of D. annectens in their simple morphology and size. They are
considerably smaller than molars of the type specimen of D. armatum (YPM
10003). Those from Toledo Bend have no crista and the crochet is small compared
with the strong crochets so diagnostic of Menoceras. The M3 (V-2266) has a
strong anterior and posterior cingulum, the latter of which continues lingually until
ending at the posterolingual surface of the protocone. There is a prominent, non-
plicated crochet that arcs anterolabially and a weak crista. M3 of Menoceras
differs in having a distinct ridge that ascends the posterolabial surface of the
hypocone which is probably a reduced, posterior extension of the ectoloph such as
that seen in Hyracodon. No such ridge is seen in Diceratherium or on the M3 from
Toledo Bend.
Lower premolars from Toledo Bend also compare closest to those of AMNH
7346 from the John Day Formation. Like all other Diceratherium premolars


examined from both the John Day Formation and from the High Plains, those from
Toledo Bend have a lingual cingulum that remains low on the crown and extends
around the anterior surface of the tooth to connect with the labial cingulum. This
differs from the condition that appears to be characteristic of Menoceras in which
the lingual cingulum ascends the lingual surface of the paraconid. This character is
particularly comparable in p3s (Fig. 7C, D). Lower molars ofD. annectens and the
Toledo Bend species are also relatively less hypsodont than those of Menoceras.
Scapula fragments include only the proximal ends. The glenoid surface
measures about 44 mm AP in both fragments and about 35 mm TR. There is no
coracoid process and the scapular spine does not extend distally to the glenoid
surface. Only a slight indication of the spine is apparent about 45 mm from the
The magnum from Toledo Bend referred to D. annectens (Fig. 9A) is virtually
identical to those in the Frick collection from the 77 Hill Quarry and also quite
similar but slightly smaller than that referred to D. niobrarense (AMNH 14212).
The Toledo Bend specimen measures 63.0 mm AP; 27.3 mm of this length includes
the anteroposteriorly concave Mc III facet which is also about 22 mm wide.
Compared with the magnum of Menoceras, the posterior tuber of Diceratherium is
longer and the Mc III facet more concave.
On the proximal Mc III fragment, the strongly convex facet for articulation
with the magnum measures about 38 mm AP by 32 mm TR. Maximum transverse
width of this element is about 44 mm. The lateral surface has two prominent facets,
one dorsally (or anteriorly) and one ventrally (or posteriorly). The anterior facet is
divided into a large, slightly convex, proximolaterally facing facet for articulation
with the uniform and a smaller anteroposteriorly-expanded, distolaterally facing
facet for articulation with the Mc IV. The former is separated from the facet for the
magnum by a prominent ridge. A deep, rugosely-textured sulcus separates the
anterior Mc IV-unciform facets from the posterior Mc IV facet. The latter is slightly
concave and oval-shaped with its long axis extending proximodistally. This element
compares well with F:AM 132057 from 77 Hill Quarry. It is larger than that of
Menoceras barbouri (MCZ 7449). Also, the lateral volar facet in the Toledo Bend
specimen is proximodistally oval, whereas in M. barbouri this facet is
dorsoventrally oval.
The astragalus measures 62.0 mm TR across the trochlea and 53.0 mm TR
across the navicular and cuboid facets. The medial trochlea meets the navicular
facet. The cuboid facet is separated from the navicular facet by a distinct ridge. The
large size of this element and the distal calcaneum fragment suggests that they may
belong to the larger Toledo Bend rhino.
Wood and Wood (1937) described a maxillary fragment with highly worn
teeth (USNM 6573) from the Cedar Run locality in Washington County, Texas, and
referred it to Caenopus cf. premitis (the "Derrick Farm rhino"). Prothero and
Manning (1987) considered the specimen representative of Menoceras arikarense.


Figure 9. (A) Diceratherium annectens, magnum, LSUMG V-2556; (B) Diceratherium armatum, left
magnum, V-2567.

The specimen was originally reported from the Catahoula Formation, but a personal
communication from J. A. Wilson to Prothero and Manning (1987:391) indicated
that it "could easily be derived from a channel from the overlying Oakville
Formation." Although USNM 6573 is similar in size to M. arikarense, it is also
quite similar to D. annectens and even shares with the latter species the close
proximity of the metaloph and protoloph, unlike the more separated condition found
in M. arikarense. In fact, Loomis (1908:54-55), referring to D. annectens, noted
that "this small species is readily distinguished by the fact that on the molars the
protoconule and the hypocone are so closely placed that on a partly worn tooth they
actually join and the intervening valley between the protoloph and metaloph is
interrupted." Similarities between the Cedar Run and Toledo Bend local faunas, in
addition to the similar size and morphology of the Derrick Farm rhino with D.
annectens, suggests the possibility that the former represents D. annectens rather
than M. arikarense.

Diceratherium armatum Marsh, 1875
Figures 7, 9

Diceratherium armatum Marsh. Peterson, 1920
Diceratherium lobatum Troxell, 1921
Diceratherium armatum Marsh. Green, 1958
Diceratherium armatum Marsh. Macdonald, 1963
Diceratherium armatum Marsh. Macdonald, 1970


Holotype.-YPM 10003, complete skull with associated foot bones, "lower
John Day Formation" (Peterson, 1920:414).
Referred Specimens.-LSUMG V-2560, right P4; V-2561, skull fragment
(frontal); V-2562, posterior right zygoma fragment; V-2563, left lateral cranium
fragment with post-glenoid process; V-2564, right post-glenoid process fragment;
V-2565, proximal right radius; V-2566, left cuboid fragment; V-2567, left magnum;
V-2569, lateral proximal phalanx; V-2570, distal right femur; V-2571, distal right
Description.-P4 (V-2560) has a cingulum that completely surrounds the
tooth except at the anterolabial corer. It begins weakly on the anterolabial surface
but is prominent and shelf-like anteriorly. There is no crochet, crista, or connecting
mure, but the metaloph shows four plicae on its anterior surface. V-2560 measures
35.5 mm AP by 44.5 mm TR (Fig. 7A).
A large, abraded, nearly flat skull fragment, undoubtedly part of the frontal of
a large animal, is referred to D. armatum primarily because no other animal of this
time had a skull with enough flat surface area to produce such a fragment.
Interiorly, the fragment shows a honey-combed pattern reminiscent of that seen
internally in elephant skulls in order to reduce weight without compromising
V-2563 is a fragment of the left lateral portion of the skull with a large post-
glenoid process and the internal continuation of the glenoid surface. The post-
glenoid process is directed strongly medially. A fragment of the right posterior
zygoma (V-2562) is dorsoventrally broadest above the glenoid surface (about 48
mm) and tapers anteriorly. The external surface is highly textured.
The partial cuboid (V-2566) consists only of the dorsal portion and measures
34.6 mm proximodistally at its anterior surface. The magnum (V-2567) measures
50 mm transversely and about 36 mm proximodistally at its anterior-most point. Its
maximum proximodistal height is about 58 mm. The articular surface for the Me III
measures 44 mm anteroposteriorly and has a maximum transverse width of 42 mm.
The posterior tuber is reduced (Fig. 9B).
The lateral proximal phalanx measures 36.5 mm long, 29 mm transversely
across the proximal end, and 24 mm across the distal end. The proximal articulating
surface is a nearly round depression with a small notch ventrally for the keel of the
distal metapodial. The distal articulating surface is smooth, flat, and asymmetrical.
The two distal femur fragments are large and show a round, slender shaft.
Maximum transverse width measures about 102 mm across the distal condyles.
Discussion.-This material is referred to D. armatum because of its similarity
to material referred to this species from the 77 Hill Quarry. The P4 is virtually
identical to the larger specimens from 77 Hill (e. g., F:AM 132056) in size, degree
of hypsodonty, and in morphology, including the occurrence of plications on the
anterior surface of the metaloph and the prominent lingual cingulum. Absence of


the long, anteriorly extending crochet and presence of a prominent lingual cingulum
are features that distinguish it from large species of Menoceras.
The zygoma and post-glenoid fragments match those of D. armatum (F:AM
112176) from north of Keeline, Niobrara County, Wyoming, and those of AC 1828
and AC 4509, which are labeled D. armatum but, as noted previously, are more
similar to D. annectens in teeth size. The Toledo Bend elements also compare well
with those of the type specimen of D. armatum, YPM 10003, and with AMNH
7321, another specimen from the John Day Formation. The zygoma of the male
Menoceras, unlike that of Diceratherium, ends in a bulbous, blunt knob, which is
only textured at its posterior-most point. That of the female Menoceras tapers
posteriorly as in Diceratherium, but is considerably smaller, more gracile, and non-
The cuboid and magnum from Toledo Bend are larger than those referred to D.
annectens (e. g., YPM 12493). The cuboid is similar morphologically, but slightly
larger than that of D. niobrarense. The magnum is substantially larger than any
referred to Menoceras and slightly larger than many referred to D. armatum.

Order ARTIODACTYLA Owen, 1848
Family ENTELODONTIDAE Lydecker, 1883
?Dinohyus sp.
Figure 10

Type Species.-Dinohyus hollandi Peterson, 1905a (also see Peterson, 1905b).
Holotype.-AMNH 7387, symphysis of lower jaw with roots of incisors and
canine, John Day Formation, Bridge Creek, Wasco County, Oregon (Peterson,
Referred Specimens.-LSUMG V-2415, incisor fragment; V-2417, left M2
fragment; V-2268, left M3; V-2416, left p4 fragment; V-2418, edentulous anterior
right maxillary fragment with C, P1, and partial P2 alveoli; V-2578, right periotic;
V-2419, right m3 fragment; V-2579, ramal fragment; V-2420, premolar fragments;
V-2575, partial atlas; V-2576, cervical vertebra fragment; V-2577, distal humerus;
V-2426, distal right humerus; V-2421, left magnum; V-2422, distal metatarsal; V-
2423, distal right tibia; V-2424, right astragalus; V-2425, left astragalus.
Description.-The edentulous maxillary fragment (LSUMG V-2418) is
broken, exposing the medial surface of the large canine alveolus. The two alveoli of
P1 are situated medial to the posteromedial corer of the canine alveolus. A short
(29 mm) diastema separates the posterior edge of the posterior P1 alveolus from the
anterior edge of the anterior P2 alveolus. The M2 fragment (V-2417) shows a broad
anterior cingulum that ends at the anterolabial corer of the paracone. This
fragment also shows a small portion of the protoconule. The M3 (V-2268) measures
49.3 mm anteroposteriorly, and would have measured greater than 44 mm


44 41


Figure 10. ?Dinohyus sp. (A) Stereo view of left M3, LSUMG V-2268; (B) Stereo view of right
periotic, lateral side, V-2578. (C) Same specimen, medial side; (D) Left astragalus, V-2425; (E) Right
astragalus, V-2424.


transversely were the labial surface intact (Fig. 10A). The unworn crown is nearly
complete, missing only the labial half of the paracone, and has a textured surface.
The protocone is the largest cusp. A thick, robust cingulum extends continuously
from the anterolabial corer lingually to the metaconule.
The p4 fragment (V-2416) includes the posterior half of the crown and shows a
thick, highly pustulose-textured posterior heel. From the point where the
posterolabial cingulum begins, there is a ridge that ascends the crown, forming a
corerr" effectively separating the labial surface of the tooth from the posterior
surface. The m3 fragment (V-2419) consists of only a single worn cusp.
The periotic is large and massively constructed (Fig. 10B, C). A detailed
description is provided in Albright (1991). Fragments of the atlas (V-2575) and
cervical vertebrae (V-2576) compare closely in size and morphology with those of
Dinohyus hollandi in the F:AM collection.
The distal metapodial III or IV fragment is convex on the anterior surface and
flat to concave on the posterior surface. The metapodial keel is placed somewhat
asymmetrically on the distal articular surface and, as in D. hollandi, does not extend
onto the anterior face of that surface. The distal articular surface measures 47.5 mm
The proximal surface of the magnum (V-2421) is dominated by the facet for
the scaphoid. The scaphoid facet is separated from the posterolaterally facing lunar
facet by an oblique ridge. On the medial side of the magnum is a rugosely textured
depression where the trapezoid articulates. The lateral side has a deep, rhomb-
shaped excavation centrally located between the two facets for the uniform. The
anterior facet for the uniform is flat and the posteror one concave. The distal
surface for articulation with Mc III is concave anteroposteriorly and convex
transversely. The magnum from Toledo Bend is smaller than that figured by
Peterson (1909:120, fig. 72) for D. hollandi (maximum AP diameter = 52.8 mm vs.
70 mm).
The distal tibia (V-2423) is flat anteriorly with a small, but prominent medial
malleolus. The astragalar facets are aligned parallel to the sagittal plane. The
medial facet is slightly narrower than the lateral. The distal fibula was not
coossified with the tibia. The astragalar facet measures about 90 mm TR and about
79 mm AP. The smaller of the two astragali (V-2425) measures 92 mm long by
62.5 mm wide. The larger, heavily worn specimen (V-2424) measures 110 mm
long by 66 mm wide (Fig. 10D, E).
Discussion.-The only described late Oligocene-early Miocene North
American artiodactyls to which material of this size can be referred are the
entelodonts Daeodon shoshonensis Cope (1878), Ammodon leidyanum Marsh
(1893), Dinohyus hollandi Peterson (1905a) (see also Peterson, 1905b), Dinohyus
(?) mento Allen (1926), or Archaeotherium trippensis Skinner et al. (1968). In
describing the poor type specimen of Daeodon, D. shoshonensis, from the John Day
Formation, Cope (1878) referred to it as one of the largest species of North
American perissodactyls. Studying additional material published by Sinclair (1905),


