ChapterPDF Available

A reconstruction of the skull of Megalotragus priscus (Broom, 1909), based on a find from Erfkroon, Modder River, South Africa, with notes on the chronology and biogeography of the Species

Authors:

Abstract and Figures

A reconstruction of the skull of the giant alcelaphine bovid, Megalotragus priscus, is provided based on a brain case and horn cores discovered and excavated at the late Florisian locality of Erfkroon on the Modder River, central Free State Province, South Africa. The sedimentary context of the M. priscus specimen can be correlated with fluvial deposits dated previously by luminescence to the Last Interglacial. Electron Spin Resonance (ESR) analyses of dental specimens from various localities at Erfkroon indicate a terminal Middle Pleistocene and Late Pleistocene age for these deposits. The skull reconstruction of M. priscus is aided by an upper jaw and mandible from the Late Pleistocene locality of Mahemspan. The M. priscus materials from Erfkroon, Mahemspan and other localities allow a re-evaluation of the morphological affinities of the species and it appears to be closer to wildebeest-like alcelaphines (genus Connochaetes) than to hartebeest-like alcelaphines (genera Alcelaphus and Damaliscus). Variability in the fossil horn cores suggests sexual dimorphism and some degree of territorial behaviour. It also suggests geographic variability in the populations of M. priscus in central southern Africa during the later part of the Middle Pleistocene and Late Pleistocene, before its extinction at the end of the Late Pleistocene and early Holocene.
Content may be subject to copyright.
CHAPTER 6
A
reconstruction
of the skull of Megalotr
agus
priscus
(Broom,
1909),
based
on a
find
from
Erfkroon, Modder River,
South Africa, with
notes on the
chronology and
biogeography
of the species
Ja mes S. Brink
Florisbad Quaternary Research Department, National Museum,
Bloemfontein, South Africa
Centre for Environmental Management, University of the Free State.
Bloemfontein, South Africa
C. Britt Bousman
Department of Anthropology, Texas State University, San Mar
cos
,
Texas, USA
GAES, University of the Witwatersrand, Johannesburg, South Africa
Rainer Grün
Research School of Earth Sciences, The Australian National Uni
ver
sity,
Canberra,
A
ustr
alia
ABSTR ACT: A reconstruction of the skull of the giant alcelaphine bovid, Megalotr
agus
priscus, is provided based on a brain case and horn cores discovered and excavated at the late
Florisian locality of Erfkroon on the Modder River, central Free State Province, South Africa.
The sedimentary context of the M. priscus specimen can be correlated with fluv ial deposits
dated previously by luminescence to the Last Interglacial. Electron Sp in Resonance (ESR)
analyses of dental specimens from various localities at Erfkroon indicate a termin al Middle
Pleistocene and Late Pleistocene age for these deposits. The skull reconstruction of M. priscus
is aided by an upper jaw and mandible from the Late Pleistocene locality of Mahemspan.
The M. priscus materials from Erfkroon, M ahemspan and other localities allow a r
e-e
valua-
tion of the morphologi cal affinities of the species and it appears to be closer to wildebeest-lik
e
alcelaphines (genus Connochaetes) than to hartebeest-like alcelaphines (genera Alcelaphus and
Damaliscus). Variability in the fossil horn cores suggests sexual dimorphism and some degr
ee
of territorial behaviour. It also suggests geographic variability in the populations of M. priscus
in central southern Africa during the later p art of the Middle Pleistocene and Late Pleistocene,
before its extinction at the end of the Late Pleistocene and early Holocene.
Brink, J.S., C.B. Bousman & R. Grün. 2015. A reconstruction of the skull of Megalotragus
priscus (Broom, 1909), based on a find from Erfkroon, Modder River, South Africa, with
notes on the chronology and biogeography of the species. Palaeoecology of Africa, 33:71-94.
72 James S. Brink et al.
6.1 INTRODUCTION
The terraces of the Vaal River are well-known for producing abundant vertebrate
fossils that span the Plio-Pleistocene and younger geological ages (Helgren, 1977).
The tributaries of the Vaal River, such as the Modder and Riet Rivers, are equa
ll
y
rich in Quaternary fossil vertebrates and archaeological materials. The Modder and
the Riet Rivers have lower-gradient longitudinal profiles than the Vaal River (Tooth
et al., 2004), flowing over less-resistant bedrock and forming floodplain deposits con-
trolled by igneous barriers, mo stly dolerite. These floodplain deposits become eroded
as the rivers incise, cutting through dolerite barriers, and in the proces s the fossil
contents of the floodplain deposits are exposed in erosional areas, locally known as
dongas. Such erosional areas can be extremely rich sources of vertebrate fos sils and
archaeological materials. On two adjacent farms on the Modder River, Erfkroon
and Orangia” (Figure 1; Churchill et al., 2000; Lyons et al., 2014), ext e ns ive foss il
exposures were dis covered in June 1996 during routine reconnaissance by the Floris-
bad Quaternary Research Depart ment of the Natio nal Museum, Bloemfontein. One
of the first majo r dis coveries on the farm Erfkroon” was a s kull and horn cores of a
giant alcelaphine antelope, Megalotragus priscus (Broom, 1909). This find is unusua
ll
y
complete with a well-preserved braincase attached to the horn cores, which provides
the opportunity of improved understanding of the skull morphology and affin ities
of the species. The aim of this contribution is to provide a description of th is find, its
sedimentary context and age estimates based on Electron Spin Resonance ana
lyses
(ESR). A reconstruction of the skull of M. priscus is proposed, which is based on the
Erfkroon specimen and assisted by a complete mandible and a maxillary fragment
from a pan site in the central Free State Province, known as “Mahe ms p a n” (Figure 1).
