A new perspective on the geohydrological and surface 
processes controlling the depositional environment at 
the Florisbad archaeozoological site 
 
Rodney Malcolm Douglas 
 
 
Submitted in fulfilment of the academic requirement for the degree of 
 
Philosophiae Doctor 
 
In the Faculty of Natural and Agricultural Sciences 
 
Department of Geology and Department of Geography 
University of the Free State 
Bloemfontein 
 
 
Promotors: Prof. M. Tredoux 
Prof. P. J. Holmes 
 
 
April  2009 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Index 
 
 
 
 
 
 
 i
DECLARATION 
 
I declare that the thesis hereby submitted by me for the Philosophiae Doctor degree at the 
University of the Free State is my own independent work and has not previously been 
submitted by me to another University. Where use has been made of the work, or 
assistance, of others, this has been duly acknowledged in the text. I further more cede 
copyright of the thesis in favour of the University of the Free State. 
 
 ii
 
DEDICATION 
 
Dedicated to my parents, Ed and Kath who, from beyond, gave to me an awareness of 
that of which I was unaware, as well as the insight, inspiration, and guidance to embark 
upon this project; and to both hominid and beast, who were sustained over the millennia 
by the spring and its unique formation, and who in their passing, ultimately created the 
opportunity for this research. 
 
 iii
ACKNOWLEDGEMENTS 
 
I would like to acknowledge, with appreciation, the following persons and institutions for 
their support and the contributions  they have made towards this research and thesis: 
 
My supervisors, Prof. M. Tredoux and Prof. P. J. Holmes, who were more than just my 
promoters, for invaluable comment, advice, discussions, encouragement, supervision, and 
an extra-ordinary eye for detail, all of which contributed significantly towards this thesis. 
 
The University of the Free State evaluation committee, Prof. H. van Schalkwyk (Dean of 
the Faculty), Prof. W. A. van der Westhuizen (Head, Department of Geology), Prof. P. J. 
Holmes (Head, Department of Geography), Prof. M. Tredoux (Department of Geology), 
and Prof. C.D.K. Gauert (Department of Geology), who through their deliberations, and 
confidence in my abilities, brought this thesis to reality. 
 
The former Council and Directors of the National Museum, Bloemfontein, Drs. C.M. 
Engelbrecht and C.D. Lynch for allowing me to diverge into this stimulating field of 
research, and for allowing it to expand to its ultimate conclusion. 
 
The present Council and Directors of the National Museum, Bloemfontein, Messes R. 
Nuttall and A. Clementz for supporting previous decisions and for allowing me to finalise 
my research to its ultimate goal. 
 
All those persons and institutions that contributed in so many ways towards this study are 
acknowledged in the appendices and in text, and in particular, those journal editors and 
referees, whose faith and constructive comments contributed so greatly to the publication 
in the appendices. 
 
Mrs L. Cronje is thanked for so meticulously translating the English abstract into 
Afrikaans. 
 
 iv
Dr. D Vermeulen and Mr E. Lucas of the Institute for Groundwater Studies, University of 
the Free State, are thanked for running the data through the Institutes hydrogeology 
WISH Program, and for their help with the Piper and Stiff Diagrams. 
 
Messers C. Venter, C. Barlow, and Mrs E. Schoeman of the Design Department, National 
Museum, Bloemfontein, are thanked for their invaluable help and patience with the 
various computer programs used. 
 
 
 
 v
ABSTRACT 
 
The Florisbad Quaternary Research Station and archaeozoological site is located 45 km 
north-west of Bloemfontein, Free State Province, South Africa (28° 46` 05.4”S, 26° 04` 
10.7”E), and is sited around a series of highly saline, warm water spring vents. The site is 
partially covered by a large sand dune. The site is significant for three important reasons. 
Firstly, the discovery of the Florisbad skull (Homo helmei) in 1932 by Prof. T. Dreyer, 
secondly, a collection of faunal fossil remains representing at least 31 taxa, including 
extinct and extant species, and referred to as the Old Collection and, thirdly, a Middle 
Stone Age (MSA) human occupation horizon representing a temporary butchering site 
with evidence of a hearth, butchering tools, and faunal fossil remains. 
 
Spring- and excavation pit water samples were taken and analysed in 1988 during a high 
rainfall period, and in 1999 during an average rainfall period. In relation to the spring 
water, the results show that the total dissolved solids (TDS) of the excavation pit water 
were, in relation to the spring water, higher during the high rainfall period and lower 
during the average rainfall period. This was contrary to the norm, where it is expected 
that high rainfall periods should produce a decrease in TDS due to a dilution effect. The 
TDS of the spring-water remained stable throughout both high and average rainfall 
periods. Further analysis showed considerable TDS increases between the excavation pit 
waters, and between the pit waters and the spring-water. It is concluded that the pit waters 
were not directly related to the spring water and that the two water bodies were separate 
entities with the pit water being recognized as groundwater. An analysis of rainfall in 
relation to the TDS of the spring- and groundwater indicated that short-term rainfall 
affected the quality of the groundwater, but not the quality of the spring-water, while 
long-term rainfall had little effect on the quality of the spring-water. 
 
The question arose as to why the TDS of the groundwater was so much higher than that 
of the spring-water, and what factors were causing these differences? Organic-clay (peat) 
samples from the walls of the excavation pits as well as the walls of the open excavation 
area were analysed. The results of the analyses, and an examination of the stratigraphy, 
strongly suggested that minerals had accumulated in the organic-clay layers due to 
 vi
organic matter having a similar colloidal organization to that of clay, with the ability to 
adsorb large quantities of minerals on their outer surfaces. A comparison of the 
groundwater and organic-clay analyses results showed that the TDS of the decomposed 
Peat II organic-clay layer was considerably higher than that of the groundwater, with the 
same being true for the far less decomposed Peat IV organic-clay layer. By analysing and 
combing the water and organic-clay layer results with the many factors, mechanisms, and 
processes involved, it is concluded that the salinization of the organic-clay layers, and the 
flushing of ions from the organic-clay layers by percolating water during rainfall periods, 
is responsible for the increased mineralization of the groundwater. Other factors, 
mechanisms, and processes, such as rainfall, aeolian deposition, evaporation, capillarity, 
wind, temperature, matrix-suction, pH, Eh, PCO2, PO2, DOC, and biomineralization, all 
of which support the accumulation of free salts in a semi-arid environment such as 
Florisbad, were also investigated. 
 
Of primary importance was the question as to whether the spring-water was actually 
responsible for fossilization of the faunal remains, and could fossilization have taken 
place within the environs of the spring vents, or in the spring vents themselves? Previous 
research has suggested that the spring-water was calcium-carbonate rich, with evidence 
of calcium-carbonate deposition further suggesting that faunal remains of the Old 
Collection must have been in contact with the spring-water in spring vents for some time. 
An analysis of the spring-water analysed over the past 84 years indicated that there had 
never been sufficient Ca (under-saturation) in the spring-water for fossilization to occur, 
and this is confirmed by the current analyses. The contemporary lack of Ca in the spring-
water, combined with other environmental factors within the environs of spring vents, 
such as the lack of organic matter and clay, combined with a high Eh environment, also 
strongly indicated that, historically, fossilization could not have taken place within the 
environs of the springs. Contrary to earlier hypotheses, it is concluded that the spring 
water and spring flow would directly assist in the de-mineralization of faunal remains. 
 
A detailed investigation of the site, along with an analysis of the stratigraphy and 
sedimentation, revealed that previous theories on the formation of the site did not 
 vii
sufficiently accommodate the current stratigraphy in the context of the organic-clay 
layers, the salinization process, and fossilization. From this deduction all the existing and 
pre-existing evidence was revisited in an attempt to provide a hypothesis which would 
accommodate the existing morphology of the site, sedimentation, and fossilization. It is 
hypothesised that the spring site formed around a large drainage-impeded pan which was 
largely covered by a sand dune that had migrated from the area of the extensive salt pan 
to the north and north-west (Soutpan). The arms of the dune eventually came to rest up 
against the windward slope of a dune belt located just south of the spring site, and a dam 
began to form. High rainfall periods produced organic-clay layers, while sandy layers 
were produced during drier windy periods. This led to the formation of alternating 
horizontal layers of organic-clay and sand, eventually building up to almost the top of the 
sand dune on the leeward face. When the water level in the dam reached the top of the 
arms of the sand dune, it broke through the eastern arm. The dam water and sediments 
then evacuated the dam in a flash flood. This flash flood eroded the area to the east of the 
site to such an extent that the drainage was diverted, and a wide flat-bottomed vlei was 
formed where much of the dam sediments were deposited. This hypothesis provides an 
alternative for the formation of the spring site, accommodating all aspects of 
sedimentation, salinization, and fossilization. 
 
The dating of the Florisbad deposits and fossils has been subject to an ongoing debate 
since the first 14C dating was carried out in 1954. The ages and depths of recently 
published profiles did not appear to correspond to the assumption of greater compaction 
with depth and time. In an attempt to resolve this issue, linear, exponential, and 
logarithmic mathematical trend lines were then experimentally applied to the published 
profiles of electron spin resonance (ESR) and optical stimulated luminescence (OSL) 
dates in order to test the theory of compaction, and to validate the results. The 
hypothetical effect of manipulating ages on trend lines was also tested. A discussion on 
some possible shortfalls regarding the dating methods used is undertaken. 
 
A best logarithmic fit to data was obtained by holding the ESR Middle Stone Age Human 
Occupation Horizon (MSA) age at 127 ka, and advancing the lower deposit age from 250 
 viii
ka to 420 ka. The next best fit to data occurred by regressing the ESR MSA age from 127 
ka to 78 ka, and holding the lower deposit age at 250 ka. The application of exponential 
and linear trend lines produced poor fits to data. A suggested compaction trend line was 
also introduced which produced an ESR MSA age of 75 ka and a lower deposit age of 
384 ka. In the final analysis, trend line results suggested an MSA age of 92 ±12 kyr and a 
basal deposit age of 400 ±20 ka. The logarithmic and suggested compaction trend line 
ages for the lower deposits both produced ages similar to the suggested Florisain – 
Cornelian faunal boundary of c. 400 ka. The exercise confirmed that the ages in the 
published profiles were disjunct and that this disjunction may be related to a number of 
different physical forces. 
 
 
 ix
KEY WORDS 
 
Key Words: Florisbad, Archaeozoological site, Spring-water, Groundwater, Organic-clay 
layers, Salinization, Fossilization, Chemistry, Geohydrology, Geology, 
Depositional environment, Formation of site, Dating. 
 x
OPSOMMING 
 
Die Florisbad Kwaternêre Navorsingstasie en argeosoölogiese terrein is 45 km noordwes 
van Bloemfontein, Vrystaat Provinsie, Suid-Afrika geleë (28° 46` 05.4”S, 26° 04` 
10.7”E) langs ‘n reeks uiters sout- en swaelryke warmwater fonteine.  Die terrein is 
gedeeltelik bedek deur ‘n groot sandduin.  Die terrein is betekenisvol vanwee drie 
belangrike redes:  Eerstens die ontdekking van die Florisbad-skedel deur prof. T. Dreyer 
in 1932, tweedens `n versameling van dierlike fossieloorblysfsels wat ten minste 31 taksa 
van uitgestorwe en bestaande spesies verteenwoordig (bekend as die Ou Versameling), en 
derdens ‘n menslike bewoningshorison uit die middel-Steentydperk (MST) wat ‘n 
tydelike slagplaas, tekens van ‘n vuurherd, slaggereedskap en dierlike fossieloorblyfsels 
toon. 
 