Peterson (1909) removed it from the Perissodactyla and placed it into the artiodactyl
family Entelodontidae. In the same publication, Peterson listed additional traits that
distinguished Daeodon from Dinohyus hollandi, another large entelodont that he
had previously described.
Peterson (1909) also considered Dinohyus hollandi, from the Agate fossil beds,
distinct from Ammodon leidyanum, as did Troxell (1920:441) who considered "A.
leidyanus" generically separate "from all the specimens found on the Great Plains."
Wilson (1957), on the other hand, felt that the generic distinction of Ammodon was
probably dubious, adding that its relationship to Dinohyus or Daeodon would have
to await the recovery of more material. Brunet (1975) similarly noted that
additional material was needed for validation of the various purported species
because many type specimens were undiagnostic. He cited the following examples:
the type specimens of Daeodon shoshonensis and Dinohyus (?) mento are both
mandibular symphyses, and Boochoerus Cope (1879) from the John Day Formation
is only known from post-crania (see Foss and Fremd, 1998). Brunet (1975) noted
that because Boochoerus and Daeodon are both from the John Day Formation, and
because Boochoerus and Dinohyus share a magnum and uniform separated by a
semi-lunar and both lack the trapezium, the four genera (including Ammodon based
on tooth similarities to D. hollandi) could probably be grouped into one genus. He
preferred Dinohyus to Daeodon because Dinohyus hollandi is known from a
complete skeleton, whereas the others (Daeodon, Ammodon, Boochoerus) are
known from much less complete material. He concluded, however, that specific
relationships between them would be nearly impossible to determine. Similarly,
Parris and Green (1969:1278), in their description of Dinohyus sp. from the Sharps
Formation, South Dakota, refrained from naming a new species "because of the
uncertain state of entelodont taxonomy." Lucas et al. (1997, 1998) recently reached
a conclusion similar to that of Brunet (although they did not mention Boochoerus),
but placed the above named species in Daeodon, rather than Dinohyus, based on
nomenclatural priority. Brunet (1975) listed numerous traits suggesting that these
large earliest Miocene forms are of Asiatic origin unrelated to the older, smaller,
North American Oligocene forms.
The other large entelodont of slightly older age is Archaeotherium trippensis.
The type specimen, F:AM 42937, includes a complete skull of an immature
individual from the Wewela Local Fauna, South Dakota, which Tedford et al.
(1987:171) considered "of Monroe Creek age." Similar in size to Dinohyus
hollandi, A. trippensis is considerably larger than other species of Archaeotherium.
Skinner et al. (1968) cited numerous characters that distinguish it from D. hollandi.
The Toledo Bend entelodont is questionably referred to Dinohyus, because, in
my opinion, the type specimen of Daeodon is "inadequate for definitive diagnosis"
and the genus, therefore, should be considered a nomen vanum (see Mones,
1989:232). Specific distinction of the Toledo Bend species is also ambiguous due to
the paucity of material. The questionable referral is also based on some major


morphological differences seen in the Toledo Bend specimen; these are discussed
The Toledo Bend species and D. hollandi share M3s that are similar in size,
that lack a metaconule, and that have a broad anterior cingulum and no posterior
cingulum. The Toledo Bend p4 fragment resembles that described of Ammodon
leidyanum in its large size and in having the broad, posteriorly extending, pustulose
textured posterior margin. In this character, the Toledo Bend species also resembles
A. trippensis.
The most diagnostic feature of the Toledo Bend species, and the most obvious
difference between it and other large entelodonts, is the position of P1 relative to the
canine. In the Toledo Bend form, the PI, as shown by its two alveoli, is situated
nearly medial to the canine rather than posterior to it. Only the posterior-most
margin of the posterior alveolus extends slightly posterior to the canine alveolus. In
Dinohyus hollandi, the P1 is directly posterior to the canine and separated from it by
a very short diastema. Dinohyus sp. from the Sharps Formation, which Parris and
Green (1969) noted more closely resembled the Whitneyan Pelonax lemleyi
Macdonald (1951), also has a P1 posterior to the canine alveolus. In the type
specimen of A. trippensis, the P1 is also clearly posterior to the canine. However, in
comparing A. trippensis with Dinohyus, Skinner et al. (1968) noted that the former
had a long, slender rostrum that was "not greatly expanded in the area of the incisors
and canines as in Dinohyus hollandi ... ." The canine of A. trippensis, therefore, had
to curve abruptly in a posterior direction because of its slender rostrum (Skinner et
al., 1968:421, fig. 15). If the lateral side of the alveolus for the canine was broken
away, it might appear that the P1 alveoli were somewhat medial to the canine
The other major difference between the Toledo Bend species and D. hollandi
concerns the distal ankylosis of the fibula to the tibia. In D. hollandi the tibia and
fibula are fused. The Toledo Bend form appears to lack the fused condition. The
postcranial anatomy of A. trippensis is unknown.
Entelodont material from the Gulf Coastal Plain, in addition to that from
Toledo Bend, includes two specimens (TMM 40223-1 and TMM 40224-1) from the
Garvin Gully Fauna that Wilson (1957) referred to Dinohyus hollandi. Although he
pointed out several differences between the Texas specimens and the type specimen
of D. hollandi, he did not cite any specific reasons for referral of these specimens to
that species. The lower jaw, TMM 40224-1, from "beds close to the base of the
Fleming formation" in San Jacinto County, Texas, does not show the tuberosities on
the chin supposedly diagnostic of Dinohyus, although there are prominent
tuberosities located more posteriorly (Wilson, 1957:644). Degree of tuberosity
expression may be a function of sexual dimorphism. TMM 40223-1 was found near
the base of the Oakville Formation in Washington County, Texas. A direct
comparison is not possible as the TMM specimen includes only the P4 and Ml.
Unfortunately, there is no postcranial material from Garvin Gully faunas that might
provide important comparisons. Another specimen that includes a left mandibular


ramus with pl-m3 was recently reported by Westgate (1993) as having been found
in the Catahoula Formation of Fayette County, Texas. With a molar row length of
102 mm, this entelodont may have been slightly smaller than that from Toledo
Bend. Elsewhere in the Gulf Coastal Plain, two broken teeth and a radio-ulna
fragment from the Franklin Phosphate Pit No. 2 Local Fauna, Florida, were referred
to Dinohyus (Simpson, 1930).
The position of P1 relative to the canine and the lack of an ankylosed distal
tibia-fibula are features that would seem to indicate distinctiveness for the Toledo
Bend entelodont. But definitive taxonomic referral is declined pending recovery of
additional material and a more thorough understanding of the relationships between
these large, earliest Miocene "genera" and "species."

Family TAYASSUIDAE Palmer, 1897
?Floridachoerus olseni White, 1941
Figure 11A

Floridachoerus olseni White, 1941
Desmathyus olseni (White). MacFadden and Webb, 1982.

Holotype.-MCZ 3657, partial skull with right P3-M3 and left P4-M3, from
Thomas Farm, Gilchrist County, Florida.
Referred Specimens.-LSUMG V-2504, right Il; V-2505, right upper C; V-
2267, left ramal fragment with ml-3; V-2506, distal left and right tibia fragments;
V-2507, astragalus.
Description.-I1 is relatively large and bulbous with a weak labial cingulum.
It is somewhat excavated lingually with a small, centrally situated ridge that runs
from the base of the crown to slightly less than half the distance to the apex. The
dagger-like upper canine has a crown length of 36.1 mm and measures 18.5 mm AP
by 11.5 mm TR. It is nearly flat lingually and convex labially. The anterior surface
is worn flat from contact with the posterior surface of the lower canine while the
posterior surface is sharp and straight-edged. Wright and Eshelman (1987:609)
stated that "identification of tayassuid species based on canine size [is] practically
The ml is highly worn. The m2 is also quite worn; the transverse median
valley is blocked internally by a transversely expanded conulid. There is also a
small tubercle at the labial entrance of the transverse median valley that is
continuous with an anterior cingulum. This cingulum weakly wraps around the
labial surface of the protoconid and joins it (Fig. 11A). The m3 has an anterior and
a labial cingulum which may have been barely continuous before the tooth was
worn. A paracristid extends lingually from the protoconid and joins with the
anterior cingulum. At the entrance to the valleys between the protoconid and
hypoconid and between the hypoconid and hypoconulid, the labial cingulum


Table 6. Lower molar measurements for various species of early Miocene peccaries.

ml m2 m3

Toledo Bend
LSUMG V-2346 20.0 11.0
LSUMGV-2267 15.8 12.3 17.9 14.0 22.3 13.1

Floridachoerus olseni
MCZ 7302 15.9 12.7
MCZ 7304 21.5 13.6

Thinohyus decedens
UCMP 1989 13.0 9.0 22.0 12.0

Van Tassell species
ACM7035 16.3 12.6 18.5 14.3 25.2 14.4

"Thinohyus" siouxensis
CM 1423 (type) 18.9 15.9 26.8 16.2

Hesperhys pinensis
AMNH 12936 (type) 17.8 13.6 22.6 16.4

H. cf. H. pinensis
UNSM 62604 17.0 13.3 20.2 16.1 25.5 16.7

thickens to form tubercle-like structures. The hypoconulid is asymmetrically
bilobate with the labial lobe being larger. There is a transversely expanded conulid
centrally located within the tranverse median valley and another one between the
hypoconulid and the hypoconid-entoconid lophid. Although the protoconid,
metaconid, hypoconid, and entoconid are separate and distinct, the two anterior
cusps and the two posterior cusps, respectively, would have united to form
transverse lophids once the tooth was worn.
Discussion.-Although lower molars and postcrania are not particularly
diagnostic, late Oligocene-early Miocene North American tayassuids of similar size
to the larger Toledo Bend species include Thinohyus decedens (Cope, 1879) from
the "Diceratherium beds, Middle John Day" (Sinclair, 1905:134), Hesperhys
pinensis (Matthew, 1907), "Thinohyus" siouxensis (Peterson, 1905c; see Wright,
1998:395), and the "Van Tassell species" (Wright, 1991:80) from the late
Arikareean of the Great Plains, and Floridachoerus olseni White (1941) from
Thomas Farm (Table 6). All of the above except T. decedens are members of
Wright's (1991, 1998) Hesperhys-"Cynorca" social clade of peccaries.


The Toledo Bend species is closest in size to Floridachoerus olseni. Although
White (1941, 1942) reported no lower molars for F. olseni, Wright's (1991:100)
more recent analysis does include some. The Toledo Bend m3 resembles that
described by Wright in having a large and bilobed hypoconulid. There is no
mention, however, of interloph structures such as those seen in the Toledo Bend
tooth, nor is the cingulum morphology described. Wright (pers. comm., 1990)
found the labial cingulum to be variable within the same species and therefore
generally undiagnostic. Although the anterior half of a worn tooth from the Cedar
Run Local Fauna labeled Floridachoerus (TMM 40068-14) is of similar size to the
m2 in the Toledo Bend jaw fragment, the specimen is in poor, taxonomically
undiagnostic condition.
The lack of more diagnostic material, such as premolars, hinders an
unequivocal identification. Referral to F. olseni would extend the temporal range of
that taxon from the early Hemingfordian to the late Arikareean, whereas Hesperhys
pinensis, "Thinohyus" siouxensis, and the Van Tassell species are likely of similar
age to the Toledo Bend species.

?Hesperhys sp.
Figure 11B

Type Species.-Hesperhys vagrans Douglass, 1903
Referred Specimens.-LSUMG V-2346, right m3.
Description.-This isolated m3 has an anterior cingulum and small cingular
segments between the protoconid and hypoconid and between the hypoconid and
hypoconulid (Fig. 11B). There is no continuous labial cingulum in contrast to the
m3 of the larger species described above. A paracristid extends lingually from the
protoconid and joins with the anterior cingulum. The hypoconulid is
asymmetrically bilobate; the labial lobe is larger. The lingual lobe appears to have
been made up of three different cuspules before wear reduced them. There is a
transversely expanded conulid centrally located within the tranverse median valley
and another one between the hypoconulid and the hypoconid-entoconid loph. The
tooth measures 20.0 mm AP by 11.0 mm TR.
Discussion.-As noted previously, lower molars are not particularly diagnostic
in the hesperhyine group of peccaries. Although this tooth may represent the
smaller end of any variation that may have existed in the large Toledo Bend species,
it is about 10% shorter and more gracile than that of the latter. As sexual
dimorphism is not detectable, except in the canines, of these early peccaries (D.
Wright, pers. comm., 1990), taxonomic referral is questionable at this time.


0 2

iB iE

Figure 11. (A) ?Floridachoerus olseni, stereo view of left ramal fragment with ml-3, LSUMG V-2267;
(B) ?Hesperhys sp., stereo view of right m3, V-2346; (C) "Cynorca" social, stereo view of left MI, V-
2617; (D) "Cynorca" social, stereo view of right ml, V-2271.


Genus Cynorca Cope, 1867
"Cynorca" social (Marsh, 1875)
Figure 11C, D

Thinohyussocialis Marsh, 1875
Palaeochoerus socialis (Marsh). Cope, 1879
Thinohyus socialis Marsh. Stehlin, 1899
Thinohyus socialis Marsh. Sinclair, 1905
Thinohyus (Bothrolabis) socialis (Marsh). Merriam and Sinclair, 1907
Perchoerus socialis (Marsh). Matthew, 1909
Cynorca social (Marsh). Woodbure, 1969
"Cynorca" social (Marsh). Wright and Eshelman, 1987

Holotype.-YPM 11785, associated right and left M2-M3, John Day
Formation, Oregon (Woodburne, 1969:289).
Referred Specimens.-LSUMG V-2617, left M1; V-2271, right ml.
Description.-M1 has a nearly square occlusal outline with a continuous
anterior, labial, and posterior cingulum (Fig. 11C). There is no lingual cingulum.
All four cusps are similar in size and there is a metaconule located slightly anterior
to the metacone and hypocone, but still posterior to the transverse median valley.
The tooth measures 9.4 mm AP by 9.2 mm TR.
The ml is anteroposteriorly elongate (Fig. 1 ID). The metaconid is the largest
cusp, the protoconid is only slightly larger than the hypoconid, and the entoconid is
the smallest. The protoconid is situated slightly anterior to the metaconid and is
separated by a larger distance from the hypoconid than the metaconid is from the
entoconid. A paralophid originates from the anterior surface of the protoconid and
extends lingually a short distance across the anterior surface of the metaconid.
There is both an anterior and posterior cingulum and a prominent hypoconulid.
There is also a prominent entoconulid (see Wright, 1991:41) in the center of the four
cusps, blocking the transverse median valley, and a tiny cingular shelf at the labial
entrance of that valley. The tooth is in an early stage of wear and measures 9.4 mm
AP by 6.9 mm TR.
Discussion.-Wright and Eshelman (1987) and Wright (1991, 1998) regard
Cynorca a nomen dubium because of the inadequacy of the genoholotype (C.
proterva) to distinguish the taxon; hence the quotation marks in the systematic
heading above.
Ml clearly belongs to a species in Wright's (1991, 1998) Hesperhys-Tayassu
clade because the paraconule lies slightly anterior to the paracone and protocone,
and similarly, the metaconule is slightly anterior to the metacone and hypocone. In
less derived taxa, the paraconule and metaconule "lie between the bases of the labial
and lingual principal cusps ... (Wright, 1991:271). That the four principal cusps of
the tooth form the vertices of a square further indicates that this species can be
placed within Wright's (1991:271, 282-283) Hesperhys-"Cynorca" social clade. In
Wright's "Prosthennops" xiphodonticus-Tayassu clade, the four principal cusps form
the vertices of an anteroposteriorly elongate rectangle. The Toledo Bend tooth is


similar in size to Wright and Eshelman's (1987:606) "Tayassuid sp. A" (the "Pope's
Creek species" of Wright, 1991) from "bed 2 of the Calvert Formation," Maryland,
and to specimens reported by Woodburne (1969:294) from Wheeler County, Oregon
(e.g., UCMP 66861).
The lower first molar from Toledo Bend resembles that of MCZ 17744,
referred by Woodburne (1969:297) to Cynorca cf. C. social, from the "?Early
Miocene (?Arikareean), 'Loup Fork Tertiary,' Nebraska." TMM-TAMU 2894, from
the Garvin Gully Fauna and also listed in Woodburne's (1969:289) hypodigm for C.
social, is somewhat larger than LSUMG V-2271. Wright (pers. comm., 1990)
considered this specimen insufficiently diagnostic to be unambiguously referred to
"Cynorca" social.

