~8S~18~ ___ B_U_L_L_E_T_IN __ _ 
OF NATURAL HISTORY 

CRANIUM OF DINOHIPPUS MEXICANUS 
(MAMMALIA: EQUIDAE) 

FROM THE EARLY PLIOCENE (LATEST HEMPHILLIAN) 
OF CENTRAL MEXICO,AND THE ORIGIN OF EQUUS 

Bruce J. MacFadden and Oscar Carranza-Castafieda 

Vol. 43, No.5, pp. 163-185 2002 

UNIVERSITY OF FLORIDA GAINESVILLE 



Numbers of the BULLETIN OF THE FLORIDA MUSEUM OF NATURAL HISTORY are published at 
irregular intervals. Volumes contain about 200 pages and are not necessarily completed in anyone calendar year. 

The end of a volume will be noted at the foot of the first page of the last issue in that volume. 

DAVID W. STEADMAN, Editor 
MARGARET E.B. JOYNER, Managing Editor 

Send communications concerning purchase or exchange of the publication and manuscripts queries to: 

ISSN: 0071-6154 

Publication date: August 7, 2002 

Managing Editor of the BULLETIN 
Florida Museum of Natural History 

University of Florida 
PO Box 117800 

Gainesville, FL 32611-7800, U.S.A. 
Phone: 352-392-1721, x457 

Fax: 352-846-0287 
e-mail: mjoyner @ flmnh.ufl.edu 

CODEN: BF 5BA5 

Price: $5.00 



CRANIUM OF DINOHIPPUS MEXICANUS (MAMMALIA: EQUIDAE) 
FROM THE EARLY PLIOCENE (LATEST HEMPHILLIAN) 

OF CENTRAL MEXICO, AND THE ORIGIN OF EQUUS 

Bruce J. MacFadden! and Oscar Carranza-Castafieda2 

ABSTRACT 
A newly discovered skull of Dinohippus is described from the latest Hemphillian (early Pliocene) Rancho El Ocote locality of 
Guanajuato, Mexico, which is dated at 4.8 Ma. This cranium is referred to D. mexicanus, the senior synonym for the latest 
Hemphillian species otherwise known from several localities in northern Mexico and the southern United States. Although 
crushed, this is the most complete skull known for this extinct species. With the diagnostic configuration of the dorsal preorbital 
fossa, distinctive dental pattern, and moderate tooth curvature, this cranium demonstrates a morphology similar to, although 
slightly more primitive than, that of closely related and slightly more derived Blancan Equus, such as E. simplicidens. As 
represented by occurrences in central Mexico and southern California, D. mexicanus coexisted with Equus during the middle 
Blancan from about 4.5 to 3 million years ago. Despite traditional interpretations of anagenetic speciation, the current study 
demonstrates that primitive species of Equus originated from D. mexican us by cladogenesis. 

RESUMEN 
En este trabajo, se describe el reciente descubrimiento de uno cnineo de Dinohippus de la localidad de Rancho El Ocote, 
Henfiliano tardio (Plioceno temprano) del estado de Guanajuato, que ha sido fechado en 4.8 Ma. Este cnineo es referido a 
Dinohippus mexicanus, senior sin6nimo de las especie del Henfiliano tardio, que se conoce en diferentes localidades del norte 
de Mexico y sur de los Estados Unidos. El cnineo es mas completo que se conoce de esta especie D. mexicanus. Con la 
configuraci6n diagn6stica de la fosa preorbital dorsal, los diferencias en el patr6n oclusal y la moderada curvatura de los 
molariformes, este cnineo demuestra una morfologia similar a Dinohippus, ligeramente mas primitive que aquellos que mas 
relacionados con Equus del Blancano, como E. simplicidens. Los registros en el centro de Mexico y el sur de California, 
indican que D. mexicanus coexisti6 con Equus durante el Blanceano medio desde 4.5 hasta 3 Ma. A pesar de los interpretaciones 
de especiaci6n anagenica, el presente estudio demuestra que los especies primitivas de Equus, se originaron de D. mexicanus 
por cladogenesis. 

Key words: Dinohippus, mexicanus, Equus, Hemphillian, Blancan, Mexico 

TABLE OF CONTENTS 

Introduction ............................................................................................................ 164 
Acknowledgments .................................................................................................. 164 
Materials, Methods, Terminology, and Abbreviations ........................................... 164 
Systematic Paleontology ......................................................................................... 165 
Discussion .............................................................................................................. 168 

Nomenclature, geographic distribution, and age of Dinohippus mexicanus .... 168 
Craniofacial morphology and transitional dental characters .. ........ .......... ....... 170 
Origin of Equus ................................................................................................. 175 

Summary and Conclusions .................................................... ................................ 181 
Literature Cited ...................................................................................................... 181 
Appendix. Practical method for determining the radius of tooth curvature.. ....... 184 

! Associate Director of Exhibits and Public Programs and Curator of Vertebrate Paleontology. Florida Museum of Natural History. University of Florida, 
Gainesville FL 32611-2710, email: bmacfadd@flmnh.ufl.edu. 
2Professor of Paleontology, Unidad de Ciencias de la Tierra, Universidad NacionalAut6noma de Mexico, Campus Juriquilla, Queretaro, c.P. 76230, Mexico. 

B. J. MacFADDEN and O. CARRANZA-CASTANEDA. 2002. Cranium of Dinohippus mexicanus (Mammalia: Equidae) from the early Pliocene (latest Hemphillian) 
of central Mexico, and the origin of EqllUS. Bull. Florida Mus. Nat. Hist. 43(5):163-185. [End of Vol 43.] 



164 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

INTRODUCTION 
The latest Hemphillian (early Pliocene) about 5 million 
years ago was a very interesting time in equid evolution 
when four to six sympatric species coexisted at many 
localities in North America. Several studies have asserted 
that the latest Hemphillian species Dinohippus mexicanus 
(Lance) 1950 is of prime importance in understanding 
the origin of the modern genus Equus. In fact, D. 
mexicanus is hypothesized to be the closest sister-species 
of primitive species of Equus, such as E. simplicidens 
(Cope) 1893 (e.g., Bennett 1980; MacFadden 1984; 
Prado and Alberdi 1996; Kelly 1998). 

The species Dinohippus mexicanus originally was 
described as Pliohippus mexican us based on a large 
collection of latest Hemphillian horses from the Yep6mera 
Local Fauna of Chihuahua, Mexico (Lance 1950). This 
horse has since been found at other localities in Mexico 
and the United States, and although these occurrences 
have sometimes been given new species names, they all 
seem close to, or conspecific with, D. mexicanus 
(MacFadden 1984; Carranza-Castaneda 1992). Over the 
past several decades, intensive geological and 
paleontological excavations have yielded an excellent 
assemblage of latest Hemphillian fossil mammals from 
the state of Guanajuato, Mexico. One of the specific 
localities in this region, Rancho EI Ocote, has a rich and 
diagnostic fauna, including abundantly represented 
horses, particularly D. mexicanus and another 
monodactyl species Astrohippus stockii (Lance) 1950, 
but also including the rarer tridactyl Neohipparion 
eurystyle (Cope) 1893 and Nannippus aztecus Mooser 
1968 [= N. minor (Sellards) 1916; Carranza-Castaneda 
and Ferrusquia-Villafranca 1978; also see Hulbert 1990, 
1992]. Despite many years of field work at Rancho EI 
Ocote, the large collection of horses from this locality 
has until now consisted of isolated teeth, dentitions, and 
postcranial remains. The facial morphology of 
Dinohippus mexicanus, which is of fundamental 
importance to understanding Neogene equid systematics, 
was previously unknown from Rancho EI Ocote, poorly 
represented from Yep6mera, and unknown from other 
latest Hemphillian localities (Lance 1950; MacFadden 
1984,1986; Azzaroli 1988). 

During a field trip in 1997, the authors visited Rancho 
El Ocote and collected a nearly complete, although 
crushed, skull of Dinohippus mexicanus, one of only 
two known from this species as it is broadly defined (e.g., 
MacFadden, 1984). This new skull is significant because 
D. mexican us is hypothesized to be the closest relative 

of primitive species of Equus (Bennett 1980; MacFadden 
1984; Prado and Alberdi 1996; Kelly 1998), and the 
evolution of craniofacial morphology is critical to 
understanding the systematics of Neogene equids (see 
review in MacFadden 1992). The new skull from Rancho 
El Ocote therefore elucidates previous phylogenetic 
hypotheses concerning the origin of Equus. In addition 
to the skull of D. mexicanus described here from Rancho 
El Ocote, new teeth of equine horses collected from an 
early Blancan locality in Jalisco, Mexico, elucidate the 
mode of speciation that gave rise to primitive Equus. 

ACKNOWLEDGMENTS 
This project was supported by NSF EAR 99-02898. We 
thank Gerardo Alvarez for specimen preparation, Antonio 
Altamira-Gallardo and Erika Simons for photography, 
Merald Clark for preparing the line drawings, Dr. Sam 
McLeod, Dra. Maria del Carmen Perrilliat, and Dr. 
Richard Tedford for access to, respectively, the LACM, 
IGM, and AMNH collections, and Richard C. Hulbert 
Jr., Everett H. Lindsay, Charlotte M. Porter, and Fred G. 
Thompson for helpful comments that improved the 
manuscript. This is University of Florida Contribution 
to Paleobiology number 520. 

MATERIALS, METHODS, TERMINOLOGY, 
AND ABBREVIATIONS 

The following vertebrate paleontology collections were 
examined during this study and are abbreviated in the 
text as follows: 
AMNH, American Museum of Natural History, New 

York. 
F:AM, Frick:American Mammals, part of the AMNH. 
IGM, Instituto de Geologia Museum, Ciudad 

Universitaria, Universidad Nacional Aut6noma de 
Mexico. 

LACM, Natural History Museum of Los Angeles County. 
UP, University of Florida. 
UMNH, Utah Museum of Natural History, University 

of Utah. 
All measurements are in millimeters, and are reported 

to the nearest tenth mm (teeth) or mm (cranial 
measurements). Statistical calculations were done using 
Microsoft Excei™. 

The following abbreviations and/or codes are used 
in the text: 
A; adult wear stage, occlusal pattern moderately worn. 
APL; greatest anteroposterior tooth length, excluding 

cement. 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 165 

35 mm Radius 

, Curvature index 35 

Figure 1. Method for determining the radius of curvature, 
CURV, of upper cheek teeth of fossil horses (modified from 
Skinner and Taylor [1967]; also see Appendix). Reproduced 
courtesy of the American Museum of Natural History. 