The fos sil materials from Mahemspan were discovered and excavated in the late 1930s
and 1940’s by staff of the National Museum (Ho ffman, 1953), and in this contribu-
tion we also pro vide a summary taxonomic list and ESR age estimates for this fossil
occurrence.
Although the general awareness of the importance of the tributaries of the
V
aa
l
River as sources of Quaternary fossils became somewhat diminished in the second
half of the 20th century (Cooke, 1964; Klein, 1984), the early development of the f
ie
ld
of Quaterna ry Palaeontology in South Africa was clos ely linked to the Modder River.
During the 19th and early 20th centuries the Modder River depos its produced the
first recorded dis coveries of Pleistocene fossils in South Africa. In 1839 a giant long-
horned buffalo (Syncerus antiquus) was found in the banks of the Modder River and
reported to the Geological Society of London (Seeley, 1891). A few years later Broom
described the second ma jor find of an extinc t Pleistocene ma mmal from the Modder
River sequence—a giant alcelaphine, which he na med Bubalis priscus (Broom, 1909),
now referred to the genus Megalotragus Van Hoepen (Gentry and Gentry 1978).
The stretch of the Modder River that produced these fossils is situated between the
present-day Krugers Drift Da m and the confluence of the Modder and the Riet
Rivers (Figure 1), which is in the same general area where the Erfkroon fossil sites
are situated. Thus, it is noteworthy that both the Erfkroon find reported here and the
type specimen of M. priscus are from essentially the same stretch of fossil deposits of
the Modder River and they are probably of comparable geological age. Broom’s type
specimen from the Modder River lacks much of the distal parts of the horn cores and
the braincase is not complete. The Erfkroon specimen co mp le men ts the type specimen
and extends our understanding of the species.
A reconstruction of the sku l l of Megalotragus priscus (Bro om, 1909) 73
Figure
1. A map illust rating the geogr aphic po sit ion of the localities with
Megalorragus,
as referred to
in
the text (A) and a plan view the fossil-bearing exposures on the farms Orangia an d Erfkroon (B).
A
B
74 James S. Brink et al.
6.2 MATERIALS AND MET HODS
6.2.1
Mahemspan
In the late 1930s and early 1940’s E. C. Van Hoepen and A.C. Hoffman fro m the
Nat ion al Museum, Bloemfontein, excavated a large faunal sample fro m the lun ette
of Mahemspan (27o 45’ 50” S; 26o 08’ 50E; Figure 1), situated between Hoopstad and
Wesselsbron, Free State Province (Unpublished reports in library of the Nation a
l
Museum). In 1994 one of us (JSB) visited the site and the approximate position of the
excavation was relocated through the help of Mrs. De Villiers, the owner of the farm.
She was present when Van Hoepen and Hoffman conducted their excavations. Fossil
specimens in the s ame state of preservat ion as in the collections at Florisba d Quater-
nary Research Station were found on the surface. Unfo rtunately a trial excavation did
not reveal any in situ materials. The s ite is now covered by aeolian sand, as was the
case before the material was exposed by wind action during the intense droughts of the
1930’s. The site is now part of a ploughed land.
The in situ fossil materials collected by Van Hoepen and Hoffman were found
at the base of the lunette and the original death ass emblage appears to have been
deposited in a mars h-like area in the presence of calcium carbonate-rich ground
w
ater.
This is evident from the ext e n s ive presence of calcrete deposits in and on the fossils.
The fossil matrix, which is still attached to many specimens in the old collection due to
cementation by calcium carbonate, is a pale-brown, partly calcretised sand. This
w
as
used to establish the background radiation for the ESR measurements.
6.2.2
Erfkroon
After the discovery of foss il-bearing deposits on the farms Erfkroon and Orangia
(Figure 1), collectively referred to as Erfkroon”, luminescence dating and sediment
analys is (Churchill et al., 2000; Tooth et al., 2013; Lyons et al., 2014; Table 1) suggest
that the bulk of the deposits on the farm Orangia represents a Late Pleistocene fluv ia
l
terrace. We have named it the “Orangia” Terrace (Figures 2 and 3), and it has both
channel and overbank facies. The channel facies sit unconfor mably on Ecca Bedrock
and make up the lower two horizons of the terrace, the Lower Gravel and the Green
Sand. The Lower Gravel is composed of small-to-large subrounded to rounded sha
le
and calciu m carbonate gravels supported by a dark olive-brown clayey-sand matrix.
Table 1. A comparative list of OSL an d IRSL dates (k a) from the Oran gia Terrace,
as given in previous studies.