Subfamily BOTHRIODONTINAE Scott, 1940
Genus Arretotherium Douglass, 1901

Arretotherium, the last of the North American anthracotheres, is characterized
by the loss of the paraconule in upper molars and a shortened face. The type
species, A. acridens, was recovered from the Blacktail Deer Creek beds in the upper
Renova Formation, western Montana (Douglass, 1901), considered lower Miocene
by Hibbard and Keenmon (1950:198) and possibly equivalent to the Harrison
Formation by Tedford et al. (1987:164). The type specimen, CM 704, includes
parts of the skull with most of the upper dentition and various post-cranial elements.
A second species, Arretotherium leptodus (Matthew, 1909), was recovered
from the "Lower Rosebud beds" on the Pine Ridge Reservation, Shannon County,
South Dakota, now considered equivalent to the early to "medial" Arikareean
Sharps, Monroe Creek, and Harrison formations (Tedford et al., 1987:171; see
Skinner et al., 1968, for a discussion of problems pertaining to the "Rosebud beds").
The type specimen of A. leptodus (AMNH 13005), originally described as Ancodon
(?=Bothriodon) leptodus, consists of a skull with lower jaws and some post-cranial
elements. Macdonald (1956:616) considered Ancodon leptodus a nomen vanum
because he claimed so little information could be gained from the poor condition of
the type. Later, however, Macdonald (1963:229) placed Ancodon leptodus in the
genus Arretotherium based primarily on its age, and additionally noted "upper
molars with reduced mesostyle, transverse groove not prominent labially, short
diastema between upper canine and P1, [and] canine moderately compressed." A
shortened snout was previously noted by Matthew (1909).
Macdonald's nomen vanum designation was due partly to the molars of
AMNH 13005 being too worn to determine presence or absence of the paraconule.
However, another skull, F:AM 132055, closely resembling the type of A. leptodus
and collected in the same area, clearly shows the absence of the paraconule leaving
little doubt as to the generic status of the type. The nomen vanum status of
Arretotherium leptodus should therefore be disregarded. Furthermore, A. acridens


and A. leptodus may represent the same species, although further study is required
to confirm this.
A third species, Arretotherium fricki Macdonald and Schultz (1956), was
described from a badly crushed female skull, UNSM 5764, from the "Upper
Marsland" Formation, Box Butte County, Nebraska. The upper part of the
Marsland Formation in Box Butte County is equivalent to the early Hemingfordian
Runningwater Formation. These rocks are not equivalent to strata in Sioux County,
Nebraska, and Goshen and Niobrara counties, Wyoming, that Schultz (1938)
termed the Marsland Formation and considered equivalent to those in Box Butte
County. The latter strata are equivalent to Peterson's (1907) Upper Harrison beds
(for a discussion on the above, see McKenna, 1965, and Tedford et al., 1987:166,
167, 186).
Macdonald (1956) noted another possible species based on a lower molar,
SDSM 53440, from the Flint Hill quarry in the Batesland Formation of Bennett
County, South Dakota, but Stirton (pers. comm. with Macdonald and Schultz, 1956)
subsequently referred that specimen to A. fricki (Harksen and Macdonald, 1967:7).
Recently, Macdonald and Martin (1987) described additional material of A. fricki
from Flint Hill, including a well preserved skull (UCMP 32369).
Arretotherium acridens and A. leptodus are distinguished from A. fricki by
their heavier, deeper, more robust jaw morphology in contrast to the slender, more
gracile rami of the latter. Although the anterior portion of the mandible is spatulate
in all species, it is much more so in A. fricki. In general, the upper molars of A.
acridens are larger than those of A. fricki. A broadly separated M3 mesostyle
supposedly distinguishes A. acridens from A. fricki (Macdonald and Schultz, 1956),
but the M3 mesostyle of the latter species is also well divided.

Arretotherium acridens Douglass, 1901
Figures 12, 13

Arretotherium acridens Douglass, 1901
Arretotherium acridens Douglass. Macdonald, 1956

Holotype.-CM 704, partial skull with upper dentition and associated post-
crania, Blacktail Deer Creek beds, Renova Formation, Montana (Douglass, 1901).
Referred Specimens.-Toledo Bend: LSUMG V-2351, left upper C; V-
2352, dPl?; V-2353, P1; V- 2354, two right, one left P2; V-2355, left P3; V-2356,
right P4; V-2357, right P4; V-2358, right P4; V-2359, left P4; V-2360, right dP4;
V-2361, left dP4; V-2362, two left Mls; V-2363, right M1; V-2487, right M1 or
dP4; V-2364, two right M2s; V-2365, two left M2s; V-2366, right M3; V-2367, left
M3; V-2368, left upper molar "ghost"; V-2269, right maxilla fragment with M2-
M3; V-2474, left lateral basicranial fragment; V-2370, three incisors (13/3?); V-
2371, lower left C, male; V-2372, lower left C, female; V-2373, two right p4s; V-
2374, two left m2s; V-2375, right m3; V-2376, m3 fragments; V-2377, lower tooth


fragments; V-2378, tooth fragments; V-2379, right mandibular symphysis with
canine fragment, adult male; V-2380, left mandibular symphysis with canine, adult
male; V-2381, left mandibular symphysis with canine and pi fragment, adult male;
V-2382, right mandibular symphysis fragment with canine root; V-2383, right
mandibular symphysis, edentulous, juvenile male; V-2384, left mandibular
symphysis, edentulous, juvenile female; V-2347, left ramal fragment with p4-m2;
V-2270, right ramal fragment with ml-m3; V-2348, left ramal fragment with m3;
V-2349, right ramal fragment with partial m3 root; V-2350, mandibular angle
fragments; V-2385, proximal right scapula; V-2386, proximal left humerus; V-
2387, proximal right radioulna; V-2458, proximal right radius fragment; V-2388,
right lunar; V-2389, left lunar; V-2476, left Mc I; V-2390, proximal metacarpal III;
V-2391, three left, one right proximal Mc IV; V-2392, two proximal Mc IV
fragments; V-2477, one right, one left Me V; V-2478, left proximal Me V; V-2393,
distal metapodial; V-2394, distal metapodial; V-2395, four proximal phalanges plus
fragments; V-2396, medial phalanges; V-2397, lateral digit distal phalanx; V-2699,
medial digit distal phalanx; V-2480, sacrum fragment; V-2481, patella; V-2482,
tibia shaft fragments; V-2398, one left, two right distal tibia fragments; V-2475,
juvenile right distal tibia; V-2399, five calcanea; V-2400, four astragali; V-2401,
five naviculars; V-2402, left cuboid; V-2403, right cuboid; V-2479, proximal left
Mt II; V-2404, three proximal Mt III; V-2405, proximal Mt IV; V-2483, distal
metapodial fragments; Garvin Gully: TMM 31084-168, protocone of upper left
molar; TMM 31084-60, anterior half of lower molar; Mann Place: TMM-TAMU
3065, left M3.
Description.-Dentition: Measurements of teeth are provided in Table 7.
Incisors are spatulate with a well developed internal cingulum on some specimens
(Fig. 12A, B). The male upper canine (V-2351) is laterally compressed, nearly flat
on the lingual surface, convex on the labial surface, with a prominent wear facet on
the anterior surface. Canines of female Arretotherium are more rounded than
laterally compressed (Macdonald and Martin, 1987).
P2 is triangular in occlusal outline with a cingulum, heavy and rounded at the
posterior lingual angle, that completely surrounds the base of the tooth. The apex
of the laterally compressed blade is anterior of center. P3 is similar to P2 but larger,
higher crowned, and the apex of the thin blade is more centrally placed. P4 has a
lingual and labial cusp with prominent anterior and posterior cingula. A lingual
cingulum is prominent on one specimen, reduced on another, and nonexistent on
two other specimens.
The upper molars have a deep transverse median valley with strong, shelf-like
anterior and posterior cingula (Fig. 13A). There is a short, prominent lingual
cingular segment between the protocone and metaconule that extends from the
internal surface of the protocone toward, but not abutting, the interior surface of the
metaconule. There is no internal blockage of the transverse median valley between
the lingual cingular segment and the mesostyle. The median buttresses of the

Table 7A. Comparative measurements of upper cheek teeth of various species and specimens of Arretotherium (* type specimens).

A. acridens A .leptodus* A. leptodus A. fricki A. fricki A. fricki A. fricki
Toledo Bend CM 704* AMNH 13005 F:AM UNSM 5764 UCMP 32369 UCMP 37440 F:AM 132053
L W L W L W 132055 L W L W L W L W
P2 15.9 11.9 13.5 7.9 7.9 ---
15.1 12.4
16.3 11.9
P3 19.2 --- 17.0 14.3 18.5 13.6 16.1 12.7 15.4 11.9
15.9 11.8
P4 13.5 19.7 15.5 18.8 13.6 17.0 14.6 16.2 12.6 15.3 12.3 15.9 11.7 16.8
13.4 17.1 14.0 17.1 13.0 16.5 12.2 15.5
13.7 20.3
13.3 17.6
M1 22.5 20.0 18.3 20.0 18.0 ---- 18.1 20.4 19.6 20.7 19.5 20.2 19.25 18.25
24.4 23.3 18.0 20.0+ 20.1 20.0 18.7 20.4 20.2 20.1
22.1 22.9
M2 24.6 26.9 24.7 26.4 22.9 24.5 26.0 24.1 24.4 23.3 23.9 22.8 24.4 25.3
27.1 25.0 25.8 26.8 22.3 23.7 24.7 24.2 24.6 24.4 23.7 22.9 23.8 25.1
25.4 25.6
25.4 --
25.0 23.5
26.4 24.6
M3 28.1 26.8 28.3 29.4 24.9 26.6 27.6 27.0 25.7 24.9 25.9 26.1 25.5 27.0
26.5 26.2 25.8 28.7 27.4 27.0 25.3 24.75 25.5 26.9
27.5 26.5


Table 7B." Comparative measurements of lower cheek teeth of various species and specimens of

Toledo Bend UCMP 32373 SDSM 53440 SDSM 6826

p4 16.8 10.5 15.6 9.4
17.4 11.4
18.0 11.5
ml 19.6 12.5 18.2 10.5 19.8 11.3 20.3 10.3
m2 24.0 14.4 23.5 12.7
24.3 14.5
m3 35.1 15.7
35.0 14.6

paracone and metacone meet very low, near the base of the crown, and just barely
block the labial entrance to the transverse median valley. Mis are heavily worn and
slightly smaller than M2-3s. The interior surface of the paracone and metacone is
highly crenulated.
The deciduous P4s are molariform and smaller than the molars. V-2360
measures 19.4 mm AP by 17.8 mm TR; V-2361 measures 18.3 by 16.3 mm.
Lower dentition: From the size and shape of alveoli in a mandibular
symphysis fragment (V-2380), i2 was larger than il, which in turn was larger than
i3. The canine in V-2380 is round in cross-section at the crown-root junction. Its
anterior edge is pinched into a carina that, beginning at the apex of the tooth,
descends the anterior edge, curving medially as it approaches the base of the crown
where it formsan anterolingual cingulum. This cingulum has numerous
protuberances giving it a rough, almost serrated texture (Fig. 12C). There is a
prominent wear facet on the posterior edge. Between the canine and pi is a short
diastema. In the adult male specimen (V-2380) there is a long diastema (32.5 mm)
between pi and p2 (Fig. 12C). Juvenile and female mandibular symphysis
specimens maintain similar c-pl diastemata, but that between pl-p2 is reduced
(Table 8). Males also possess a larger alveolus for the canine. All mandibular
symphysis fragments show a laterally directed flare at the anterior end that begins
just posterior to the pi alveolus. Alveoli show that pi was single rooted; p2-3 were
double rooted.
The single-cusped p4 has an anterior, posterolabial, and lingual cingulum.
The labial surface is convex, the lingual surface slightly concave. There is a tiny
paraconid and a ridge, the protocristid, that extends down the lingual side of the
main cusp, the protoconid, to the lingual cingulum.


Figure 12. Arretotherium acridens. (A) Incisor, lingual view (left), labial view (right), LSUMG V-
2370; (B) Incisor, labial view (left), lingual view (right), V-2370; (C) Left anterior ramus with canine,

Lower molars are simple with a high, pointed protoconid and hypoconid that
are V-shaped at their apex and rounded at their base (Fig. 13B, C). Lingual to the
protoconid and hypoconid, respectively, are the similarly high and pointed, slightly
anteriorly curving, metaconid and entoconid. The mis are highly worn with a small
labial cingular segment between the protoconid and the hypoconid. The ml-2 have
a prominent, rugosely textured anterolabial cingulum and an even more robust
posterior cingulum. Between the protoconid and hypoconid, within the median
valley, is a single, small, cone-like tubercle, and the labial entrance to the median


Table 8. Mandibular diastema measurements for Arretotherium.

Specimen age/sex c p pl -p2

LSUMG-V 2380 adult male 10.3 32.5
LSUMG-V 2381 adult male 7.1 -----
LSUMG-V 2379 adult male 15.0 -----
LSUMG-V 2383 juvenile male 8.0 17.3
LSUMG-V 2384 juvenile female 10.4 14.0

valley is blocked by a small cingular segment. The m3 has a prominent
hypoconulid and a strong, rugosely textured cingular segment between the
hypoconid and hypoconulid labially and between the entoconid and hypoconulid

Post-crania: Table 9 provides measurements of the more complete post-cranial
elements. The radius and ulna are solidly fused except for a short (approx. 1 cm)
portion about 25 to 30 mm distal to the humeral facet of the radius. Other proximal
radii fragments indicate that degree of fusion to the ulna may be age dependent.
On the distal tibia, facets for the astragalus are parallel to the sagittal plane
with the medial facet being longer and narrower than the lateral. This morphology
contrasts with the description of A. acridens given by Douglass (1901:274) who
described the same facets as being "very oblique." The anterior surface of this
element is flattened and separated from the lateral surface by a thin ridge. The
distal-most part of this ridge forms the anterior surface of the concavity into which
the fibula articulated. The posterior surface of the distal tibia is rounded. The tip of
the anterior tuberosity is bifurcated.
The distal articulating surface of the metapodials is asymmetrical with a strong
keel that does not extend onto the anterior face. The proximal articulating surface
of the proximal phalanx is slightly dished with a strong notch on its posterior edge
into which articulates the keel of the distal metapodial. The distal end of the
phalanx has two slightly asymmetrical condyles for articulation with the medial
The medial phalanges are shorter and more "stubby" than the proximal ones.
They can be divided into two groups. The proximal articular surface of the group
with the larger elements is transversely broader than it is anteroposteriorly deep and
it shows two slightly asymmetrical, very shallow facets matching the distal articular
surfaces of the proximal phalanges. Elements of the second group are smaller and
have a proximal surface that is equally as broad as it is deep. The distal articular


Figure 13. Arretotherium acridens. (A) Stereo view of right M2-3 in maxillary fragment, LSUMG V-
2269; (B) Stereo view of right ml-3 in ramal fragment, V-2270; (C) Labial view of same.