C; upper canine. 
CURV, curvature of mesostyle, as measured in the radius 

of circle inscribed (Fig. 1, taken from Skinner and 
Taylor 1967; also see Appendix). 

DPOF; dorsal preorbital fossa. 
I; upper incisor. 
J; juvenile wear stage; tooth unworn or little worn; this 

is used to measure individuals that demonstrate maxi­
mum potential crown height (see MSTHT, below). 

L; left side. 
L.F.; local fauna; a geographically and temporally 

restricted fossil assemblage. 
M; m; molar (upper, lower). 
Ma; megannum, millions of years ago, in reference to a 

point in time. 
MSTHT; mesostyle crown height. 
0; old age, tooth heavily worn; these individuals were 

removed from pooled analyses representing 
characteristic dental measurements (Table 1). 

P, p; premolar (upper, lower). 
R; right side. 
TRN; greatest transverse width, excluding cement. 

SYSTEMATIC PALEONTOLOGY 
Class Mammalia Linnaeus, 1758 
Order Perissodactyla Owen, 1848 

Family Equidae Gray, 1821 
Genus tDinohippus Quinn, 1955 

tDinohippus mexicanus (Lance) 1950 
(Figs. 2-11; Tables 1-2) 

Synonomy of Rancho EI Ocote referred sample: 
large Pliohippus sp. Arellano 1951, p. 613 
Hippotigris ocotensis Mooser 1958, p. 360 
Protohippus muelleri Mooser 1965, p. 157 
Dinohippus muelleri Mooser 1973, p. 258 
Equus (Dolichohippus) mesamericanus Mooser 1973, 

p.261 
Pliohippus mexicanus Carranza-Castaneda and 

Ferrusquia-Villafranca 1978, p. 165 
Dinohippus ocotensis Dalquest and Mooser 1980, p. 8 
Dinohippus mexicanus Carranza-Casteneda 1992, p. 186 

Holotype.-"Pliohippus" mexicanus, LACM-CIT 
(California Institute of Technology) 3697, partial L 
maxilla with P2-M3 and part of zygomatic arch, from 
CIT locality 286, Yep6mera L. F., Chihuahua, Mexico 
(Lance 1950). 

Referred specimen.-IGM 7596, cranium with R 
13, P2-M3, L 12-13, P2-M3; also cast UF 206861. 

Locality, Age, and Collector.-Rancho EI Ocote, 
Ravine de La Caretta, IGM locality GTO 2b, 
Guanajuato, Mexico, latest Hemphillian (early Pliocene), 
ca. 4.8 Ma. Collected by the authors on 29 May 1997. 

Specific Diagnosis.-Medium-sized monodactyl 
equine horse, basilar skull length ca. 430 mm; mean 
moderately worn (ontogeny = A) MI-M2 APL = 24.5 
mm and TRN = 24.0 mm (Table 1). Nasal notch retracted 
to a position lying dorsal to the P2. Moderately hypsodont 
with mean unworn (ontogeny = J) M12 MSTHT = 69.0 
mm (Table 1). Cheek teeth moderately curved 
transversely, with mean CURV = 73 mm. DPOF 
moderately well developed dorsally, ventrally, and 
posteriorly, although characteristically lacking a distinct 
rim. Long preorbital bar between DPOF and orbit. Malar 
fossa very poorly developed, or absent. Cheek tooth enamel 
pattern generally simple. Upper cheek tooth protocones 
oval and moderately elongated, especially posterior to 
the connection with the protoloph, and pre- and 
postfossettes crescentic in occlusal cross section with few 
plications. Hypoconal groove well developed and persists 
until late wear. Lower cheek tooth metaconids and 
metastylids have rounded borders and are well-separated, 
ectoflexid moderately deep in the premolars and deep in 



166 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

Table 1. Comparison of dental measurements from Dinohippus interpolatus (from Miami = Coffee Ranch Quarry, Texas, in 
AMNH collection), Dinohippus mexicanus (pooled sample from Florida, Texas, and Mexico in F:AM, IGM, LACM and UF 
collections), and primitive Equus (pooled sample from Florida, Idaho, Nebraska, Texas, and Mexico in AMNH, F:AM, IGM, 
LACM, UF, and UMNH collections). Tooth measurements are taken on Ml or M2. Measurements are reported in the 
following sequence for each entry: N, number of specimens measured; x, mean; s, standard deviation; range, i.e., observed 
minimum to maximum. See text for abbreviations. 

Character Dinohippus interpolatus Dinohippus mexicanus primitive Equus ( ANOVAProb! Different?! 

APU 12, 25.8, 0.8, 24.6-27.3 21, 24.5, 1.1, 22.5-26.7 12, 27.2, 1.4, 24.8-29.0 <0.001 Yes 

TRN2 12,25.1, 1.1,22.8-26.4 20, 24.0, 1.6, 20.7-27.8 12,28.2, 1.9,25.1-31.3 <0.001 Yes 

MSTHT3 9,69.0,4.9,62.1-78.1 9,69.0,6.0,60.1-77.4 5, 90.3, 5.5, 86.3-99.8 <0.001 Yes 

CURV 5,60, 0, 60-60 6, 73, 6.1, 65-80 11,115.5,25.4,90-185 <0.001 Yes 

TRL 3, 159.6, 1.1, 158.4-160.6 3, 157.0,6.2, 151.0-163.3 34,188.5,7.9,168.8-204.34 

I ANOVA probability level for the three species and whether or not these samples are statistically different. 
2 Adult wear stage (A), i.e., juveniles (J) and old age (0) individuals removed from pooled sample. 
3Juvenile wear stage (J) in which tooth crowns are unworn or little worn, to indicate maximum MSTHT. 
4Data for Equus taken from MacFadden (1989) for sample of Equus simplicidens from Idaho. 

the molars, and pli caballinids poorly developed or absent 
(also see Lance 1950; MacFadden 1984). 

Dinohippus mexicanus differs from contem­
poraneous Astrohippus stockii because of its larger size, 
lack of a ventral (malar) fossa, and details of the enamel 
pattern (e.g., shape of the protocone and less flared 
metaconids and metastylids). D. mexicanus differs from 
more primitive species within this genus (such as D. 
interpolatus) in less transversely curved cheek teeth and 
possibly slightly less defmed DPOF. D. mexicanus differs 
from primitive species of Equus, such as E. simplicidens, 
in smaller size, shorter crown heights, more transversely 
curved upper cheek teeth, less elongated protocones with 
rounded enamel, and less expanded metaconids and 
metastylids with rounded enamel borders. 

Specimen Description.-With a mean cheek tooth 
row length of 157.9 mm and a mean M12 APL of 23.3 
mm (Table 2), IGM 7596 represents a moderately large 

equine horse. As evidenced by tooth measurements, IGM 
7596 falls at the lower end of the observed range for the 
current, geographically broader concept of the species 
Dinohippus mexicanus (Table 1). The occipital condyles 
are not preserved, but the basioccipital region just anterior 
to the condyles indicates an approximate basilar skull 
length (i.e., tip of foramen magnum to anterior-most 
portion of symphysis; Osborn 1912) slightly greater than 
430 mm (Fig. 2). The skull is badly crushed but preserves 
many important characters of this species, in particular 
the morphology of the facial region and dorsal preorbital 
fossa (DPOF). The nasal notch is retracted to a position 
that lies dorsal to the posterior half ofP2. The infraorbital 
foramen appears to lie dorsal to P3. An apparent 
depression directly anterior to the DPOF seems to have 
resulted from crushing during fossilization. The DPOF 
is best preserved on the left side (Fig. 3). It is positioned 
high on the cheek -75 mm above the dorsal part of the 

Table 2. Dental measurements (excluding cement) of Dinohippus mexicanus, IGM 7596 (UF 206861) 
from Rancho EI Ocote (GTO 2b), Guanajuato, Mexico. 

Measurement 13 P2 P3 P4 M1 M2 M3 P2-M3 

R anteroposterior length (APL) 18.3 34.3 26.3 25.8 22.5 22.6 23.8 152.4* 
R transverse width (TRN) 10.6 24.9 26.0 25.0 24.2 23.8 21.6 
L anteroposterior length (APL) 18.5 35.0 26.0 24.8 25.0 163.3 
L transverse width (TRN) 9.9 27.0 26.8 25.5 21.0 

*Postmortem gap between M2 and M3 (3.4 mm) subtracted from 155.8 mm to yield 152.4 mm, the latter of which is taken 
as the actual P2-M3 APL. 



F
ig

ur
e 

2.
 

V
en

tr
al

 v
ie

w
 o

f 
cr

an
iu

m
 o

f 
D

in
oh

ip
pu

s 
m

ex
ic

a
nu

s,
 I

G
M

 7
59

6 
(=

 U
F

 2
06

86
1,

 c
as

t)
, 

fr
o

m
 R

an
ch

o 
E

l O
co

te
, 

la
te

st
 H

em
ph

ill
ia

n 
o

f 
G

ua
na

ju
at

o,
 M

ex
ic

o.
 

3::
 '" (

)
 ~
 

o o tT
l 

Z
 '" ::
I 

0
- n ~ z N
 » n » ~I o » n ;;J
 

~
. 

c :3 o ~
 s·
 

o ~
 :g'
 

t::
 '" ~ ~.
 

2 " ~ '" ::I 0
- s- (I
) o ::J
. 

<!9
. 

::
I o ....,

 
J;1

 
~ '" -0- -.