OSL IRSL IRSL
OSL
(Lyons et al., 2014)
(Tooth
et al., 2013)
(Tooth
et al., 2013)
et al., 2000)
Brown 0.83
±
0.096.32
±
0.67
Upper Grey
11.3
±
0.9819.5
±
1.04
±
Red
20.0
±
1.1932.2
±
1.74
32
±
2
29
±
2
Lower Grey
40.1
±
2.3043.5
±
3.18
42
±
2
42
±
3
Green Sand
118
±
35
±
Lower Gravel
163
±
7
A reconstruction of the sku ll of Megalotragus priscus (Broom, 1909) 75
Figure
2. A plan view of the Erfkroon sedimentary deposits, showing the approximate lateral
e
xtent
of the Erfkroon and Orangia Terraces, the ESR sampling localities and the find locality
of the Megalotragus priscus specimen. The num be r s of t he E SR lo calities correspond
to the num be rs in Table 4 and in Figure 3.
It is truncated by erosion which forms an unconformity on its upper surface. The
Lower Gravel is classified as a C soil horizon. Above this is the Green Sand,
w
hich
is composed of small Ecca shale pebbles supported in a silty sand to very fine sand
matrix. Iron staining is common and highlights tilted la mella and thin beds. Both of
these layers reflect fluvial channel deposition. The top of the Green Sand also is trun-
cated by erosion to form another unconformity. This is also a C soil horizon.
Above this unconformity are the overbank facies consisting of the Lower Gr
ey
,
Red, Upper Grey, and Brown palaeosols. The Lower Grey sediments are y
e
llowish
brown to strong brown alte rnat ing sand and clay laminae grading up to a clay loam.
Calciu m carbonate nodules increase upwards in the profile in frequency and size. The
Lower Grey grades into the lower part of the Red, which shifts to a yellowish red very
firm sandy loam with coarse moderate subangular blocky structure and with no ca
l-
ciu m carbonate nodules. It is bound by unconformities on the top and botto m.
The upper part of the Red is a yellowish red sandy loam with coarse, weak sub-
angular blocky structure. It grades into the Upper Grey from a strong brown to light
yellowis h brown firm loams to sandy loams. Calcium carbonate nodules are absent in
the bottom of the Upper Grey, slowly increase in density and size upwards in the pro-
file and then decline at the top of the horizon. The Brown palaeosol overlies the Up per
Grey and grades from a strong brown to dark brown loam with declining a mounts of
calcium carbonate nodules upwards in the prof
ile
.
The Orangia Terrace is covered by a thin (
25 cm) aeolian deposit, which corr
es-
ponds with the “Sandy Cap” of Lyons et al. (2014) and we have named it the “Soetdoring
Terrace”. The Soetdoring Terrace has not been described in detail, but pr
e
liminary
76 James S. Brink et al.
Figure
3. A diagram illustrat ing our current int erpret ation of the terrace deposits of the Modder River,
as seen on the farms “Erfkroon” and Orangia” (A). The Erfkroon, Oran gia and Soetdoring terraces
reflect sequential and possibly, marginally overlapp ing periods of deposition (B). We are at present
uncertain of the correct st ratigraphic p o sition of the Shelly Gravel as either a lateral facies
of the Upp er Grey of the Orangia Terrace or the basal horizon of the Soetdoring Terrace.
The numbers and arrows indicate ESR sampling points, as illustrat ed in Figure 2.
A reconstruction of the sku ll of Megalotragus priscus (Broom, 1909) 77
observations indicate that it is a light brown sandy loam with at least two buried soils
that consist of dark brown to strong brown loams. This deposit becomes deeper to
w
ards
the present-day channel of the Modder River, where it appears to be underlain by a
shelly, foss il-rich gravel deposit, referred to as “Sandy gravel with bivalves (Lyons
et al., 2014). The bivalves in this deposit occur articulated and, therefore, they are untrans-
ported and not redeposited and must post-date the depositional event that produced the
gravel horizon. For the present purpose we refer to this horizon informally as the “She
ll
y
Gravel” (Figure 3).We are uncertain whether it forms the basal horizon of the Soetdoring
Terrace or whether it is a lateral facies of the Upper Grey of the Orangia Terrace.
Upstream from Orangia, on the farm of Erfkroon, the Late Pleistocene overbank
component of the sedimentary package is less well developed, but numerous fluvia
l
gravel outcrops proved to be highly fossiliferous. Some of these outcrops are thought to
be older than the Orangia Terrace deposits and thes e are collectively referred to as the
Erfkroon Terrace”. These can easily be distinguished by the lack of carbonate c lasts
so common in the Lower Gravel of the Orangia Terrace. The Erfkroon Terrace deposits
include four lithostratig raphic units. From bottom to top these are the Basal Grav
e
l,
Cross-Bedded Sand, Silty Sand and Pedocalcrete (Figure 3). These deposits have not
been described in detail, but it is important to note that the Basal Gravel complete
l
y
lacks carbonate clasts, which are a significant component of the Orangia Terrace Lo
w
er
Gravel. We believe that the overlying Pedocalcrete horizon of the Erfkroon Terrace is
the source of the carbonate clasts in the Lower Gravel of the Orangia Terrace.
The M. priscus specimen was found emb edded in the Lower Gravel at the contact
with the overlying Green Sand (Figure 4). Its original orientation of deposition
w
as
Figure
4. The exposed sedimentary profile of the outcrop, showing the in situ po sit io n of the skull of
M. priscus (A), a diagram of the profile, illustrating the in situ posit ion of the skull at the cont act between
the Lower Gravel and the Green Sa n d (B) and the skull being exposed during excavation (C).