Monsters van die fonteinwater en sypelwater in die uitgrawings is in 1988 gedurende ‘n 
tydperk van hoë reënval, en in 1999 gedurende ‘n tydperk van gemiddelde reënval, 
geneem en ontleed.  Die resultate dui aan dat die totale opgeloste vastestowwe (TOV) van 
die sypelwater in die uitgrawings, in vergelyking met dié van die fonteinwater, hoër was 
gedurende die nat periode en laer gedurende die gemiddelde reënvalperiode.  Dit is 
teenstrydig met die norm wat sou verwag dat hoë reënvalperiodes ‘n afname in TOV sal 
oplewer as gevolg van ‘n verdunningseffek.  Die TOV van die fonteinwater het stabiel 
gebly deur beide die hoë en gemiddelde reënvaltydperke. Verdere ontleding het ‘n 
aansienlike TOV toename tussen die sypelwater van verskillende uitgrawings en tussen 
die sypelwater en fonteinwater getoon.  Die gevolgtrekking is dat sypelwater in 
uitgrawings nie direk verband hou met fonteinwater nie en dat die twee waterliggame 
aparte entiteite is waarvan die sypelwater as grondwater beskou kan word.  ‘n Analise 
van reënval in verhouding tot die TOV van die fontein- en grondwater dui aan dat 
korttermyn reënval die kwaliteit van die grondwater beïnvloed maar nie dié van die 
fonteinwater nie.  Langtermyn reënval het weinig invloed op die kwaliteit van 
fonteinwater. 
 
 xi
Die vraag het ontstaan waarom die TOV van die grondwater soveel hoër is as dié van die 
fonteinwater en watter faktore hierdie verskille veroorsaak.  Organiese klei (veen) 
monsters van die kante van uitgrawings asook die kante van die oop uitgrawing is 
geanaliseer.  Die resultate van die  analises en ‘n ondersoek van die stratigrafie het sterk 
aanduidings getoon dat minerale in die organiese kleilae versamel het, as gevolg van die 
feit dat organiese materiaal dieselfde kolloïdale struktuur as klei het en die vermoë besit 
om groot hoeveelhede minerale in hulle buitenste lae te adsorbeer.  ‘n Vergelyking tussen 
die resultate van die grondwater- en organiese klei analise het getoon dat die TOV van 
die ontbinde Veen II organiese kleilaag aansienlik hoër was as dié van die  grondwater, 
terwyl dieselfde vir die veel minder ontbinde Veen IV organiese kleilaag geld.  Deur die 
resultate van die water en organiese kleilaag te vergelyk met die baie faktore, 
meganismes en prosesse betrokke, word die gevolgtrekking gemaak dat versouting van 
die organiese kleilae en die loging van ione uit die organiese kleilae deur sypelwater 
gedurende reënval periodes, verantwoordelik is vir die toenemende mineralisasie van die 
grondwater.  Ander faktore, meganismes en prosesse soos reënval, aeoliese neersetting, 
verdamping, kapillariteit, wind, temperatuur, matriks-suiging, pH, Eh, PCO2, PO2, DOC 
en biomineralisasie wat almal bydra tot die opeenhoping van vry soute in ‘n semi-ariede 
omgewing soos Florisbad, is ook ondersoek.  
 
Van primêre belang was die vraag of fonteinwater eintlik verantwoordelik was vir 
fossilering van die dierlike oorblyfsels en of fossilering in die omgewing van fonteine of 
in die fonteine self kon plaasvind.  Vorige navorsing het daarop gedui dat fonteinwater 
kalsiumkarbonaatryk was, met aanduidings van kalsiumkarbonaat afsetting wat verder 
daarop dui dat dierlike oorblyfsels van die Ou Versameling vir ‘n geruime tyd in kontak 
met die fonteinwater in fonteinne moes gewees het.  Wateranalise van die fonteinwater 
oor die afgelope 84 jaar het aangedui dat daar nog nooit voldoende Ca (onderversadiging) 
in die fonteinwater was vir fossilering om plaas te vind nie en dit word bevestig deur die 
huidige analise.  Die hedendaagse gebrek aan Ca in die fonteinwater, in kombinasie met 
ander omgewingsfaktore in die omtrek van fonteine, soos die gebrek aan enige organiese 
materiaal of klei en ‘n hoë Eh omgewing, is ‘n sterk aanduiding dat fossielisering nie in 
die verlede in fonteine kon plaasgevind het nie. In teenstelling met vorige hipoteses word 
 xii
die gevolgtrekking gemaak dat fonteinvloei bydraend is tot die demineralisasie van 
dierlike oorblyfsels. 
 
‘n Gedetailleerde ondersoek van die terrein, saam met ‘n analise van die stratigrafie en 
sedimentasie, het aan die lig gebring dat vorige teorieë oor die ontstaan van die terrein nie 
die huidige stratigrafie ten opsigte van die organiese kleilae, die versoutingsproses en 
fossilisering, genoegsaam in ag geneem het nie. Met hierdie afleiding in gedagte is al die 
bestaande en vooraf bestaande getuienis weer nagegaan in ‘n poging om met ‘n hipotese 
voor ‘n dag te kom wat die bestaande morfologie van die terrein, sedimentasie en 
fossilering sou kon akkommodeer.  Daar word gehipoteseer dat die fontein gevorm het in 
die omgewing van‘n groot pan met beperkte dreinering.  Hierdie pan was grootliks bedek 
deur ‘n sandduin, wat migreer het van die oorspronklike terrein in ‘n noord- en 
noordwestelike rigting (Soutpan).  Die arms van die duin het uiteindelik tot ruste gekom 
teen die windkanthang van ‘n duingordel wat net suid van die terrein van die fontein 
geleë is en het ‘n dam gevorm.  Hoë reënval periodes het organiese kleilae gevorm, 
terwyl sanderige lae gedurende droër winderige periodes gevorm is.  Dit het gelei tot die 
vorming van afwisselnde horisontale lae van organiese klei en sand, wat uiteindelik tot 
amper by die kruin aan die lykant van die sandduin opgebou het.  Die stygende watervlak 
het deur die oostelike arm van die sandduin gebreek en water en sediment in die dam is 
d.m.v. ‘n blitsvloed gedreineer.  Hierdie blitsvloed het die area oos van die terrein tot so 
‘n mate geërodeer dat die dreinering herlei is en ‘n wye vlei gevorm het waar baie van die 
sedimente van die dam gedeponeer is.  Hierdie hipotese verskaf ‘n alternatiewe 
verklaring vir die vorming van die terrein om die fonteine en sluit alle aspekte van 
sedimentasie, versouting en fossilering in. 
 
Die ouderdom van die Florisbad afsettings en fossiele is sedert die eerste 14C 
ouderdomsbepaling gedoen in 1954, onderworpe aan ‘n voortgesette debat.  Die 
ouderdomme en dieptes van onlangs gepubliseerde profiele het skynbaar nie 
ooreengestem met die aanname van hoër kompaksie met diepte en tyd.  In ‘n poging om 
hierdie kwelvraag op te los, is lineêre, eksponensiële en logaritmiese wiskundige 
tendenskrommes op die gepubliseerde profiele van ESR en opties gestimuleerde 
 xiii
luminessensie (OSL) ouderdomme gebruik om die teorie van kompaksie te toets en die 
resultate daarvan te bekragtig.  Die hipotetiese effek van die manipulering van 
ouderdomme op tendenskrommes is ook getoets.  Die moontlike tekortkominge van die 
dateringsmetodes wat gebruik is, word ook bespreek. 
 
’n Beste logaritmiese datapassing is verkry deur die ESR Middel Steentydperk Menslike 
Bewoningshorison (MST) ouderdom van 127 ka konstant te hou en die laer 
afsettingsouderdom van 250 ka na 420 ka te verander.  Die volgende beste datapassing is 
verkry deur die ESR MST ouderdom van 127 ka na 78 ka terug te skuif en die laer 
afsettingsouderdom op 250 ka konstant te hou.  Die aanwending van eksponensiële en 
lineêre tendenskrommes het swak datapassings opgelewer.  ‘n Voorgestelde kompaksie-
tendenskromme is ook toegepas.  Dit het ‘n ESR MST ouderdom van 75 ka en ‘n laer 
afsettingsouderdom van 384 ka opgelewer.  In die finale analise het die tendenskromme 
resultate ‘n MST ouderdom van 92 ±12 ka en ‘n basale afsettingsouderdom van 400 ±20 
ka voorgestel.  Die logaritmiese en voorgestelde kompaksie tendenskromme 
ouderdomme vir die laer afsettings het beide ouderdomme opgelewer soortgelyk aan die 
voorgestelde Florisiaans – Corneliaanse fauna grens van c. 400 ka.  Die oefening het 
bevestig dat die ouderdomme in die gepubliseerde profiele disjunk was en dat hierdie 
disjunksie verwant kan wees aan verskilende fisiese kragte. 
 
 xiv
 
TABLE OF CONTENTS 
Index i 
 
Declaration ii 
 
Dedication iii 
 
Acknowledgements iv 
 
Abstract vi 
 
Key Words x 
 
Uittreksel xi 
 
Table of Contents xv 
 
List of Tables xxi 
 
List of Figures xxiv 
 
List of Appendices  xxx 
 
Chapter 1 Introduction 1 
1.1 An Historical Overview of the Florisbad Site 2 
1.2 A Brief Description of the Florisbad Area 5 
1.3 Previous Scientific Research at Florisbad 9 
1.4 The Florisbad Faunal Deposits 14 
1.3.1 The Old Collection 14 
1.3.2 The MSA Human Occupation Horizon 18 
 xv
1.5 The Florisbad Skull 20 
1.6 Dating of the Florisbad Deposits 24 
1.7 Rationale for the Study 26 
1.8 Key Areas to be Addressed 28 
1.9 Synopsis 29 
 
Chapter 2 The Physical Environment. Part I: Geology and 
 Geomorphology 30 
2.1 Introduction 31 
2.2 Regional Geology 32 
2.2.1 Precambrian Strata 32 
2.2.2 The Karoo Basin 35 
2.2.3 The Karoo Supergroup 38 
2.2.3.1 Dwyka Group 38 
2.2.3.2 Ecca Group 39 
2.2.3.3 Beaufort Group 40 
2.2.3.4 Stormberg Group 40 
2.2.3.5 Drakensberg Group 42 
2.2.4 The Karoo Dolerite Suite 42 
2.2.4.1 Dolerite sills and Ring-Complexes 43 
2.2.4.2 Dolerite Dykes 46 
2.2.5 Other Post Karoo Intrusions 46 
2.2.5.1 Breccia Plugs 46 
2.2.5.2 Volcanic Vents 47 
2.2.5.3 Kimberlites 48 
2.2.6 Cenozoic Deposits and Soils  48 
2.2.6.1 Alluvial and Colluvial Deposits 49 
2.2.6.2 Calcretes 49 
2.2.6.3 Sands and Soils 49 
2.3 Geology of the Florisbad Area 53 
2.4 Geomorphology 55 
 xvi
2.4.1 Topography 55 
2.4.2 The Western Free State Panveld 57 
2.4.3 The Florisbad Sand Dune 62 
2.5 Synopsis 69 
 
