Systematics and dietary evaluation of a fossil
equid from South Africa
Tamara A. Franz-Odendaal , Thomas M. Kaiser and Raymond L. Bernor
Hipparionine history and South African localities
The Langebaanweg ‘E’ quarry locality consists essentially of
two fossiliferous deposits, which form part of the Varswater
Formation: the Quartzose Sand Member (QSM) and the overly-
ing Pelletal Phosphate Member (PPM). The QSM is a floodplain
and salt marsh deposit and PPM is a river channel deposit
(Fig. 1). Two river channels, beds 3aS and 3aN, were identified;
however, the temporal relationship between them is not clear.
Bed 3aS appears to have been the earlier river channel and may
contain some fossils reworked from the underlying QSM.1,2
The Varswater Formation has yielded a large number of terres-
trial fossils. It was deposited during an early Pliocene transgres-
sion1,2 between 4.2 and 5.2 million years ago (Myr), which was
also responsible for reflooding the Mediterranean basin and
resulting in a regressive ice-phase in Antarctica.3The age of the
QSM and PPM deposits is therefore estimated to be c. 5 Myr.1The
Langebaanweg ’E‘ quarry fossils are uniquely situated in time
potentially to document some of the changes that were occur-
ring in taxa in southern Africa. One example is that of the
hipparionine horse, which is part of an extended African evolu-
tion of three-toed horses from the early late Miocene into the
Hipparionines (three-toed horses) originated in North America
during the Miocene. During the early Miocene (20–15 Myr),
horses in North America underwent rapid diversification with
the evolution of taxa that were more hypsodont.4After this, the
diversity declined. Primitive, lower-crowned horses declined
first (during the middle Miocene, ~15–10 Myr) probably as a
result of the succession from forested habitats to more open
country woodlands and savannas.5In North America, horse
extinctions during the late Miocene are believed to have been
the result of the expansion of C4ecosystems,6with hipparionines
becoming extinct at the end of the Pliocene. A sharp shift in
stable carbon isotopes obtained from North American horse
teeth in the later Miocene (after c. 7 Myr) indicates that a change
from a C3diet of browse and/or grasses to a C4grass diet took
place at around this time.4
Hipparionines entered the Old World near the base of the late
Miocene, between 11.1 and 10.8 Myr.7–10 In Europe, the Near East
and India, hipparions disappeared from the fossil record during
the early Pleistocene, but persisted until the Middle Pleistocene
in China. Hipparionines coexisted with Equus in Eurasia from
2.6 Myr (Lindsay11 with adjustment to magnetostratigraphic
African hipparionine fossils were first discovered in the
Maghreb of North Africa during the 19th century.13 Since then,
hipparionines have been found to comprise a significant
element of the North African fossil fauna. They have been
recovered from throughout the late Miocene and Pliocene and
perhaps early Pleistocene of North Africa,14,15 the late Miocene
Pleistocene of East Africa (Ethiopia, Uganda, Kenya, Tanzania,
Malawi)16–18 as well as South Africa.19–23 Hipparionines entered
EastAfricaatthebeginning of the late Miocene7–9 and, in contrast
Research Articles South African Journal of Science 99, September/October 2003 453
aDepartment of Zoology, University of Cape Town, Rondebosch, 7700, South Africa.
Present address: Department of Biology, Dalhousie University, Halifax, B3H 4J1,
Canada. E-mail: email@example.com
bErnst-Moritz-Arndt University Greifswald, Institute and Museum of Zoology, Johann-
Sebastian-Bach Str. 11-12, D-17489 Greifswald, Germany.
cCollege of Medicine, Department of Anatomy, Laboratory of Evolutionary Biology,
Howard University, 520 W St. N.W., Washington, D.C. 20059, U.S.A.
*Author for correspondence.
In this study we present the first palaeodietary investigation of
three-toed horses from southern Africa and a systematic revi-
sion. The dietary regime of ‘Eurygnathohippus’ cf. baardi from
Langebaanweg (South Africa) was evaluated using the mesowear
method. This hipparion was originally identified as Hipparion cf.
baardi.However, recent evidence discussed here suggests that it
belongs to the Eurygnathohippus clade. Cluster analysis compar-
ing this equid to other fossil hipparionines from central Europe and
North America indicates that ‘E.’ cf. baardi was a dedicated grazer at
Langebaanweg. Subtle differences in the feeding preference
between populations of ‘E.’ cf. baardi from the two river channel
deposits [Pelletal Phosphate Member (PPM), Beds 3aS and 3aN]
and from the Quartzose Sand Member (QSM) were also found.
‘Eurygnathohippus’ cf. baardi from the two PPM deposits have
similar grazing dietary signals, which are most similar to those of
extant Connochaetes taurinus (wildebeest) and Alcelaphus
buselaphus (hartebeest). ‘Eurygnathohippus’ cf. baardi from the
QSM, a floodplain and salt marsh deposit underlying the river
channel, clusters separately and is more similar to the grazing
bovid, Damaliscus lunatus (topi). This study shows that
‘Eurygnathohippus’ cf. baardi was an eclectic feeder with a strong
grazing signal. The presence of high-crowned dentitions in the
‘Sivalhippus’ Complex, to which ‘Eurygnathohippus’ cf. baardi
belongs, can be considered an exaptation of the group. Our results
also provide some evidence for either differential habitat or habitat
change during the late Miocene/early Pliocene.
Fig. 1. Stratigraphy of ‘E’ quarry Langebaanweg (modified from Hendey1). PPM =
Pelletal Phosphate Member, QSM = Quartzose Sand Member,CSM = Calcareous
to the situation in Europe, they persisted in East and South
Africa well into the later Pleistocene24 (R. Bernor, unpubl. obs.).