Table 9. Measurements of various post-cranial elements of Arretotherium from Toledo Bend.
(Number of elements).

Distal tibia (3)
AP diameter
Transverse diameter
Metacarpal I (1)
Metacarpal V (1)
Proximal phalanx (3)
Width, proximal end
Lunar (2)
Dorsoventral diameter
Calcaneum (5)
Length, end of tuber to astragalar fac
Anterior-posterior diameter of tuber
Transverse diameter of tuber
Astragalus (4)
Width, distal end
Cuboid (2)
Width across astragalar facet
Navicular (3)
Dorso-plantar thicknes

35.8 33.0 36.5
40.1 41.1 44.1



32.3 35.2 41.1
17.1 18.5 18.6

28.0 23.5

49.2 68.7 59.7 66.0
22.7 28.2 27.8 28.7
13.2 14.1 15.0 15.0

52.6 56.5 49.0 54.8
27.6 30.5 27.2 29.0

26.1 27.0
13.4 13.5

30.4 30.0 30.3
19.8 19.4 19.3
20.5 19.0 18.8

surface in both groups, as in the proximal phalanges, is divided into two slightly
asymmetrical condyles. The larger elements are probably associated with digits III
and IV while the smaller ones likely represent lateral digits. Two distal phalanges
(claw cores) mirror the situation above. One (V-2397) is small with a highly
asymmetrical articular surface, whereas the other (V-2699) is large and
Discussion.-Absence of a paraconule in the upper molars clearly places the
Toledo Bend species in the genus Arretotherium. Although there is some degree of
overlap in molar measurements between A. acridens and A. fricki, particularly in
Mis, Table 7 and Fig. 14 A-D show that the Toledo Bend species, the type
specimen of A. acridens, the type specimen of A. leptodus (AMNH 13005), and
F:AM 132055 (= A. leptodus), are generally larger than specimens referred to A.
fricki. They also have a deeper, more robust mandible than the gracile morphology
of the Batesland and Runningwater specimens assigned to A. fricki. The latter
include Batesland specimen UCMP 32373 (see Macdonald and Martin, 1987) and


TR Width
-. ta -I C --
~ --D


- .






TR Width
I 0 I CO I

TR Width
N ,

I- N)





i TR Width g


o 0

*n 1> > > I 1 > 1>1

oc -nc e
o0o z C
M CD >-" S
-4i N) N) -4 I ^
o C.) CD C. C) -
8 8 g

If D





Runningwater specimens F:AM 132053 and F:AM field number (AINS) 475-287.
Although sexual dimorphism occurs in Arretotherium (Macdonald and Martin,
1987), this is not the cause of the difference in robustness reported here. The small,
round, lower canine alveoli and the short pl-2 diastema indicate that the type
specimen of A. leptodus is a female, yet this specimen is considerably more robust
than UCMP 32369, a male A.fricki.
P4 of the type specimen of A. acridens has a small tubercle, or style, at the
labial end of the preprotocrista. This feature is probably variable, as it is also seen
in the left P4 of the type specimen of A. fricki, but not in any other specimens
examined. No P4s from Toledo Bend have this feature.
In the type specimen of A. acridens, the paracone and metacone of M3 are
completely separated such that the two median buttresses of these cusps that make
up the mesostyle do not join. As a result, there is a labial entrance to the transverse
median valley. This character was previously used to distinguish A. acridens from
A. fricki (Macdonald, 1956; Macdonald and Shultz, 1956). However, in upper
molars of all other specimens examined, the median buttresses of the paracone and
metacone abut against one another very low such that a labial entrance to the
transverse median valley is just barely blocked. These specimens include: the type
of A. leptodus and F:AM 132055 from Shannon County, South Dakota; the
Batesland specimens UCMP 32369 and 37440 referred to A. fricki (Macdonald and
Martin, 1987); another specimen reported here for the first time and referred to A.
fricki, F:AM 132053, from the Runningwater Formation; and the Toledo Bend
Macdonald and Schultz (1956), in describing M1-3 of A. fricki, noted a
reduced anterior and posterior cingulum. Although personal observation of the type
agrees with their description, the Batesland and Runningwater specimens 'referred
to A. fricki have strong, shelf-like anterior and posterior cingula similar to those of
A. acridens. Macdonald and Schultz (1956) also described Ml as having a lingual
cingular segment between the protocone and metaconule, but they made no note of
this feature on M2 or M3 and their fig. 2 does not show it. Macdonald and Martin
(1987), on the other hand, noted "a small cuspule" at the lingual opening of the
medial valley of each molar. The anterior and posterior cingula of the Toledo Bend
upper molars are quite prominent and all three molars have a lingual cingular
segment between the protocone and metaconule.
Plate IX, fig. 2 of Douglass (1901) and text fig. 13 of Macdonald (1956),
shows the M3 anterior cingulum of A. acridens wrapping around the lingual surface
of the protocone, then strengthening between the protocone and metaconule before
terminating against the anterior surface of the metaconule. Two of four Toledo
Bend M2s (but none of the M3s) possess a weak remnant of a lingual protocone
cingulum, but it is not nearly as prominent as that illustrated in the figures noted
above. Examination of a cast of the illustrated specimen showed that the lingual
cingulum does not continue unbroken around the protocone to the extent the
illustrations suggest.


Referring to M3 of the type specimen of A. fricki, Macdonald and Schultz
(1956) noted that the transverse median valley was blocked by spurs from the
protocone and "hypocone" (= metaconule). Macdonald and Martin (1987:58) also
noted that "the medial ends of the selene of the protocone and hypocone barely
meet in the medial valley." In the Toledo Bend upper molars, the medial ends of
the protocone and metaconule selenes are broadly separated and the transverse
median valley shows no tendency toward being blocked internally.
The Toledo Bend lower molars are of similar size and of virtually the same
morphology as both the Shannon County specimens and the more recent
Runningwater specimens. However, lower molars from Batesland show some
complexities not found in the others. Slightly smaller than those from Toledo Bend,
they differ from the rest in having a strong, posteriorly-directed spur that originates
on the posterior surface of the protoconid and effectively blocks the median valley.
In m3, this posteriorly-directed spur is met by an anteriorly-directed one originating
from the anterior surface of the hypoconid. Additionally, the Batesland m2s have a
strong, conical style on the posterolingual corer.
Arretotherium has not been previously reported from the Texas Coastal Plain
or from Florida. The only other specimens of the genus currently known from the
Gulf Coastal Plain include a fragmented M3 from the Mann Place site in San
Jacinto County, Texas, of early Hemingfordian age, and two molar fragments from
the similarly aged Garvin Farm Local Fauna in Grimes County. The Mann Place
M3 (TMM-TAMU 3065) is only slightly smaller in size than those from Toledo
Bend. It measures 25.1 mm AP by about 27 mm TR. The Garvin Farm specimens
consist of a protocone of a left upper molar (TMM 31084-168) and the anterior half
of a lower molar (TMM 31084-60).
Also noteworthy is a report by C. D. Frailey (pers. comm., 1992; unpubl. ms)
of the older (Chadronian to Orellan) anthracothere Bothriodon from a site
stratigraphically below the Suwannee Limestone in Pasco County, Florida.
Although not yet published, this record is particularly important with respect to the
earliest land mammal faunas in Florida and the timing of the subaerial emergence of
the Florida Platform. It is with anticipation, therefore, that we await further
confirmation of this report.

Suborder TYLOPODA Illiger, 1811
Family CAMELIDAE Gray, 1821
Subfamily "NOTHOKEMATINAE" Honey et al., 1998
Genus Nothokemas White, 1947
Nothokemas sp.
Figure 15A

Type Species.-Nothokemas floridanus (Simpson, 1932; see Patton,


Referred Specimens.-LSUMG V-2406, left ramal fragment with m2-3; V-
2407, right ml.
Description.- The broken m2 measures 10.9 mm AP by 8.0 mm TR. Actual
measurements would be somewhat larger, as the ectoloph and the posterior edge are
missing. Both m2 and m3 have an intercolumnar tubercle between the protoconid
and hypoconid. The m3 measures 17.5 mm AP by 8.9 mm TR. The entoconid of
m3 extends posteriorly and overlaps about one-half the length of the labial surface
of the hypoconulid. The hypoconulid is not the "double enamel loop" type of
protoceratids; it consists of a single cusp. In occlusal view, the hypoconid and
hypoconulid appear separated from the rest of the tooth (Fig. 15A). The enamel is
distinctly crenulated.
Discussion.-In the master's thesis from which this report originates (Albright,
1991), LSUMG V-2406 was referred to Floridatragulus sp. This specimen is also
responsible for placement of "Floridatragulus sp." on the faunal lists of Albright
(1998a) and Manning (1990), and it is the latter citation that likely resulted in the
report of "Floridatragulus sp." at the Toledo Bend site (as locality GCS) in Honey et
al. (1998:446). However, more recent study of the specimen has led to the
conclusion that V-2406 resembles Nothokemas and Delahomeryx brown Stevens (in
Stevens et al., 1969) more closely than any of the Floridatragulinae, i.e.,
Aguascalientia and Floridatragulus. The basis for this revision rests primarily on
the morphology of the m3. Although Nothokemas, Delahomeryx, the Toledo Bend
species, and the floridatragulines have intercolumnar tubercles between the
protoconid and hypoconid and an entoconid that overlaps the hypoconulid, the
entoconid overlap in the former three taxa is not nearly as pronounced as in
Aguascalientia and Floridatragulus, where even greater posterior extension of the
entoconid, together with transverse inflation, has resulted in the distinctive, divided
m3 hypoconulid so diagnostic of the Floridatragulinae.
Because ancestors of the Floridatragulinae would have likely gone through a
stage similar in morphology to that of Nothokemas, Delahomeryx, and the Toledo
Bend species, one might argue that the latter is a primitive member of the former
family. Unfortunately, the ancestry of this group of putative camelids is unknown
and lack of additional material of the Toledo Bend species, such as premolars,
precludes comparisons that might otherwise provide refined insights into the
relationship of this taxon with either the "Nothokematinae" or the Floridatragulinae.
The oldest floridatragulid, Aguascalientia sp., is known only from the early
late Arikareean Castolon Local Fauna, Trans-Pecos Texas (Stevens, 1977).
Aguascalientia sp. is smaller than the late Hemingfordian A. wilsoni, which,
likewise, is known only from the Zoyatal Local Fauna, Mexico (Dalquest and
Mooser, 1974; Stevens, 1977). Floridatragulus ranges from the early
Hemingfordian (F. nanus, F. dolichanthereus, and F. barbouri) through the
Barstovian (F. texanus [early], F. hesperus [late]) of the Gulf Coastal Plain (Patton,
1969; Honey et al., 1998). Differences that Stevens (1977) noted between
Aguascalientia and Floridatragulus involve characteristics of the anterior portion of


the mandible and premolars which, as noted above, cannot be directly compared
with the Toledo Bend specimen. The Toledo Bend species is slightly larger than
Aguascalientia sp., but smaller than all species of Floridatragulus.
Nothokemas and Delahomeryx brown are also first recorded in the early late
Arikareean. Nothokemas waldropi Frailey (1978) occurs in at least two of Florida's
Arikareean sites (SB-1A and Franklin Phosphate Pit no. 2), and another even
smaller species that appears closely related to N. waldropi is abundantly represented
at several sites (Albright, 1998a); Delahomeryx, like Aguascalientia sp., is known
only from the Castolon Local Fauna. Nothokemas also occurs in the early
Hemingfordian Thomas Farm and Garvin Gully faunas as N. floridanus and N.
hidalgensis (Patton, 1969). The Toledo Bend species is larger than N. waldropi and
smaller and lower crowned than N. floridanus and N. hidalgensis.
The rare appearance of what is likely a new species of Nothokemas at Toledo
Bend, together with the abundance of the small, undescribed species noted above
found at several Florida localities (yet entirely absent at Toledo Bend), re-
emphasizes the restricted, low latitude distribution of these morphologically
primitive camelids.

Family PROTOCERATIDAE Marsh, 1891

Although Wilson (1974) described a number of very early protoceratids from
the late Eocene of southwest Texas, the Toledo Bend Local Fauna records the first
appearance of protoceratids in the Gulf Coastal Plain prior to the early
Hemingfordian. Moreover, at Toledo Bend are two different size classes of teeth
indicating the presence of two different protoceratids: (1) a small form similar in
size to Protoceras from the Whitneyan to late Arikareean of the northern Great
Plains, and to Paratoceras wardi from the early Barstovian Trinity River Pit 1 Local
Fauna, Texas, and (2) a larger species indistinguishable from the Coastal Plain
endemic Prosynthetoceras texanus. (Note: Patton and Taylor [1971] considered
Lambdoceras a subgenus of Prosynthetoceras which would then place the latter in
the northern Great Plains. Following Webb [1981] and Prothero [1998:436],
Lambdoceras is here considered a valid genus, thus restricting Prosynthetoceras to
the Gulf and Atlantic Coastal Plains.) The small species is the most common
mammal in the fauna, accounting for nearly 22% of all mammalian remains
recovered (second to anthracotheres). The larger species, represented by only three
molars, is rare.
Complicating matters are three small horn fragments representing three distinct
morphologies. Two are clearly attributable to frontal, or orbital, horns. The
morphology of the third is suggestive of a rostral tine, but close examination reveals
that it, too, is likely that of an orbital horn. The fragments belonged to an animal
substantially smaller than previously described species of Prosynthetoceras, or of

Table 10. Comparative measurements of various species of Protoceratidae. All measurements except Toledo Bend from Patton and Taylor (1973).