J 



168 BULLETIN of the FLORIDA MUSEUM OFNATURALIDSTORYVol. 43 (5) 

c=-===.:: ---

o 5 10 - - -em- - I 

Figure 3. Left lateral, reconstructed view of cranium of Dinohippus mexicanus, IGM 7596 (= UP 206861, cast) from Rancho 
EI Ocote, latest Hemphillian of Guanajuato, Mexico. 

tooth row (above M1). The DPOF has well-developed 
edges, including the dorsal margin (also see Kelly 1998 
for the significance of this morphology). Posterior to the 
crushed region that lies anterior to the DPOF, the 
anteroposterior length of the DPOF is greater than ~65 
mm, the dorsoventral height is ~25 mm, and this fossa is 
~1O mm deep in the center (also see Eisenmann et al. 
1988 for measurement conventions), although this depth 
may be accentuated by crushing. The posterior margin 
of the DPOF is situated far forward of the orbit, i.e., 
there is a long preorbital bar of ~80 mm. The sutures of 
individual bones are not well preserved, but it appears 
that the DPOF lies on the nasal and maxillary bones 
anterior to the lacrimal bone. The region of the malar 
facial fossa, located ventrally, is not preserved on the 
left side, but there is an indication of a faint depression 
on the right side (not illustrated). As represented on the 
right side, a strong transverse malar crest is located on 
the ventral region of the face. The muzzle region is very 
broad (74 mm) and robust. There is a slight postcanine 
constriction (so that the corresponding transverse width 
just posterior to the canine is 56 mm). 

Although the R & L 11 and 12 are not preserved, as 
inferred from the alvoeli and R & L 13, the shape of the 
incisor series is moderately curved (Fig. 2), i.e., it is 
neither very rounded, as in some browsing mammals, 
nor is it very linear, as in such extinct grazing horses as 
CaUppus. Neither of the canines is preserved. The post­
canine diastema (R = 54.0 mm, L is not preserved) is 
much larger than the precanine diastema (R = 15.8 mm, 

L is not preserved). The dPl is absent. The cheek teeth 
represent a mature adult in late middle wear. The 
perimeter of the tooth crowns is covered with thick 
cement. Although direct measurements of CURV cannot 
be taken, as represented by the alveolus for the left M1 
and adjoining P4 and M2, the cheek teeth are moderately 
curved in 10M 7596 (see Table 1 for CURV for other 
specimens of Dinohippus mexicanus, D. interpolatus, 
and primitive Equus). The enamel forming the exterior 
of the tooth and internal fossettes is relatively thick (Fig. 
4). There are prominent parastyles and mesostyles on 
the ectoloph. The fossette borders are relatively simple 
(in contrast to Equus or some advanced hipparionines) 
with no, or one, plication on the anterior border of the 
prefossette and posterior border of the postfossette. In 
contrast, the posterior half of the prefossette and anterior 
half of the postfossette contain one or two plications. The 
hypoconal groove is moderately developed. Because the 
protocone is at a relatively advanced stage of wear, the 
pattern exposed on the occlusal surface is only moderately 
distinct. The characteristic advanced protocone seen in 
equines, which consists of a wooden-shoe shape with 
angular posterior border, is particularly evidenced in M2 
andM3. 

DISCUSSION 
Nomenclature, geographic distribution, and age of 
Dinohippus mexicanus.-The genus Dinohippus was 
proposed by Quinn (1955) to encompass those species 
previously referred to Pliohippus that lack a well­
developed ventral pre orbital fossa. Pliohippus sensu 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 169 

o 5 

em 

Figure 4. Occlusal view of right (reversed) P2-M3 of Dinohippus mexicanus, IGM 7596 (= UP 206861, cast) from Rancho EI 
Ocote, latest Hemphillian of Guanajuato, Mexico. 

stricto is retained for horses with very complex facial 
pits, including those with well-developed dorsal and 
ventral fossae. As currently envisioned, Dinohippus is 
first known from middle Miocene (?Barstovian, but 
certainly Clarendonian) localities from California (D. 
leardi (Drescher) 1941 from Black Hawk Ranch; Kelly 
1998) and specimens in the AMNH/F:AM collections 
from the mid-continent (e.g., Webb 1969; Skinner and 
Johnson 1984). Thereafter, Dinohippus is relatively 
widespread until the end of the Hemphillian (MacFadden 
1992; also see discussion of Blancan localities below). 
Several species names have been applied to large 
monodactyl horses within the concept of advanced, i.e., 
late Hemphillian, Dinohippus Quinn 1955. During the 
early late Hemphillian, as exemplified by localities in 
the Texas panhandle (e.g., Coffee Ranch = Frick Miami 
Quarry), Dinohippus has traditionally been assigned to 
the species D. interpolatus (Cope) 1893 (Matthew and 
Stirton 1930, as Pliohippus) andD. leidyanus (Osborn) 
1918 from the Snake Creek Formation and equivalent 
units in Nebraska (Osborn 1918; Matthew 1924a; Stirton 
1940, asPliohippus). Kelly (1998) describes Dinohippus 
interpolatus from Oakdale, California. In addition to 
these localities, large, essentially undescribed, early-late 
Hemphillian collections of Dinohippus, either referable 
to D. interpolatus or D. leidyanus, are represented in 
the AMNH/F:AM collections from elsewhere in North 
America, including large quarry samples from Edson, 
Kansas, Optima (= Frick Guymon), Oklahoma, and 
Redington, Arizona, as well as smaller collections from 

South Dakota, New Mexico, and Nevada (MacFadden 
pers. observ. 2001). All these samples are early-late 
Hemphillian, ca. 6 Ma (Tedford et al. 1987). Early-late 
(ca. 6 Ma) and latest (ca. 5 Ma) Hemphillian mammalian 
faunas in North America contain different equid species 
assemblages. The typical early late Hemphillian equid 
fauna consists of Dinohippus interpolatus or D. 
leidyanus, Astrohippus ansae (Matthew and Stirton) 
1930, Neohipparion eurystyle (or N. gidleyi), and 
Nannippus lenticularis (Cope) 1892 (sensu Hulbert 
1988, 1993, = Hipparion lentieulare of Matthew and 
Stirton, 1930). In contrast, typical latest Hemphillian 
faunas contain Dinohippus mexicanus, Astrohippus 
stockii, Neohipparion eurystyle, and Nannippus aztecus 
(= N. minor, see Hulbert 1990). 

Lance (1950) originally described the species 
Pliohippus mexieanus from the Yep6mera L.P. of 
Chihuahua, Mexico. Quinn (1955) erected the genus 
Dinohippus for larger "pliohippine" horses from the late 
Miocene and early Pliocene of North America that lack 
the complex facial fossae seen inPliohippus sensu stricto. 
As his study was mostly confined to earlier Miocene 
faunas from the Texas Gulf Coastal Plain, he did not 
address the generic allocation of "Pliohippus" mexicanus 
within the genus Dinohippus. Mooser (1973) followed 
Quinn's (1955) generic designation, although not 
accepting Lance's (1950) species for the large latest 
Hemphillian horse from Rancho El Ocote. Mooser first 
allocated the Rancho El Ocote Dinohippus to Hippotigris 
oeotensis (Mooser 1958), then Protohippus muelleri 



170 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

(Mooser 1964), and finally, following Quinn (1955), to 
the genus Dinohippus as D. muelleri. Dalquest and 
Mooser (1980) assert that D. ocotensis from Rancho El 
Ocote differs from D. mexicanus from Yepomera in 
having more elongated, angular, and deeply grooved· 
protocones, and "in having an anterior extension or spur 
extending in advance ofthe isthmus" (p. 10). The exact 
significance of this latter character is ambiguous; it seems 
to imply the difference in development of the pli caballinid 
in the lower cheek teeth. Carranza-Castafieda and 
Ferrusquia-Villafranca (1978) and Carranza-Castafieda 
(1992) did not accept the validity of Dinohippus muelleri 
or D. ocotensis from Rancho El Ocote and other 
equivalent-aged localities in Guanajuato and referred 
these latest Hemphillian horses to the species mexican us 
(first as Pliohippus, and more recently to Dinohippus). 
Other workers (e.g., Bennett 1980; MacFadden 1984; 
Prado and Alberdi 1996; Kelly 1998) have mostly 
included mexicanus in Dinohippus. Statistical analyses 
of measured dental characters of early late and latest 
Hemphillian Dinohippus seem to represent continuous 
variation within two morphologically similar, closely 
related species, i.e., D. interpolatus and D. mexicanus 
(the latter sensu lato, i.e., encompassing a pooled sample 
from Ocote, Yepomera, and Florida; Fig. 5). We therefore 
assert that the differences used by Dalquest and Mooser 
(1980) to distinguish D. ocotensis and D. mexicanus 
represent individual variation that can be seen within 
different wear stages within a population. Thus, the latest 
Hemphillian species of large mono dactyl horse is 
referable to D. mexicanus and it was widely distributed 
throughout southern North America. 

With the revised, g~ographically more inclusive, 
concept presented here, Dinohippus mexicanus is known 
from several latest Hemphillian fossil localities in North 
America (Fig. 6). In central Mexico, D. mexicanus occurs 
at Rancho El Ocote (including Arroyo de Carretta, GTO 
2b, location of IGM 7596), as well as several other 
localities in Guanajuato, including Rinconada (IGM 
locality GTO 43; Carranza-Castafieda 1992), Arroyo 
Tepalcates (GTO 52), and Rancho San Martin (GTO 
42). In Chihuahua, D. mexicanus is known from the 
extensive LACM collection from Yepomera (Lance 
1950). D. mexicanus also occursJrom the Bone Valley 
deposits collected from open-pit phosphate mines in 
central Florida (MacFadden 1986). In the Texas 
panhandle, D. mexicanus occurs at the Christian Ranch 
L.P. and Rentfro Pit 1 locality (Schultz 1977; Tedford et 
aL 1987; MacFadden pers. observ. 2001). As far as is 

known, D. mexicanus does not occur in the western U.S. 
at any localities in California, Oregon, or Washington. It 
also does not occur in the northern Great Plains of the 
U.S. and Canada (north of the Texas panhandle). Whether 
this limited geographic distribution is a result of a paucity 
of latest Hemphillian sites (see Tedford et aL 1987), or 
represents the actual biogeographic distribution of D. 
mexicanus, cannot be determined at the present time. 

Craniofacial morphology and transitional dental 
characters.-In the context of the present paper, many 
essential cranial characters pertain to the development 
of preorbital pits, or fossae, in the cheek region on either 
the maxillary or nasal bones. The function of these 
preorbital fossae, which do not occur in modem Equus 
and lack any direct functional analog, is not certain. They 
have been purported to be sexually dimorphic and some 
workers have argued that these structures are not of 
taxonomic utility. We do not intend to rehash this 
argument here as it has been amply addressed in the 
literature (see review in MacFadden 1992). Suffice it to 
say that, as also discussed in MacFadden (1984) for the 
Yepomera horses, the preorbital facial fossa (in this case 
the DPOF) is of fundamental importance in understanding 
the morphological differences between, and phylogenetic 
interrelationships of, late Cenozoic equids. As far as is 
known, Pliohippus and Astrohippus have very complex 
facial fossa that consist of multiple pits located both 
dorsally and ventrally on the cheek region (e.g., Bennett 
1980; MacFadden 1984; Kelly 1998). With the exception 
of a weakly developed DPOF in Equus grevyi (Skinner 
and Hibbard 1972), extant Equus lacks any preorbital 
fossae. With the addition of the new specimen (IGM 
7596) from Rancho El Ocote described here, the evolution 
of pre orbital facial fossa in HemphillianDinohippus and 
primitive Blancan Equus can be further clarified. 