Hipparionines may have persisted later in Africa due to their
extreme adaptation to grazing with very high crowned teeth,
and in some species a very broad symphyseal gape for cropping
grass short.10 Their disappearance in Africa is not well correlated,
but it is evident that it is much later than in Eurasia (R. Bernor,
unpubl. obs.). Beginning about 2.4 Myr, Equus entered East
Africa and coexisted with hipparionines until the latter’s extinc-
tion.10,25 The most recent review of African hipparionine evolu-
tion was provided by Bernor and Armour-Chelu,10 but
unfortunately no cladogram is available. We draw upon their re-
view in developing this contribution on the evolution of
hipparionine palaeodiets; however, very little is known about
the dietary adaptations of hipparionines from southern Africa,
or indeed Africa as a whole.
The systematics of African hipparionines remains unresolved
for much of their record. Churcher and Richardson24 recognized
eight species: Hipparion (they did not recognize this as
Hippotherium primigenium), Hipparion albertense, Hipparion baardi,
Hipparion namaquense, Hipparion sitifense, Hipparion turkanense,
Hipparion afarense and Hipparion libycum. None of these taxa is
plausibly referable to Hipparion s.s.7,10 Of these, Hippotherium
primigenium has not been definitively shown to occur in Africa,10
Hipparion sitifense is best considered a nomen dubium because no
type was assigned, and the original material has been lost
(V. Eisenmann, pers. comm.),26 H. albertense and H. libycum are
poorly defined, and H. turkanense and H. afarense have been
definitively placed in the genus Eurygnathohippus, to which
H. albertense, H. baardi, H. namaquense and H. libycum all probably
The late Miocene localities at Lothagam Hill in East Africa
include late Miocene age horizons in the Upper and Lower
Nawata members as well as the Apak member.27 A rare species of
hipparionine from the Lower Nawata appears to be a holdover
from earlier late Miocene horizons. Apart from this rare species,
there is a large, massive-limbed form, Eurygnathohippus
turkanense, and a slenderly built smaller form, Eurygnathohippus
n. sp., both of which appear to be distinct lines of East African
hipparionines.27 The early Pliocene is poorly represented by
hipparionine material, but by the middle Pliocene a large
hipparion with long, strongly built metapodials is known from
the Hadar Formation Denen Dora Member, which Bernor and
Armour-Chelu10 have referred to as Eurygnathohippus hasumense.
This species has also been identified from correlative geological
horizons in the Manonga Valley10. At the end of the Pliocene and
beginning of the Pleistocene, East and South Africa apparently
supported a succession of hipparions that underwent a marked
increase in maximum cheek tooth height, to as high as 90 mm,
exactly double that of the first occurring hipparionine at Sinap,
Turkey, dated 10.7 Myr.28 Bernor and Armour-Chelu10 have
provisionally applied the nomen Eurygnathohippus ‘ethiopicus’ to
this series of hipparions, fully recognizing the need for their
Eurygnathohippus cornelianus identified from Olduvai Bed II as
well as from the South African site of Cornelia (early Pleistocene)
is remarkable for its greatly expanded mandibular symphysis,
hypertrophied and straight, procumbent i1’s and i2’s, and
sharply atrophied mandibular canine. These may be hallmarks
of an extreme grazer.10 Evidently, late-occurring East and South
African hipparionines underwent a strong shift to dedicated
grazing like the wildebeest, and probably depended on nutri-
ent-rich grasses that grow on volcanic or clay-rich soils in East
Eurygnathohippus baardi and E. namaquense are both recorded
from the late Miocene/early Pliocene of the southwestern region
of South Africa.10,24 The hipparionine from Langebaanweg was
first described from Baard’s Quarry (~2 Myr, late Pliocene) as a
new subspecies of H. albertense, Hipparion albertense baardi, by
Boné and Singer22 but was later revised by Hooijer23 as Hipparion
cf. baardi. In his revised species list for Langebaanweg, Hendey2
refers to the hipparionine from ‘E’ quarry as Hipparion cf. baardi.
The early part of Baard’s Quarry may represent fluviatile facies
of the Varswater Formation,30 suggesting a close relationship
with the early hipparionine forms at ‘E’ quarry. Both Hendey31
and Eisenmann,25 however, believe that the hipparionine from
‘E’ quarry is different from Hipparion baardi. According to
Hooijer,23 the maximum crown height of the ‘E’ quarry
hipparion is 75–80 mm, clearly outside the range of Hippotherium
primigenium, which has a maximum crown height of 50 mm.
Bernor and Armour-Chelu10 stated that the nearly complete
skull recorded from ‘E’ quarry is similar in its size, snout and
much of its dental morphology to Eurygnathohippus turkanense.