Chapter 3 The Physical Environment. Part II: Climate and Vegetation 70 
3.1 Introduction 71 
3.2 Palaeoclimate 71 
3.3 Present Climate 76 
3.3.1 Rainfall 76 
3.3.1.1 Convergent Rainfall 76 
3.3.1.2 Orographic Rainfall 77 
3.3.1.3 Frontal Systems 77 
3.3.1.4 Rainfall of the Florisbad Area 77 
3.3.2 Temperature 78 
3.3.3 Evaporation 78 
3.3.4 Wind 80 
3.4 Vegetation 82 
3.5 Synopsis 85 
 
Chapter 4 The Physical Environment. Part III: Geohydrology 87 
4.1 Introduction 88 
4.2 Background 89 
4.2.1 Aquifers and Aquitards 92 
4.2.1.1 Intergranular Aquifers 93 
4.2.1.2 Fractured Aquifers 93 
4.2.1.3 Karstic Aquifers 93 
4.2.1.4 Intergranular and Fractured Aquifers 94 
4.3 Geohydrology of the Karoo Supergroup 94 
4.3.1 Dwyka Group 94 
4.3.2 Ecca Group 96 
 xvii
4.3.3 Beaufort Group 97 
4.3.4 Stormberg Group 97 
4.3.4.1 Molteno Formation 97 
4.3.4.2 Elliot Formation 98 
4.3.4.3 Clarens Formation 98 
4.4 Geohydrology of the Dolerite Suite 98 
4.4.1 Dolerite Dykes 98 
4.4.2 Dolerite Sills and Ring-Complexes 100 
4.5 Geohydrology of Other Post Karoo Intrusions 101 
4.5.1 Breccia Plugs and Volcanic Vents 101 
4.5.2 Kimberlites 101 
4.6 Geohydrology of the Recent Deposits 102 
4.7 Non-Intrusive Tectonic Features 102 
4.8 Synopsis 103 
 
Chapter 5 The Florisbad Springs 104 
5.1 Introduction 105 
5.2 Classification of South African Springs 105 
5.3 Geohydrology of the Florisbad Area 108 
5.4 The Florisbad Spring Aquifer 109 
5.4.1 Control of the Springs by Geological Features 112 
5.4.2 Temperature 114 
5.4.3 Discharge Rate 116 
5.5 Chemistry of the Spring- and Groundwater 118 
5.6 The Depositional Environment and Sedimentation 129 
5.6.1 Theories on the Formation of the Florisbad 
 Sediments 129 
5.6.2 Current State of the Florisbad Sediments 133 
5.6.3 Salinization of the Organic-Clay Layers 140 
5.6.4 A New Perspective on the Fossilization of the 
Faunal Remains 147 
 xviii
 
5.6.5 A New Perspective on the Formation of the 
 Florisbad Spring Site 153 
5.7 Medicinal Properties 163 
5.8 Synopsis 165 
 
Chapter 6 The use of Mathematical Trend Lines in Evaluating ESR 
and OSL Dating at Florisbad 167 
6.1 Introduction 168 
6.1.1 Stratigraphy 170 
6.1.2 Archaeozoological and Archaeological Deposits 172 
6.1.3 Previous Dating 173 
6.2 Methods 174 
6.2.1 The Application of Trend Lines to the ESR and 
OSL Ages 174 
6.3 Results 177 
6.4 Discussion 183 
6.5 Synopsis 189 
 
Chapter 7 Discussion and Conclusions 192 
 
References 203 
 
Appendices 238 
 
Appendix I 239 
 
Appendix II 250 
 
Appendix III 275 
 
 xix
Appendix IV 285 
 xx
LIST OF TABLES 
 
Table 1 Examples of previous research carried out at Florisbad Research 
Station. Subject fields are referenced alphabetically with sources 
being referenced at the end of the table 10
  
Table 2 A comparison of fauna recorded between the MSA Assemblage and 
the Old Collection based on relative abundance (minimum number 
of individuals) (after Brink, 1987; 1988). 15
  
Table 3 Changes and updates to the original taxonomic list of fauna 
occurring at Florisbad Research Station (Brink, 1987), compiled 
from Churchill et al. (2000). 16
  
Table 4 A history of the dating of the Florisbad deposits. 25
  
Table 5 A comparison of Free State soil classifications after MacVicar 
(1973), MacVicar et al. (1977), and Hensley et al. (2006). 52
  
Table 6 Examples of global and local palaeoclimates as determined by 
various factors and authors (log scale). 72
  
Table 7 Natural transitional zones that have their boundaries passing 
through the Florisbad area, as defined by various authors. 86
  
Table 8 Interrelated factors contributing to the complex interactions of 
weathered thickness, porosity, permeability, and groundwater flow 
(after Vegter, 2001). 92
  
Table 9 Classification of South African spring waters and the criteria used 
in their classification. 106
 xxi
Table 10 A comparison of water quality between the Glen artesian borehole 
and the Florisbad spring (after Baran, 2003 [B03]), Kent, 1971 
[K71]) and Appendix 1). 111
  
Table 11 The chronological order of flow rate determinations at the Florisbad 
spring site showing the decrease in flow rate as a possible result of 
the dewatering of the Free State goldfields. 117
  
Table 12 The difference in water quality (TDS, mg/l) between the spring-
water and exploration pit 1 water during high and average rainfall 
periods (after Appendix 1). 121
  
Table 13 Some examples of variations in the increase of anion and cation 
concentrations between the spring water (1999) and the exploration 
pit waters (1999) (after Appendix 1). 122
  
Table 14 The difference in ion concentrations (mg/l) between the 
groundwater and the Peat II organic-clay layer in exploration pits 2 
and 4 during 1999, with a comparison to water from the spring eye 
(after Appendix I; Appendix II). 141
  
Table 15 A comparison between the salinization of the various Peat II 
samples and the organic-clay from the vlei area showing the high 
salinization of the vlei area organic-clay.  162
  
Table 16 A tabular interpretation of age estimates and depths for the third test 
pit of Grün et al. (1996) (illustration (a)) (refer to Pit 2 for this 
study) and the lower spring sediments, Grün et al. (1996) 
(illustration (c)). 175
  
 xxii
 
Table 17 A summary of the various ages and depths discussed in text and the 
affect of applying linear, exponential, and logarithmic trend lines. 182
 
 xxiii
LIST OF FIGURES 
 
Figure 1 A Google Earth satellite image of the Florisbad area showing 
Florisbad in relation to Soutpan and the large areas of red aeolian 
sand deposition in the surrounding areas. Note the evaporative salt 
industry that has established itself on the pan. 3
  
Figure 2 Plan of the Florisbad farm (after Douglas, 1992). 6
  
Figure 3 An aerial photograph of the Florisbad spring site indicating some 
key features ( Photo: National Museum, Bloemfontein). 7
  
Figure 4 A simplified schematic cross-section of the Florisbad sedimentary 
deposits showing the major organic-clay (peat) and sand horizons 
with their silt content (after Kuman et al., 1999; Appendix IV). 8
  
Figure 5 A graphic reconstruction of the head of the extinct giant buffalo 
Pelorovis antiquus from the Florisbad Old Collection. The horns of 
this species attained a spread of over 2 meters. (Graphic: E. 
Russouw, National Museum, Bloemfontein). 17
  
Figure 6 A section of the Middle Stone Age Human Occupation Horizon 
excavation site at Florisbad (Photo: National Museum, 
Bloemfontein). 19
  
Figure 7 Reconstruction of the Florisbad skull (Homo helmei). Note the 
hyena tooth impression on the forehead. (Photo: National Museum, 
Bloemfontein). 21
 
 xxiv
 
Figure 8 Simplified geology of the Free State showing geological formations 
mentioned in the text. (after Keyser (a); Keyser, 1997; Catuneanu et 
al. 2005; Rubidge, 2005). 33
  
Figure 9a The extent of the Karoo Supergroup rocks over South Africa (after 
McCarthy and Rubidge, 2005). 36
  
Figure 9b Schematic cross-section across the Karoo Basin, from Port 
Elizabeth in the south to Nelspruit in the north-east, showing the 
stratigraphy of the Karoo Supergroup rocks and the variation in 
depth of the Karoo Basin. The Karoo Basin also reflects the 
asymmetry of the Karoo Sea (after McCarthy and Rubidge, 2005). 36
  
Figure 10 The mechanism of the emplacement of dolerite sill and ring 
complexes (ring dykes). The saucer shape and ring dyke model after 
Chevallier and Woodford (1999). 44
  
Figure 11 Different types of fractures associated with dolerite sill and ring 
complexes (after Woodford and Chevallier, 2002). 44
  
Figure 12 Soil map of the Free State (after MacVicar, 1973; MacVicar, 1977; 
Lynch, 1983). 51
  
Figure 13 The geology of Florisbad and surrounding areas (after Loock and 
Grobler, 1988). 54
  
Figure 14 The topography of the Florisbad area showing the location of 
Florisbad and other features discussed in the text (after Appendix 
IV). 56
  
 xxv
Figure 15 Drainage systems of the Free State Province showing the Western 
Free State Panveld as defined by Geldenhuys (1982) and Holmes 
and Barker (2006). 58
  
Figure 16 Google Earth images giving a comparison of south and east pan 
fringe lunette development between Morgenzon Pan, Sunnyside 
Pan and Soutpan. Note the low degree of pan fringe lunette 
development at Soutpan and Varspan, as well as the orientation of 
the pans. Images have been adjusted to define the fringe lunettes in 
more clearly. (Not to scale). 66
  
Figure 17 Twenty-three year annual rainfall at Florisbad showing the variation 
in annual rainfall between the lowest (1965) and highest (1988) 
recorded rainfall periods. 79
  
Figure 18 Monthly mean maximum and minimum temperatures for the 
Florisbad area (taken at Glen Agricultural College, 1987). 79
  
Figure 19 Wind roses for Bloemfontein, 45 km south-east of Florisbad, 
showing wind direction and velocities for the months of January, 
April, July, and October (after Kruger, 2002). 81
  
Figure 20 Vegetation of the central-western Free State showing the veld types 
after Acocks (1988), panveld boundaries after Geldenhuys (1981) 
and main drainage systems. 83
  
Figure 21 Some comparative examples of borehole yields between the Karoo 
Supergroup Formations and dolerite sills (after Baran, 2003). 95
 
 xxvi
 
Figure 22 Plan of Florisbad showing the deduced dolerite sill and the extent of 
the organic deposits (after Rubidge and Brink, 1985; Grobler and 
Loock, 1988b). 113
  
Figure 23 A detailed plan of the present day Florisbad spring site showing 
spring-water, groundwater, and organic-clay material sample 
localities (after Douglas, 1992; Appendix I; Appendix 1). 119
  
Figure 24 Piper diagram of the spring eye and exploration pit waters at 
Florisbad. For comparative purposes the spring eye water analysis 
of Rindl (1915) and Bloemfontein rainwater (after Littehauer, 2007) 
have been plotted 124
  