In overall morphology of the pre orbital fossae, 
Dinohippus interpolatus andD.leidyanus are generally 
similar (also see Kelly 1998). Although the type of 
"Pliohippus" interpolatus Cope 1893 is based on a single 
M2 from near Goodnight, Texas (Osborn 1918), large 
samples referred to this species are described from Coffee 
Ranch, Texas (Matthew and Stirton 1930), and there is 
a similarly large, essentially undescribed, collection from 
the equivalent Frick Miami Quarry at the AMNH. As 
exemplified by F:AM 116171 and 116172 (the latter 
being the type of Pliohippus bakeri Azzaroli 1988 
(Azzaroli 1988, plate 1), a concept not followed here), 
the preorbital facial morphology of D. interpolatus can 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexican LIS and the origi n of Equus 

32 
A 

30 

28 
---. 
E 
-S26 
Z 
0::: 
~ • 24 • 

22 

20 
22 

o 

o 

o o 8 
eo DO 

18..0 [] 00
0 

L:J nc.D "7 r'iJ 0 
'-tJ O Ltp 0 

o 
24 

~D ca D 

o 

26 
APL (mm) 

o 

28 

o 
o 

o 

30 

Do. interpo/atus Do. mexican us .IGM 7596 0 primitive Equus 

100 B 

E 
E 

90 

j: 80 
~ 
(f) 

~ 

70 

o 

o 
D 

o 

o 
oObo o 0 

1"'1 
OLJ 

o 

o 
o 

o 

60+------r~---,------~-----r----_.----__, 
24 26- 28 30 32 34 36 

APL (mm) 

171 

Figure 5. Bivariate plots of dental characters for Dinohippus interpofatus, Dinohippus mexicanus, and primitive Equus. (A) 
Anteroposterior M I or M2 length versus transverse M I or M2 width , middle wear Uuveniles and old age individuals removed). 
One outlier of Equus from Sand Draw, Nebraska (APL = 33. 1 mm, TRN =28.7 mm) is removed from the plot. (B) Anteroposterior 
M 1 or M2 length versus unworn to little worn Uuvenile) M I or M2 crown he ight. 



172 BULLETIN of the FLORlDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

-;
!\ 

... , 

\_ -- t-------J - - , ----._ 1 

......, '-"'"r- _ ; 0 Hagerman --""1 1 
,'-..\- . I I 

I '!'--'I~_ r----C5~~-----
I T -- ~- _ .... __ Sand Draw 
; i I --~- ..... 
I ~--------

\ J ~ r 
\ ,~--_ I I . , r --t-~- I I 

\ f -'-,f=~--"'::;~~~:~-~\-----l '---'~ __ ... _-, 
IChristian-Rentfro l ' -- ...... - --(---:... 
, J"\.. I ' J ' ....... ' '-'_ \ 

; I' Mt Blanco '\ __ _ 
--,~ I 

\, 
I 

Guanajuato • 
Dinohippus 0 
primitive Equus 0 

Hemphillian (black) 
Blancan (open) 
combined (shaded) 

Figure 6. Selected late Hemphilli an and early Blancan localities from North America with occurrences of Dinohippus and 
Equus, as discussed in thi s paper. Christian-Rentfro refers to the Christian Ranch L. F. and Rentfro Pit 1 locality in the Texas 
panhandle (Sch ultz 1977; Tedford et a l. 1987). 

be characterized as having a poorly developed ventral 
(malar) fossa and a moderately well developed DPOF. 
The ventral fossa is essentially a faint, exceedingly 
shallow depression lying principally on the maxillary 
bone, but with its posterior-most extent developed on the 
maxillary-malar suture. The DPOF is a well defmed oval­
shaped depression that lies on the nasal and maxillary 
bones. Although the general configuration of the DPOF 
is consistently developed in different individual s, there 
is some variation in the development of the margins of 
the DPOF. For example, in F:AM 116171 and 128143 
of D. inlerpo/atus from Miami Quarry, the DPOF 

margins are well defined dorsally, ventrally, and 
posteriorly, and these margins are delinated by a curving 
of the bone into the pit. In the other well preserved 
specimen (F:AM 116172; see Azzaroli 1988, plate 1) 
from Miami Quarry, the general configurations of the 
fossae are similar, although the dorsal and posterior 
margins of thi s structure are delineated by a di stinct rim. 
In both specimens, as also seen in other species of 
Dinohippus, the anterior part of the DPOF is weakly 
defined. 

The other early late Hemphillian species of 
Dinohippus was originally described as Pliohippus 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 173 

leidyanus Osborn 1918 from the Snake Creek Fauna of 
Nebraska. The holotype, AMNH 17224, consists of a 
beautifully preserved cranium and mandible of a subadult 
female (with the M3 almost fully erupted but not in wear, 
Fig. 7 A). Another specimen from the Snake Creek Fauna, 
AMNH 18972 (not illustrated here), represents an adult 
male in which the malar fossa is essentially absent. 
Because the ventral region of the holotype is distorted 
by post-mortem crushing, it is not possible to see any 
original morphology of the malar preorbital fossa. In these 
specimens the DPOF is well developed and its general 
morphology is similar to that for D. interpolatus, with 
well developed dorsal, posterior, and ventral margins. 

Given the description above, the two species of early 
late Hemphillian Dinohippus, i.e., D. leidyanus and D. 
interpolatus, are very similar in facial morphology, as 
they also are in such other characters as size and dental 
pattern. In addition to those described above from Coffee 
Ranch/Miami and Snake Creek, excellent quarry samples 
exist in the AMNH for early late Hemphillian 
Dinohippus, e.g., from Optima = Guymon, Oklahoma, 
and Edson, Kansas. These large quarry samples could 
be used in the future to further assess the intrapopulational 
variation in the DPOF (as Skinner and MacFadden 1977 
did for Cormohipparion) , as well as those characters 
that could be used to differentiate the closely related D. 
interpolatus from D. leidyanus. 

MacFadden (1984) described a partial skull of Dino­
hippus mexicanus (LACM 275/3732; Fig. 7B) from the 
latest Hemphillian of Yepomera (sublocality 275) that 
preserves the preorbital region. The Yepomera fossil 
horizons occur within a normally magnetized zone 
correlated to the early Gilbert magnetic chron. Lindsay 
et al. (1984) present an extrapolated age of ca. 4.6 Ma 
for the locality, placing it in the latest Hemphillian, an 
age that is roughly the same as for the Guanajuato sites 
that contain the same mammal faunas, including the other 
skull of D. mexican us (IGM 7596). Both skulls have 
well-preserved pre orbital regions in which the DPOF is 
well defined dorsally, ventrally, and posteriorly, whereas 
the anterior region is confluent with the cheek region. 
The malar region is not preserved in IGM 7596, but in 
LACM 275/3733 this region is essentially smooth and the 
malar fossa is absent. The overall morphology of the 
preorbital region in D. mexicanus is therefore similar to 
that demonstrated in D. interpolatus and D. leidyanus. 

The late Pliocene (Blancan) species of primitive 
Equus have been given many names, and there currently 
is not universal agreement as to their nomenclature and 

distribution. The resolution of the nomenclature of 
"primitive Equus" species is outside the intended scope 
of this paper. We therefore refer to it here as primitive 
Equus, but note that it embodies the concept of, or is 
close to, Equus simplicidens, as it was first known from 
the Texas panhandle. Primitive Equus, which has been 
referred variously to E. simplicidens, E. shoshonensis 
Gidley 1930, andPlesippus idahoensis (Merriam) 1918 
(Gazin 1936; Skinner and Hibbard 1972; Repenning et 
al. 1995), is also known from the extensive collection of 
skeletons from the Horse Quarries at Hagerman, Idaho. 
One skull of E. simplicidens, AMNH 22077, from 
Crawfish Draw, Mt. Blanco, is the type locality for this 
species. This nearly complete skull preserves the 
preorbital region, including the DPOF and malar fossa. 
In this specimen the malar fossa is weakly developed as 
a shallow depression on the maxillary and malar bones. 
The DPOF has well-developed dorsal and posterior 
margins, whereas the ventral and anterior parts of this 
structure are confluent with the adjacent cheek region. 
By far the most comprehensive sample known for 
primitive Equus is from Hagerman, as exemplified by 
four adult specimens in the AMNH/F:AM (32555 
[female], 32553 [male], 32551 [female], and 32556 
[male]). In all these specimens the configuration of the 
facial fossae is similar. The malar fossa is usually poorly 
developed and represented by a small pit on the maxillary 
bone just anterior to the maxillary/zygomatic suture. The 
DPOF is present, although it seems slightly less well 
defined than in Dinohippus. In particular, the ventral 
and anterior margins are not distinct, thus these parts of 
the fossa are confluent with the adjoining part of the 
cheek region. In the Hagerman sample, the mean length 
of the DPOF is 95 mm and the mean height is 36 mm 
(see Eisenmann et al. 1988, p. 11, for exact location of 
measurements, i.e., B33 and B35). The relative size of 
the DPOF in primitive Equus is therefore similar to that 
of Dinohippus mexicanus. Hence the morphological 
changes in the development of the DPOF from 
Dinohippus to Equus involve the weakening of the 
anterior and ventral margins, and not a relative reduction 
in size. 

Several important characters related to size and 
dental morphology distinguish Dinohippus mexican us 
from primitive Equus, such as E. simplicidens. Primitive 
Equus is a larger horse, with a mean basilar skull1ength 
of 540 mm (N = 4; specimens AMNH 20077,32551, 
and F:AM 32553,32555), whereas D. mexicanus (IGM 
7596) has a basilar length of 430 mm. Although there is 



o 
5 

1
0

 
_

_
 

""!
!' 

_ 
J 

_.
n 

8 
D

 /
~
~
~
'
t
'
~
-

/;
/~

r?
?'