Comparable features include its large size, unretracted nasals
and apparently strongly reduced preorbital fossae placed dor-
sally as in ‘Sivalhippus’ perimense, and distinct from Hippotherium
primigenium (contra Hooijer;23 Churcher and Richardson;24 see
also Bernor et al.32,33). Of further importance is the apparent
presence of lingual grooving on the mandibular incisors
(Hooijer:23 specimen SAM PQL 20553, plate 6, Fig. 1), a
synapomorphy for more advanced members of the Eurygna-
thohippus clade.10 The maximum crown heights reported by
Hooijer exceed those known from any genuine African late
Miocene Hipparion; indeed, they slightly exceed the highest
recorded crown heights known from the Hadar Formation,
about 3.4–2.9 Myr and are equivalent to those of the Omo
Shungura F Member, and younger.17 Finally, the metapodial
proportions are not as robust as in Eurygnathohippus turkanense,
but at the same time are not as elongate as early middle Pliocene
hipparion from Hadar and the Manonga Valley, Tanzania.10
The mosaic of characters exhibited in the Langebaanweg ‘E’
quarry hipparion assemblage, with a late Miocene East African
horse morphology (particularly skull characters and lack of
ectostylids) occurring alongside a decidedly post-Hadar later
Pliocene morphology (cheek tooth maximum crown height),
suggests that the ‘E’ quarry horse might not be as temporally
homogeneous as has been previously asserted. Alternatively,
likely younger than the Nawata Formation hipparions. The skull
and associated dentition of the Langebaanweg ‘E’ quarry
postcrania, are plausibly derived from a late Miocene/early
Pliocene horse assemblage belonging to the Eurygnathohippus
lineage. Although the absence of ectostylids in the Langebaanweg
hipparion is problematic, it should be noted that Hooijer23
reported permanent mandibular cheek teeth with ectostylids
from the uppermost part of the PPM that are similar to
‘Hipparion’ cf. namaquense. Hendey,30 however, suggests that
H. cf. namaquense may be ancestral to H. albertense. The similarity
between H. cf. namaquense and stratigraphically later material
from ‘the uppermost part of the PPM’ argues against its being
the ancestor of H. albertense, provided that no re-mixing of fossils
from earlier deposits took place. Hendey30 did, however, note
that such mixing might have taken place. The importance of the
presence or absence of ectostylids is relevant as it would bring
them all within the Eurygnathohippus genus in keeping with the
454 South African Journal of Science 99, September/October 2003 Research Articles
taxonomic evidence presented here. This site is considered
part of the uppermost level of Bed 3 (that is, upper PPM).23
Accompanying the ectostylids, specimen SAM PQL 2419723 also
exhibits a more advanced, angular metaconid and metastylid
morphology typical of later Pliocene African hipparions. The
mix of morphologies among Langebaanweg hipparions and the
occurrence of Equus from Baard’s quarry suggest that Lange-
baanweg’s chronology is not well controlled.
We provisionally refer the Langebaanweg ‘E’ quarry skull to
‘Eurygnathohippus’ cf. baardi. Despite its apparent lack of ecto-
stylids, a number of synapomorphies of the face and dentition
unite this horse with other late Miocene-Pleistocene species of
the Eurygnathohippus clade.
In order to obtain a maximum of stratigraphic control, we
analysed ‘Eurygnathohippus’ cf. baardi from the QSM and PPM
separately (Fig. 1). These two deposits are correlated as being c.
5 Myr and together span about 0.5 Myr.1,2
Materials and methods
Twenty-four maxillary teeth from ‘Eurygnathohippus’ cf. baardi
from the late Miocene/early Pliocene Langebaanweg (LBW)
(18°9’E, 32°58’S), South Africa, ‘E’ quarry sample were examined
for mesowear analysis (Table 1). All specimens are housed at the
South African Museum, Cape Town. Nineteen teeth come from
the river channel deposit, Pelletal Phosphate Member, while five
are from the underlying Quartzose Sand Member. Three teeth
from the PPM were not distinguishable into either of the two
bed deposits and are included in the assessment of the PPM
The mesowear method, developed by Fortelius and Solounias,34
treats ungulate tooth mesowear as two variables, namely,
occlusal relief and cusp shape. Occlusal relief (OR) is classified as
high (h) or low (l), depending on how high the cusps rise above
the valley between them. Occlusal relief is used in the analyses
as percentages: % high and % low (Table 2). The second
mesowear variable, cusp shape, includes three scored attributes:
sharp (s), round (r) and blunt (b) according to the degree of facet
development. Cusp shape is used as a percentage and is given in
Table 2 as the three variables % sharp, % round and % blunt. We
use the scorings for the sharpest cusp for our analysis as in
Fortelius and Solounias.34
We applied the extended mesowear method introduced by
Kaiser and Solounias35 in selecting all P4–M3s. In this study, we
analysed two populations of ‘Eurygnathohippus’ cf. baardi from
Langebaanweg — the population QSM from the Quartzose
Sand Member and the population PPM from the Pelletal Phos-
Research Articles South African Journal of Science 99, September/October 2003 455
Table 1. Maxillary cheek teeth attributed to ‘
from Langebaanweg ‘E’ quarry analysed in this investigation.
Spec. ID Tooth Side Stratum OR CS (A) CS (P)
SAM-PQL-41467 txM1 r PPM h s r
SAM-PQL-11982 txP4 r PPM l b b
SAM-PQL-43031 txP4 r PPM h s r
SAM-PQL-52910 txM1 r PPM 3aN h s r
SAM-PQL-11222 txM1 l PPM 3aN l r r
SAM-PQL-50238 txM2 l PPM 3aN h r s
SAM-PQL-50237 txM2 r PPM 3aN h r r
SAM-PQL-69093 txM2 r PPM 3aN l b b
SAM-PQL-46136 txP4 l PPM 3aN h r –
SAM-PQL-52050-B txP4 r PPM 3aN l b r
SAM-PQL-69079 txP4 r PPM 3aN l b b
SAM-PQL-52050-A txP4 r PPM 3aN h r –
SAM-PQL-4768 txM1 l PPM 3aS h r –
SAM-PQL-40619-I txM1 l PPM 3aS l b r
SAM-PQL-55676 txM1 l PPM 3aS h – r
SAM-PQL-40941-A txM2 r PPM 3aS h r r
SAM-PQL-41324 txM3 r PPM 3aS l b b
SAM-PQL-40818-E txP4 r PPM 3aS l b r
SAM-PQL-11222 txP4 r PPM 3aS h r r
SAM-PQL-5353 txM1 r QSM l b –
SAM-PQL-21558 txM2 r QSM l – s
SAM-PQL-24842 txM2 l QSM l r b
SAM-PQL-25027 txM3 r QSM h – r
SAM-PQL-21066 txM3 r QSM l b b
Spec. ID — Specimen number (SAM = South African Museum, Cape Town);Tooth = tooth position (tx = maxillary tooth); Side = side (r = right, l = left); OR = occlusal relief mesowear variable scoring (h = high, l = low); CS = cusp shape
mesowear variable scoring (s = sharp, r = round, b = blunt). (A) = anterior cusp; (P) = posterior cusp.