Figure 25 Stiff diagrams for the Florisbad spring eye water (E88, E99), 
exploration pit 1 water (P188, P199), and the waters of exploration 
pits P299, P399 and P499. The spring eye water analysis of Rindl 
(1915) and Bloemfontein rainwater (Litthauer, 2007) are presented 
for comparison. For further comparison purposes all samples have 
been plotted on the maximum scale. For further detail on sample 
codes see Figure 19.  125
  
Figure 26 A simplified schematic cross-section of the Florisbad sedimentary 
deposits showing the major organic-clay (peat) and sand horizons 
with their silt content (after Kuman et al., 1999; Appendix III). 135
  
Figure 27 Meiring’s (1952) excavations at Florisbad showing the continuous 
stratigraphy of the sediments at this particular location, with 
Meiring’s (1952) descriptions of the various layers (Photo: Meiring, 
1952). 137
  
 xxvii
Figure 28 The upper section of excavation Pit 2 on the leeward face of the 
Florisbad sand dune showing the horizontal stratigraphy of the 
sedimentary deposits which appear to have been deposited in an 
aquatic environment (Photo: P.J. Holmes). 139
  
Figure 29 Schematic plan of the Florisbad spring site showing the movement 
of ions and stages in the mineralization of the groundwater and the 
salinization of the organic clay layers, as well as other factors 
contributing to the fossilization of the Old Collection (after 
Appendix III). 146
  
Figure 30 Legend for Figures 22 and 23. 156
  
Figure 31 Schematic plan of the proposed developmental stages in the 
formation of the Florisbad spring site (revised after Appendix IV). 157
  
Figure 32 Schematic profile of the proposed developmental stages in the 
formation of the Florisbad spring site (revise after Appendix IV). 158
  
Figure 33 A plot of linear, logarithmic, and exponential trend lines to recent 
Florisbad ESR and OSL ages from Test Pit 3 over a depth of 8 
metres. The best fit to data of a trend line is represented by R2=1. 15 
= Grün et al. 1996; (a) represents the illustration identifier. 179
  
Figure 34 A plot showing the effect on linear, logarithmic, and exponential 
trend lines by regressing the MSA age to 78 ka and holding the 
lower sediment age at 250 ka. The best fit to data of a trend line is 
represented by R2=1. 15 = Grün et al., 1996; (a) represents the 
illustration identifier. 180
 xxviii
 
Figure 35 A plot showing the effect on linear, logarithmic, and exponential 
trend lines by extending the age of the lower sediments to 420 ka 
and holding the MSA age at 127 ka. The best fit to data of a trend 
line is represented by R2=1 15 = Grün et al., 1996; (a) represents 
the illustration identifier. 181
 
 xxix
LIST OF APPENDIXES 
 
Appendix I. Douglas, R.M. 2001a. The quality of the Florisbad spring-
water in relation to the quality of the groundwater and the 
effects of rainfall. Water SA 27:39-48. 239 
 
Appendix II. Douglas, R.M. 2001b. Salinization of the Florisbad organic 
layers, clay and Groundwater. Navorsinge van die Nasionale 
Museum, Bloemfontein 17(1): 1-24. 250 
 
Appendix III. Douglas, R.M. 2006a. Is the spring-water responsible for 
fossilization of faunal remains at Florisbad, South Africa. 
Quaternary Research 65: 87-95. 275 
 
Appendix IV. Douglas, R.M. 2006b. Formation of the Florisbad spring and 
fossil site – an alternative hypothesis. Journal of 
Archaeological Science 33: 696-706. 285 
 