. 
r 

.,n
\l~

~ 
.;:

$':
:/1

##
,';

';'"
 

;
#

P
 

--
--

--
--

--
-.. -

... ,
 

,-,-
--

I 

t/
 

t' I I 

F
ig

ur
e 

7.
 C

om
pa

ri
so

n 
of

la
te

ra
l v

ie
w

s 
o

f c
ra

ni
a 

of
 (A

) 
D

in
oh

ip
pu

s 
le

id
ya

nu
s,

 A
M

N
H

 1
72

24
, h

ol
ot

yp
e,

 m
od

if
ie

d 
fr

om
 O

sb
or

n 
(1

91
8)

, (
B

) 
D

in
oh

ip
pu

s 
m

ex
ic

an
us

, 
L

A
C

M
-C

IT
 2

75
/3

72
3,

 f
ro

m
 Y

ep
6m

er
a 

(f
ro

m
 M

ac
F

ad
de

n 
19

84
),

 (
C

) 
D

in
oh

ip
pu

s 
m

ex
ic

an
us

, 
IG

M
 7

59
6 

(=
 U

F
 2

06
86

1,
 c

as
t)

, 
fr

om
 

R
an

ch
o 

E
l 

O
co

te
, 

an
d 

(D
) 

E
qu

us
 s

im
pl

ic
id

en
s,

 F
:A

M
 3

25
55

, 
fr

om
 H

ag
er

m
an

, 
Id

ah
o.

 

.....
. 

-..
..l 
~
 

to
 i o ....,
 

So
 

(0
 i >
 ~ C

Il ~ o 'T
I Z
 ~ ~ eJ C

Il ""' o ~ ~ -I>
­

w
 :9
 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 175 

some overlap of individual specimens (Fig. 5A), the MI 
and M2 APL and TRN of primitive Equus are statistically 
larger than in D. mexicanus (Table 1). Primitive Equus 
also has significantly higher crowns, with juvenile, little 
worn M1 or M2s having a mean MSTHT of 90.3 mm, 
whereas D. mexicanushas the corresponding mean 
MSTHT of 69.0 mm (and D. interpolatus also has a 
mean M1 or M2 MSTHT of 69.0 mm; Table 1, Fig. 
5B). The upper cheek tooth crowns are most highly 
curved in D. interpolatus (mean CURV = 60 mm), 
moderately curved in D. mexicanus (mean CURV = 75 
mm), and less curved in primitive Equus (mean CURV 
= 115.5 mm; Table 1). This represents a morphocline 
from the more primitive, highly curved crowns found in 
Dinohippus during the early late Hemphillian on the one 
hand to the relatively straight crowns found in primitive 
Equus during the Blancan on the other hand. 

The protocone also shows a morpho cline from 
Dinohippus to Equus (also see Kelly 1998). In relatively 
primitive mono dactyl horses (pliohippines and 
dinohippines, i.e., excluding Equus), the protocone is 
small and connected to the protoloph (Fig. 8A). In 
"typical" or middle wear of D. mexicanus, the protocone 
is characteristically elongated posterior to the connection 
to the protoloph and has the advanced "wooden-shoe" 
shape in which the lingual margin of the protocone is 
concave (Figs. 4, 8B; also see Dalquest 1988). In more 
advanced wear this wooden-shoe shape is weakened and 
the lingual margin of the protocone is anteroposteriorly 
straight. In both primitive and advanced Equus, the 
anterior margin of the protocone is greatly expanded 
towards the anterior of the tooth (Fig. 8C). In the lower 
cheek teeth the metaconids and metastylids of D. 
mexicanus are typically rounded, whereas in primitive 
Equus they typically have angular borders (also see 
discussion below). 

Origin of Equus.-The fossil evidence documenting the 
origin of primitive Equus during the Pliocene has both 
fascinated and perplexed paleontologists since the second 
half of the nineteenth century. Marsh (1879) not only 
produced his orthogenetically arranged chart of equid 
morphological and presumed evolutionary stages, but 
also recognized that Pliohippus was a "near ally of the 
modem horse" (p. 504). Gidley (1907) revised the 
Miocene and Pliocene Equidae from North America and 
allocated Cope's Equus simplicidens to Pliohippus 
because he believed that this species had more of an 
overall resemblance to that genus than to Equus. Based 
on some then-recently excavated equid skeletal material 

from Mt. Blanco, Texas, Matthew (1924b) proposed the 
generic name Plesippus, and also assigned Cope's 
material of Pliohippus simplicidens from the same 
locality to this new genus. This Pliocene equid genus 
was purported to be morphologically similar to (e.g., in 
tooth curvature), although more advanced than, 
Pliohippus, but also shared similarities (e.g., the great 
reduction in the facial fossa) with Equus sensu stricto. 
Matthew (1926) arranged all North American fossil 
Equidae into nine morphological levels, or grades, starting 
with Hyracotherium ("Eohippus") and ending with 
Equus. As evidenced from the resulting phylogenetic chart 
(Matthew 1926, p. 167), Plesippus is depicted as having 
been descended anagenetic ally (i.e., with no temporal 
overlap) from P liohippus in the early Pliocene, and Equus 
from P lesippus at the beginning of the Pleistocene. 

Stirton (1940, 1942) did not follow Matthew's 
designation of Plesippus as a valid genus. Rather, he 
considered it a subgenus within Equus. Interestingly, in 
these same papers, Stirton mentions the possibility that 
primitive Equus was descended polyphyletically from two 
or more species of Pliohippus. Stirton (1940, p. 194) 
noted: "More careful work, however, needs to be done to 
trace the descent of the different species from the Lower 
to Middle Pliocene forms." Based on several characters 
of the upper and lower cheek tooth morphology, Dalquest 
(1978, 1988) and Bennett (1980; reproduced as Fig. 9) 
also supported the notion of a polyphyletic origin for the 
genus Equus from Astrohippus on the one hand (giving 
rise to Asinus) and Dinohippus ("Pliohippus") on the 
other hand (giving rise to zebras and caballines). 

In a significant departure from the accepted dogma 
of the time, Quinn (1955) hypothesized that living equids 
are represented by three extant genera, Hippotigris 
(zebras), Asinus (asses), and Equus sensu stricto (horses), 
and that these were descended from a common ancestor, 
Eoequus wilsoni Quinn 1955, from the middle Miocene 
Hemingfordian Cold Spring Fauna of the Texas Gulf 
Coastal Plain. This radial vertical taxonomy has not met 
with subsequent acceptance. A legacy of Quinn's (1955) 
work, however, was his creation of the new genus 
Dinohippus for pliohippine horses lacking the complex 
facial fossa seen in Pliohippus. Interestingly, and also at 
odds with subsequent studies, Quinn (1955, Fig. 3) 
indicated that Dinohippus became extinct in the early 
Hemphillian and is not involved in the ancestry of Equus. 

Skinner (in Skinner and Hibbard 1972) allocated the 
Pliocene species simplicidens to Equus, and included it 
in the subgenus Dolichohippus, which has otherwise been 
used to denote Grevy's zebra Equus (Dolichohippus) 



176 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

A B c 

Dinohippus 
interpo/atus 

Din ohipp us 
mexicanus 

Equus 
sp. 

Figure 8. Comparison of protocone shapes and hypsodonty (taken from Dalquest 1988, Fig. 4; scale 
not indicated, but presumed to be at, or close to, original scale). Occlusal (top) and lateral (bottom) 
views of (A) Dinohippus interpolatus from the Coffee Ranch (= Frick Miami) L. F., late Hemphillian 
of Texas, (B) Dinohippus mexican us from the Rancho El Ocote L. F., latest Hemphillian of Guanajuato, 
Mexico, and (C) Equus sp. from the Cedazo L. F., Pleistocene of Aguascalientes, Mexico. 

grevyi. Skinner's belief was that the Blancan Equus 
species E. simplicidens was widespread throughout North 
America, including Mt. Blanco (the type locality), 
Hagerman, Idaho (previously referred to as E. 
shoshoensis by Gazin 1936), Kansas, and numerous 
localities in Nebraska. 

Bennett (1980) was the first worker to produce a 
cladogram of the interrelationships of Equus and its 
closest sister taxon Dinohippus. In the pre-computer 
days, this cladogram was "hand drawn," and therefore 
did not benefit from the rigorous phylogenetic character 
analysis that has developed since her study. In this 
scheme, the species D. mexicanus was the closest sister­
species of Equus "shoshoensis." MacFadden (1984) 

discussed the interrelationships among Dinohippus, 
Astrohippus, and Equus. In contrast to what previous 
workers had said about the polyphyletic origins of Equus, 
MacFadden (1984) indicated that, based on the very 
complex facial fossa morphology, Astrohippus is not 
closely related to Equus. If this were the case, as has 
been previously asserted based on dental characters, then 
Astrohippus with its complex DPOF and malar fossa, 
would have to undergo a considerable morphological 
transformation in order to be closely related to the origin 
of Equus.MacFadden (1984; also 1986) states that 
Dinohippus is close to Equus simplicidens, in particular 
the advanced species D. mexican us. Azzaroli (1988) 
analyzed the morphology oflate Hemphillian monodactyl 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 177 

horses from Coffee Ranch. Although we do not agree 
with the designation of a new species, Pliohippus bakeri, 
from this important locality (also see Kelly 1998), several 
conclusions of Azzaroli (1988) are relevant here, 
including: (1) the taxonomic importance of the facial pits 
in understanding the interrelationships of late Cenozoic 
equine horses, and (2) the close relationship between 
Dinohippus (including D. leidyanus [= interpolatus] and 
D. mexicanus) and Equus. Although the possible 
polyphyletic origin of Equus is still arguable, the current 
consensus is that Dinohippus, in particular D. mexicanus, 
is the closest outgroup to some segment of primitive 
Equus that existed in North America during the Blancan. 