Table 2. Mesowear variable distribution in the populations of ‘
from Langebaanweg ‘E’ quarry.
l h s r b % l % h % s % r % b
Eubaa (QSM) 5411 228020204040
Eubaa (PPM) 19 8 11 4 11 4 42 58 21 58 21
Eubaa (PPM 3aN) 9452 524456225622
Eubaa (PPM 3aS) 7340 61435708614
Eubaa (QSM/PPM) P2= 1.770, d.f. = 2,
h, s, b
-value = 0.413
Eubaa (PPM 3aN)/PPM 3aS) P2= 1.417, d.f. = 2,
h, s, b
-value = 0.492
Eubaa (QSM) = ‘
from the Quartzose Sand Member,Eubaa (PPM) = entire data set of ‘
from the Pelletal Phosphate Member.Eubaa (PPM 3aN) =
from the Pelletal Phosphate Member,
Bed 3aN only,and Eubaa (PPM 3aS) =
from the Pelletal Phosphate Member,Bed 3aS only;
= number of specimens available. Mesowear variables: l = absolute scorings low,h = absolute scorings high, s = absolute scorings
sharp, r = absolute scorings round, b = absolute scorings blunt; % l = percentage low occlusal relief, % h = percentage high occlusal relief, % s = percentage sharp cusps, % r = percentage rounded cusps, % b = percentage blunt cusps.
Chi-square statistics are also shown for QSM and PPM and for the two deposits 3aN and 3aS.
phate Member. Within the PPM two sub-populations were also
analysed, PPM 3aN and PPM 3aS. These derive from beds 3aN
and 3aS, respectively.
Specimens included in this study include wear stages 2 and 3
only (after Kaiser et al.36): 0 = unerupted, unworn; 1 = just
erupted with early wear, but occlusal surface not yet entirely in
wear; 2 = tooth with entire occlusal face in wear but pre- and
postfossette still fused with anterior or posterior enamel band,
respectively; 3 = pre- or postfossette isolated from anterior or
posterior enamel band, but tooth not yet worn to less than 30%
of maximum crown height). Unworn teeth were also excluded,
as were specimens in very early wear (as in Fortelius and
Solounias34). Very early wear shows the ontogenetic signal of the
‘original’ (unworn) cusp apex morphology, which means that
the equilibrium of attrition and abrasion has not yet been
established. This equilibrium, however, has to be stable before
mesowear can be expected to reflect diet rather than the unworn
Following Fortelius and Solounias,34 we exclude very late wear
by omitting specimens with less than ~20 mm of crown height.
As comparative data for dietary classification, we used 27 extant
species reported by Fortelius and Solounias.34 This data set
includes two fossil equid species in addition to the data sets from
Langebaanweg. We rank this set of fossil taxa within a nested set
of recent ungulate species of known diet in order to evaluate
their relative dietary behaviours.
We used Systat 9.0 and Axum 6 software to compute chi-square
statistics and test for significance of differences observed
between individual data sets. The absolute frequencies of
mesowear variables (low, high, sharp, and round) were tested.
Hierarchical cluster analysis with complete linkage (furthest
neighbours) was applied following the standard hierarchical
amalgamation method of Hartigan.37 The algorithm of Gruvaeus
and Wainer38 was used to order the tree. We analysed the three
mesowear variables (% high, % sharp and % blunt). For this
analysis we used Fortelius and Solounias’34 original data set,
information on central European Miocene hipparionines
published by Kaiser et al.36 and Kaiser,39 and the data presented
here for the first time from Langebaanweg (Fig. 2).
In all samples compared in this analysis (QSM, PPM, PPM 3aN
and PPM 3aS), occlusal relief ranged between 20% high (QSM)
and 58% high (PPM). Cusp shape scorings ranged between 22%
(PPM 3aN) and 0% (PPM 3aS) sharp, 86% (PPM 3aS) and 40%
(QSM) round. Blunt cusps ranged between 40% (QSM) and 14%
(PPM 3aS) (Table 2).
Chi-square analysis of mesowear parameters high, sharp and
round for the populations QSM and PPM indicated low levels of
significance of differences (P= 0.41) (Table 2). Similar low levels
of significance were also found if the comparison was undertaken
between the two sub-populations of individuals deriving from
the different beds of PPM.
Cluster analysis polarized the entire set of 27 ‘typical’ recent
species and the fossil data sets into a pattern with grazers and
browsers at the extremes and with mixed feeders in between
(Fig. 2). The cluster diagrams computed show relations of data
sets by linking them to the same clusters. The closer the data are
statistically, the smaller is the normalized Euclidean distance
(NED) at the branching point. The exact sequence and direction
of species arrangement in the diagram, however, may not be
interpreted as an expression of sequential differences. There are
four main clusters, one containing the most abrasion-dominated
grazers and the QSM population of ‘E.’ cf. baardi, one containing
the remaining grazers including the PPM population and the
two sub-populations PPM 3aS and PPM 3aN, one containing the
mixed feeders and one the browsers. The QSM population of ‘E.’
cf. baardi shares a cluster of fifth order with the topi (Damaliscus
lunatus). The PPM and PPM 3aN populations are closest linked
to the wildebeest (Connochaetes taurinus) and the PPM 3aS
population is classified next to the hartebeest (Alcelaphus
buselaphus). All these species are grazers.
Histograms of the mesowear variables of the different popula-
tions of ‘E.’ cf. baardi compared with the recent species clustering
closest to each population are shown in Fig. 3. The mesowear
variables for Equus burchelli (Burchell’s zebra) are also shown for
comparative purposes. Data for E. burchelli are based on 122
individuals, deriving from a variety of habitats.34 Differences
observed in the extant comparison species are therefore across
the species distribution and related to species feeding adapta-
tions/strategies and not to environmental variation.