78 James S. Brink et al.
with the horns down and the snout facing upwards. However, the upper dentitions
and premaxillae were not preserved and their position would have coincided with the
contact between the Lower Gravel and the overlying Green Sand, suggesting that thes e
parts of the skull were eroded away s o me time after deposition or during the deposi-
tion of the Green Sand. The left horn core was broken off and was recovered first.
The rest of the specimen was in situ and was excavated by firs t cutting away some of
the deposit to create a step. The step was used as a platform from which to work and
to allow excavation in plan view. In this way the whole specimen was exposed and con-
solidated. Wooden struts were used to maintain the connection between the horn cor
es
and the braincase. Later, in the fossil laboratory at the Flo ris bad Quaternary Research
Station the preparat ion was completed and the left horn core was glued back.
In 2002 and in 2010 we took ES R s a mples and measurements at various localit ies
in the Orangia Terrace (Figures 2 and 3). In order to establish the gamma dose rate for
the various localities, readings were taken with a portable gamma detector. Excavated
dental samples of known provenance were used for ESR testing, with the exception of
the dental specimens from the Upper Grey palaeosol, which were collected from the
surface. From recent test excavations it is clear that these specimens originally deriv
e
from the Upper Grey palaeosol.
6.3 RES ULTS
6.3.1 M
ahemspan
Comparative taxonomic lists are given in Table 2 for the Las t Interglacial fossil assem-
blages from Erfkroon, the fossil assemblage of Mahemspan and for the Florisbad
Spring assemblage. The Mah ems p an fossil materials are unusually complete. The giant
alcelaphine, M. priscus, is the predominant ele ment in the collection and is represented
by virtually all skeletal ele men ts. The Mahemspan collection of M. priscus represents
probably the largest and mos t complete sample of this species in southern Africa. The
completeness of the material and the predominance of large-bodied taxa in the assem-
blage rais e the question of selective recovery of the material. However, from the car
e
that was taken in the preparat ion of the fossils before excavation and the met iculous
accessioning of the material it is evident that the excavat ors took considerable troub
le
to recover the material as completely as possible and it appears unlikely that recovery
was selective. The presence of sun cracks and porcupine gnawing on specimens, the
absence of hyaena coprolites and the paucity of carnivores in the fossil assembla
ge
argue against the likelihood that it repres ents the contents of ancient hyaena burro
ws
.
This and the fact the bones were originally deposited in ma rshy conditions, may point
to the Mahemspan assemblage representing natural deaths or carnivore kills on the
edge of the pan. This is analogous to the taphonomic reconstruction of the Florisbad
Spring assemblage, which is considered to represent the remains of carnivore hunting
and scavenging around the ancient spring pools (Brink, 1987, 1988; Grün et al., 1996).
Various authors made reference to the Mahe ms p an material and formed opinions
on its geological age (Cooke, 1974; Gentry and Gen t ry, 1978). Van Hoepen (1947) and
Hoff man (1953) used cranial ele ments from this assemblage for taxonomic descrip-
tions. Cooke (1974) considered Mahemspan material to be Florisian in age, but some-
what older than the Floris b ad spring material. However, based on the shape of the
black wildebeest horn cores Gen try and Gen try (1978) suggested an age younger than
the Flo risb ad spring material. The ESR results for an early uranium uptake mo d e
l
(EU) suggest an age of around 12,000 years BP and for a linear uranium uptake mode
l
Table 2. Taxonomic list of Florisian faunas: a comparison bet ween Erfkroon La st Interglacial levels (L/I),
Mahemspan and Florisbad Spring*.
Primates
Florisbad
spring Erfkroon
L/I Mahemspan
Homo helmei 1
Lagomorpha
Lepus sp. 6 4
Roden tia
Hystrix africaeaustr
alis
Pedetes sp. cf. P. capensis
Carnivora
Aonyx capensis
Cynictis penicillata
Gal erella
sanguinea
Atilax paludinosus
Canis mesomelas
Vulpes chama
Lycaon pictus
Crocuta crocuta
Panthera
leo
Perissodactyl a
Equus capensis††
Equus
l
ylei††
Equus quagga subsp.
Ceratotherium simum
Artiodactyla
Hippopotam
us
amphibius
Phacoc
hoerus
africanus/aethiopicus
Taurotragus oryx
Syncerus antiquus††
Kobus leche
Kobus sp.†
Hippotragus sp.†
Damaliscus niro††
Damaliscus pyg
argus
Alcelaphus buselaphus
Connochaetes gnou
Connochaetes taurinus
Megalotragus priscus††
Antidorcas bondi††
Antidorcas marsupialis
Raphicerus campestris
1
8
3
2
1
3
6 2
2
3 1
7
1 1
73 6
61 1
97 2
3
333 2
33 2
24 1
25 1
60 2
4
16
111 11
9 12
8
284 10
30 2
889 34
107 2
4
7
1
80
cf
.
48
4
17
21
28
21
2
11
19
10
24
9
241
1
6
††Ext inct.
†Regionally
e
xtinct.
*Faunal lists (given according to MNI) are modified and adapt ed from Brin k ( 1987, 1994 ; in press).