Within the past decade, several phylogenetic studies 
that include the interrelationships of Dinohippus and 
Equus have been presented for Neogene equids from 
North America (Evander 1989; Hulbert 1989; Prado and 
Alberdi 1996; Hulbert 1996; Kelley 1998). Of these, the 
study by Kelly (1998) presents the most comprehensive 
cladistic analysis using 40 cranial, dental, and post­
cranial characters for the Tribe Equini, including the 
species of Dinohippus, i.e., D. leardi, D. interpolatus, 
D. leidyanus, D. mexicanus, and Equus simplicidens. 
The synapomorphies that Kelly (1998) uses to justify 

his cladogram (Fig. 10) are mostly confirmed, or further 
demonstrated, by the cranium of D. mexican us IGM 7596 
described here. In particular, the depth of the nasal notch 
(character/state 1.2), configuration of the DPOF (10.2, 
5.0, 9.0, 6.0), deep hypoconal groove (26.0), oval 
protocone (18.2), and protocone never connected to 
hypocone (21.0) are all characters that are seen in the 
Dinohippus species morphocline, includingD. mexicanus 
from Rancho EI Ocote. 

While we are in general agreement with the obser­
vations presented by Kelly (1998), the development of 
the malar fossa deserves some comment here. Kelly 
(1998) indicates that the malar fossa is present and well 
separated from the DPOF (8.1) in D. leardi and D. 
leidyanus, but this feature is absent in D. leidyanus and 
D. mexicanus (8.0). Studies ofthe popUlations referred 
toD. interpolatus andD.leidyanus attheAMNHindicate 
that there may be more variation in the development of 
the malar fossa and its separation from the DPOF than 
is coded in Kelly's (1998) cladistic analysis. Other than 
this, Kelly's (1998) study presents a solid basis for 
understanding the morphological and inferred phylogetic 
tranSformations seen in the species of Dinohippus. 

Kelly (1998) also discusses the synapomorphies that 

Table 3. Synapomorphies used to support nodes in cladogram presented by Kelly (1998, Fig. 10 here) for Neogene 
equine horses from North America and the character states demonstrated in D. mexicanus from Rancho EI Ocote. 

Node Ocote D. mexicanus 

Dinohippus-Equus clade (Node 8) 
Nasal notch dorsal to posterior half of P2 (character 1, state 2) 
Preorbital bar long (character 10, state 2) 
TRL> 160 mm (character 37, state 4) 

Dinohippus interpolatus (Node 9) 

1.2 
10.2 

* 

Shallow DPOF (character 5, state 0) 5.0 
DPOF posterior margin without rim, no pocket (character 9, state 0) 9.0 
Hypoconal groove open to base of crown (character 26, state 0) 26.0 
Presence of intertubercular crest (INT) on humerus (character 40, state 1) ? 

Dinohippus leidyanus (Node 10) 
Dorsal margin DPOF rounded (character 6, state 0) 6.0 
Malar fossa absent (character 8, state 0) 8.0 
Protocone never connects to hypocone (character 21, state 0) 21.0 

Dinohippus mexicanus-Equus simplicidens (Node 11) 
Protocone oval (character 18, state 2) 18.2 

*With a mean P2-M3 TRL of 157.9 mm for IGM 7596, this is transitional between character/state 37.3 (TRL between 
126-160 mm) and 37.4 (TRL> 160 mm). 



178 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

a 
z 
-< 
z 
-< 
(J 
z 
-< 
-' 
CD 

w 
Z 
w 
(J 

o 
-' o 
J: 

w 
z 
w 
(J 

0 
I-
~ 
w 
-' 
~ 

Z 
-< 
::; 
-' OinohiClClUS "A" (E ~·Iike) DinOhip-M "Eo (!; zebra · like) 

---- '---DinohiPj:!US~ 
J: 
~ 

~ 
w 
J: 

Figure 9. Orig in of Equus fro m Dinohippus, from Bennett (1980) and reproduced with permission of 
the Society of Systematic Biologists. 

justify the Dinohippus mexicanus-Equus simp/icidens 
(= primitive Equus here) node (hi s 11 ) with 18.2, the 
shared presence of an oval protocone. In addition to this , 
our present study indicates that a decrease in CURY can 
also be used to justify thi s dichotomy. Kelly (1998) does 
not discuss the synapomorphies that separate Equus from 
more primitive Dinohippus, but numerous characters are 
presented in Bennett (1980), Hulbert (1989), and Prado 
and Alberdi (1996) . The results of the present study 
indicate, or further conf irm previous studies, that the 
following sy napomorphi es se parate Equus from 
Dinohippus: 

1. DPOP more poorly defined; 
2. Increased overall size (e.g., as represented by basilar 

length, tooth row length, or upper molar dimensions; 
3. Increased relative hypsodonty; 
4. Reduced CURY; and 
S. More flared protocone; metaconids and metastylids 

wi th angular enamel borders. 
The question ari ses as to the mode of speciation that 

occurred between Dinohippus mexicanus and primitive 
Equus, such as Equus simplicidens. Early workers (e.g., 
Matthew 1926) indicated descent through grades , 
suggesting anagenesis. Dalquest (1988) believed that 
horse evolution in the Hemphillian and early Blancan 

was g rad ual and probabl y accelerated in the latest 
Blancan and Pleistocene. He further asserted that this 
increased rate of evo lution provided an example of 
punctuated eq uilibria. Hu lbert (1996) depicts Equus as 
originating from Dinohippus by anagenesis, i.e., phyletic 
speciation during the P liocene. In fact, recognition and 
calibration of the exact evolutionary transition between 
Dinohippus and primitive Equus has been difficult to 
reso lve beca use of the lack of superposed latest 
Hemphillian/earJy Blancan sites (Lindsay et al. 1984). 
Nevertheless, there are two localities that span this 
transition and potentially document evidence of the mode 
of speciation between D. mexicanus and E. simplicidens. 
One of these was previously reported (Downs and Miller 
1994) and the other is a new locality from central Mexico 
that has transitional morphology represented by isolated 
teeth of advanced equine horses peltai ning to Dinohippus 
and/or Equus. 

Downs and Miller (1994) describe late Cenozoic 
horse from the well-calibrated sequence in the Anza­
Borrego desert of southern California. Although the 
presence of Dinohippus is not surprizing from the late 
Hemphillian localities, they also describe a specimen that 
they tentatively refer to cf. Dinohippus sp. from overlying 
levels. If thi assignment to Dinohippus is cOlTect, then 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexican us and the origin of Equus 179 

this occurrence extends the range of this species well 
into the Blancan, with a local range for this species from 
about 4 to 2.7 million years, indicating temporal overlap 
with the known range of Equus simplicidens (sensu 
Downs and Miller 1994) at Anza-Borrego. 

Newly collected specimens further suggest an 
extension for the genus Dinohippus into the Blancan. 
These come from the previously unreported Jal-Teco 7 
locality from J alisco in central Mexico (Fig. 6), which is 
currently being worked by the IGM. This locality also 
has an occurrence of Pliocene equids that are of relevance 
to an understanding of the origin of Equus. One level, 
Las Gravas, within the continuous J al-Teco 7 sequence, 
which spans late Hemphillian to Rancholabrean, overlies 
an ash dated at 4.8 Ma (unpublished data), contains 

glyptodonts, capybaras, and two types of equids, and is 
interpreted to be early Blancan age. There are several 
isolated equid teeth collected in situ from Las Gravas 
that represent two distinct morphologies. The more 
primitive morphology, which is referable to Dinohippus 
mexicanus, includes shorter crowned lower dentitions 
(Fig. 11, left) with relatively rounded metaconids and 
metastylids. The more advanced morphology, which is 
referable to primitive Equus, e.g., E. cf. simplicidens, 
includes relatively more hypsodont teeth with greatly 
expanded metaconids and metastylids with angular 
borders (Fig. 11, right). There are two possible 
explanations for this very interesting co-occurrence of 
ancestral and descendant species. These two 
morphologies could represent: (1) different individuals 

Increased size, increased hypsodonty, 
reduced CURV, angular and flared 
protocones, metaconids, and metastylids 

Node 10: 6.0,8.0,21.0; reduced CURV 

Figure 10. Portion of strict consensus c1adogram presented by Kelly (1998; Fig. 10) representing Dinohippus and Equus, with 
coded synapomorphous character states justifying each node, as follows: (8) 1.2, nasal notch deep, i.e., dorsal to posterior half 
of P2 or deeper; 10.2, preorbital bar long; 37.4, mean TRL > 160mm; (9) 5.0, shallow DPOF, 9.0; posterior DPOF margin 
without pronounced rim and no pocket; 26.0, hypoconal groove open to near base of crown; 40.1, forearm intertubercular crest 
(INT) only moderately developed; (10) 6.0, DPOF with rounded dorsal margin; 8.0, malar fossa absent; 21.0, protocone never 
connects to hypocone; (11) 18.2, protocone elongate-oval. In addition to these characters, the present study indicates the 
following for Equus relative to D. mexicanus: reduction in CURV; and for Equus (node not analyzed in Kelly 1998): DPOF 
more poorly defined, increased size, increased hypsodonty, further reduced CURV, angular and flared protocones, metaconids, 
and metastylids. 



180 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

o 

A 
I 

I 
I 

~ 
I 
I 
~ 

5 

em 

Figure II. Comparison of occlusal (top) and external (bottom) views of lower molars of Dinohippus mexican us (IGM 7597, 
left m3) with Equus simplicidens (JGM 7598, right m I or m2) from the same locality, Jal-Tec07, Las Gravas, early Blancan, 
Jalisco, Mexico. 

within the same population of one of the two species, or 
(2) two sympatric, sibling species soon after the 
cladogenesis resulting in primitive Equus. In either case, 
this example, along with that from the Anza-Borrego 
desert described above, indicate that primitive Equus 

OIiginated from Dinohippus mexican us via cladogenesis 
and that there was a time during the Blancan in which 
these two sister-species co-existed. This pattern has not 
been previously recognized because of the lack of 
suitable, well calibrated Blancan localities in North 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexican us and the origin of Equus 181 

America or the possible biogeographic ally restricted 
range of sympatry, or both. 

ill addition to the mode of speciation, it is interesting 
to attempt to reconstruct the diets of Dinohippus in order 
to better understand the origin of Equus. Although high­
crowned, based on evidence from carbon isotopes and 
enamel microwear (MacFadden et al. 1999), Dinohippus 
had a variety of diets depending upon the local ecology. 
In western North America, Dinohippus (i.e., D. 
interpolatus and D. leidyanus) was principally a C4 
grazer, whereas in Florida the slightly more advanced, 
but closely related, species D. mexicanus had a mixed 
diet with a considerable proportion of C3 plant foods, 
perhaps representing browse. Ongoing studies 
(MacFadden et al. in progress) of the carbon isotopes of 
Dinohippus mexicanus from relevant late Hemphillian 
localities in Mexico, including Yep6mera and Rancho El 
Ocote, will further resolve geographical patterns of diets 
in the known southern range of this important species. 