On examination of the cusps of ‘E.’ cf. baardi, unusual fissures
456 South African Journal of Science 99, September/October 2003 Research Articles
Fig. 2. Hierarchical cluster diagram based on the reference tooth positions P4–M3
according to the ‘extended’ mesowear method (Kaiser and Solounias35). Circle =
recent browser, rectangle = recent mixed-feeder, triangle = recent grazer. The
mesowear features are percentage high occlusal relief, percentage sharp cusps
and percentage blunt cusps. NED = normalized Euclidean distance
(root-mean-squared difference). Numbers at branching points indicate distance.
Clusters are based on a set of ‘typical’ recent species after Fortelius and
Solounias.34 Fossil species (populations) included are: Eubaa (QSM) =
from the Quartzose Sand Member, LBW (South Af-
rica); Eubaa (PPM) =
from the Pelletal Phosphate Member, LBW;
Eubaa (PPM 3aN) =
from the Pelletal Phosphate Member, Bed 3aN,
LBW; Eubaa (PPM 3aS) =
from the Pelletal Phosphate Member,
Bed 3aS, LBW; cogoo (F&S) =
(after Fortelius &
Solounias34); meins (F&S) =
(after Fortelius and Solounias34);
hPri (DD) =
from the Turolian of Dorn-Dürkheim
(Germany), after Kaiser
36 hPri (HO) =
Vallesian of Höwenegg (Germany) after Kaiser;39 hPsm (DD) =
from the Turolian of Dorn-Dürkheim, after Kaiser
36 hPri (EP) =
noted and are shown in Fig. 4. These fissures can extend into the
adjoining dentine areas and have not been reported in other
hipparionines. At the occlusal surface, the fissures are widened
to form a funnel-like indentation in the ectoloph. The walls of
these indentations are clearly rounded and are therefore likely
to be the result of food abrasion, which widens a pre-existing
crack. This observation suggests that the incident(s) causing the
enamelband to crack occurred in vivo while the teeth were still in
use.In general, the preservation of fossils from Langebaanweg is
very good.2,40 The cause of these cracks is not known but most
likely has to do with an enamel structure of low resistance.
Results from mesowear analyses indicate that all populations
of ‘Eurygnathohippus’ cf. baardi from Langebaanweg were
grazers. From cluster analyses, some subtle differences between
populations and sub-populations are, however, apparent. PPM
and PPM 3aN populations are most similar to Connochaetes
taurinus (wildebeest), a variable grazer, which feeds on short
grass and is dependent on a reliable water supply or moist (green)
grasses.41 The ratio of dicotyledonous/monocotyledonous is
12%/88% for the population studied by Kingdon.41 The PPM 3aS
population is most similar to Alcelaphus buselaphus (hartebeest),
also a variable grazer that drinks frequently41 and has in one
study a dicotyledonous/monocotyledonous ratio of 20%/75% in
its food. The ‘E.’ cf. baardi population from the QSM, a floodplain
and salt marsh deposit, clusters most closely to Damaliscus
lunatus, a species which favours swamp or floodplain areas and
requires green grass as food41 (dicotyledonous/monocotyledonous
ratio = 5%/95%). These results suggest that at least the PPM
and QSM populations were encountering slightly different
dietary regimes, reflecting more readily available grass in the
Since the QSM population is the older of the two, a shift from
floodplain habitats with abundant fresh grasses to a more
extreme environment is likely. Several authors2,40,44,45,48 have
suggested that periodic flooding and droughts were common in
PPM times and this may have upset the local graze habitats.
Based on stable isotopes, the climate regime at Langebaanweg
during deposition of the PPM deposit was winter-wet/summer-
dry, similar to the present climate of the region.40 It has been
suggestedpreviouslythatclimates were changing from subtrop-
ical to temperate.2,42 In addition, the high prevalence of dental
Research Articles South African Journal of Science 99, September/October 2003 457
Fig. 3. Histograms of mesowear variables % low (l), % high (h), % sharp (s), %
round (r) and % blunt (b). a,c,e,g, Histograms of ‘
based on the values given in Table 2. b,d,f,h, Comparative histograms based
on published data by Fortelius and Solounias:34 b, grazer
(topi); d, grazer
(Burchell’s zebra); f, grazer
(wildebeest); h, grazer
Fig. 4. Macroscopic occlusal features (bottom) and ectoloph apical morphology (top) of upper cheek teeth of ‘
Arrows indicate indentations in
the anterior ectoloph, which result from food abrasion widening a crack in the enamel band of the ectoloph. a, Right txM1 SAM-PQL-52910; b, right txM1 SAM-PQL-41467;
c, left txM2 SAM-PQL-50238 (mirrored in figures), d, right txM2 SAM-PQL-50237.
pathologies in the herbivores from the PPM deposits,43,44
together with taphonomic evidence (such as burnt bones),45,46
suggests that living conditions were not favourable and that
fires and droughts were common. Our observation that ‘E.’ cf.
baardi cheek teeth frequently have cracks in the enamel, which
widened with abrasion, may offer additional evidence for physi-
ological stress reflected by a less resistant enamel structure.