 xxx
Chapter 
1 
Introduction
1 
CHAPTER 1 
INTRODUCTION 
1.1  A HISTORICAL OVERVIEW OF THE FLORISBAD SITE 
Florisbad Quaternary  Research Station, 28° 46` 05.4”S, 26° 04’ 10.7”E,  (spring 
eyes)  is an  important archaeozoological site  located 45 km north­west of Bloemfontein, 
Free  State  Province,  South  Africa.  A  large  salt  pan,  Soutpan,  located  to  the  north  of 
Florisbad, dominates the  local environment, as  illustrated  in the Google Earth  image of 
the area (Figure 1). The significance of Florisbad is threefold. Firstly, the discovery of the 
Florisbad  hominid  cranium  (Homo  helmei)  by  Prof.  T.  Dreyer  in  1932,  its  subsequent 
description in 1935 (Dreyer, 1935), and the recent dating, based on a fragment of a tooth, 
to 259 ±35 ka by electron spin resonance (ESR) (Grün  et  al., 1996), brought recognition 
to Florisbad. Secondly, there is the existence of a collection of artefacts and vast number 
of faunal fossil remains representing 26 species, which is referred to as the Old Collection 
(Brink, 1987). The Old Collection represents the Florisian Land Mammal Age (Hendey, 
1974),  or  the  Florisian­Cornelian  Faunal  Boundary,  with  an  age  of  c.  400  ka  (Klein, 
1984).  Thirdly,  a  Middle  Stone  Age  (MSA)  human  occupation  horizon,  representing  a 
temporary butchering site, at which butchering tools and faunal fossil remains have been 
excavated and identified (Brink, 1987; Henderson, 2001a; Brink and Henderson, 2001). 
Very little is known about the modern day habitation of Florisbad prior to the settlement 
of  one  Hendrik  Venter  and  his  family  on  the  farm  Rietfontein,  and  the  adjoining  farm 
Jackalsfontein,  before  the  farms  were  purchased  from  the  Land  Commission  in  1849 
(Henderson, 1995). Rietfontein is the farm on which Florisbad is situated. Because of the 
interest  in,  and  collection  of,  early  fossils  by  Martha  Venter,  Hendrik’s  wife,  many  of 
which formed part of the initial collection donated to the National Museum.
2 
3
A brief history of the farm, and its owners, is given, based on Henderson, (1995), unless 
otherwise stipulated. Prior to their settlement at Rietfontein, Hendrik and Martha Venter 
(née  Coetzer) had a  son  Floris  Johannes,  born  in 1836, who  married Renske du Plooy. 
Renske gave birth to a number of children, including a son born in the Bethulie district in 
1869,  also  named  Floris  Johannes,  after  whom  Florisbad  was  later  named.  Part  of 
Jackalsfontein, was registered in Renske’s name in 1886, and all of Rietfontein in 1894. 
When  Hendrik  Venter  died  in  1899,  Renske  remarried  one  Gert  Abraham  Coetzee.  In 
1917 Renske divided the  farms among her children, with the part of Rietfontein, which 
included the springs, going to Floris Johannes (Hendrik Venter’s grandson), This portion 
was registered as Florisbad, with word of the mineral springs healing properties spreading 
far and wide, and Floris Johannes earning the reputation of the “healer” (Anon [1], 1980). 
The  springs  became  known  as  “Floris  se  bad”,  hence  the  name  Florisbad  (Anon  [1], 
1980).  Floris  married  Martha  Johanna  Breed  and  they  had  four  children.  Later  Gert 
Abraham sold his share in the baths to a niece and her husband, who in turn were bought 
out by the brothers, Edward and John Sowden. The State bought out 93 ha of  the  farm, 
including  the  springs  in  1980,  and  handed  over  custody  of  the  site  to  the  National 
Museum in 1981. 
In 1912, due to the increase in the numbers of persons wanting to use the spring, Floris 
built a small house over  the spring eye  in order to provide some protection and privacy 
for  bathers.  While  increasing  the  size  of  the  swimming  bath,  a  number  of  fossils  were 
unearthed.  The  same  year,  a  strong  earthquake  occurred,  with  its  epicentre  near 
Koffiefontein,  125  km  southwest  of  Florisbad,  and  having  an  intensity  of  VIII  on  the 
Mercalli  scale  (Fernàndez  and  Guzmàn,  1979).  This  event  has  often  mistakenly  been 
reported  in  the  literature  as  having  occurred  at  Fauresmith  (Anon  [1],  1980).  The 
earthquake reportedly resulted  in a new spring eye erupting  in the excavations due to a 
build­up of gasses  (Grobler and Loock, 1988a),  throwing up  many  fossils and  artefacts 
(Anon  [1],  1980).  Martha  Venter  made  a  large  collection  of  these  fossils,  which  were 
later donated to the National Museum by her son, Gert.
4 
1.2  A BRIEF DESCRIPTION OF THE FLORISBAD AREA 
In general, the Florisbad area is undulating, broken occasionally by dolerite dykes which 
have  formed  low  hills.  The  area  is  underlain  by  Ecca  Group  rocks  of  the  Karoo 
Supergroup, with  no evidence of  the  rocks outcropping  in  the  immediate area which  is 
partially due  to  the area being blanketed by  red and  yellow aeolian sands.. A  large salt 
pan,  Soutpan  (Figure  1),  dominates  the  area  and  has  had  a  considerable  effect  on  the 
formation of the Florisbad spring site in that brackish wind blown material, deflated from 
the pan, has contributed to the Florisbad sand dune and the site as a whole. A belt of san 
dunes  also  occurs  south  east  of  Florisbad  indicating  that  aeolian  deposition  was  wide 
spread. Other smaller salt pans are visible in the top left hand corner and the bottom right 
hand  corner  of  the  image.  Figure  2  presents  a  plan  of  the  Florisbad  farm  showing  the 
spring and excavation sites in relation to Soutpan as well as other key features. 
Figure  3  is  an  aerial  photograph  of  the  Florisbad  archaeozoological  site  indicating  a 
number of key  features referred to in the text. The Florisbad archaeozoological site and 
excavations are centred in the area of a number of warm water spring eyes, around which 
the swimming pools were built. A number of buildings have been removed over the years 
in order  to expand  the excavations. The  MSA Human Occupation Horizon excavations 
are located within, but to the right of, the main excavation area (Figure 3). 
Figure 4 is a simplified interpretation of a section of the Florisbad sedimentary deposits 
as they are today. The original Florisbad spring site was a large pan with water supplied 
mainly from the spring eyes, which surfaced along a dolerite dyke. It was in this pan that 
faunal  remains  were  fossilized  in  association with  organic­clay (Peat)  sediments  which 
developed at the bottom of the pan. Fossil remains from the lower sediments are referred 
to as the Old Collection. During the late Middle Pliocene a sand dune developed on the 
windward side of Soutpan to the north­west, growing  in size as  it migrated towards the 
spring site. Eventually its migration was restricted by a previously existing row of largely 
static  sand  dunes  lying  to  the  south  east  of  Florisbad.  This  resulted  in  a  dam  being
5 
6
7
8
formed between the leeward face of the Florisbad sand dune and the windward face of the 
static  dunes  to  the  south­east.  High  rainfall  periods  produced organic­clay  layers  while 
sandy layers were produced during drier windy periods. This  then led to the alternating 
horizontal layers of organic­clay and sand (Figure 4) building up to almost the top of the 
dune. This stratigraphy is not part of the sand dune formation, but a separate unit. When 
the water in the dam neared the top of the eastern arm of the sand dune, it broke through 
the eastern arm, resulting in a flash flood which gouged out a wide vlei to the north­east 
of the site, where much of the evacuated dam sediments were deposited (Figure 3). The 
flash flood also left a large eroded area between the spring site and the base of the south­ 
east dune belt, a large part of which was the original dam floor. 
This brief description of the Florisbad site is based entirely on the hypotheses presented 
in this study. More detailed descriptions of the geomorphology and geology are given in 
Chapter  2,  while  the  geohydrology,  sedimentation,  and  previous  hypotheses  on  the 
formation of the site are to found in Chapter 5. 
1.3  PREVIOUS SCIENTIFIC RESEARCH AT FLORISBAD 
Over the years many and diverse scientific studies have been initiated at Florisbad, many 
of them unrelated to the fossil or archaeological material. Table 1 provides a summary of 
a large portion of this work 
The first scientific interest shown in the Florisbad fossils was by Dr Robert Broom who 
published his findings in the Annals of  the South Af rican Museum (Broom, 1913). A more 
intensive research programme was initiated by Prof. T.F. Dreyer of the Grey University 
College,  later  to  become  the  University  of  the  Free  State,  at  the  request  of  the  South 
African Museum in Cape Town. Prof. Dreyer and Captain R. Egerton Helme (Anon [1], 
1980),  who  assisted  Dreyer  with  funding  during  the  1920s  (Henderson,  1992),  made  a 
small excavation at  the side of  the pool  in 1927, and  in 1928, Prof Dreyer and Miss A. 
Lyle carried out further excavations, describing these findings in 1931 (Dreyer and Lyle,
9 
Table 1. Examples of previous research carried out at Florisbad Research Station. Subject 
fields are arranged alphabetically with sources being referenced at the end of the 
table. 
Subject  Category/field  Description of research  References 
Acarology  Fossil  oribatid  (free  living)  mites  23, 24, 25, 
with a description of a new species  26. 
and genus from Florisbad. 
Archaeology  Artefacts  Descriptions  of  the  stone  12, 13, 19, 
implements  found  at  Florisbad,  37, 53, 55, 
their  occurrence  and  position  59, 60, 61, 
within the sediments.  68, 69, 74, 
89. 
Middle Stone Age  References  particular  to  the  MSA  13, 16, 51, 
period  at  Florisbad  including  the  52, 59, 60, 
artefacts  and  the  MSA  human  61, 67. 
occupation horizon. 
Archaeozoology  Taphonomy  Reconstructing of the fossil history  13, 14. 
record in determining how animals 
became part of the fossil record. 
Alceloaphini  Revision and discussions on  fossil  13, 54, 80, 
wildebeest  and  hartebeest  from  85. 
Florisbad. 
Bovidae  Description  and  discussions  on  13, 18, 86, 
fossil  antelope  and  buffalo  at  87. 
Florisbad. 
Equus (Ainus)  Discovery of an ass from southern  15. 
Africa with range extension. 
Hominidae  These  references  represent  the  22, 34, 35, 
stages  in  the  classification  of  the  36, 37, 38, 
Florisbad  skull  and  the  taxonomic  39, 40, 41, 
conclusions  made  by  the  various  46, 69, 70, 
authors.  78. 
General  Description  and  discussion  on  the  13, 14, 28, 
faunal fossils of Florisbad.  29, 42. 
Giraffidae  Description  and  discussions  on  13, 29, 79. 
fossil  giraffe  found  at  Florisbad  – 
recorded by Cooke  (1964) but not 
by Brink (1987). 
(continued…..)
10 
Table 1. (continued) 
Subject  Category/field  Description of research  References 
Hippopotamidae  Revision and discussions on  fossil  13, 50, 55. 
hippopotami found at Florisbad. 
Mustelidae  Revision and discussions on  fossil  44. 
otters found at Florisbad 
Perissodactyla  Revision and discussions on  fossil  27. 
zebras found at Florisbad. 
Suidae  Revision and discussions on  fossil  43. 
pigs found at Florisbad. 
Dating  Dating of the sedimentary deposits  7, 11, 12, 
and  fossils.  A  list  of  ages  and  17, 30, 49, 
methods  employed  are  presented  60, 62. 
in Table 4. 
Geology and  General  Discussions  on  the  geology  of  31, 45, 47, 
Geomorphology  Florisbad including minerals found  48, 56, 58, 
in  the  sediments  as  well  as  the  59, 61, 63. 
geology and geomorphology of the 
surrounding areas. 
Lacustrine  Theories on sediment deposition at  56, 57, 59, 
deposits  Florisbad by  inundation (flooding)  84. 
of the palaeo­lake (Soutpan). 
Liquefacation  Evidence  of  earthquake  induced  83. 
liquefacation in the sediments. 
Lithostratigraphy  Discussions  on  the  stratigraphy  of  2, 4, 20, 
and sedimentary  the  sediments  as  well  as  various  39, 56, 60, 
deposits  cross­sections  of  different  aspects  67, 73, 81, 
of the deposits and drilling results.  82. 
Entomology  Arachnidae  Composition  of  surface­active  64. 
spiders  determined  by  pitfall 
trapping. 
Coleoptera  Seasonal  composition  of  beetles  65 
determined by pitfall trapping. 
Formation  Various  theories  on  the  formation  2, 3, 13, 
of the Florisbad spring site and the  19, 20, 39, 
Florisbad sand dune.  39, 59, 60, 
61. 
(continued…..)
11 
Table 1. (continued) 
Subject  Category/field  Description of research  References 
Groundwater  Salinization  Description  and  analysis  of  the  2. 
processes of  ion enrichment of the 
clay  and  organic  fractions  within 
the sediments. 
Herpetofauna  Composition  of  reptiles  and  8, 9, 10, 
amphibians  at  Florisbad  and  32, 33. 
surrounding  areas  determined  by 
pitfall traps, funnel traps and other 
methods,  including  a 
biogeography. 
Ornithology  Diets  of  various  birds  relating  to  5, 8, 9, 10. 
reptile and mammal prey. 
Palaeoenvironment  Discussions  on  the  13, 19, 60, 
palaeoenvironments  that existed at  61, 75, 76, 
Florisbad.  77. 
Palynology  Coprolites  Preservation  and  interpretation  of  75, 76, 77. 
pollen in fossilized hyaena dung. 
Pollen  Discussions  on  the  role  of  pollen  75, 76, 77, 
and  plants  in  determining  the  81, 82. 
palaeoenvironments of Florisbad. 
Phytoliths  Discussions  on  the  occurrence  of  72, 76. 
fossilized  pollen  grains  including 
those  from  the  teeth  of  antelope 
species found at Florisbad. 
Spring water  Various  analyses  of  the  Florisbad  1, 2, 3, 32, 
spring­water  and  discussions  on  45, 47, 66, 
various aspects of the springs.  71. 
Wood  Discussions  and  identification  of  6, 21. 
the only  significant piece of wood 
recovered from the sediments. 
(continued…..)