Protoceras celer P. skinneri P. neatodelpha Paratoceras wardi Prosynthetoceras orthrionanus


M1 16 11.0-14.5 12.0-15.3 1 13.5 14. 1 14. 15. 5 11.1-12.5 13.0-14.0 3 12.0-13.0 14.1-15.0

5 0 5 0

M2 16 11.8-15.4 13.8-17.0 1 15.5+ 16. 1 16. 16. 4 12.2-13.3 15.2-16.0 2 13.3-14.0 15.3-16.5

7 2 8

M3 11 12.6-15.5 13.3-17.5 -- -- 1 15. 16. 3 12.5-12.8 15.2-16.6 2 14.4-14.5 16.2-16.6

0 5

ml 16 11.0-14.0 8.9-11.5 1 12.0+ -- -- 11 10.7-11.8 -- 4 11.0-11.3 7.8-8.5

m2 16 11.0-14.2 9.6-12.0 2 14.5+ -- -- 12 11.5-12.5 -- 1 13.1 10.5

m3 15 16.0-20.3 9.9-11.1 2 20.5-21.0 11. -- -- 12 16.2-18.0 7 18.0-21.8 10.1-11.7



the earliest synthetoceratine, Syndyoceras cooki from the Harrison Formation
(Barbour, 1905), and likely represent the same species as that to which the small
teeth belong. It is also likely that all three morphologies represent only one species,
but vary, perhaps, as a function of ontogenetic stage.
Because post-cranial material from Toledo Bend compares closest to the
Whitneyan Protoceras celer, which had only small, bony knobs rather than well
developed horns, it is possible that these horns belonged to one of the two
Arikareean species of that genus, Protoceras skinneri or P. neatodelpha. However,
it is not known if the latter two species had horns because only skulls of females
have been recovered (Patton and Taylor, 1973). The small early Barstovian Texas
Coastal Plain endemic, Paratoceras wardi, on the other hand, had prominent orbital
horns as did the synthetoceratines Prosynthetoceras and Syndyoceras. But, again,
the Toledo Bend horn fragments belonged to an animal much smaller than
previously described species of the latter two taxa. Evidence provided below
suggests that this material belongs to a small, primitive new species of
Prosynthetoceras that lived alongside early P. texanus. Measurements of the
dentition of several species of protoceratids, including the small Toledo Bend
species, are provided in Table 10.

Subfamily SYNTHETOCERATINAE Frick, 1937
Genus Prosynthetoceras Frick, 1937
Prosynthetoceras orthrionanus sp. nov.
Figure 15 B-H

Blastomeryx texanus: Wood and Wood, 1937, p. 137, pl. 1, figs. 5-6.

Holotype.-LSUMG V- V-2619, right orbital horn tip
Paratypes.-LSUMG V-2630, orbital horn tip; V-2631, orbital horn tip
Type Locality.-Toledo Bend Site, Newton County, Texas
Etymology.-Named for the early temporal occurrence (Gr., orthrios, early,
dawn) and its small size (Gr., nanos, little) relative to Prosynthetoceras texanus.
Referred Specimens.-LSUMG V-2623, three upper canine fragments; V-
2624, two dP4s; V-2625, left P4; V-2626, one left, two right Mis; V-2627, right
M2; V-2770, left M2; V-2628, one left M3 and one left M3 fragment; V-2771, left
M3; V-2629, upper molar fragments; V-2632, il?; V-2633, left i2?; V-2634,
anterior right ?p2; V-2635, right ramal fragment with ml; V-2636, one right, two
left mls; V-2637, right m2; V-2638, six m3s; V-2772, right m3; V-2639, m3
fragments; V-2640, two edentulous anterior right ramal fragments; V-2641, two axis
fragments; V-2642, four distal scapula fragments; V-2643, distal humerus
fragments; V-2644, proximal and distal radius fragment; V-2645, right scaphoid; V-
2646, one right, two left lunars; V-2647, left cuneiform; V-2648, femur head
fragment; V-2649, distal tibia fragments; V-2650, thirteen astragali; V-2651, four
calcanea; V-2652, calcanea fragments; V-2653, six naviculars; V-2654, cuboid; V-


2655, two proximal right Mt III; V-2656, proximal right and left Mt IV; V-2657,
distal metapodial fragments; V-2658, proximal phalanx fragments; V-2659, medial
phalanges; V-2660, ungual phalanx.
Diagnosis.-Differs from Prosynthetoceras texanus in smaller size, lower
crowned teeth, and in retaining unfused distal radius and ulna; lower m3s differ
from P. texanus in retaining stylids between protoconid and hypoconid and between
hypoconid and hypoconulid; similar in size to Paratoceras wardi but differs in
lacking a continuous lingual cingulum on upper molars and in retaining unfused
distal radius and ulna; similar in size to Protoceras celer but differs in having orbital
horns and in lacking a continuous lingual cingulum on upper molars; similar in size
and tooth morphology to Protoceras skinneri and Protoceras neatodelpha, but
postcrania and cranial armament unknown for latter two taxa.
Description.-Horns: V-2619, a fragment consisting of only the distal part, is
about 42 mm long, transversely compressed, and weakly curved toward the sagittal
plane (Fig. 15B). It is anteroposteriorly broad proximally, tapers distally, and the
forward edge is rounded while the posterior edge is pinched. As in Paratoceras and
the synthetoceratines, it shows longitudinal vascular grooving. It most closely
resembles the distal part of a Paratoceras wardi orbital horn, but lacks the bulbous
knob at the tip. Whether this horn had the medial strut described for that of P. wardi
(Patton and Taylor, 1973) is indeterminable from this fragment. It is not yet known
if Protoceras skinneri or P. neatodelpha had orbital horns.
In contrast to V-2619, specimen V-2631, also a fragment representing only the
distal part, is not transversely compressed. It is slightly more recurved than V-2619
and has a bulbous tip. A prominent sulcus ascends the posterior edge. On the
anterior surface is another, less prominent, sulcus. Several minor grooves
additionally sculpt the surface, particularly medially. The length of this fragment
along the outside curvature is about 39 mm; both the anteroposterior and transverse
diameters measure about 8.5 mm. With the grooves and bulbous tip, V-2631 most
closely resembles the orbital horns of Prosynthetoceras texanus and Syndyoceras
cooki except for its tiny size (Fig. 15C).
The third horn fragment, V-2630, is neither transversely compressed like V-
2619, nor does it show the bulbous tip and prominent posterior sulcus of V-2631.
V-2630 is relatively long and slender, not as strongly recurved as the others, and
tapers to a blunt, rounded point. It measures 55 mm long, is rounded labially, and
flattened medially (Fig. 15D).
Upper dentition: Terminology follows Patton and Taylor (1973:355). Three
small fragments represent the long, recurved upper canine typical of protoceratids.
The P4 is triangular in occlusal outline and wider labially than lingually. It has a
well developed parastyle and metastyle and a lingual cingulum that extends a short
distance labially on both the anterior and posterior surface. The dP4 is molariform
and smaller than Ml. Upper molars are low crowned with strong styles and cingula
(Fig. 15E, F). They are broader transversely than long anteroposteriorly and the





1,, ,' -

Figure 15. (A) Nothokemas sp., left ramal fragment with m2-3, LSUMG V-2406; (B) Prosynthetoceras
orthrionanus n. sp., frontal horn fragment, V-2619; (C) P. orthrionanus, frontal horn fragment, V-2631;
(D) P. orthrionanus, frontal horn fragment, V-2630; (E) P. orthrionanus, left M2, V-2770; (F) P.
orthrionanus, left M3, V-2771; (G) P. orthrionanus, right m3, V-2772; (F) Prosynthetoceras texanus,
left M3, V-2662; (G) P. orthrionanus, right m3, V-2772, occlusal view; (H) Same specimen, labial
view. 3 mm scale bar for A only; 1 cm scale bar for B-H..




anterior half is set farther lingually than the posterior half, particularly on M3. Both
the paracone and metacone have a strong labial rib and there is a prominent
mesostyle. The crescents are generally V-shaped. Between the protocone and
metaconule is a well developed blade-like style characteristic of protoceratids.
There is an anterior and posterior cingulum; the former large and blade-like
lingually. In some molars the posterior cingulum continues lingually around the
metaconule to connect with the blade-like intercolumnar style.
Lower dentition: The crown of an incisor tentatively referred to this species is
believed to be an il. It is smooth, relatively symmetrical, and convex labially.
There is a cingulum that extends upward from the base of the crown along both
lateral edges. The ventral edge of the crown of i2? is convex while the dorsal edge
is concave. The lingual surface is slightly excavated. The i3 is asymmetrically
spoon-shaped. Like i2, the ventral edge of the crown is convex while the dorsal
edge is concave. The anterior ?p2 fragment shows a distinct notch between the
paraconid and protoconid, but is otherwise undiagnostic. The ml and m2 show a
variably small to prominent intercolumnar tubercle between the protoconid and the
hypoconid. In one specimen the tubercle is blade-like as in upper molars. The
posterior edge of the metaconid barely overlaps the anterior edge of the entoconid
forming only a slight metastylid. The parastylid, however, is relatively prominent.
The metaconid and entoconid are both ribbed and there are no cingula. The m3s
also have a prominent parastylid and similar metaconid-entoconid overlap (Fig.
15G). Characteristic of protoceratids, the hypoconulid of m3s shows the double
enamel loop which encloses a fossettid with wear (Patton, 1969). Five of eight
complete m3s and one anterior fragment have an anterolabial cingular segment.
Some also have an intercolumnar stylid between the protoconid and hypoconid and
also between the hypoconid and hypoconulid.
Mandible: Two small, worn, anterior ramal fragments (V-2640) have a pi
alveolus, the anterior margin of which lies directly above the distinct, downwardly
directed "hook" on the ventral edge of the ramus that marks the posterior-most
position of the symphysis. Posterior to the pl alveolus is a relatively long, sharp-
ridged diastema. Both specimens are broken anterior to the p2.
Vertebrae: The axis fragments consist primarily of the odontoid process.
Although slightly smaller, they closely resemble those ofProtoceras celer from the
Poleside Member of the Brule Formation (F:AM 53527).
Forelimb: All postcranial elements compared were nearly identical to those of
Protoceras celer and only slightly different from those of Paratoceras wardi. The
scapula fragments consist only of the glenoid surface and part of the neck, and
compare well with F:AM 53521, a scapula of P. celer. The neck is narrow and there
is no indication that the scapular spine extended onto it. The coracoid process is
short but prominent.
The distal humeri fragments range from 21.5 to 27.0 mm transversely across
the distal trochlea. These measurements would be slightly greater if the specimens
were unworn. There is no entepicondylar foramen. The proximal end of the radius


measures 22.0 mm transversely (vs. 23.2 mm in P. celer, F:AM 53521) and, as in P.
celer specimens F:AM 53521 and 40879, is not fused to the ulna. As a result, the
distal end of the radius has only a scaphoid and lunar facet; the cuneiform facet is on
the distal end of the ulna. In Paratoceras wardi and the Synthetoceratinae, the distal
ulna is strongly fused to the distal radius. The scaphoid, lunar, and cuneiform are
virtually indistinguishable from those of F:AM 40879. These elements have been
adequately described by Osborn and Wortman (1892) and Scott (1940).
Measurements taken in the proximodistal aspect are 11.0 mm for the scaphoid, 12.0
mm for the lunar, and 11.0 mm for the cuneiform; much smaller than the same
elements of Prosynthetoceras texanus.
Hindlimb: The head of the femur is indistinguishable from that of both
Protoceras celer and Paratoceras wardi. Distal tibia fragments range from about 18
to 20 mm transversely. The prominence of the medial malleolus and intercondylar
ridge result in a deep, narrow, dorsal notch (the medial astragalar condylar groove),
as well as a deep and narrow medial facet for the astragalus. The lateral facet for the
astragalus is broad and shallow. Astragali vary in size from 19.8 mm to 30.5 mm
long. The astragalus of P. celer (F:AM 53521) and Paratoceras wardi (F:AM
40870) are virtually identical in size and morphology to those from Toledo Bend
and measure 30.0 and 29.0 mm, respectively. The calcanea from Toledo Bend vary
from relatively robust specimens at 55.3 mm long to gracile ones at 50.6 mm.
Astragali referred to Paratoceras wardi (F:AM 40868 and 40869) measure 56.5 to
59.2 mm, respectively, and differ from P. celer and the Toledo Bend specimens only
in a slightly narrower sustentacular facet. The naviculars, which range from
approximately 15 to 19 mm anteroposteriorly, are not fused to the cuboid as they are
in ruminants, and are indistinguishable from the same elements in P. celer.
Metatarsals III and IV are not coossified. The width of the proximal Mt III from
Toledo Bend is 9.4 mm, and 9.7 mm for Mt IV, compared with 10.8 and 10.2 for the
same elements, respectively, of P. wardi (F:AM 40872 and 40874). More detailed
descriptions of the postcrania of Protoceras celer and Paratoceras wardi can be
found in Osbom and Wortman (1892) and Patton and Taylor (1973), respectively.
Discussion.-Were it not for the recovery of the orbital horn fragments
described above, the size and morphology of the teeth and, particularly, the
primitive morphology of the postcrania of this small protoceratid would suggest
referral to Protoceras, two species of which occurred in the Arikareean: P. skinneri
from the early Arikareean Sharps and Gering formations of South Dakota and
Nebraska, and P. neatodelpha from beds equivalent in age to the Harrison
Formation, north of Keeline, Niobrara County, Wyoming (Patton and Taylor,
1973:367; Note: recent work by R. Hunt [pers. comm., 1998] indicates that Arikaree
deposits north of Keeline may be slightly older than the stratotype Harrison
Formation in Sioux County, Nebraska). In fact, it was to P. neatodelpha that I
originally assigned this material in the Master's thesis from which this report is
derived (Albright, 1991), because post-cranial material from Toledo Bend compares
closest to the Whitneyan Protoceras celer, the upper molars closely resemble those