The finding here that ‘Eurygnathohippus’cf.baardi appears to
have changed in the dietary regime from feeding in a wetter
habitat (such as floodplain areas) during QSM times to inhabit-
ing a more extreme, drier habitat later, is in agreement with
previous findings, and suggests that ‘E.’ cf. baardi underwent an
adaptation to eating (drier) graze possibly as a result of changing
have maintained its need for moist vegetation or frequent drink-
ing throughout QSM and PPM times. As a member of the
‘Sivalhippus’ Complex,10,47 ‘E.’ cf. baardi was larger and had
higher-crown teeth than all other, more primitive clades of Old
World hipparionines. Increased crown height for the group was
an adaptation for abrasive diets, particularly ones dominated by
C4grasses, which has a record extending back as early as 8 Myr in
Africa (R. Bernor and T. Kaiser, unpubl. obs.). However, in the
Western Cape winter rainfall area, C3vegetation dominates48
anda recent study40 indicates that this C3dominant environment
was also present in the area at QSM and PPM times (c. 5 Myr.)
‘E.’ cf. baardi may therefore have developed increased crown
height elsewhere, in a C4environment, and later migrated into
the C3environment of the LBW area.
‘Eurygnathohippus’ cf. baardi from the QSM is more similar in its
dietary preference to extant Equus than many fossil equids of
similar ages in central Europe, which have been previously
investigated in respect of their dietary preferences36,39 (Fig. 2).
The PPM population (and sub-populations from Beds 3aS and
3aN) demonstrates a palaeodietary preference more similar to
the Miocene (16 Myr) Cormohipparion goorisi (after Fortelius and
Solounias34) from Trinity River Pit (Texas) and the early Vallesian
(c.10.5 Myr) Hippotherium primigenium from Eppelsheim
(Germany). Kaiser39 showed that populations of H. primigenium
from Germany ingested a fairly abrasive diet in the proposed
flooding areas of the Miocene River Rhine, which caused this
population to show a dietary signal in the grazer range.
However, this certainly did not include the tougher C4grasses
found in tropical Old World areas after 7 Myr. The same species,
however, behaved like a browser in the mesophytic forest envi-
ronment of Höwenegg, Germany [hPri (HO)].39 Kaiser’s finding
suggests a high degree of opportunistic feeding in these central
European Vallesian horses, which implies that differential
dietary signals should be considered as reflecting differences in
food availability and access in a given habitat, and do not neces-
sarily indicate a different dietary niche determination. A similar
situation may be operating at Langebaanweg. We therefore con-
sider the observed differences in the dietary regime of ‘E.’ cf.
baardi at Langebaanweg rather to reflect local food availability in
the different habitats sampled by the individual stratigraphic
members and thus to represent the ecological conditions in the
Comparisons with other African hipparionines indicate that
the three-toed horse from Langebaanweg belongs to the
Mesowear analyses on populations of these horses from two
fossiliferous deposits indicate that ‘Eurygnathohippus’ cf. baardi
appears to have been able to change its dietary strategy from
feeding predominantly on moist (green) grasses during QSM
times to consuming drier grasses with frequent drinking in PPM
times. This subtle, apparent shift in diet, however, is more likely
to reflect changes in the local habitat or alterations in seasonality
based on similar findings for other equids by Kaiser and
colleagues,35,39 as well as at Langebaanweg as suggested here.
Changes in the local habitat conditions possibly forced these
horses to adapt to extreme (possibly drier) conditions, eating
grasses with less moisture and drinking frequently. There is
evidence to suggest that the local river flooded periodically,45,48
perhaps forcing these horses to vacate the floodplain areas and
to inhabit drier ground. Recently, Franz-Odendaal43,44 provided
evidence based on stable isotope analyses and dental pathologies
that droughts were common at least during accumulation of
Bed 3aN (PPM) fossils. These changing environmental
conditions must have placed several constraints on the animals,
anditisapparentthatonlythemostadaptable species, including
Eurygnathohippus, were able to survive.
We thank the Deutsche Akademische Austauschdienst for a scholarship awarded
to T.A.F-O. The South African Museum is gratefully acknowledged for providing
permission to loan the Eurygnathohippus specimens from Langebaanweg.
Additional thanks go to Graham Averyand James Brink for comments on an earlier
version of this manuscript. Mrs K. Meyer (University Greifswald) is thanked for her
assistance in casting the fossil teeth. R.L.B. thanks the National Science Foundation
(grant EAR-0125009) for supporting his work on Old World hipparionine horse
Received 27 June. Accepted 17 October 2003.
1. Hendey Q.B. (1981). Geological succession at Langebaanweg, Cape Province,
and global events of the late Tertiary. S. Afr. J. Sci. 77, 33–38.
2. Hendey Q.B. (1984). Southern African late Tertiary vertebrates. In Southern
African Prehistory and Palaeoenvironments, ed. R.G. Klein, pp. 81–106. A.A.
3. Van Couvering J.A. (1976). The terminal Miocene event. Mar. Micropalaeontol. 1,
4. MacFadden B.J., Solounias N. and Cerling T. (1999). Ancient diets, ecology
and extinction of 5-million-year-old horses from Florida. Science 283, 824–827.
5. Webb S.D. (1977). A history of savanna vertebrates in the New World. Part I:
North America. Annu. Rev. Ecol. Syst. 8, 355–380.
6. Ehleringer J.R., Cerling T.E. and Helliker B.R. (1997). C4photosynthesis,
atmosphere CO2, and climate. Oecologia 112, 285–299.
7. Bernor R.L., Koufos G.D., Woodburne M.O. and Fortelius M. (1996). The
evolutionary historyandbiochronologyof European andSouthwestAsianlate
Miocene and Pliocene hipparionine horses. In The Evolution of Western Eurasian
Neogene Mammal Faunas, eds R.L. Bernor, V. Fahlbusch and H-W. Mittmann,
pp. 307–338. Columbia University Press, New York.
8. Swisher C.C. (1996). New 40Ar/39Ar dates and their contribution toward a
revised chronology for the late Miocene non-marine of Europe and West Asia.
In The Evolution of Western Eurasian Neogene Mammal Faunas, eds R.L. Bernor, V.