12 
Table 1. (continued) 
References 
(1) Appendix 1; (2) Appendix 2; (3) Appendix 3; (4) Appendix 4; (5) Avenant (2005); 
(6) Bamford & Henderson (2003); (7) Barendsen  et  al. (1957); (8) Bates (1988a); (9) 
Bates,  (1988b);  (10)  Bates  et  al.  (1992);  (11)  Beaumont  &  Vogel  (1972);  (12) 
Beaumont  et  al. (1978); (13) Brink (1987); (14) Brink (1988); (15) Brink (1994): (16) 
Brink & Henderson (2001); (17) Broeker et  al. (1956); (18) Broom (1913); (19) Butzer 
(1984); (20) Butzer (1988); (21) Clark (1955); (22) Clarke (1985); (23) Coetzee (2001): 
(24) Coetzee (2002); (25) Coetzee (2003); (26) Coetzee and Brink (2003); (27) Cooke 
(1950);  (28)  Cooke  (1952);  (29)  Cooke  (1964);  (30)  Deacon  (1966);  (31)  De  Bruiyn 
(1971); (32) Douglas (1992); (33) Douglas (1995): (34) Drennan (1935); (35) Drennan 
(1937);  (36)  Dreyer  (1935);  (37)  Dreyer  (1936a);  (38)  Dreyer  1936b);  (39)  Dreyer 
(1938a);  (40)  Dreyer  (1938b);  (41)  Dreyer  (1947);  (42)  Dreyer  &  Lyle  (1931);  (43) 
Ewer  (1957);  (44)  Ewer  (1962);  (45)  Fourie  (1970);  (46)  Galloway  (1937);  (47) 
Grobler & Loock (1988a); (48) Grobler & Loock (1988b); (49) Grün et al. (1996); (50) 
Henderson  (1996);  (51)  Henderson  (2001a);  (52)  Henderson  (2001b);  (53)  Hoffman 
(1953); (54) Hoffman (1955); (55) Hooijer (1958); (56) Joubert (1990); (57) Joubert & 
Visser  (1991);  (58)  Joubert  et  al.  (1991);  (59) Kuman  (1989);  (60) Kuman  & Clarke 
(1986);  (61)  Kuman  et  al.  (1999);  (62)  Libby  (1954);  (63)  Loock  &  Grobler  (1988); 
(64)  Lotz  et  al.  (1991);  (65)  Louw  (1987);  (66)  Mazor  &  Verhagen  (1983);  (67) 
Meiring (1956); (68) Oakley (1954); (69) Protsch (1974); (70) Rightmire (1978); (71) 
Rindl  (1915);  (72)  Rossouw  (1996);  (73)  Rubidge  &  Brink  (1985);  (74)  Sampson 
(1974);  (75)  Scott  &  Brink  (1992);  (76)  Scott  &  Rossouw  (2005);  (77)  Scott  et  al. 
(2003); (78) Singer (1958); (79) Singer & Boné (1960); (80) Thackeray  et  al.  (1996): 
(81)  Van  Zinderen  Bakker  (1957);  (82)  Van  Zinderen  Bakker  (1989);  (83)  Visser  & 
Joubert  (1990);  (84)  Visser  &  Joubert  (1991);  (85)  Wells  (1959);  (86)  Wells  (1965); 
(87) Wells (1967); (88) Wells (1972). 
(1931).  In  1932, Prof.  Dreyer  and  Gert  Venter  uncovered  parts  of  a  hominid  skull,  the 
Florisbad skull (Anon [1] 1980). For the location of the skull see Figure 3. Prof. Dreyer 
donated his collection to the National Museum in 1936 when, due to a lack of funding, all 
excavations were halted. With funding from the then Council for Scientific and Industrial 
Research,  Dr  A.C.  Hoffman,  Director  of  the  National  Museum,  and  A.J.D.  Meiring, 
Assistant Director of  the National Museum, recommenced excavations  in 1952, but  this 
was  short­lived  as  the  then  owners  of  the  farm,  the  Sowden  brothers,  halted  all 
excavations later that year.
13 
In  order  that  the  site  be  preserved,  and  that  further  research  could  be  carried  out 
unhindered,  the  State  bought  93  ha  of  the  farm,  including  the  springs,  in  1980,  and 
handed  this  over  to  the  custody  of  the  National  Museum.  The  resort  was  then 
permanently closed to the public and new research initiatives were begun by the National 
Museum  in  1981.  Recent  work  at  Florisbad  has  related  to  extensions  to  the  previous 
excavations  and  detailed  work  on  the  MSA  occupation  horizon  (Henderson  2001a, 
2001b;  Brink  and  Henderson  2001),  as  well  as  extensive  taphonomic  studies  (Brink, 
1987). 
1.4  THE FLORISBAD FAUNAL DEPOSITS 
Two  distinct  archaeozoological  and  archaeological  deposits,  which  are  separated  from 
each other both horizontally  and  vertically,  have  been excavated at Florisbad.  Figure 4 
provides  a  generalized  cross­section  of  the  present  day  Florisbad  sedimentary  deposits 
showing the major organic­clay and sand horizons and their silt content. Further detail on 
the depositional environment and sedimentation is given in Chapter 5, subsection 5.6. 
1.4.1  The Old Collection 
The first collection is referred to as the Old Collection (Table 2; Table 3), comprising a 
basal  accumulation  of  fossilized  faunal  remains  in  the  areas  of  spring  activity,  and 
representing  a  death  assemblage  resulting  largely  from  carnivore  killings  (Brink,  1987, 
1988). Table 2 represents a list of the faunal interpretation by Brink (1987), while Table 3 
represents  changes  made  to  this  faunal  list,  by  Churchill  et  al.  (2000).  Figure  5  is  an 
example of the extinct giant buffalo,  Peloro vis  an tiquus, from the Old Collection. Bones 
in this assemblage are characterized by evidence of hyaena damage and, to a lesser extent 
by porcupine gnawing, with no indication of cut marks (Brink, 1987, 1988). Owing to the 
relatively  low  incidence  of  hyaena  damage  to  the  bones  (21.34%)  (Brink,  1987),  it  is 
postulated  here  that  the  death  of  animals  resulting  from  various  diseases  contributed 
significantly  to  the  Old  Collection.  Africa  has  a  large  number  of  bovine  and  equine
14 
15
Table  3.  Changes  and  updates  to  the  original  taxonomic  list  of  fauna  occurring  at  Florisbad 
Research Station (Brink, 1987), compiled from Churchill et al. (2000). 
Now included  Now excluded 
in Churchill  from Churchill 
et al. (2000)  et al. (2000) 
Amphibia – indet.  X 
Reptilia – indet. X 
Aves – indet.  X 
Mamalia 
Lagomorpha 
Leporidae  X 
Lepus sp.  X 
Rodentia 
Murinae – indet.  X 
Pedites capensis  X 
Pedites sp.  X 
Carnivora 
Hyaenidae – indet. (coprolites)  X 
Gallerella sanuinea  X 
Herpestes sanguineus  X 
Perissodactyla 
Equus spp. (possibly two species, E. [Asinus] 
sp. [=E. lylei], and a plains zebra 
similar to E. quagga.  X 
Equus burchelli  X 
Equus quagga  X 
Ceratotherium simum  X 
Artiodactyla 
Phacochhoerus sp.  X 
Phacochhoerus aethiopus  X 
Phacochhoerus africanus  X 
Kobus ellipsipprymnus  X 
Kobus sp.  X 
Alcelaphus buselaphus  X 
Raphicerus sp.  X
16 
17
diseases  such  as  foot  and  mouth  disease,  anthrax  and  African  horse  sickness,  some  of 
which  are  highly  contagious  and  may  reach  epidemic  proportions.  It  is  suggested  here 
that these, or similar diseases, took a significant toll on herds of animals at various times 
in  the  past,  with  sick  animals  remaining  close  to  a  water  source,  which  may  well 
explanation  the  large  number  of  faunal  remains  in  the  Old  Collection.  Should  this 
hypothesis  be  correct,  the  incidence  of  bone  damage  would  largely  reflect  hyaena 
scavenging, and not carnivore predation. The Old Collection comprises a vast number of 
faunal fossil remains comprising 26 species, and represents the Florisian Land Mammal 
Age (Hendey, 1974), or the Florisian­Cornelian Faunal Boundary, with an age of  c. 400 
ka (Klein, 1984). 
1.4.2  The MSA Occupation Horizon 
The  second  collection  is  the  MSA  human  occupation  horizon  which  occurs 
approximately 3.5 metres above basement, where species diversity is far  less than in the 
Old Collection (Table 2). Figure 6 give an  idea as  to what the MSA human occupation 
horizon excavations looked like. This horizon represents a temporary butchering site and 
has delivered artefacts  in the  form of butchering tools as well as  faunal remains (Brink, 
1987;  Henderson  2001a;  Brink  and  Henderson,  2001).  These  were  found  in  horizontal 
deposits  on  a  sandy  substrate,  which  had  been  deposited  in  an  aqueous  pan  type 
environment  with  little  disturbance  (Kuman  and  Clarke,  1986;  Kuman  et  al.,  1999; 
Henderson, 2001a; Brink and Henderson, 2001). In this assemblage, signature marks on 
bones indicate slicing, scraping and cutting, as well as bone­breaking, and show no signs 
of carnivore damage (Brink, 1987). MSA faunal remains are also more friable, in relation 
to  remains  from  the  Old  Collection  (Kuman  and  Clarke,  1986;  Brink,  1987).  It  is 
important to note here that the MSA faunal remains are not directly associated to a peat 
layer, but lie above the Peat II layer, which may explain their friability.
18 
19
1.5  THE FLORISBAD SKULL 
The  position  of  the  Florisbad  skull  in  the  evolution  of  modern  man  is  of  considerable 
importance.  Drennan  (1937)  considered  the  Florisbad  skull  (Figure  7)  to  be  the  most 
primitive,  and  therefore  the  most  important  modern  lineage  human  fossil  thus  far 
unearthed in South Africa. Dating of the Florisbad skull at 259 ±35 (Grün  et  al., 1996), 
supports a recent African origin for all modern people, and is considered the single most 
important find at Florisbad (Brink, 1987). Further to this, the skull  is the only relatively 
complete  example  of  the  late  archaic  phase  of  modern  development  in  Africa  (Brink, 
1997). Mr G. Venter discovered a hominid right upper third molar in the debris of one of 
the western spring eyes in 1932, and shortly thereafter, on 25 July 1932. Only later, did 
Prof.  T.F.  Dreyer  find  a  hominid  skull­cap  and  other  facial  bones  in  the  same  debris 
(Dreyer,  1938a;  Clarke,  1985;  Henderson,  1992).  The  remains  of  the  Florisbad  skull 
represent  but  a  portion  of  the  total  skull,  comprising  part  of  the  parietal  bone,  frontal 
bone, nasal bone, maxilla, and a tooth. The original reconstruction of the Florisbad skull 
is  given  in  Figure  7.  Dreyer  made  a  brief  announcement  of  this  find  in  The  Friend 
newspaper  dated  26  July  1932,  with  a  full  description  of  the  skull  appearing  in  The 
Friend,  dated  13  August  1932  (Henderson,  1992).  A  more  detailed  description  of  the 
skull  was  published  in  1935  (Dreyer,  1935),  placing  Florisbad  firmly  on  the 
archaeological and physical anthropology map. 
Interpretations on the taxonomic status of the Florisbad skull have been made by Dreyer 
(1935,  1936a,  1938b,  1947),  Ariëns  Kappers  (1935),  Drennan  (1935,  1937),  Galloway 
(1937),  Singer  (1958),  Wells  (1969,  1972),  Rightmire  (1976,  1978)  and  Clarke  (1985). 
With the exception of Ariëns Kappers (1935) and Wells (1969, 1972) who considered the 
skull  to  be  neoanthropic  (relating  to  more  modern  forms  of  humankind),  all  the  other 
researchers  recognized  archaic  features  of  the  skull  and  considered  it  to  be 
palaeoanthropic  (relating  to earlier  forms of  humankind). Despite general agreement on 
this particular point, there was much contention around the morphological  interpretation 
of  the skull. Dreyer, (1935) considered the Florisbad skull  to be more closely related to
20 
21
Homo  sapiens  than  H.  primigeniu s.  (250  to  20  ka),  the  latter  being  a  predecessor  of 
Neanderthal Man (130 to 24 ka). It has been hypothesised that  H.  primigenius coexisted 
with Neanderthal  man, possibly  interbreeding with Neanderthal Man  to evolve  into  the 
modern  day  H  sapiens  (Brodrick,  1960;  Pfeiffer,  1969).  It  was  also  noted  by  Dreyer 
(1935)  that  the  flange  of  the  malar  projected  outwards  and  backwards,  similar  to  the 
condition  seen  in  Bushmen  skulls.  Dreyer  (1935)  placed  the  Florisbad  skull  in  the 
Hominidae,  below  H.  sapiens  and  H.  primigeniu s,  giving  it  the  sub­genus  status  of  H. 
(Af ricanthropus)  helmei, an archaic form of  Homo  sapiens. Both Ariëns Kappers (1935) 
and  Dreyer  (1935)  agreed  that  the  Florisbad  skull  differed  from  Rhodesian  Man  (H. 
rhodesianus)  (Synonyms:  Broken  Hill  cranium;  Kabwe  cranium)  from  Zambia  (300  to 
125 ka). 
Drennan (1935, 1937) considered the skull  to be more characteristic of a “low” type of 
Homo  sapiens  represented  by  Rhodesian  Man,  and  concluded  that  it  was  an  African 
variant of the Neanderthal race with minor modern features, naming it  H.  f lorisbadensis 
(helmei). Both Dreyer (1935) and Drennan (1935) agreed that artefacts from the site were 
of  a  Mousterian  nature,  being  representative  of  a  Middle  Palaeolithic,  or  Neanderthal 
culture. Dreyer (1936a)  then proceeded to show similarities between the Florisbad skull 
and Bushmen skulls, concluding that the Florisbad skull belonged to a pre­historic race of 
Bushmen. From descriptions of the Steinheim skull, H. steinheimensi s, (250­350 ka) from 
Germany  by  Weinert  (1936),  Dreyer  (1936b)  concluded  that  the  Florisbad  skull  was 
similar,  but  a  more  primitive  member  of  the  same  race.  Drennan  (1937)  placed  the 
Florisbad  skull  before  the  Rhodesian  skull  in  the  evolutionary  series,  while  Galloway 
(1937)  placed  the  Florisbad  skull  following  the  Rhodesian  skull,  seeing  it  as  a  link 
between the Broken Hill and Boskop skulls. 
In  his  interpretation  between  the  Rhodesian  skull  and  the  Bothaville  skull  (a  modern 
European skull), Dreyer (1938b) found them to be so similar that there was no reason to 
assume that the former did not belong to  H.  sapiens. Dreyer (1938b) concluded that the 
differences  in  the  frontal  lobe  of  the  Florisbad  cast  were  so  conspicuous  that  separate 
specific  rank  must be accorded  to the Florisbad skull.  In comparing  the Maatjies River
22 
(MR) skull to that of the Florisbad skull Dreyer (1947) concluded that the Florisbad skull 
“fits almost  ideally” as a predecessor to the MR skull. In his taxonomic classification of 
South  African  skulls,  Dreyer  (1947)  further  concluded  that  the  race  to  which  he  had 
originally  assigned  the  Florisbad  skull,  should  be  maintained  as  H.  (Af ricanthropus) 
helmi,  that  the  MR  stage  should  be  assigned  to  H.  (Af ricanthropus)  dreyeri,  and  the 
Bushmen race to H. (Af ricanthropus) austroaf ricanus. 
Singer  (1958)  suggested  that  the Florisbad skull  belonged  to the Rhodesian – Saldanha 