SUMMARY AND CONCLUSIONS 
The discovery of the new skull of Dinohippus mexicanus 
described here adds to knowledge of the previously poorly 
represented facial morphology of this important late 
Cenozoic equid species. D. mexicanus is morphologically 
transitional in the distinctive facial morphology and 
dentition with respect to more primitive late Hemphillian 
D. interpolatus and D. leidyanus on the one hand and 
primitive Blancan Equus on the other hand. While classic 
interpretations of the evolution from advanced 
pliohippines (Dinohippus in the more recent literature) 
to primitive Equus mostly have advocated anagenesis, 
the co-occurrence of Dinohippus and Equus in the 
Blancan indicates cladogenesis. Dinohippus mexicanus 
was widespread in the southern u.s. and Mexico during 
the latest Hemphillian and it also ranged into the early 
Blancan during this time, although its latter distribution 
may have been more restricted. Ongoing studies from 
central Mexico will further resolve the calibration and 
paleoecology of the Dinohippus/ Equus transition. 

LITERATURE CITED 
Arellano, A. R. V. 1951. Research on the continental 

Neogene of Mexico. Amer. Jour. Sci. 249:604-616. 
Azzaroli, A. 1988. On the equid genera Dinohippus 

Quinn 1955 andPliohippus Marsh 1874. Boll. Soc. 
Paleontol. Italiana 27 :61-72. 

Bennett, D. K. 1980. Stripes do not a zebra make. Part I: A 
cladistic analysis of Equus. Syst. Zool. 29:272-287. 

Carranza-Castaneda, 0., 1992. Una nueva localidad del 
Henfiliano tardio en la Mesa Centrale de Mexico. 
illst. GeoL, Univ. Nac. Aut6noma Mexico, Revista 
10: 179-196. 

___ -----', and!. Ferrusquia-Villafranca.1978. Nuevas 
investigaciones sobre la fauna Rancho El Ocote, 
Plioceno medio de Guanajuato, Mexico; Informe 
Preliminar. Inst. Geol., Univ. Nac. Aut6noma 
Mexico, Revista 2: 163-166. 

Cope, E. D. 1892. A contribution to the vertebrate 
paleontology of Texas. Proc. Amer. Philos. Soc. 30: 
123-131. 

____ . 1893. A preliminary report on the vertebrate 
paleontology of the Llano Estacado. Fourth Ann. 
Rep. Geol. Survey Texas. pp. 1-136. 

Dalquest, W. W. 1978. Phylogeny of American horses 
of Blancan and Pleistocene age. Acta Zool. Fennica 
15: 191-199. 

____ .1988. Astrohippus and the origin of Blancan 
and Pleistocene horses. Occas. Papers, The Museum 
Texas Tech Univ. 116:1-23. 

____ , and O. Mooser. 1980. Late Hemphillian 
mammals of the Ocote Local Fauna, Guanajuato, 
Mexico. Pearce-Sellards Series, Texas Mem. Mus. 
32:1-25. 

Downs, T. and G. J. Miller. 1994. Late Cenozoic equids 
from the Anza-Borrego Desert of California. Nat. 
Hist. Mus. Los Angeles Co., Contrib. Sci. 440: 1-90. 

Drescher, A. B. 1941. Later Tertiary Equidae from the 
Tejon Hills, California. Carnegie Inst. of Washington 
Pub. 530: 1-23. 

Eisenmann, v., Alberdi, M. T., De Giuli, C., and U. 
Staesche. 1988. Studying fossil horses. In Collected 
Papers after the "New York International Hipparion 
Conference, 1981," eds. M. O. Woodburne andP. Son­
daar. Volume I: Methodology. Leiden, E. 1. Brill. 71 pp. 

Evander, R. L. 1989. Phylogeny of the Family Equidae. 
In The Evolution ofPerissodactyls, D. R. Prothero 
and R. M. Schoch, eds. Pp. 109-127. Oxford Univ. 
Press, New York. 

Gazin, C. L. 1936. A study of the fossil horse remains 
from the upper Pliocene ofIdaho. Proc. U. S. Nat. 
Mus. 83:281-320. 

Gidley, 1. W. 1907. Revision of the Miocene and Pliocene 
Equidae of North America. BulL Amer. Mus. Nat. 
Hist. 23:865-934. 

____ . 1930. A new Pliocene horse from Idaho. 
Jour. Mammal. 16: 52-60. 

Gray, J. E. 1821. On the natural arrangement of 



182 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

vertebrose animals. London Med. Repository Rev. 
15:296-310. 

Hulbert, R. C. 1988. Cormohipparion and Hipparion 
(Mammalia, Perissodactyla, Equidae) from the late 
Neogene of Florida. Bull. FI. State Mus., BioI. Sci. 
33:229-338. 

_____ 1989. Phylogenetic interrelationships of 
North American late Neogene Equinae. In The 
Evolution of Perissodactyls, D. R. Prothero and R. 
M. Schoch, eds. Pp. l76-196. Oxford Univ. Press, 
New York. 

_____ 1990. The taxonomic status of Hipparion 
minus Sellards, 1916 (Mammalia, Equidae). Jour. 
Paleontol. 64:855-856. 

_____ 1992. A checklist of the fossil vertebrates 
of Florida. Papers FI. Paleontol. 6:1-35. 

__________ 1993 . Late Miocene Nannippus 
(Mammalia: Perissodactyla) from Florida, with 
description of the smallest hipparionine horse. Jour. 
Vert. Paleontol. 13:350-366. 

_____ 1996. The ancestry of the horse. In Horses 

through Time, S. L. Olsen, ed. Pp. 11-35. Roberts 
Rinehart Publishers, Boulder, Colorado. 

Kelly, T. S. 1998. New middle Miocene equid crania 
from California and their implications for the 
phylogeny of the Equini. Nat. Hist. Mus. Los Angeles 
Co., Contrib. Sci. 473:1-43. 

Lance, J. F. 1950. Paleontologfa y estratigraffa del 
Plioceno de Yep6mera, Estado de Chihuahua-I" 
parte: Equidos, excepto Neohipparion. lost. Geol., 
Univ. Aut6noma Mexico, Bol. 54: 1-81. 

Lindsay, E . H., Opdyke, N. D., and N. M. Johnson. 1984. 
Blancan-Hemphillian land mammal ages and late 
Cenozoic dispersal events. Ann. Rev. Earth Planet. 
Sci. 12:445-488. 

Linnaeus, C. 1758. Systema Naturae per Regna Tria 
Naturae, Secundum Classes, Ordines, Genrea, 
Species cum Characteribus Differentis Synonymis 
Locis. Edita decima, reformata. Laurenti Salvii, 

Stockholm I, 824 pp. 
MacFadden, B. J. 1984. Astrohippus and Dinohippus 

from the Yep6mera Local Fauna (Hemphill ian , 

Mexico) and implications for the phylogeny of one­
toed horses . Jour. Vert. Paleontol. 4:273-283. 

_____ 1986. Late Hemphillian monodactyl horses 
(Mammalia, Equidae) from the Bone Valley 
Formation of central Florida. Jour. Paleontol. 
60:466-475. 

_____ 1989. Dental character variation in 

paleopopulations and morphospecies offossil horses 
and extant analogs. In The Evolution of 
Perissodactyls. D. R. Prothero and R. M. Schoch 
(eds.), Pp. 128-141. Oxford Univ. Press, New York: 

_____ 1992. Fossil horses: Systematics, Paleo­
biology, and Evolution of the Family Equidae. 
Cambridge Univ. Press, New York. 369 pp. 

____ , Solounias, N., and T. E. Cerling. 1999. 
Ancient diets , ecology, and extinction of 5 million­

year-old horses from Florida. Science 283:824-827. 
Marsh, O. C. 1879. Polydactyle horses, recent and 

extinct. Amer. Jour. Sci. 17:499-505 
Matthew, W. D. 1924a. Third contribution to the Snake 

Creek Fauna. Bull. Amer. Mus. Nat. Hist. 50:59-210. 
_____ 1924b. A new link in the ancestry of the 

horse. Amer. Mus. Novitates 131: 1-2. 
_____ 1926. The evolution of the horse. A record 

and its interpretation. Quart. Rev. BioI. 1: 139-185. 
____ , and R. A. Stirton. 1930. Equidae from the 

Pliocene of Texas. Univ. California Pub., Bull. Dept. 
Geol. Sci. 19:349-396. 

Merriam, J. C. 1918. New mammals from the Idaho 
Formation. Univ. California Pub. , Bull. Dept. Geol. 
Sci. 10: 523-530. 

Mooser, O. 1958. La fauna "Cedazo" del Pleistoceno en 
Aguascalientes. Anales lost. BioI. Mexico 29:409-452. 

_____ 1965. Una nueva especie de equido del 
genero Protohippus del Plioceno medio de la Mesa 
Central de Mexico. Anales Inst. BioI. Mexico 
35:157-158. 

____ 1968. Fossil Equidae from the middle 
Pliocene of the Central Plateau of Mexico. 
Southwestern Naturalist 13: 1-12. 

_____ 1973. Pliocene horses of the Ocote local 
fauna, central plateau of Mexico. Southwestern 
Naturalist 18:257-268. 

Osborn, H. F. 1912. Craniometry of the Equidae. Mem. 
Amer. Mus. Nat. Hist. 3:55-100. 

____ 1918. Equidae of the Oligocene, Miocene, and 

Pliocene of North America. Iconographic type revision. 
Mem. Amer. Mus. Nat. Hist. (n.s.) 2:1-326. 

Owen, R. 1848. Description of teeth and portions of jaws 
of two anthracotheroid quadripeds (Hyopotamus 
vectianus and H . bovinus) discovered by the 
Marchioness of Hastings in the Eocene deposits on 
the N. W. coast ofthe Isle of Wight, with an attempt 
to develop Cuvier's idea of the classification of 
pachyderms by the number of toes. Quart. Jour. Geol. 
Soc. 5:380-383. 



MacFADDEN and CARRANZA-CASTANEDA: Cranium of Dinohippus mexicanus and the origin of Equus 183 

Quinn, J. H. 1955. Miocene Equidae of the Texas Gulf 
Coastal Plain. Univ. Texas Bur. Econ. Geol. 5516:1-
102. 