Table 3. Electron Spin Resonance result s on dental specimens from Mahemspan.
Sample
Dose
De
error
U
(EN)
U
(DE)
TT
U
(SED)
Th
(SED)
K
(SED)
EU
Age
(ka)
EU-
error
LU
Age
(ka)
LU
error
1411
5.5 1.75 0.45
1412 AM
24.9
0.9
1.86
88.2
1230
13.6
0.8
17.6
1
1412 AS1
24.3
0.4
13.3
0.7
17.2
0.8
1412 BM
23.7
0.4
2.04
58
1170
14
0.7
7.6
0.9
1412 BS1
22.2
0.2
13.1
0.6
16.5
0.8
1413 AS1
16.7
0.7
0.89
24.5
700
11.6
0.7
13.3
0.8
1413 BS1
19
0.2
1.08
32.4
730
12.5
0.6
14.7
0.7
1413 CS1
19.5
0.2
1.36
54.8
830
11.4
0.5
14.2
0.7
1414 AS1
18.7
0.2
0.98
26.9
730
12.8
0.6
14.8
0.7
1414 BS1
19.2
0.3
1.33
40.1
700
11.7
0.6
14.2
0.7
1415 AM
16.7
0.3
0.7
20.94
930
12.8
0.6
14.3
0.7
1415 AS1
16.8
0.3
21
12.9
0.6
14.4
0.7
1415 BS1
15.5
0.3
0.6
25.8
830
11.4
0.6
12.9
0.7
1415 CS1
15.6
0.2
0.58
28.4
1130
12.1
0.6
13.6
0.7
(LU) of around 1317,000 years BP (Table 3). These age estimates support the sugges-
tion of Gentry and Gentry (1978) that the Mahemspan assemblage postdates that of
the Florisb ad Spring (Grün et al., 1996).
6.3.2 Erfkroon
In Table 2 the taxonomic list for Erfkroon represents the fossil assemblages from the
Last Interglacial levels. These are the Lower Gravel and Green Sand from the Orang ia
Terrace, as recovered on the farms Erfkroon and Orangia (see Table 1). The ESR a
ge
es timates for the various horizons of the Orangia Terrace are given in Table 4 and
are bas ed on the testing of the dental specimens recovered from the farms Erfkroon
and Orangia. The numbers of the sample batches in Table 4 correspond to the sam-
pling localities given in Figures 2 and 3. The EU age estimates from the Upper Gr
ey
(Orangia Terrace) seem to correspond fairly well with the luminescence estimates, as
referred to above, but there is s ome disagreement between the luminescence and the
ESR estimates from the Green Sand of the Orangia Terrace. The ESR es timates for
the Shelly Gravel suggest a terminal Late Pleistocene age. We are at pres ent uncertain
whether the Shelly Gravel is a lateral facies of the Upper Grey of the Orangia Terrace
or the basal horizon of the Soetdoring Terrace. The ESR estimates accord well with
either stratig raphic interp retation (Figure 3)
The M. priscus skull and horn cores were found in an outcrop which we corr
e
late
with the Lower Gravels and Green Sands of the Orangia Terrace (Figure 4). Based on
the luminescence dating of the Orangia Terrace this would i mp ly a Las t Interglac ia
l
age for the M. priscus outcrop. We were not able to recover dental specimens from
the M. priscus outcrop for ESR testing, but the presence of abundant rolled calcrete
blocks together with clasts of shale gives confidence to the correlation of this deposit
with the Lower Gravel of the Orangia Terrace.
6.3.3 Horns an d braincase of M. priscus
6.3.3.1 The Erfk roon specimen
The Erfkroon s pec imen consists of a braincase and both horn cores (Figu re 5;
Table 5). The left horn core tip is not preserved. The ho rns are dorso-ventra
ll
y
co mp ressed near the bases and in mid -course, but become rounded towards the
las t third of their cou rse. Nea r the base of the horn on the anterior surface there is
a s light swelling, als o present in the Florisbad spec imen FLO 2274. Horn pedice
ls
are fused and overhang the occip ital surface. Horns are bent down sharply and
diverge with a mutual angle o f around 150°. In mid-course they are sub-horizonta
l
before they curve up and forward and there is c lockwise torsion on the right. The
horns have faint transverse ridges near the bases, an indicat ion of nodes on the
horn sheath. The frontals suture appears less ext remely fused than in C. gnou.
There is no postcornual fossa. The braincase appears antero-posteriorly shortened
in lateral view. This is due ma in ly to the reduction in the parietals, which are vis-
ible only in lateral view so that the occipital makes contact with the frontal. The
combined effect of the reduction in the braincase and the posterior projection of
the fused pedicels is that the braincase appears partly hidden beneath the horn
bases. This config u rat ion of the braincase appears to have caused the nuchal crest
to have become inverted to form a concave structure in order to allow sufficient
82 James S. Brink et al.
Table 4. Electron Spin Resonance result s on dental specimens from Erfkroon.
1.