Fahlbusch and H-W. Mittmann, pp. 64–77. Columbia University Press, New
9. Woodburne M.O., Bernor R.L. and Swisher C.C. (1996). An appraisal of the
stratigraphic and phylogenetic bases for the ‘Hipparion Datum’ in the Old
World.In The Evolution of Western Eurasian Later Neogene Faunas, edsR.L.Bernor,
V. Fahlbusch and H-W. Mittmann, pp. 124–136. Columbia University Press,
10. Ber nor R.L. and M. Armour-Chelu M. (1999). Towards an evolutionary history
of African hipparionine horses. In African Biogeography, Climate and Human
Evolution, eds T.G. Bromage and F. Schrenk, pp. 189–215. Oxford University
11. Lindsay E.H., Opdyke N.D. and Johnson N.M. (1980). Pliocene dispersal of the
horse Equus and late Cenozoic mammalian dispersal events. Nature 287,
12. Cande S.C. and Kent D.V. (1995). Revised calibration of the geomagnetic
polarity timescale for the late Cretaceous and Cenozoic. J. Geophys. Res. 100,
13. Pomel N.A. (1897). Les équidés. Carte Géologie Algérie, Paléontologie Monographie
14. Hopwood A.T. (1926). FossilMammalia. The geology and palaeontology of the
Kaiso Bone Beds, Pt II. Palaeontology. Geological Survey Department, Uganda
Protection, Occasional Papers 2, 13–36.
15. Joleaud L. (1933). Un nouveau genre d’équidé quaternaire de l’Omo
(Abyssinie): Libyhipparion ethiopicum. Bull. Soc. Geol. (France) 3(5), 7–28.
16. Hooijer D.A. and Maglio V.J. (1973). The earliest hipparion south of the Sahara
458 South African Journal of Science 99, September/October 2003 Research Articles
in the late Miocene of Kenya. Proceedings of the Koninklijke Nederlandse Akademie
van Wetenschappen Series B, 76, 311–319.
17. Hooijer D.A. and Maglio V.J. (1974). Hipparions from the late Miocene and
Pliocene of northwestern Kenya. Zoologische verhandelingen 142, 1–80.
18. Eisenmann V. (1976). Nouveaux crânes d’hipparions (Mammalia, Perisso-
dactyla) Plio-Pléistocenes d’Afrique orientale (Ethiopie et Kenya): Hipparion sp.
Hipparion cf. ethiopicum et Hipparion afarense nov. sp. Geobios 9(5), 577–605.
19. Van Hoepen E.C.N. (1930). Fossiele perde van Cornelia, O.V.S. Palaeontologie
Navorsings Nasionale Museum Bloemfontein 2, 13–24.
20. Van Hoepen E.C.N. (1932). Die stamlyn van sebras. Palaeontologie Navorsings
Nasionale Museum Bloemfontein 2, 25–37.
21. Haughton S.H. (1932). The fossil Equidae of South Africa. Ann. S. Afr. Mus. 28,
22. Boné E.L. and Singer R . (1965). Hipparion from Langebaanweg, Cape Province
and a revision of the genus in Africa. Ann. S. Afr. Mus. 48, 273–397.
23. Hooijer D.A. (1976). The late Pliocene Equidae of Langebaanweg, Cape
Province, South Africa. Zoologische verhandelingen 148, 1–39.
24. Churcher C.S. and Richardson M.L. (1978). Equidae. In Evolution of African
Mammals, eds V.J. Maglio and H. B. S. Cooke, pp. 379–422, Harvard University
Press, Cambridge, MA.
25. Eisenmann V. (1983). Family Equidae. In The Fossil Ungulates: Proboscidea,
Perissodactyla, and Suidae, ed. J.M. Harris, pp. 156–214, Koobi Fora Research
Project, vol. 2. Cambridge University Press, Cambridge.
26. Bernor R L. and Scott R.S. (2003). New interpretations of the systematics,
biogeography and palaeoecology of the Sahabi hipparions (latest Miocene),
Libya. Geodiversitas 25, 297–319.
27. Ber nor R L. and J. Harris J. (2003). Systematics and evolutionary biology of the
late Miocene and early Pliocene hipparionine horses from Lothagam, Kenya.
In The Lothagam Site, Kenya, eds M. Leakey and J. Harris, pp. 387–438. Columbia
University Press, New York.
28. Bernor RL.,Scott R.S., ForteliusM.,Kappelman J. andSenS.(2003). Systematics
and evolution of the late Miocene hipparions from Sinap, Turkey. In The Geol-
ogy and Palaeontology of the Miocene Sinap Formation, Turkey, eds M. Fortelius, J.
Kappelman, S. Sen and R.L. Bernor, pp. 220–281. Columbia University Press,
29. Owen-Smith R.N. (1988). Megaherbivores: the influence of very large body
size on ecology. Cambridge University Press, Cambridge.
30. Hendey Q.B. (1974). The Late Cenozoic Carnivora of the south-western Cape
Province. Ann. S. Afr. Mus. 63, 1–369.
31. Hendey Q.B. (1978). The age of the fossils from Baard’s quar ry, Langebaanweg,
South Africa. Ann. S. Afr. Mus. 75, 1–24.
32. Bernor R L., Kovar J., Lipscomb D., Rögl F. and Tobien H. (1988). Systematic,
stratigraphic and palaeoenvironmental contexts of first appearing Hipparion
in the Vienna basin, Austria. J. Vert. Palaeontol. 8, 427–452.
33. Bernor R L., Tobien H. and Woodburne M.O. (1989). Patterns of Old World
hipparionine evolutionary diversification. In European Neogene Mammal
Chronology, eds E. Lindsay, V. Fahlbusch and P. Mein, pp. 263–319. NATO
Advanced Research Workshop, Schloss Reisensberg. Plenum, New York.