group, while Rightmire (1976) stated that sub­Saharan archaic man could not simply be 
grouped as an African Neanderthal. As Rightmire (1978) considered the Florisbad skull 
to  be  older  than  Neanderthal,  representing  a  Middle  Pleistocene  period,  he  aligned  the 
Florisbad skull to the Broken Hill lineage. 
After  decades  of  speculation  as  to  the  taxonomic  status  of  the  Florisbad  skull,  Clarke 
(1985)  carried  out  a  further  reconstruction  of  the  cranium.  During  the  reconstruction 
Clarke  (1985)  came  to  a  number  of  conclusions,  with  the  most  significant  being  that 
Dreyer  (1935)  had  incorrectly  reconstructed  the  facial  bones  in  three  major  areas.  As 
there were no positive  joints between any of  the  facial  bones, Clarke  (1985) concluded 
that the plaster reconstruction between the bones was educated conjecture, and therefore 
any  facial  measurements  would  be  meaningless  (Clarke,  1985).  Clarke  (1985)  further 
concluded that the Florisbad skull had a much more archaic and  larger appearance than 
that of Dreyer’s reconstruction, and considered it to have strong similarities to the Broken 
Hill cranium. However, Clarke (1985) also noted that subsequent interpretations based on 
the Dreyer’s (1935) reconstruction by authors such as, Drennan (1935), Galloway (1937), 
and Rightmire (1978), were unreliable. 
In 1997 Foley and Lahr (1998) resurrected the name  H.  helmei based on mode 3/Middle 
Palaeolithic industries and archaeological remains, which resulted in the establishment of 
the  immediate  ancestors  to  the  Neanderthals.  This  classification  of  the  Florisbad  skull 
appears  to have  been accepted by a  number of researchers  (Deacon and  Deacon, 1999; 
Kuman et al., 1999) as an archaic form of H. sapie ns. However, from an anatomical point
23 
of view there still remain many questions. Lieberman  et  al. (2002) placed  H.  helmei in a 
group of recent, or modern, human fossil records that has a confusing pattern of variation, 
with  many  vaguely  defined  taxa  which  are  not  widely  accepted.  This  confusion, 
Lieberman  et  al.  (2002)  stated  was  born  out  of  a  lack  of  established  unique  derived 
features  (autapomorphies) between anatomical  modern  H.  sapiens  (AMHS) and archaic 
Homo  species  (AH).  White  et  al.  (2003)  felt  that  in  addition  to  the  difficulties  in 
partitioning lineages, many of the available species names were based on inadequate type 
specimens,  such  as  H.  helmei,  and  a  lack  of  substantial  and  accurately  dated  hominid 
fossils  between  300  and  100  ka.  Greater  accuracy  and  detail  is  important  in  light  of 
hypotheses  like  the  one  proposed  by  Lahr  and  Foley  (1998)  which  propose  that. 
Neanderthal  and  modern  lineages  share  a  common  ancestor  in  an  African  population 
between 350 and 250 ka ago. In summary,  it would appear  that the incomplete cranium 
and  lack of an entire brain case  in the  H.  helmei  type specimen is a major  factor  in not 
being  able  to  pin  point  its  precise  taxonomic  lineage.  Therefore,  unless  further,  or 
complete, H. helmei skulls are found, the H. helmei question may never be truly resolved. 
1.6  DATING OF THE FLORISBAD DEPOSITS 
Dating of the Florisbad deposits has been an ongoing exercise over decades, with the age 
of the Florisbad skull being a matter of conjecture until it was ESR dated in 1996 at 259 
±35 ka (Grün  et  al., 1996). It will be noted in Table 4 that 14 C dating ages of the lower 
deposit ages are usually greater than the upper limit of the dating method which was used 
at the time, and therefore these ages are very inaccurate: the true ages being much higher 
than the greater­than­ages given. Laboratory codes for the 14C   dates are given in Table 4. 
However,  the  greater­than  ages  may  also  have  been  as  a  result  of  there  being  no 
measurable radioactivity  in the  lower Florisbad sediments (Libby, 1954; Oakley, 1955). 
Greater detail on the modes of fossilization of the faunal remains and details on the dating 
are discussed in Chapter 5: subsection 5.6.4 and Chapter 6 respectively. Table 4 provides 
a history of the dating of the Florisbad sediments compiled from Deacon (1966), Kuman 
and Clarke (1986), and Grün et al. (1996). More detail on dating and the influences of the 
environment on dating is given in Chapter 6.
24 
Table 4. History of the dating of the Florisbad deposits. 
Stratigraphic  Age  Sample Ref.  Method  Reference 
Layer/ Item  yr  number 
Peat IV  3 530 ±80  Pta­1128  14C    Kuman & Clarke 1986 
3 550 ±60  Pta­3617  14C    Kuman & Clarke 1986 
3 580 ±60  Pta­3631  14C    Kuman & Clarke 1986 
Sand above Peat III  4 370 ±70  Pta­1127  14C    Kuman & Clarke 1986 
10 000 ±100  Pta­1125  14C    Kuman & Clarke 1986 
11 700 ±110  Pta­3609  14C    Kuman & Clarke 1986 
Peat III  6 700 ±500  C­852  14C    Libby 1954 
19 530 ±650  L­271D  14C    Broecker  et al. 1956 
White sand below Peat  >44 700  Pta­3465  14C    Kuman & Clarke 1986 
III 
Organic layer above  >47 200  Pta­3623  14C    Kuman & Clarke 1986 
Peat II 
White sand above Peat  146 000 ±15 000  OSL  Grün et al. 1996 
II  128 000 ±22 000 
133 000 ±31 000 
MSA Human  121 000 ±6 000  ESR  Grün et al. 1996 
Occupation horizon 
Peat II  9 104 ±420  C­851  14C    Libby 1954 
28 450 ±2 200  L­271C  14C    Broecker  et al. 1956 
>42 800  Pta­1108  14C    Beaumont  &  Vogel 
1972 
Olive green sand above  157 000 ±21000  OSL  Grün et al. 1996 
Peat I 
Peat I  >35 000  L­271B  14C    Broecker  et al. 1956 
>41 000  C­850  14C    Libby 1954 
>44 000  Y­103  14C    Barendsen et al. 1957 
>48 100  GrN­4208  14C    Beaumont  &  Vogel 
1972 
281 000 ±73 000  OSL  Grün et al. 1996 
Hominid tooth  259 000 ±35 000  ESR  Grün et al. 1996 
Basal brown sand layer  279 000 ± 47 000  OSL  Grün et al. 1996 
14 C = Radiocarbon dating 
ESR = Electron Spin Resonance dating 
OSL = Optically Stimulated Luminescence
25 
1.7  RATIONALE FOR THE STUDY 
The research reported on here originated in 1988 when the ecology and biogeography of 
the  herpetofauna  of  Florisbad  was  examined  (Douglas,  1992,  1995).  Amphibians  were 
found to occur in most of  the waters related to the Florisbad spring site. In  light of  this 
finding,  the  water  quality  of  the  spring,  vlei,  an  exploration  pit  that  amphibians  were 
inhabiting,  the  farm  dam,  and  Soutpan  (Figure  2),  were  analysed,  and  the  results 
examined  in  light  of  the  water  bodies  being  suitable  as  habitats  and  breeding 
environments for amphibians (Douglas, 1992, 1995). A re­examination of the 1988 water 
analysis results by the author indicated that the historically  low Ca concentration  in the 
spring water did not support theories in the literature (Brink, 1987) that faunal remains at 
the  site  had  become  fossilized  in  a  spring  context.  Further  to  this,  the  TDS  (total 
dissolved  solids)  of  the  water  in  the  exploration  pit  was  27%  higher  than  that  of  the 
spring­water, only 22m away, and this disparity was questioned (Douglas, 1992, 1995). 
This disparity in the water quality was not identified, or questioned, in the original study 
due  to  these  waters  only  being  examined  in  relation  to  their  suitability  as  habitats  for 
amphibians. These  two aspects of  the water quality  led  to a number of questions being 
asked in four main areas, for which no answers could be found: 
•  Firstly: was there actually a disparity between the spring­water and the excavation 
pit water, or was  this an  isolated occurrence peculiar  to the sampling period?  A 
fairly  accurate  assessment  of  the  spring  water  quality  could  be  made  from 
sporadic water sampling over the past 84 years (Douglas, 1992), but no sampling 
of the excavation pit waters had been carried out, and therefore no information on 
the quality of the excavation pit waters was available. 
•  Secondly:  should  there  be  a  disparity  between  the  quality  of  the  spring­  and 
excavation pit waters, questions arose as to why there was a difference, what was 
causing  the  difference,  and  what  processes  and  factors  were  involved. 
Furthermore, in light of the low spring­water Ca content, which aspect of the site
26 
had sufficient mineralization  for fossilization of  the  faunal remains,  if  it was not 
the spring­water, from where did these minerals originate? 
•  Thirdly: in light of the low spring­water Ca content the assumption was made that 
the spring­water was not responsible for fossilization. If this were the case, where 
was  fossilization  taking place,  how was  it  taking place,  and what processes and 
factors were involved? 
•  Fourthly:  Existing  hypotheses  on  the  formation  of  the  Florisbad  site  could  not 
account for a number of unanswered questions. These included the absence of the 
upper  red­brown  sand  units  on  the  eastern  side  of  the  “mound”  (Rubidge  and 
Brink,  1985;  Brink  1987),  the  seven  metres  of  clay  deposits  in  the  vlei  area 
(Butzer 1984), the size of  the vlei  in relation to the ephemeral drainage  line,  the 
erosion  of  the  east  bank  of  the  vlei,  the  two  almost  right  angle  dog  legs  in  the 
current ephemeral drainage  line,  and  the pinching out and height of  the Peat  IV 
layer on the west wall of the excavations. 
Dating of the Florisbad deposits were a significant factor in the literature review leading 
up to these  investigations. However,  illustration (a)  in Grün  et  al.  (1996) suggested that 
when  the  depth  of  the  sediments  was  plotted  against  time,  this  would  reflect  equal 
compaction  and  compression  of  the  sediments  at  all  depths,  which  appeared  to  be 
illogical. It was hypothesised that the application of mathematical trend lines to depth/age 
plots would either confirm or disprove  this  theory.  It was  further hypothesised  that  the 
application  of  a  linear  trend  line  would  reflect  equal  compaction  over  depth  and  time, 
while  logarithmic  and exponential  trend  lines would  reflect a  more  logical  and gradual 
degree of compaction over depth and time. It was hoped that from the results it would be 
possible  to  verify,  or  disprove,  the  dates of  depths  given  in  illustration  (a),  Grün  et  al. 
(1996), as well as the dates given by Grün et al. (1996) for the Middle Stone Age Horizon 
located  in the middle of  the profile, and the much older  lower sediments,  located at the 
bottom of the profile.
27 
These  questions  then  provided  the  motivation  for  further  investigations.  It  should  be 
mentioned  that  since  the  publication  of  the  results  of  the  study  in  Appendices  I­IV, 
certain aspects of  the study have been revisited, and the hypotheses updated. Therefore, 
the hypotheses, illustrations, discussions, and conclusions, presented in this main body of 
the thesis, should they differ from those in the Appendices, should be considered as being 
the more correct. 
1.8  KEY AREAS TO BE ADDRESSED 
•  An  examination  the  physical  environment  in  order  to  determine  the  many 
components  which  may  have  contributed  to  the  existence  and  formation  of  the 
site, as well as contributing to an environment suitable for fossilization. 
•  Determine  whether  there  is  a  difference  between  the  quality  of  the  spring­  and 
subterranean waters away from the spring eyes, and determine to what extent the 
quality of the two entities may differ. 
•  Determine the degree of salinization of the organic­clay layers, and relate this to 
the degree of mineralization of the spring­ and groundwater. 
•  Determine the properties of  the organic­clay  layers  in relation to salinization, as 
well  as  possible  processes  and  factors  within  the  environment  which  would 
contribute to this salinization. 
•  Correlate the above information in order to determine whether conditions existed 
within the organic –clay layers for the initiation of fossilization. 
•  Determine whether  fossilization was capable of  taking place within the confines 
of the spring vents and eyes. 
•  Develop a theory, based on existing morphological and chemical evidence, for the 
formation of  the  spring  site  that would accommodate the geomorphology of  the 
site, salinization of the organic clay layers, and the fossilization of faunal remains. 
•  Develop a method of evaluating the variability of the ages and depths of published 
sedimentary profiles and their fossil components.
28 
1.9  SYNOPSIS 
The importance and uniqueness of the Florisbad spring site as an archaeozoological and 
archaeological site has been briefly illustrated in this chapter. Brouwer (1967) stated that 
many  detailed  palaeontological  descriptions  make  no  mention  at  all  of  the  conditions 
under  which  fossilisation  has  taken  place.  Brink  (1987)  further  stated  that,  to  fully 
understand  the  fossil  fauna  it  is  important  that  the  palaeoenvironment  from  which  the 
fossil  remains  were  extracted,  is  better  known.  Even  today,  the  priority  appears  to  be 
largely centred on the collection and description of archaeozoological evidence, with only 
modest  attention  being  paid  to  palaeo­, or  current,  environmental  and  chemical  factors 
and  processes,  which  resulted  in  the  fossilization.  In  other  words,  a  more  holistic 
approach  needs  to  be  taken  when  examining  a  site  such  as  the  Florisbad  spring  site 
because  fossilization,  for  example,  cannot  be  simply  explained,  or  determined,  by  its 
independent  component  parts.  In  the  following  chapters  it  will  be  seen  how  the  inter­ 
relationship and behavioural variability of these component parts can affect the system as 
a whole.
29 
 