Upper
Grey, Orangia Terrace
1920 AS1 56.7 2 .1 1470 50 450
42 4 50 6
1920 BS1
33
1.1
1440
40
470
3.8
4.5
0.81
744.5
25.2
2.6
29.5
3.5
1923 AS1
45.5
1.1
1130
70
340
1.45
6.5
1.29
695.9
24.9
2.1
32
3
1923 BS1
49
1.3
1110
90
270
27.2
2.3
35
3
1926 AS1
49.7
0.7
1000
90
330
3
8.5
1
873.8
22.4
1.8
31
2
1926 BS1
44.2
1.4
1130
90
160
22
1.8
29.4
2.8
2. Duke 8, Shelly
Gravel, Incertae
se di s
1929 AS1 43.4 1 .3 3.7 99 960 50 50 696.7 18 1.2 25.2 1.6
1929 BS1
48.5
1.3
3.8
115
1090
50
50
19.7
1.2
27.8
1.7
1930A
54.4
0.8
0.88
112
800
40
70
23.6
1.5
32
2
1930B
41.1
0.6
0.51
112
970
70
160
20.3
1.3
27.1
1.7
1931A
40.6
0.5
181
710
50
70
11.5
0.8
17.6
1.1
1931B
43.9
0.6
5.5
168
820
90
60
13.6
0.9
20.4
1.2
1932A
54.7
0.6
10
144
1040
70
300
14.7
1.1
23.2
1.5
1932B
63.3
1
4.5
147
1010
30
190
20.8
1.4
31
1
1932C
41.6
0.5
1.5
169
1000
160
190
17.1
1.1
24.2
1.5
3. Hippo Site,
Green
Sand, Orangia Terrace
1939A 733 17 8 102 1180 90 90
713 158 13 261 19
1939B
825
20
9.1
109
1090
70
70
698
153
12
253
19
1940A
661
13
9.3
91
690
20
20
127
10
209
15
1940B
0
11
90
530
20
20
1941A
0
2
82
2510
290
290
2.6
6.6
1.07
724.8
1941B
0
2.3
82
2620
225
225
1941C
0
2.8
82
2550
290
290
4. W Site,
Green
Sand, Orangia Terrace
1948A 580 13 3.7 65 760 50 50 135 12 201 18
1948B
796
16
13
71
730
20
20
124
11
199
17
1949A
568
10
8
23
900
90
90
135
12
201
18
1949B
522
9
4.5
21
890
110
70
161
14
223
22
1950A
587
11
5
26
730
50
50
164
15
232
22
1950B
407
6
2.3
26
860
100
90
157
14
204
21
5. Duke 13,
Lower Gravel,
Orangia Terrace
1953A
1582
65
36
90
650
35
35
123
13
218
23
1953B
1697
30
27
86
580
40
40
154
15
270
25
1954A
601
10
3.5
72
1020
60
150
183
13
277
18
1954B
613
7
4.8
77
1230
100
180
173
12
268
18
area for the attachment of the neck musc les. Although the spec imen is some
w
hat
damaged in this region, it appears that the nuchal furrow exten d s into a very pro-
nounced supra mastoid crest anteriorly, while posteriorly it lin ks with the petrosa
l
part of the te mpo ral to form a very strong structure for neck muscle attachment.
The occip ital condyles are very wid e. The basioc cipital is short, wide, approaches
being rectangular, has no med ia n ridge and has large anterior tuberos ities (tub er-
cula muscu laria). The ba s iocc ipital and the sphenoid are not on the same plane,
but have an angle of around 140°.
A reconstruction of the sku ll of Megalotragus priscus (Broom, 1909) 83
Figure
5. The comp lete braincase and horn cores of Megalotragus priscus from Erfkroon:
Frontal view (A), right lateral view (B), enlarged right lateral view of the brain case (C)
and a basal view (D). The scale is in centimetr
es
.
The bullae tympanicae are not preserved, but the spaces in which they were situ-
ated are preserved and it can be deduced that they were moderately large and rounded.
The foramina ovalia are quite large. From the remain ing part of the frontal it is
e
vident
that the angle of the braincase to the face is small, approaching 90°. There is enough
of the posterior margin of the orbits preserved to suggest that they would have been
at least moderately projecting.
84 James S. Brink et al.
Table 5. Measurements (in mm) of the Megalotragus priscus skull and horn cores from Erfkroon. The
equivalent measurements given in Von den Dr iesch (1976) for cattle (Bos taurus) are indicated in bracket s.
1 Least frontal breadth (32) 143.5
2
Greatest breadth across the orbits (33)
172.8
+
3
Greatest diameter of the ped ice
l
107.7
4
Least diameter of the pedice
l
83
5
Greatest diameter of the base of the horn core (45)
108.5
6
Least diameter of the base of the horn core (46)
83
7
Greatest diameter of the horn core at mid-section
90.5
8
Least diameter of the horn core at mid-section
65.1
9
Greatest breadth of occipital condyles (26)
144*
10
Greatest breadth of the foramen magnum (28)
40.5*
11
Least breadth between the bases of the horn cores (31)
57
+
12
Least breadth of the braincase (30)
142.2
+
These measurement s are est i mates due to incom plete preservation of the relevant parts.
* These measurements are minimum values, because of damage to the sp ecimen .