34. Fortelius M. and Solounias N. (2000). Functional characterisation of ungulate
molars using abrasion-attrition wear gradient: a new method for reconstruct-
ing palaeodiets. Am. Mus. Novit. 3301, 1–36.
35. Kaiser T. M. and Solounias N. (2003). Extending the tooth mesowear method to
extinct and extant equids. Geodiversitas 25, 321–345.
36. Kaiser T.M., Bernor R.L., Franzen J.L., Scott R. and Solounias N. (2003). New
interpretations of the systematics and palaeoecology of the Dorn-Dürkheim 1
Hipparions (Late Miocene, Turolian Age [MN11]), Rheinhessen, Germany.
Senckenbergiana lethaea 83, 103–133.
37. Hartigan J.A. (1975). Clustering Algorithms. John Wiley, New York.
38. Gruvaeus G. and Weiner H. (1972). Two additions to hierarchical cluster
analysis. Br. J. Math. Stat. Psychol. 25, 200–206.
39. Kaiser T.M. (2003). The dietary regimes of two contemporaneous populations
of Hippotherium primigenium (Perissodactyla, Equidae) from the Vallesian (up-
per Miocene) of southern Germany. Palaeogeogr. Palaeoecol. Palaeoclim. 198,
40. Franz-Odendaal T.A., Lee-Thorp J.A. and Chinsamy A. (2002). New evidence
for the lack of C4grassland expansions during the early Pliocene at
Langebaanweg, South Africa. Palaeobiology 28, 378–388.
41. Kingdon J. (1979). East African Mammals: An atlas of evolution in Africa. University
of Chicago Press, Chicago.
42. Scott L. (1995). Pollen evidence for vegetational and climatic change in
southern Africa during the Neogene and Quaternary. In Palaeoclimate
and Evolution with Emphasis on Human Origins, eds E.S. Vrba, G.H. Denton,
T.C. Partridge and L.H. Burckle, pp. 65–76. Yale University Press, New Haven,
43. Franz-Odendaal T.A., Chinsamy A. and Lee-Thorp J.A. (in press). High
prevalence of enamel hypoplasia in an early Pliocene giraffid (Sivatherium
hendeyi) assemblage. J. Vert. Palaeontol.
44. Franz-Odendaal T.A. (2002). Analysis of dental pathologies in the herbivores of
Langebaanweg and their palaeoenvironmental implications. Ph.D. thesis, University
of Cape Town.
45. Hendey Q.B. (1976). The Pliocene fossil occur rences in ‘E’ quarry, Langebaan-
weg, South Africa. Ann. S. Afr. Mus. 69, 215–247.
46. Klein R.G. (1984). The large mammals of southern Africa: Late Pliocene to
Recent. In Southern African Prehistory and Palaeoenvironments, ed. R.G. Klein,
pp. 107–148. A.A. Balkema, Rotterdam.
47. Bernor RL.and Hussain S.T. (1985). An assessmentofthe systematic, phylogen-
etic and biogeographic relationships of Siwalik hipparionine horses. J. Vert.
Palaeontol. 5, 32–87.
48. Vogel J.C., Fuls A. and Ellis R.P. (1978). The geographical distribution of Kranz
grasses in southern Africa. S. Afr. J. Sci. 74, 209–214.
49. Klein R.G. (1982). Patterns of ungulate mortality and ungulate mortality
profiles from Langebaanweg (early Pliocene) and Elandsfontein (middle
Pleistocene), South-Western Cape Province, South Africa. Ann. S. Afr. Mus. 90,
Research Articles South African Journal of Science 99, September/October 2003 459
News from the National Research
NRF ratings to be published
The National Research Foundation will
publish the rating categories of individual
researchers, who were successfully rated by
the organization, on its website (www.nrf.ac.
za/evaluation) from 2004. The information to
be made available will include the name of
the researcher, the employing institution, the
individual’s research interests as provided to
the NRF for evaluation purposes, as well as
the rating category (A, B, C, P, Y or L).
Ten years of proton therapy
The 10th anniversary of using protons to
treat cancer patients was celebrated at
iThemba LABS (the former National Accelera-
tor Centre) outside Cape Town in October.
Ten years ago, the centre pioneered this form
and was responsible for introducing various
novel technologies for treating patients in this
way, including a device that permits highly
accurate positioning of the proton beam in
relation to the target tumour.
A dedicated proton beam facility, based on a
230-MeV cyclotron, is currently being planned
aspartoftheproposedMajor Radiation Medi-
cine Centre to augment the present service.
New head of SAAO
Philip Charles has been appointed managing
director of the South African Astronomical
Observatory on a 5-year contract starting in
the new year. Professor Charles is currently
with the School of Physics and Astronomy at
the University of Southampton in the U.K.
Royal Society/NRF Programme to
This bilateral agreement, which has been in
operation since the mid-1990s, has been
extended for another five years, from 2004.
The first phase of the programme consisted of
five collaborative projects between scientists
at historically black universities in South
Africa and their counterparts, all senior
researchers, in the U.K. South Africa will
host a meeting around the middle of next
year at which the scientific achievements of
the programme will be presented, its opera-
tion reviewed, and the opportunity offered
to potential new participants to join the
Plans are in hand to publish a suite of papers
in the South African Journal of Science, dedi-
cated to the new science, notably in the fields
of materials modelling, nanotechnology,
proteomics and conservation biology, created
by this particularly successful example of
international collaboration. These articles
should appear towards the end of next year.
Incidentally,the Royal Society, which is one of
Britain’s longest established and principal
publishers of scientific research, continues to
extend its journals cataloque and the ser-
vices it provides via its website (www.pubs.
royalsoc.ac.uk). A growing body of Royal
Society journal content is free online. In
addition, papers published by the society
from 1665 to 1997 are now available from the
JSTOR archive at www.jstor.org.