of P. skinneri and P. neatodelpha in lacking the prominent continuous lingual
cingulum seen in P. celer, and P. neatodelpha is likely of similar age to the Toledo
Bend species. However, P. celer had only small, bony knobs rather than well
developed horns. On the other hand, the early Barstovian Paratoceras wardi had
prominent orbital horns. Considering the sister taxon relationship of Protoceras
with Paratoceras (Patton and Taylor, 1973) one might argue that prominent orbital
horns likely developed in Protoceras neatodelpha as well, given its approximately
9-10 million year occurrence following Protoceras celer. Currently, however, the
cranial armaments of P. neatodelpha and P. skinneri are unknown because only
skulls of females have been recovered (Patton and Taylor, 1973).
Although Patton and Taylor (1973:397) found Paratoceras wardi more closely
related to Protoceras than to any of the synthetoceratines, they considered the origin
of Paratoceras to be "the greatest gap in our knowledge of protoceratine evolution."
Temporally the Toledo Bend species falls within that gap and biogeographic data
would also support the idea that the Toledo Bend species might be an evolutionary
precursor of P. wardi. (Interestingly, Whitmore and Stewart [1965] referred a
primitive protoceratid from the Hemingfordian of Panama to Protoceras. Patton
and Taylor [1973], however, later assigned the specimens to Paratoceras.) One
significant morphological argument against this interpretation, however, rests on the
prominent lingual cingulum retained in upper molars of Paratoceras wardi (as in P.
celer). This condition in an early Barstovian taxon would require an evolutionary
reversal from that in the late Arikareean Toledo Bend species where the lingual
cingulum is weak to absent.
It is unfortunate that only a single worn P4 represents the premolar dentition of
the small Toledo Bend protocertid, because the primary means of distinguishing P.
skinneri from P. neatodelpha, and both of those species from P. celer and
Paratoceras, relies on premolar morphology and size relative to the molars (Patton
and Taylor, 1973). Lower premolars are also helpful in distinguishing Protoceras
from Paratoceras, but none were recovered. Lower molars from Toledo Bend, as in
P. skinneri and P. neatodelpha, have a less prominent anterior cingulum than those
ofP. celer. Paratoceras wardi lower molars also have a weak anterior cingulum.
The Toledo Bend species is substantially smaller than the first
synthetoceratines recorded, Syndyoceras cooki and Prosynthetoceras texanus,
although upper molars are similar in lacking lingual cingula. Lower molars are also
smaller and lower crowned than those of P. texanus and, as in Protoceras and
Paratoceras, they show weak intercolumnar stylids between the protoconid and
hypoconid (and between the hypoconid and hypoconulid in the m3s), and they have
a weak anterior cingular segment. Lower molars of P. texanus generally lack
intercolumnar stylids and an anterior cingulum (Patton, 1969:181; Patton and
Taylor, 1973:382). Recovered postcranial elements lack derived features found in
Prosynthetoceras texanus, and in Paratoceras wardi, such as a distally fused radius
and ulna, and are virtually identical to those of Protoceras celer. Horn fragments,


however, although much smaller than those of Prosynthetoceras texanus and
Paratoceras wardi, closely resemble those of the latter two taxa in morphology.
The most reasonable hypothesis therefore, given the evidence currently
available, is to consider the Toledo Bend form a small, new, primitive species of
Prosynthetoceras that lived alongside early P. texanus. Although the post-crania
most closely resemble those of Protoceras celer, this resemblance is likely the
retention of primitive traits that carry, in a cladistic sense, no phylogentic
significance. The similarity in morphology of the teeth with those of P. skinneri
and P. neatodelpha, whereby the continuous lingual cingulum in upper molars is
lost, may, likewise, be a convergent trait. The possibility that the species represents
an evolutionary precursor to Paratoceras wardi based on similar size and horn
morphology is also dismissed because of the retention of prominent lingual cingula
in upper molars of P. wardi-a character absent in the older Toledo Bend species.
It is important, however, to consider the possibility that Protoceras neatodelpha
made its way to the Texas Coastal Plain from the northern Great Plains in the early
late Arikareean, as both Arretotherium acridens and Nexuotapirus marslandensis,
two taxa previously known only from the northern mid-continent, are abundant at
Toledo Bend. But this hypothesis cannot be tested until a male skull of Protoceras
neatodelpha and/or P. skinneri is recovered. Such a find would shed considerable
light on the evolution of frontal horns in this lineage and obviously help clarify the
relationship of the small Toledo Bend protoceratid with both the protoceratinae and
the synthetoceratinae.
In light of the above, the report by Tedford et al. (1987:176) of "early
Prosynthetoceras" in the Cedar Run Local Fauna is likely attributable to P.
orthrionanus, particularly considering the revised age of the Cedar Run Local
Fauna as equivalent with Toledo Bend (Albright, 1998a). The two Cedar Run teeth
(AMNH 30084) were referred to "Blastomeryx texanus" by Wood and Wood
(1937), a taxon later synonymized with P. texanus by Patton and Taylor (1971).
Patton and Taylor (1971) did not, however, include AMNH 30084 in their
hypodigm of material referred to P. texanus, possibly because the teeth are
considerably smaller than specimens they referred to the latter. This conclusion is
made in consideration of Patton and Taylor's (1971) findings through measurements
of lower third molars that P. texanus increased in size through the progressively
younger Aiken Hill, Garvin Farm, and Burkeville local faunas.
One final note should be included regarding the biogeography of this group. It
is particularly intriguing, given the abundance of Prosynthetoceras orthrionanus at
Toledo Bend, that protoceratids are absent prior to the early Hemingfordian in
Florida where the Arikareean is much better represented than in Texas.
Furthermore, Paratoceras is known only from the early Barstovian and
Clarendonian of Texas and the Hemingfordian of Panama (Whitmore and Stewart,
1965; Patton and Taylor, 1973:367) and Prosynthetoceras, which ranges from the
late Arikareean to the late Barstovian, is also known only from the Gulf and Atlantic
Coastal Plains. The scarcity of protoceratids in the extensivley sampled Arikareean


deposits of the northern Great Plains, their absence in the thick Arikareean and
Hemingfordian deposits of the John Day Valley, Oregon, together with an
abundance of endemic forms from the Arikareean through Clarendonian along the
Gulf Coastal Plain supports ideas that evolution of the group was largely centered in
the tropical to subtropical regions of Central America and southern North America.

Prosynthetoceras texanus (Hay, 1924)
Figure 16A

See Patton and Taylor (1973) for synonymy.

Holotype.-TAMU 2387, right M3, Garvin Gully Fauna, Grimes County,
Referred Specimens.-LSUMG V-2661, left M2; V-2662, left M3; V-2663,
right i3; V-2664, partial right m3.
Description.-M2 and M3 are typically protoceratid in their prominent
parastyle and mesostyle, ribbed paracone and metacone, and intercolumnar blade
between the V-shaped protocone and metaconule (more prominent on the posterior
side of the median valley than on the anterior side). M2 measures 18.7 mm AP by
21.7 mm TR; M3 measures 17.0 mm AP by 19.0 mm TR (Fig. 16A).
The m3 fragment is high crowned with a prominent parastylid. The posterior
edge of the metaconid strongly overlaps the anterior edge of the entoconid. Both the
labial and the lingual surface of the metaconid and entoconid are ribbed. There is
no anterior cingulum and there is no intercolumnar tubercle between the protoconid
and the hypoconid. Because the hypoconulid is broken off and missing, the 18.8
mm AP length is about 4.5 to 5 mm less than its total length would have been were
the hypoconulid intact.
Discussion.-The teeth referred here to Prosynthetoceras texanus are larger
and higher crowned than those of P. orthrionanus sp. nov. discussed above, and the
m3 lacks an anterior cingulum and stylids between the labial cusps.
It is interesting that P. texanus, which is common in early Hemingfordian
faunas of the Coastal Plain such as Garvin Gully and Thomas Farm, is absent from
Arikareean sites in Florida such as the Buda, SB-1A, Franklin Phosphate Pit No. 2,
and Brooksville local faunas. The rare occurrence of P. texanus at Toledo Bend,
therefore, which is considered nearly age equivalent to the Florida Arikareean sites
(Albright, 1998a), provides one of the oldest records of the species, in turn
suggesting an earlier arrival in Texas than in Florida.


Suborder RUMINANTIA Scopoli, 1777
Genus Nanotragulus Lull, 1922
Nanotragulus sp.
Figure 16 B, C

Type Species.-Nanotragulus loomisi Lull, 1922
Referred Specimens.-LSUMG V-2492, right ml; V-2493, right m2; V-
2254, right m2; V-2494, proximal right scapula fragment; V-2495, two distal
humeri; V-2496, distal tibia-fibula; V-2497, three astragali; V-2498, two right
cubonaviculars; V-2499, one left, one right ectocuneiform; V-2500, two distal
metapodials; V-2501, proximal phalanx.
Description.-In lower molars the labial surface of the metaconid and
entoconid are weakly ribbed, there is a small anterior cingulum situated high on the
crown; there is a stylid at the labial entrance of the valley between the protoconid
and the hypoconid; and the labial surface of the labial cusps form narrow, sharp Vs
(Fig. 16B). There is no metastylid and no overlap of the entoconid by the
metaconid. V-2254 measures 7.0 mm AP by 4.4 mm TR with a slightly worn crown
height of 7.1 mm. V-2492 measures 7.0 mm AP by 3.9 mm TR with a worn crown
height of 6.9 mm. V-2493 measures 7.4 mm AP by 4.9 mm TR with a worn crown
height of 4.4 mm.
The scapula (V-2494) has a strong coracoid process and no coracoscapular
notch. The distal tibia shows three prominent projections formed by the medial
malleolus, the intercondylar ridge, and a strongly fused distal fibula (Fig. 16C)
which Scott (1940:519) noted was "highly characteristic" of Hypertragulus. The
astragular facets of the distal tibia are both deep and narrow. The astragali (V-2497)
are narrow and elongate with no distal keel on the lower ginglymi. The
ectocuneiforms (V-2499) measure about 10 mm AP, are about 6 mm deep
dorsoventrally, and about 8 mm TR. The dorsal articular surface is concave
anteriorly and convex posteriorly. Ventrally, the opposite is true. The
cubonavicular (V-2498) is fused and the central metapodials (V-2500) are unfused.
The metapodial keel is strong but confined to the ventral surface.
Discussion.-The "triple projection" morphology due to fusion of the distal
tibia and fibula clearly places this small species within the Hypertragulidae, as this
feature is not present in other ruminants (Webb and Taylor, 1980). Unfused
metapodials preclude referral to the Leptomerycidae or Blastomerycinae.
Hypertragulus retains an "extremely reduced but apparently continuous" fibula
(Webb and Taylor, 1980:144; Scott, 1940). A continuous fibula is evidently lost in
Nanotragulus, leaving only the fused proximal and distal remnants. Hence referral
of the Toledo Bend species to Nanotragulus.
The Toledo Bend species is slightly smaller than N. albanensis Frick (1937)
and similar in size to N. ordinatus (Matthew, 1907) and N. matthewi Cook (1934)





1 cm
Figure 16. (A) Prosynthetoceras texanus, LSUMG V-2662, left M3; (B) Nanotragulus sp., stereo view
of right m2, LSUMG V-2254; (C) Nanotragulus sp., fused distal tibiofibula, V-2496; (D) Small
artiodactyl, undet., left partial upper molar, V-2253. 3 mm scale bar for B and C; 1 cm scale bar for A
and D.

from the Harrison Formation. The latter two species have more rounded labial
cusps than the sharp, distinctly V-shaped labial cusps of the Toledo Bend species.
Nanotragulus loomisi Lull (1922) is smaller and also has more rounded labial cusps.
Although hinted at by Voorhies (1973), Frailey (1979) showed that material
representing N. loomisi, N. intermedius Schlaikjer (1935), and N. lulli Frick (1937)
should be referred to the single species N. loomisi, abundant in Arikareean faunas of


the Great Plains (e.g., Macdonald, 1963) and in the Buda Local Fauna, Florida
(Frailey, 1979).
Stevens et al. (1969) and Stevens (1977) reported N. ordinatus from the
Castolon Local Fauna of Big Bend National Park, Texas. This fauna is of similar
age to the Toledo Bend Local Fauna and the size of N. ordinatus and the Toledo
Bend hypertragulid is similar. But the figures of the Castolon specimens (1969, fig.
13; 1977, fig. 17B) showing cut-away sections of the mandible indicate that m2 and
m3 were considerably more hypsodont than in the Toledo Bend species.
In his discussion of hornless ruminants, Webb (1998:473) listed the
blastomerycine Machaeromeryx gilchristensis as occurring in the Toledo Bend
Local Fauna, evidently following the preliminary faunal list of Manning (1990).
More detailed study revealed that this material represented Nanotragulus. Thus, the
record of M. gilchristensis at Toledo Bend is in error.

Artiodactyla, undet.
Figure 16D

Referred Specimen.-LSUMG V-2253, partial left upper molar.
Description and Discussion.-This very low crowned tooth was compared
with nearly all small artiodactyls in the F:AM collection, but eludes identification.
It is missing most of its two posterior cusps (Fig. 16D). There is a thin but
prominent anterior cingulum at the base of the crown extending from the
anterolabial surface of the ectoloph to a point where the protocone is broken
lingually. It cannot, therefore, be determined whether this cingulum continued
around the protocone to connect with a prominent, transversely elongate style
between the two lingual cusps, although it appears to have done so. The labial
surface of the paracone has a median rib and there is an anteriorly-directed
mesostyle. There does not appear to have been a parastyle. The posterior crescent
of the protocone does not join the anterior crescent of the metaconule. The anterior
crescent of the metaconule is strongly connected to the mesostyle. Most peculiar are
the greatly offset lingual cusps and the robust lingual cingulum protecting the
metaconule. This cingulum abuts the transversely elongate style mentioned above.
The length of the tooth cannot be determined but its transverse width would have
been only slightly greater than 8 mm.
Nanotragulus and Hypertragulus generally lack a mesostyle in upper molars.
Leptomeryx, Pronodens, Pseudoparablastomeryx, Blastomeryx, Machaeromeryx,
and Longirostromeryx do not have lingual cingula. Dremotherium has a continuous
anterior and lingual cingulum, but the posterior lingual cingulum is not nearly as
developed as in the Toledo Bend tooth. Upper molars of Heteromeryx show a
lingual cingulum, but this species is larger, with heavier enamel and a more labially
rounded paracone and metacone. Barbouromeryx has no lingual cingula, is larger,
and has crista-like structures on the anterior limb of the metaconule. Bouromeryx is


also larger and has complicated cristae. Archaeomeryx is smaller. Paracosoryx has
hypsodont upper molars. The tooth is reminiscent of Eocene forms such as
Leptoreodon, Leptotragulus, and Poabromylus, but these taxa have a more
prominent lingual cingulum on the protocone than on the metaconule and they have
a prominent parastyle. The tooth does not resemble deciduous P4s of hypertragulids
or leptomerycids. On the other hand, the presence of distinct ribs on the paracone
and metacone, a prominent mesostyle, a prominent style between the protocone and
metaconule, and what appear to be prominent lingual cingula suggest the possibility
that the tooth may be a dP4 of the small Toledo Bend protoceratid or Nothokemas.
However, the dP4 of Paratoceras wardi figured by Patton and Taylor (1973:380)
differs from V-2253 in its larger size, prominent parastyle, and absence of strong
lingual cingula.
It is unfortunate that only this single partial tooth has been recovered.
Although it may represent a new Gulf Coastal Plain taxon, additional material is
required before an unambiguous taxonomic designation is forthcoming.