6.3.3.2 Comparisons
It is a commonly held view that all very large-bodied alcelaphine antelope in Africa
belong to one genus, Megalotragus Van Hoepen, 1932, and that the various fossil spe-
cies of this genus are closely related (Ge ntry, 1978, 2010; Gentry and Gent ry, 1978;
Gentry et al., 1995; Vrba, 1979, 1997). Previously a smalle r-bodied species from
Rusinga Island, Kenya, Rusingoryx atopocranion Pickford and Thomas, 1984,
w
as
referred to the genus Megalotragus (Vrba, 1997), but more recently in a revision of
the morphological and phylogenetic relationships of this species it was re-assigned
to Rusingoryx as a genus distinct from Megalotragus, although closely related (Pick-
ford and Thomas, 1984; Faith et al., 2011). Previously it was suggested that an Eas t
African species, M. kattwinkeli, is ancestral to the southern African M. priscus,
w
hich
is considered to include two temporal fo rms, M. priscus eucornutus and M. priscus
priscus (Gentry and Gen t ry, 1978). However, this hypothesis seems no longer to be
supported (Gentry, 2010). There is no fossil evidence yet for the species M. eucornu-
tus and M. priscus outside of southern Africa, although Faith et al. (2011) mentions
the presence of large wildebeest-like alcelaphine from the Wasiriya Beds on Rusinga
Island, which they refer to the genus Megalotragus. M. eucornutus is known only from
Cornelia-Uitzoek, the type locality of the Cornelian Land Mammal Age (LMA) and
from Cornelia-Mara, a nearby locality of similar age (Brink et al., 2012; Brink, in
press). It has been noted that there are grounds for maintain ing the specific distinc -
tion between M. priscus and M. eucornutus, based on horn core morphology (Bender
and Brink, 1992), but also on dental proportions (Brink, 2005). Thus, for the sake of
clarity the distinction between names M. eucornutus Van Hoepen, 1932 and M. priscus
(Broo m, 1909) is maint ained her
e
.
Originally the Flo ris bad giant alcelaphine material was named Bubalis helmei by
Dreyer and Lyle (1931). Van Hoepen (1932) described an isolated find of an incom-
plete horn core pair with intact pedicels from the farm Doornberg, on the Sand River
near Kroonstad, Free State Province, as Pelorocerus elegans. Van Hoepen (1947)
described the Mahe ms p an material in itially as Pelorocerus mirum. He also redescribed
the Florisba d B. helmei as Lunatoceras mirum (Van Hoepen, 1947), which was later
referred to Alcelaphus helmei by Cooke (1952) and again to Pelorocerus helmei by
Hoff man (1953). Subsequently Cooke (1964) referred the Flo risb ad material to both
A reconstruction of the sku ll of Megalotragus priscus (Broom, 1909) 85
Pelorocerus helmei and to Lunatoceras mirum. Eventually Gentry and Gentry (1978)
included the Florisbad material and all other Florisian giant alcelaphine materials in
M. priscus. Th is was followed by Vrba (1979, 1997), Klein (1984) and Brink (1987).
The synonymy of the genus Megalotragus and the species Megalotragus priscus
is as follo
ws:
Genus MEGALOTRAGUS Van Hoepen 1932
1932 Megalotragus Van Hoepen
1932 Pelorocerus Van Hoepen
1953 Lunatoceras Hoffman
1965 Xenocephalus Le ak
ey
Type species. Megalotragus priscus (Broom 1909)
The type specimen (SAM 1741) is a cranial fragment with part of the left horn core
preserved and it is housed in the Iziko South African Museum in Cape Town.
1909 Bubalis priscus Broom
1931 Bubalis helmei Dreyer & L
yle
1932 Pelorocerus elegans Van Hoepen
1947 Lunatoceras mirum Van Hoepen
1951 Connochaetes grandis Cooke & W
ells
The Erfkroon specimen is very similar to the type specimen, but more complete.
An almos t complete set of horn cores from Flo risbad , FLO 2274, is virtually identica
l
to the Erfkroon specimen in terms of size, horn shape and horn curvature, but lacking
the braincase. In FLO 2274 the frontals’ suture is partially fused, as in the Erfkroon
specimen and similar to the Barbary sheep, Ammotragus lervia. In another Florisbad
specimen, FLO 2273, the horns tend to be more sharply curved, they exte nd furt her
backwards at their bases before curving sideways and forwards. The bases of the horn
cores lack the protuberance on the cranial surface, as seen in the Erfkroon specimen
and in FLO 2274. The mutual angle between the horn bases is somewhat smaller,
w
hile
the specimen is generally more gracile. This specimen is probably a female, while the
Erfkroon specimen, the type specimen and FLO 2274 appear to be of ma
les
.
In addition to the above the horn core of M. priscus from the Ongers River near
Britstown, central Karoo (Brink et al., 1995), has a base that is not antero-posterior
l
y
extended and appears to be somewhat rounded in cross section. There is no bas al pro-
tuberance and, while it is difficult to estimate the degree of pedicel fusion in this speci-
men, it appears to have had a reasonably wide mutual angle between the horn cor
es
.
The specimen is very gracile and in size co mparable to the specimen from Doornberg,
C. 1711. Of all the specimens ass igned to M. priscus the Doornberg specimen have the
smalles t mutual angle between the horn core bas es, a condition that is considered to
be plesiomorphic for Megalotragus (Gentry and Ge ntry, 1978). For this reason it is
probable that the Doornberg specimen is geologically older than the other specimens
of M. priscus under consideration here. The Ongers specimen resembles the Doo rn -
berg specimens in gracility, but has a wider mutual angle between the horn bases. The
horn core bases of the Doornberg specimen are also less expanded antero-posterior
l
y
than the Ongers River specimen and have no basal protuberances, which is als o a
plesiomorphic condition for M. priscus. Both these specimens are likely to be fema
le
.