Prado, J. and M. T. Alberdi. 1996. A cladistic analysis 
of the horses of the Tribe Equini. Palaeontology 
39:663-680. 

Repenning, C. A., Weasma, T. R, and G. R Scott. 1995. 
The early Pliocene (latest Blancan-earliest 
Irvingtonian) Froman Ferry Fauna and history of 
the Glenns Ferry Formation, southwestern Idaho. 
U. S. Geol. Surv. Bull. 2105:86 pp. 

Schultz, G. (ed.) 1977. Guidebook: Field conference on 
late Cenozoic biostratigraphy of the Texas panhandle 
and adjacent Oklahoma, August 4-6, 1977. Kilgore 
Research Center, Spec. Pub. 1, West Texas State 
Univ., Canyon, Texas, 160 pp. 

Sellards, E. H. 1916. Fossil vertebrates from Florida. A 
new Miocene fauna , new Pliocene species, the 
Pleistocene fauna. Florida Geol. Survey, 8th Ann. 
Report, pp. 87-119. 

Skinner, M. F. and F. W. Johnson. 1984. Tertiary 
stratigraphy and the Frick Collection of fossil 
vertebrates from north-central Nebraska. Bull. Amer. 
Mus. Nat. Hist. 178:215-368. 

____ " and C. W. Hibbard. 1972. Early Pleistocene 
pre-glacial and glacial rocks and faunas of north-central 

Nebraska. Bull. Amer .. Mus. Nat. Hist. 148:1-148. 
, and B. J. MacFadden. 1977. Cormo­

hipparion n. gen. (Mammalia, Equidae) from the 
North American Miocene (Barstovian­
Clarendonian). Jour. Paleontol. 51 :912-926. 

_____ , and B. E. Taylor. 1967. A revision of the 
geology and paleontology of the Bijou Hills, South 
Dakota. Amer. Mus. Novitates 2300:1-53. 

Stirton, R. A. 1940. Phylogeny of North American 
Equidae. Univ. California Pub., Bull. Dept. Geol. 
Sci. 25:165-198. 

____ , 1942. Comments on the origin and generic 
status ofEquus. Journal of Paleontology. 16:627-637. 

Tedford, R H., Skinner, M. E, Fields, R W., Rensberger, 
J. M., Whistler, D. P., Galusha, T., Taylor, B. E., 
Macdonald, J. R, and S. D. Webb. 1987. Faunal 
succession and biochronology of the Arikareean 
through Hemphillian interval (late Oligocene through 
earliest Pliocene epochs) in North America. In 
Cenozoic Mammals of North America: 
Geochronology and Biostratigraphy. M. O. 
Woodburne, ed. pp. 153-210, Univ. California Press, 
Berkeley. 

Webb, S. D. 1969. The Burge and Minnechaduza 
Clarendonian mammalian faunas of north-central 
Nebraska. Univ. California Pub. Geol. Sci. 78:1-191. 



184 BULLETIN of the FLORIDA MUSEUM OF NATURAL HISTORY Vol. 43 (5) 

APPENDIX 
Practical method for determining the radius 

of tooth curvature 

As is now well established, the evolution of transverse 
curvature of upper cheek teeth (CURV; Fig. 1) is of 
principal importance in distinguishing late Cenozoic 
horses and their phylogenetic interrelationships, in this 
case, the transition from D. interpolatus (and D. 
leidyanus), to D. mexican us, to primitive Equus. As 
described by Skinner and Taylor (1967), a practical 
method for measuring this character was developed with 
a characteristically simple, decidedly "low-tech" method 
by M.P. and S.M. Skinner of the AMNH. Oftentimes, 
simpler is better and intuitively obvious to the practical­
minded, and such is the case here. The Skinners produced 
both: (1) a series of stiff cards (made of oak-tag), each 
having one curve (or in some cases two or three curves 

for the smaller curvatures) cut out with radii of curvatures 
varying from 10 to 310 mm (Fig. AI, left), and (2) a 
glass plate with the equivalent (to those of the cards) 
curves defined by increasing radii of curvature (Fig. AI, 
right). In either ofthese cases, the researcher can take a 
particular tooth and slide it up or down along the glass 
plate, or determine the best fit from the cards, so that the 
curvature can be determined. Also, the oak-tag cards can 
be used for teeth still in maxilla and crania. The method 
works wonderfully, and produces measured data that can 
be quantified and statistically analyzed, as was done 
above (Table 2). Both the "prototypes" described here 
are available for use on the second "Horse" Floor in the 
Frick Collection at the AMNH. 



MacFADDEN and CARRANZA-CASTANEDA: Craniulll of IJiIlO/ripl}/{1 //Iexicanll .1 and the origin of Eqllll\ 185 

-

~6' 

, 0 

.'? 
;c 

.. -
l·J 

~ 

....... 

__ em 

Figure A I. Card cutouts (left) and glass plate (right) used to measure radius of curvature (CURY) of the upper eheek teeth of 
fossil horses. 



The BULLETIN OF THE FLORIDA MUSEUM OF NATURAL HISTORY publishes original biological research. Manu­
scripts dealing with natural history or systematic problems involving the southeastern United States or the neotropics are 
especially welcome. Submitted papers should be of medium length, ca. 10,000-60,000 words. Authors should include the names 
of three suggested reviewers. The BULLETIN receives manuscripts in confidence and protects the confidentiality of the content. 

The BULLETIN is distributed worldwide through institutional exchanges and standing orders. Fifty free copies are 
sent to the first author who may order additional reprints at cost (see below) . 

. ,PREPARATION OF MANUSCRIPT 
Format-The BULLETIN encourages the following format. Contributors also should consult recent numbers of the BULLETIN 
for style and format, as well as the book Scientific Style and Format, CBE Style Manual for Authors, Editors, and Publishers, 
6th Edition, 1994 (published by the Council of Biology Editors). Authors should submit three hard copies of the manuscript 
with complete tables, table legends, figures (one set original, two photocopies), and figure legends, plus a floppy disc or CD 
compatible with Word for Windows. Manuscripts must be typed, one side only, on 8-1/2 x 11" (A4) white bond paper, and 
double-spaced throughout. All pages must be numbered consecutively, including references, figure legends, footnotes, and 
tables. All margins should be at least 1 inch (25 rom) wide, and not justified on the right margin. Manuscripts not submitted 
in the format will be returned to 'the authors. 

In most cases the parts of the manuscript should be as follows. 
a. Title page-A separate sheet with the title, name(s) and addressees) of author(s), and a suggested running title. 
b. Abstract-A separate sheet with 200 words or less summarizing the paper, and listing five boldfaced key words. 
c. Conventions-Introduction, material and methods, results, discussion, and acknowledgments. Taxonomic reviews: the 
first mention in the text of a binomial species name should include the taxonomic authority and the year of publication [e.g., 
Notogillia wetherbyi (Dall, 1865)]. Thereafter the full generic name and species epithet must be written out each time the 
name first appears in a paragraph. The generic name may be abbreviated in the remainder of that paragraph as follows: N. 
wetherbyi. The reference need not be cited when author and date are given only as authority for a taxonomic name. Use 
italics for species names and metric units for measurements. A figure is cited in the text as "Fig." (e.g., Fig. 3). Authors 
should cite figures and tables in the text. 
d. Literature cited-References in the text should give the name of the author(s) followed by the date of pUblication: for one 
author (Smith, 1999), for two authors (Schultz and Whitacre, 1999), and for more than two (Britt et al., 1983). Only papers 
cited in the text may be included. All authors' last names are given in full. Initials are used for first and middle names. Each 
citation must·be complete, with all journal titles unabbreviated, and in the following ferms: 
Periodicals: 

Eisenberg, J. F. and J. R. Polisar. 1999. The mammal species of north-central Venezuela. Bulletin ofthe Florida Museum 
of Natural History, 42 (3): 115-160. 

Books: 
Lundell, C. L. 1937. The Vegetation of Peten. Carnegie Institute of Washington, Publication 478. Washington, D.C. 

Composite Works: 
Gardner, A. L. 1993. Didelphimorpha. pp. 15-24 in D. Wilson and D. M. Reeder, eds. Mammal Species of the World. 

Washington, D.C.: Smithsonian Institution Press, 1206 pp. 
e. Legends-Legends should be typed on a separate sheet for each table and each set of figures. Abbreviations used in the 
figures should be included in the figure legends. 
f. Tables-Each table should be on a separate sheet. Tables must be numbered with Arabic numerals. Each table should be 
headed by a brief legend. Avoid vertical rules. 
g. Figures-All illustrations are referred to as figures and should be numbered consecutively in arabic numbers. Authors 
should submit one set of original figures, plus clear photocopies for reviewer's copies of submitted manuscript. All figures 
must be ready for publication and comply with the following standards: 

Photographs and halftone reproductions must be on glossy high-contrast paper, grouped as appropriate, and in 
focus with good contrast. Any not meeting these criteria:will be returned or be reshot for higher contrast at author expense. 
A scale bar may be used in the photograph; otherwise, the figure legend should give the size. If the background of photographs 
(especially those of specimens) is not desired, amberlith or other means may be used to mask the background. 

Text figures should be printed or drawn in black and completely lettered. Drawings (unless computer generated) 
should be made with dense black waterproof ink on quality paper or illustration board. All lettering must be sans serif in 
medium weight (bold for small type that prints white on black), using cutout, dry transfer, or lettering guide letters. Lettering 
must be not less than 2 rom high or greater than 5 mmhigh after reduction for publication. 

Maximum size for figures is .7x9" (18 x 22.5 cm). Figures should be approximately typepage width (7"; 
18cm) or column width (3-3/8"; 8.5 cm), and should comply with page format. Figures should be labeled on the back to 
indicate the top of the figure and to identify author's name, manuscript title, and figure number. 
Proofs-Page proofs will be sent to the first author to check for printer's errors and returned to the Editor promptly. 
Page charges-Publication necessitates page charges to the first author, which are determined after consultation with the first 
author upon final acceptance of the manuscript. The first author is responsible for any charges incurred for alterations made 
by him on galley or page proofs, other than corrected printer's errors. The MUSEUM will send an invoice to the first author 
for all charges upon completion of publication. Color illustrations will be published at extra cost to the author(s). 
Reprints-Order forms are sent to the first author with the page proofs. The first author is responsible for placing all orders.