The heterogeneous nature of the deposit from which the Toledo Bend
assemblage was recovered implies that nearly all fossils found are at least to some
extent reworked. Fossils from the Paleozoic through the earliest Miocene have
been recovered (Manning, 1990). Nevertheless, those mammals that comprise the
Toledo Bend Local Fauna are considered contemporaneous primarily because all
have overlapping ranges with similar taxa in comparably aged faunas elsewhere,
because few vertebrate fossils are known from the underlying Catahoula Formation
(considered Oligocene in age), and because early Oligocene marine strata of the
Vicksburg Group underlie the latter. In other words, only older strata lie upstream
from the Toledo Bend locality, and there is no evidence that those older strata could
have contributed the mammalian taxa that accumulated in the paleochannel deposit.
Taxa from a younger age are also unlikely to have been reworked into the deposit
because the fossil bearing conglomerate appears to have been "encapsulated" within
the Carnahan Bayou Member siltstones shortly after the high energy event that
formed the deposit subsided.
Although Albright (1998a) discussed several problems that discourage high
resolution geochronology for Arikareean assemblages of the Gulf Coastal Plain, the
age of the Toledo Bend Local Fauna can be relatively well constrained by a number
of biochronologically diagnostic taxa. As Manning (1990) noted, the presence of
an entelodont and an anthracothere indicates that the site can be no younger than
early Hemingfordian, as that is their last occurrence in the North American fossil
record (Tedford et al., 1987). On the other hand, the site can be no older than early
Arikareean because that is when Diceratherium, Moropus, and Nanotragulus make
their first appearance (Tedford et al., 1987, 1996). Diceratherium, Nanotragulus,


and Daphoenodon also last occur in the late Arikareean (Tedford et al., 1987,
1996), although Miohippus last occurs in the early Arikareean. The presence of
Prosynthetoceras, however, provides a later Arikareean aspect to the fauna as does
Although material representing the same tapir as that from Toledo Bend is also
found in the early Arikareean Monroe Creek Formation, the type specimen of
Nexuotapirus marslandensis is thought to have come from the early Hemingfordian
Runningwater Formation. Arretotherium, too, ranges from the early Arikareean to
early Hemingfordian in the High Plains, although the Toledo Bend species
resembles most closely the older Arretotherium acridens.
The mutual occurrence of Daphoenodon notionastes, Moropus sp., and
Texomys sp. in both the Toledo Bend and Buda local faunas suggests age
equivalence. Both faunas also have the characteristic Arikareean hypertragulid
Nanotragulus. As shown in Figure 17, the Buda Local Fauna is considered "medial"
Arikareean in age (see Albright, 1998a, for a summary of Arikareean faunas in
Noticeably absent from Toledo Bend are taxa found in the Gulf Coastal Plain
during the early Hemingfordian such as the amphicyonid Amphicyon longiramus,
the canid Tomarctus, the horses Parahippus leonensis and Anchitherium clarencei,
and the rhino Menoceras barbouri. Instead, at Toledo Bend the amphicyonid is
Daphoenodon, the common horse is Anchippus texanus, and two of the three
rhinoceroses are species of Diceratherium. Although oreodonts are present in the
Texas Coastal Plain during the early Hemingfordian, and camels are both diverse
and abundant, the absence of oreodonts and rarity of camels at Toledo Bend, in
contrast to their presence in Florida during the Arikareean and Hemingfordian, is
likely due to ecological factors rather than temporal (see below). Figure 17 shows
the chronostratigraphic position of the Toledo Bend Local Fauna relative to some of
the other Arikareean and Hemingfordian faunas of North America with which it has
been compared.
One final mention should be made regarding the reports of Machaeromeryx
gilchristensis in the Toledo Bend Local Fauna by Webb, of Floridatragulus sp. by
Honey et al., and of Menoceras barbouri and Floridaceras white by Prothero in the
recently published volume on Tertiary mammals of North America edited by Janis
et al. (1998). These records are evidently based on a preliminary faunal list
published by Manning (1990) prior to the completion of the present author's more
detailed studies. More recent study determined that these identifications were
erroneous and that the material referred to Machaeromeryx actually represents
Nanotragulus, that material referred to Floridatragulus more likely belongs to a
new species of Nothokemas, and that the material referred to Menoceras and
Floridaceras belongs to Diceratherium.


Figure 17. Correlation chart of early Miocene faunas of the Gulf Coastal Plain and northern Great
Plains. Chronostratigraphy of northern Great Plains follows MacFadden and Hunt (1998); ABB =
Agate Bone Bed, Agate Fossil Beds National Monument, Nebraska; AH = Aikin Hill Local Fauna; GG
= Garvin Gully Local Fauna.


The Toledo Bend Local Fauna includes the most diverse representation of
Arikareean mammals yet reported from the Gulf Coastal Plain. This assemblage of
at least 25 mammalian taxa differs from other early Miocene coastal plain faunas in
the unusual abundance of otherwise rare, riparian species. Exceptional is the fact
that three riparian forms, the "hippo-like" anthracothere Arretotherium acridens, the


tapir Nexuotapirus marslandensis, and the small new protoceratid, Prosynthetoceras
orthrionanus, account for the majority of mammalian remains recovered-a likely
result of these animals living in or near a riverine environment conducive to the
preservation of their remains.
It has long been recognized that early Miocene faunas of the Gulf Coastal Plain
represented a distinct biogeographic province from the better known assemblages of
similar age in the northern Great Plains. This is particularly evident in the
Arikareean faunas of Florida and further emphasized at Toledo Bend with records of
such new, apparently endemic, forms as the small protoceratid, camelid, and
rhinoceros. In contrast to Florida, however, there is also at Toledo Bend the added
element of several taxa previously known only from the Great Plains, such as
Arretotherium, Nexuotapirus, and Diceratherium, resulting in an unusual and
diverse mixture of taxa that reflects an environment able to sustain overlapping
ranges from adjacent biomes; so called "ecotone" conditions.
There are additional differences between Toledo Bend and Florida's
Arikareean assemblages, as well. Oreodonts, for example, are present in Florida
during this interval, yet they are entirely absent at Toledo Bend-a particularly
interesting point considering that the late Arikareean was their time of maximum
diversity. Camelids, too, are relatively common in Florida's Arikareean sites, yet
they are extremely rare at Toledo Bend. In contrast, the three most abundantly
represented mammals at Toledo Bend, noted above, are entirely unknown in Florida,
and rhinos, also common at Toledo Bend, are exceptionally rare in Florida until the
These differences may be due to physiographic factors, as the east Texas
region of the Gulf Coastal Plain would certainly have been influenced by a west-to-
east migrating Mississippi River system during the early Miocene (Galloway et al.,
1991) in contrast to what may have been a somewhat less fluvially influenced
environment in Florida. Furthermore, prior to the late Oligocene the Florida
Platform persisted as a marine shallow-water carbonate bank that was separated
from the continental mainland by the Suwannee Strait, which flowed from the
northeastern Gulf of Mexico to the Atlantic Ocean (Huddlestun, 1993). Within the
Suwannee Strait was the deeper-water Gulf Trough which extended from the Gulf of
Mexico northeasterly across the eastern part of the Florida panhandle region far into
southern Georgia. During the Late Oligocene, according to Huddlestun (1993),
there began a series of progressively lower low stands punctuated by lower high
stands which finally culminated in the extreme low stand at the beginning of
depositional cycle TB1.1 of Haq et al. (1987). Not until then, during the early part
of the Chattian Stage, did the Florida Platform finally become subaerially
continuous with the continental mainland, at least in eastern Georgia beyond the
easternmost extent of the Gulf Trough. During the latest Oligocene, Huddlestun
(1993:140-142) speculates that the Suwannee Straits region would have been "a
vast, low-elevation karst plain" and that the Gulf Trough would have been a large
river valley during low stands and a "vast estuary" during high stands. Subsequent


early Miocene sea-level high stands and low stands would result in further episodes
of Florida peninsula isolation and continental connection until the connection was
permanently established upon burial of the Gulf Trough during the early Middle
Miocene (Huddlestun, 1993). Such a scenario provides ample opportunity for
filtered dispersal into northern Florida, south of the Suwannee Strait, as well as for
insular isolation and the consequent affects including endemism and dwarfism.
Another probable cause for differences between the western and eastern Gulf
Coastal Plain is the location of the Toledo Bend site nearer to the Great Plains than
is Florida, particularly considering the environmental corridor between the Great
Plains and the Texas Coastal Plain likely provided by a paleo-Mississippi River
system. Temporal inequivalencies that are difficult to resolve due to the lack of
accurate geochronological control over this interval in the region may also have
contributed, as well as inadequate knowledge regarding truly contemporaneous
faunas elsewhere due to circumstances such as the temporal gap noted in the
Arikaree Group of northwestern Nebraska from about 22-27 Ma. It is with
anticipation, therefore, that we await the results of new geochronological studies
currently underway in fossil-bearing strata of the John Day Formation, Oregon, that
span this important interval.
Nevertheless, the identification of at least three, and perhaps more, new
endemic ungulates in the Toledo Bend Local Fauna, in addition to the presence of
other species apparently restricted to the Gulf Coastal Plain, such as the small
chalicothere (it has yet to be determined if this and the John Day species are the
same) and amphicyonid (see Albright, 1996, 1998a), serves to reinforce the
previously espoused concept of faunal provinciality of this southeastern region
during the earliest Miocene (see Webb, 1977; Prothero and Sereno, 1982; Tedford et
al., 1987; and Webb et al., 1995). Further study of the mammals noted in this
report, in addition to continued paleontological work in the Gulf Coastal Plain in
general, can only add to our understanding of this region as a unique biogeographic
province during the early phase of large herbivore diversification in North America.


Albright, L. B. 1991. The vertebrate paleontology, taphonomy, and paleoecology of a new early
Miocene site in the Fleming Formation near Toledo Bend Dam, Newton County, Texas, and
its importance to Gulf Coastal Plain Biostratigraphy. M. S. thesis, Louisiana State Univ.,
Baton Rouge, 319 pp.
1994. Lower vertebrates from an Arikareean (earliest Miocene) fauna near the Toledo Bend Dam,
Newton County, Texas. J. Paleon. 68:1131-1145.
1996. Insectivores, rodents, and carnivores of the Toledo Bend Local Fauna: an Arikareean
(earliest Miocene) assemblage from the Texas Coastal Plain. J. Vert. Paleon. 16:458-473.
.1998a. The Arikareean Land Mammal Age in Texas and Florida: southern extension of Great
Plains faunas and Gulf Coastal Plain endemism. Pp. 167-184 in Depositional Environments,
Lithostratigraphy, and Biostratigraphy of the White River and Arikaree Groups (Late Eocene
to Early Miocene, North America), ed. D. O. Terry, H. E. LaGarry, and R. M. Hunt. Geol.
Soc. Amer. Spec. Pap. 325.


1998b. New genus of tapir (Mammalia: Tapiridae) from the Arikareean (earliest Miocene) of the
Texas Coastal Plain. J. Vert. Paleon. 18:200-217.
Allen, G. M. 1926. Fossil mammals from South Carolina. Bull. Mus. Comp. Zool. 67:445-465.
Barbour, E. H. 1905. Notice of a new mammal from Sioux County, Nebraska. Nebraska Geol. Surv.
Behrensmeyer, A. K. 1982. Time resolution in fluvial vertebrate assemblages. Paleobiology 8:211-
Brunet, M. 1975. Donn6es nouvelles concemant la phylogdnie et les migrations des Entelodontidae:
artiodactyles primitifs de L'ancien et du nouveau monde. Colloque intematinal C.N.R.S. n.
218. Probl6mes actuels de pal6ontologie-6volution des vertebr6s, 707-715.
Cook, H. J. 1934. New artiodactyls from the Oligocene and lower Miocene of Nebraska. Amer. Mid.
Nat. 15:148-165.
,and M. C. Cook. 1933. Faunal lists of the Tertiary Vertebrata of Nebraska and adjacent areas.
Nebraska Geol. Surv. 5:1-58.
Coombs, M. C. 1974. Ein Vertreter von Moropus aus dem europaischen Aquitanien und eine
Zusammenfassung der europaischen postoligozanen Schizotheriinae (Mammalia,
Perissodactyla, Chalicotheriidae). -Sitzungsber. Osterr. Akad. Wiss., Mathem.-naturw. Kl.
1975. Sexual dimorphism in chalicotheres (Mammalia; Perissodactyla). Syst, Zool, 24:55-62.
.1978. Reevaluation of early Miocene North American Moropus (Perrisodactyla, Chalicotheriidae,
Schizotheriinae). Bull. Carnegie Mus. Nat. Hist. 4:1-62.
.1979. Tylocephalonyx, a new genus of North American dome-skulled chalicotheres (Mammalia,
Perissodactyla). Bull. Amer. Mus. Nat. Hist. 164:1-64.
S1989. Interrelationships and diversity in the Chalicotheriidae. Pp. 438-457 in The Evolution of
Perissodactyls, eds. D. R. Prothero and R.M. Schoch. Oxford Monographs on Geology and
Geophysics No. 15, Oxford University Press, Inc., New York, N.Y.
Cope, E. D. 1867. An addition to the vertebrate fauna of the Miocene period, with a synopsis of the
extinct Cetacea of the United States. Proc. Acad. Nat. Sci. Philadelphia:138-156.
.1878. On some of the characters of the Miocene fauna of Oregon. Proc. Amer. Phil. Soc. 18:63-
1879. On the extinct species of Rhinoceridae of North America and their allies. Bull. U. S. Geol.
Geog. Surv. Terr. 5:227-237.
Douglass, E. 1901. Fossil mammalia of the White River beds of Montana. Trans. Amer. Phil. Soc.
Forsten, A. 1975. The fossil horses of the Texas Gulf Coastal Plain: a revision. Pearce-Sellards Ser.
Foss, S. E., and T. Fremd. 1998. A survey of the species of entelodonts (Mammalia:Artiodactyla) of the
John Day Basin, Oregon. Dakoterra 5:63-72.
Frailey, D. 1978. An early Miocene (Arikareean) fauna from northcentral Florida (the SB-1A Local
Fauna). Occ. Pap. Mus. Nat. Hist., Univ. Kansas 75:1-20.
Frailey, D. 1979. The large mammals of the Buda Local Fauna (Arikareean: Alachua County, Florida).
Bull. Florida State Mus., Biol. Sci. 24:123-173.
Frick, C. 1937. Homed ruminants of North America. Bull. Amer. Mus. Nat. Hist. 69:1-669.
Galloway, W. E., L. A. Jirik, R. A. Morton, and J. R. DuBar. 1986. Lower Miocene (Fleming)
depositional episode of the Texas Coastal Plain and Continental Shelf: structural framework,
facies, and hydrocarbon resources. Bur. Econ. Geol., Univ. Texas, Rept. Invest. No. 150:1-
,D. G. Bebout, W. L. Fisher, J. B. Dunlap, R. Cabrera-Castro, J. E. Lugo-Rivera, and T. M. Scott.
1991. Pp. 245-324 in The Geology of North America, v. J. Cenozoic. The Gulf of Mexico
Basin, ed. A. Salvador. Geological Society of America, Boulder, Colorado.
Gidley, J. W. 1907. Revision of the Miocene and Pliocene Equidae of North America. Bull. Amer.
Mus. Nat. Hist. 23:865-934.
Green, M. 1958. Arikareean rhinoceroses from South Dakota. J. Paleon. 32:587-594.