In the type specimen of P. mirum (C. 2013) from Mahemspan the horn pedi-
cels are not as extremely fused as in the Erfkroon and Flo risb ad specimens and the
area of pedicel fusion is less elevated above the frontals (Van Hoepen, 1947). The
mutual angle between the horn bases is reduced, resembling the Doornberg specimen.
The basal parts of the horn cores are not as robust as in the large specimens from
86 James S. Brink et al.
Erfkroon and Florisbad. The cranial sides of the horn bases are not preserved and it
cannot be established whether there were protuberances. This specimen appears to be
ma
l e
.
In specimen C. 2537 from Mahemspan, a frontal frag ment with the basal parts
of the horn cores preserved, there is a marked posterior projection of the horn bas es
and the horn bases are much thinner. There is a reduced mutual angle between the
horn core bases, a reduced degree of pedicel fusion, while the frontals’ suture appears
les s fused than in the Erfkroon and Florisb ad specimens. C. 2537 is also more gracile
than Mahemspan specimen C. 2013 and very s imila r to the Doornberg specimen and,
consequently, is likely to be fema
le
.
Mah e ms pa n specimen C. 2246 has an equally narrow mutual angle between the
horn core bas es. In contrast with the Erfkroon specimen the nuchal crest forms a
convex relief, and is not inverted, and it is not as wide as in the Erfkroon specimen. In
the co-type of P. mirum from Mahemspan, C. 2292, the horn base appears not to hav
e
a protuberance and it is not antero-posteriorly extended. The curvature of the horns
is intermediate between the large forms from Flo risb ad and Erfkroon and the sma
ll
specimens from Doornberg and M ahemspan. These specimens are probably fema
le
.
6.3.3.3 Sexual dimorphism and geographic variability in the horn cores of M. priscus
The more complete and dated materials available now for M. priscus allow an appr
eci-
ation of sexu al dimorphism and geographic variability. Although there is considerab
le
variability in size, the horn core specimens of M. priscus can be separated into cat-
egories of male and fe male. Females are those with more gracile horn cores, with less
dorso-ventral extended basal parts and with slightly shorter horn curvature. Males, on
the other hand, have generally larger horn cores, with dorso-ventral expanded horn
bases and with a thickening, or protuberance on the dorsal side. Male horns tend to be
more horizontally positioned, as seen in the Erfkroon specimen, and to be less sharply
curved. This supports the observation that the mutual angle in M. priscus horn cor
es
is a sexually dimorphic character with males tending to have mo re downward pointing
horns than females and that a greater mutual angle is associated with greater robustic-
ity (Brink et al., 1995).
The type specimen and the specimens from Florisbad and Er fkroon are consid-
erably larger than those from M ahemspan, Doornberg and the Ongers River. In the
Mah e msp a n specimens there is a lesser degree of horn pedicel fusion, the fused pedice
ls
are less elevated above the frontals and the horn cores are generally more vertica
ll
y
inserted. Previously it has been suggested that morphological variability in M. priscus
horns represent a temporal cline in that horns become more downward and forward
pointing in the course of geological time (Cooke, 1974). This statement was based on
the assumption that Mahe ms p an predates the Florisbad spring assemblage. Ho
we
ver,
on the evidence presented here it appears now that Mahemspan is of terminal Late
Pleistocene age and that a more likely exp lan ation for horn core variability may be
that it reflects sexual dimorphism and geographic variability in populations. Although
undated, the Doornberg specimen, C. 1711, na med P. elegans by Van Hoepen (1947), is
probably an early and very gracile female version of M. priscus. Similarly, the two forms
of giant alcelaphine from Flo risbad, P. helmei and L. mirum, probably r
ef
lects
sexual dimorphism in M. priscus, with the forme r being male and the latter fema
le
.
On the available fossil evidence M. priscus can be divided into two morphologica
l
entities, which may have represented two geographic variants. This is a very tentativ
e
observation and will need further testing. If the variability observed truly reflects geo-
graphic variability, it would be in parallel with the variation seen in ext ant populations
A reconstruction of the sku ll of Megalotragus priscus (Broom, 1909) 87
of hartebeest, Alcelaphus buselaphus subspp., and ts ess ebe, Damaliscus lunatus subspp.
(Kingdon, 1997).
6.4 RECONSTRUCTING THE SKULL OF M. PRISCUS
The horn core pair and braincase from Erfkroon, an upper jaw fragment (C. 1804) and
a lower jaw (C. 2472), both from Mah e ms pan (Figure 5), form the basis of the recon-
struction of the skull of M. priscus (Figures 6 and 7). Because it is possible to establish
the position of the jaw articulat ion on the Erfkroon braincase, the Mah e ms p an lo
w
er
jaw allows the estimation of the length of the face. Also, in conjunction with the rem-
nant of the frontal, the mandible allows the angle of the braincase to the face to be
es timated. The extre me posterior position of the horn bas es and the position of the
occipital condyles suggest a hanging, ox-like head position. Even if the Mahemspan
materials