Haq, B. U., J. Hardenbol, and P. R. Vail. 1987. Chronology of fluctuating sea levels since the Triassic.
Science 2235:1156-1167.
Harksen, J. C., and J. R. Macdonald. 1967. Miocene Batesland Formation named in southwestern
South Dakota. South Dakota Geol. Surv., Rept. Invest. 96:1-10.
Hay, O. P. 1924. Description of some fossil vertebrates from the upper Miocene of Texas. Proc. Biol.
Soc. Washington 37:1-20.
Hibbard, C. W., and K. A. Keenmon. 1950. New evidence of the lower Miocene age of the Blacktail
Deer Creek Formation in Montana. Contrib. Univ. Michigan Mus. Paleon. 8:193-204.
Honey, J. G., J. A. Harrison, D. R. Prothero, and M. S. Stevens. 1998. Camelidae. Pp. 439-462 in
Evolution of Tertiary Mammals of North America, Volume 1: Terrestrial Carnivores,
Ungulates, and Ungulatelike Mammals, eds. C. M. Janis, K. M. Scott, and L. L. Jacobs.
Cambridge University Press, Cambridge, UK.
Huddlestun, P. F. 1993. A revision of the lithostratigraphic units of the coastal plain of Georgia the
Oligocene. Georgia Geol. Surv. Bull. 105: 1-152.
Hulbert, R. C. 1984. Paleoecology and population dynamics of the early Miocene (Hemingfordian)
horse Parahippus leonensis from the Thomas Farm site, Florida. J. Vert. Paleon. 4:547-558.
Hunt, R. M. 1972. Miocene amphicyonids (Mammalia: Camivora) from the Agate Spring quarries,
Sioux County, Nebraska. Amer. Mus. Nov. 2506:1-52.
Janis, C. M., K. M. Scott, and L. L. Jacobs. 1998. Evolution of Tertiary Mammals, Volume 1:
Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals. Cambridge University
Press, Cambridge, UK. 691 pp.
Leidy, J. 1868. Notice of some remains of extinct pachyderms. Proc. Acad. Nat. Sci. Philadelphia:23-
S1869. The extinct mammalian fauna of Dakota and Nebraska. J. Acad. Nat. Sci. Philadelphia,
2nd Ser., 7:312-313.
S1873. Contributions to the extinct vertebrate fauna of the Western Territories. Department of the
Interior. Report of the United States Geological Survey of the territories. Washington, 1:14-
Loomis, F. B. 1908. A new horse from the lower Miocene. Amer. J. Sci., 4th Ser.:163-165.
Lucas, S. G., D. P. Whistler, and H. M. Wagner. 1997. Giant entelodont (Mammalia:Artiodactyla)
from the early Miocene of southern California. Contrib. Sci., Nat. Hist. Mus. Los Angeles
Co. 446:1-9.
Lucas, S. G., R. J. Emry, and S. E. Foss. 1998. Taxonomy and distribution of Daeodon, an Oligocene-
Miocene entelodont (Mammalia:Artiodactyla) from North America. Proc. Biol. Soc.
Washington 111:425-435.
Macdonald, J. R. 1956. The North American anthracotheres. J. Paleon. 30:615-645.
S1963. The Miocene faunas from the Wounded Knee area of western South Dakota. Bull. Amer.
Mus. Nat. Hist. 125:139-238.
1970. Review of the Miocene Wounded Knee faunas of southwestern South Dakota. Bull. Los
Angeles Co. Mus. Nat. Hist. 8:1-82.
and Martin, J. E. 1987. Arretotherium fricki (Artiodactyla, Anthracotheriidae) from the
Hemingfordian (Miocene) Flint Hill local fauna in South Dakota. Pp. 57-62 in Papers in
Vertebrate Paleontology in Honor of Morton Green, ed. J. E. Martin. Dakoterra 3:57-62.
and C. B. Schultz. 1956. Arretotherium fricki, a new Miocene anthracothere from Nebraska.
Bull. Univ. Nebraska State Mus. 4:53-67.
MacFadden, B. J., and R. M. Hunt. 1998. Magnetic polarity stratigraphy and correlation of the
Arikaree Group, Arikareean (late Oligocene-early Miocene) of northwestern Nebraska. Pp.
143-166 in Depositional Environments, Lithostratigraphy, and Biostratigraphy of the White
River and Arikaree Groups (Late Eocene to Early Miocene, North America), eds. D. O.
Terry, H. E. LaGarry, and R. M. Hunt. Geol. Soc. Amer. Spec. Pap. 325.
Maglio, V. J. 1966. A revision of fossil selenodont artiodactyls from the middle Miocene Thomas
Farm, Gilchrist County, Florida. Breviora 255:1-27.
Manning, E. 1990. The late early Miocene Sabine River. Trans. Gulf Coast Assoc. Geol. Soc. 40:531-


Marsh, O. C. 1870a. (untitled notes) Remarks on Hadrosaurus minor, Mosasaurus crassidens, Liodon
laticaudus, Baptosaurus and Rhinoceros matutinus. Proc. Acad. Nat. Sci. Philadelphia, (for
Jan. 4, 1870), 22:2-3.
1870b. (untitled notes) Announcement of discovery of Meleagrus altus and of Dicotyles
antiquus. Proc. Acad. Nat. Sci. Philadelphia, (for May 8, 1870), 22:11.
1873. Notice of new Tertiary mammals. Amer. J. Sci., 3rd Ser. 105:407-411, 485-488.
1874. Notice of new equine mammals from the Tertiary formation. Amer. J. Sci. 7:247-258.
S1875. Notice of new Tertiary mammals, IV. Amer. J. Sci., 3rd Ser., 109:239-250.
S1877. Notice of some new vertebrate fossils. Amer. J. Sci., 3rd Ser., 14:249-256.
S1893. Description of Miocene Mammalia. Amer. J. Sci., 3rd Ser., 46:407-412.
Matthew, W. D. 1907. Lower Miocene fauna from South Dakota. Bull. Amer. Mus. Nat. Hist. 23:169-
1909. Observations upon the genus Ancodon. Bull. Amer. Mus. Nat. Hist. 26:1-7.
. 1913. MS (see Osbom, 1918, p. 2).
McKenna, M. C. 1965. Stratigraphic nomenclature of the Miocene Hemingford Group, Nebraska.
Amer. Mus. Nov. 2228:1-21.
Merriam, J. C. 1913. New anchitheriine horses from the Tertiary of the Great Basin area. Bull. Dept.
Geol., Univ. California 7:419-434.
Morgan, G. S. 1994. Miocene and Pliocene marine mammal faunas from the Bone Valley Formation
of central Florida. Pp. 239-268 in Contributions in Marine Mammal Paleontology Honoring
Frank C. Whitmore, Jr., ed. A. Berta and T. Dem6er. Proc. San Diego Soc. Nat. Hist. 29.
Osbom, H. F. 1918. Equidae of the Oligocene, Miocene, and Pliocene of North America, iconographic
type revision. Mem. American Mus. Nat. Hist., New Ser., 2:1-217.
,and J. L. Wortman. 1892. Characters of Protoceras (Marsh) the new artiodactyl from the Lower
Miocene. Bull. Amer. Mus. Nat. Hist. 4:351-372.
Parris, D. C., and M. Green. 1969. Dinohyus (Mammalia: Entelodontidae) in the Sharps Formation,
South Dakota. J. Paleon. 43:1277-1279.
Patton, T. H. 1966. Miocene and Pliocene artiodactys, Texas Gulf Coastal Plain. Ph.D. Dissertation,
Univ. Texas, Austin, 166 pp.
S1967a. Revision of the selenodont artiodactyls from Thomas Farm. Quart. J. Florida Acad. Sci.
S1967b. Reevaluation of Hay's artiodactyl types from the Miocene of the Texas Coastal Plain.
Texas J. Sci. 19:35-40.
S1967c. Oligocene and Miocene vertebrates from central Florida. Pp. 115-226 in Miocene -
Pliocene Problems of Peninsular Florida, eds. H. K. Brooks and J. R. Underwood, Jr.
Southeastern Geological Society 13th Field Trip,.
1969. Miocene and Pliocene artiodactyls, Texas Gulf Coastal Plain. Bull. Florida State Mus.,
Biol. Sci., 14:115-226.
,and B. E. Taylor. 1971. The Synthetoceratinae (Mammalia, Tylopoda, Protoceratidae). Bull.
Amer. Mus. Nat. Hist. 145:119-218.
and B. E. Taylor. 1973. The Protoceratinae (Mammalia, Tylopoda, Protoceratidae) and the
systematics of the Protoceratidae. Bull. Amer. Mus. Nat. Hist. 150:347-414.
,and S. D. Webb. 1970. Fossil vertebrate deposits in Florida. The Plaster Jacket, Florida State
Museum 14:1-19.
Peterson, O. A. 1905a. A correction of the generic name (Dinochoerus) given to certain fossil remains
from the Loup Fork Miocene of Nebraska. Science, New Ser., 22:719.
S1905b. Preliminary note on a gigantic mammal from the Loup Fork beds of Nebraska. Science,
New Ser., 22:211-212.
1905c. New Suilline remains from the Miocene of Nebraska. Mem. Caregie Mus. 1904-1906,
S1907. The Miocene beds of western Nebraska and eastern Wyoming and their vertebrate faunae.
Ann. Carnegie Mus. 4:21-72.
1909. A revision of the Entelodontinae. Mem. Camegie Mus. 4:41-156.
1920. The American diceratheres. Mem. Carnegie Mus. 7:399-476.


Prothero, D. R. 1998. Rhinocerotidae. Pp. 595-605 in Evolution of Tertiary Mammals of North
America, Volume 1: Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals, eds. C.
M. Janis, K. M. Scott, and L. L. Jacobs. Cambridge University Press, Cambridge, UK.
and E. Manning. 1987. Miocene rhinoceroses from the Texas Gulf Coastal Plain. J. Paleon.
and N. Shubin. 1989. The evolution of Oligocene horses. Pp. 142-175 in The Evolution of
Perissodactyls, eds. D. R. Prothero and R. M. Schoch. Oxford Monographs on Geology and
Geophysics No. 15, Oxford University Press, Inc., New York, N.Y.
and P. C. Sereno. 1982. Allometry and paleoecology of medial Miocene dwarf rhinoceroses from
the Texas Gulf Coastal Plain. Paleobiology 8:16-30.
, C. Gudrin, and E. Manning. 1989. The history of the Rhinocerotoidea. Pp. 321-340 in The
Evolution of Perissodactyls, eds. D. R. Prothero and R.M. Schoch. Oxford Monographs on
Geology and Geophysics No. 15, Oxford University Press, Inc., New York, N.Y.
E. Manning, and C. B. Hanson. 1986. The phylogeny of the Rhinoceotoidea (Mammalia,
Perissodactyla). Zool. J. Linnaean Soc. 87:341-366.
Quinn, J. H. 1955. Miocene Equidae of the Texas Gulf Coastal Plain. Univ. Texas Bur. Econ. Geol.
Publ. 5516:1-90.
Schlaikjer, E. M. 1935. Contributions to the stratigraphy and paleontology of the Goshen Hole area,
Wyoming, IV. New vertebrates and the stratigraphy of the Oligocene and early Miocene.
Bull. Mus. Comp. Zool. 76:91-189.
1937. A study of Parahippus wyomingensis and a discussion of the phylogeny of the genus.
Bull. Mus. Comp. Zool. 80:255-280.
Schultz, C. B. 1938. The Miocene of western Nebraska. Amer. J. Sci. 35:441-444.
Scott, T. M. 1997. Miocene to Holocene history of Florida. Pp. 57-67 in The Geology of Florida, eds.
A. F. Randazzo and D. S. Jones. University Press of Florida, Gainesville.
Scott, W. B. 1896. The Mammalia of the Deep River beds. Trans. Amer. Phil. Soc., New Ser., 18:55-
S1940. Artiodactyla. Pp.363-746 in W. B. Scott and G. L. Jepsen, eds. The mammalian fauna of
the White River Oligocene, Part 4. Trans. Amer. Phil. Soc., New Ser., 28.
Simpson, G. G. 1930. Tertiary land mammals of Florida. Bull. Amer. Mus. Nat. Hist. 59:149-211.
1932. Miocene land mammals from Florida. Bull. Florida State Geol. Surv. 10:7-41.
Sinclair, W. J. 1905. New or imperfectly known rodents and ungulates from the John Day series.
Univ. California Publ., Bull. Dept. Geol. 4:125-143.
Skinner, M. F. 1968. A Pliocene chalicothere from Nebraska, and the distribution of chalicotheres in
the Late Tertiary of North America. Amer. Mus. Nov. 2346:1-24.
S. M. Skinner, and R. J. Goors. 1968. Cenozoic rocks and faunas of Turtle Butte, south-central
South Dakota. Bull. Amer. Mus. Nat. Hist. 138:379-436.
Stevens, M. S. 1977. Further study of Castolon Local Fauna (early Miocene) Big Bend National Park.
Texas: Pearce-Sellards Ser. 8:1-69.
B. Stevens, and M. R. Dawson. 1969. New early Miocene formation and vertebrate local
fauna, Big Bend National Park, Brewster County, Texas. Pearce-Sellards Ser. 15:1-53.
Stirton, R. A. 1940. Phylogeny of North American Equidae. Univ. California Publ., Bull. Dept. Geol.
Sci. 25:165-198.
Tanner, L. G. 1969. A new rhinoceros from the Nebraska Miocene. Bull. Univ. Nebraska State Mus.
Taylor, B. E., and S. D. Webb. 1976. Miocene Leptomerycidae (Artiodactyla, Ruminantia) and their
relationships. Amer. Mus. Nov. 2596:1-22.
Tedford, R. H., M. F. Skinner, R. W. Fields, J. M. Rensberger, D. P. Whistler, T. Galusha, B. E. Taylor,
J. R. Macdonald, and S. D. Webb. 1987. Pp. 153-210 in Faunal succession and
biochronology of the Arikareean through Hemphillian interval (late Oligocene through
earliest Pliocene epochs) in North America. Cenozoic Mammals of North America,
Geochronology and Biostratigraphy, ed. M. O. Woodbume. Univ. California Press,

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