 
 
 
Chapter 
 
2 
 
 
 
 
 
The Physical 
Environment 
 
Part I 
 
Geology and 
Geomorphology 
 
 30
CHAPTER 2 
 
THE PHYSICAL ENVIRONMENT 
 
PART I 
 
GEOLOGY AND GEOMORPHOLOGY 
 
 
2.1 INTRODUCTION 
 
The geohydrological and surface processes controlling the depositional environment at 
Florisbad, and the resultant fossilization of faunal remains, fall under the influence of the 
physical environment, whether it is the influences of geological, geohydrological, 
depositional, climatic, vegetation, mechanical, or chemical components. 
 
The importance and influence of the physical environment on the existence and formation 
of the Florisbad spring site cannot be over emphasised. It is therefore important to 
examine, in some detail, as many aspects of the physical environment as possible in order 
to present a holistic interpretation of the Florisbad entity. For example, the geology of the 
area may influence soil types, and in conjunction with climate, will determine vegetation. 
Further more, geology, along with many factors, will ultimately determine the formation 
of aquifers and the flow of groundwater. To this end, the physical environment has been 
divided and grouped into three sections, namely, geology and geomorphology, vegetation 
and climate, and geohydrology. In Part I of the physical environment the regional 
geology is examined in order that the local geology, which is an integral part of the 
greater Karoo Basin, can be put into context. From a geomorphological perspective, the 
Western Free State panveld, and Soutpan, which lies north-west of the Florisbad spring 
site, as well as the Florisbad sand dune, which has resulted in the spring sites modern day 
morphology, have both had major influences on the depositional history of the site. 
 
 31
2.2 REGIONAL GEOLOGY 
 
The Free State lies close to the geographic centre of the southern African subcontinent 
and has been less strongly influenced and modified by the breaking up of Gondwanaland 
than the margins (Moon and Dardis, 1988). The Archaen/Precambrian geology, which is 
comprised of formations of the Witwatersrand (3000-2800 Ma) and Ventersdorp (2800-
2650 Ma) Supergroups, and includes the Central Rand and Dominion Groups (3100-3050 
Ma), are important in that they are the basis of the Free State gold mining industry 
(Holmes and Barker, 2006). The Precambrian geology is in turn blanketed almost entirely 
by rocks of the Middle Jurassic to Late Carboniferous (300-179 Ma) Karoo Supergroup. 
This Group comprises the Dwyka (300-289 Ma), Ecca (289-255 Ma), Beaufort (255-237 
Ma), Stormberg Groups (230-183 Ma), and the Drakensberg Groups (183-279 Ma) 
(Catuneanu et al., 2005), with Florisbad lying within the Ecca Group (Figure 8). Karoo 
Supergroup rocks cover approximately two-thirds of the South African land surface (see 
2.1.13), and are important for their coal bearing seams in the shallower northern parts of 
the Karoo Basin (McCarthy and Rubidge, 2005). Karoo rocks are also recognised 
globally for their wealth of fossil tetrapods, which span the stratigraphic record from the 
Mid-Permian to the Early Jurassic (Rubidge, 2005). 
 
Classification, descriptions, ages, Precambrian geology, and the Karoo Supergroup 
members have been sourced from Keyser, (1997), Bisschoff, and Mayer (1999), Anon [2] 
(2001), Vegter, (2001), Chevallier and Woodford (1999), Woodford and Chevallier 
(2002), Baran (2003), Grandstein and Ogg (2004), Neveling (2004), Catuneanu et al. 
(2005), McCarthy and Rubidge (2005), Rubidge (2005), and Holmes and Barker (2006), 
unless otherwise referenced. 
 
2.2.1 Precambrian Strata 
 
The Precambrian geology, although poorly represented in the Free State, forms a part of 
the regional geology. Precambrian strata is largely restricted to the northern Free State at  
 
 32
 
 33
the site of the Vredefort meteoric impact site with the oldest rocks of the area being 
Inlandsee gneiss and Parys granite of Swazian age (>3100 Ma). The Vredefort structure 
(c. 2023 Ma) (Figure 8) is the largest verified, and oldest, clearly visible meteorite impact 
structure on Earth and was declared a World Heritage Site by the United Nations 
Educational, Scientific and Cultural Organization on 14 July 2005 (Anon [3], 2006; Anon 
[4], 2007). The Vredefort structure was formed by the impact of a meteorite estimated to 
have been at least 10 km in diameter (Mayer, 2007). It is estimated that the impact 
vaporized some 70 cubic kilometres of rock (Anon [3], 2006) at temperatures of up to 
2000 °C (Mayer, 2007.), creating a crater ±250 km in diameter (Anon [3), 2006). 
 
In order to view the Vredefort structure in a global context, the following craters may be 
considered important. The Sudbury impact structure in Canada rates as the second largest 
impact site with a diameter of 200 km (Anon [3], 2006). The Morokweng impact crater of 
South Africa, is >70 km in diameter, and the second largest crater in Africa, estimated to 
be 145 Ma, which coincided with the Jurassic-Cretaceous boundary (Maier et al., 2006). 
It is not visible from the surface as it buried beneath a layer of Kalahari sand (Maier et 
al., 2006). With a diameter of 16 km, the Suavjärvi crater in Russia is recognised as the 
oldest impact crater at 2400 Ma, but due to regional metamorphism and weathering, it is 
no longer visibly recognisable as an impact structure (Mashchak and Naumov, 1996). 
There appears to be some confusion over the use of the term Vredefort “Dome” (Figure 
8). The Vredefort Dome represents the central uplift of a very large impact structure that 
exposed nearly a complete cross-section through the continental crust of 25-30 km thick 
(Martini, 1992). This uplift and overturn of Precambrian strata was created when a high-
speed primary shock-wave was sent vertically down into the earth, followed almost 
instantaneously by a rebound shock-wave from inside the earth (Mayer, 2007). This 
caused a central plug in the floor of the crater to undergo massive uplift and overturn 
(Mayer, 2007). The overturn at Vredefort has resulted in the deeper Precambrian rock 
layers of the Witwatersrand Supergroup being exposed almost vertically on the surface, 
with a dip of 100° to 110°, or 70° to 80° inwards (Hart et al., 1990, 1991, 1995; Tredoux, 
 34
et al., 1999). The core of the structure comprises granite-gneiss, surrounded by a collar of 
sediments and lavas, which become progressively younger towards the outer perimeter 
(Hart et al., 1990, 1991, 1995; Tredoux, et al., 1999; Bisschoff and Mayer, 1999).The 
Vredefort Dome lies in the central part of the Witwatersrand Basin and is the type locality 
for pseudotachylite (Eiko, 2001). Pseudotachylite (Shand, 1916) is a dark glassy, grey, 
granite, formed by frictional heating, and subsequent melting, along oblique impact 
surfaces during high speed faulting as a result of the meteoric impact (Nakamura, 2003). 
 
Further to the above, the only other occurrence of Precambrian strata is the Allanridge 
Formation of the Ventersdorp Supergroup which encroaches over the western border of 
the Free State in a few small areas north-west of Bothaville, north-west of Jacobsdal, 
west of Hertzogville, and at Allanridge. 
 
2.2.2 The Karoo Basin 
 
Because Florisbad lies within the Karoo Basin, and because the Karoo Basin is the 
dominant geological structure of the South African landscape, covering approximately 
two-thirds of the land surface, this is briefly discussed. 
 
While the Karoo Basin (Figures 9a, 9b) is a unique structure in itself, Karoo-age basins, 
which show clear similarities to the South African Karoo Basin, occur across Africa 
(Catuneanu et al., 2005). The Karoo Basin is a retro-arc foreland basin. Catuneanu et al. 
(2005) noted that two factors were responsible for the formation of the south-central 
African Karoo Basins during the Late Paleozoic-Early Mesozoic period when the Pangea 
supercontinent reached its maximum extent. 
 
Primary control was by tectonic mechanisms, with subsidence mechanisms ranging from 
flexural in the south, to extensional in the north, and propagating southwards from the 
divergent Tethyan margin (Catuneanu et al., 2005). These tectonic mechanisms, 
combined with influences exerted by the inherent structures of the underlying 
 
 35
 
 
 36
Precambrian basement, resulted in discrete depozones, which were further influenced by 
shifts in climatic regimes. During the Karoo interval, shifts in climatic and tectonic 
conditions from the northern and southern margins of the African continent resulted in 
the lithostratigaphic character of the Karoo Supergroup changing significantly across the 
continent. Karoo basins sensu stricto in Africa are restricted to Africa south of the 
equator, while Karoo-age successions preserved north of the equator are distinctly 
different (Catuneanu et al., 2005). Woodford and Chevallier (2002) further related 
geomorphology of the subcontinent to the African surface (±40 ma), which occurred 
before the African Plate came to rest over the mantle. This is reflected in the topography 
of the central part of the Karoo Basin when the Great Escarpment, formed during the 
Jurassic, was still a prominent feature (Woodford and Chevallier, 2002). Topography was 
further influenced by the post-African I surface (±30 ma) when there was an uplifting 
centred below Lesotho, resulting in the reactivation of the Great Escarpment (Woodford 
and Chevallier, 2002). The latter is reflected in the peripheral topography of the Karoo 
Basin. 
 
At this time, South Africa was located over the South Pole. The subduction zone 
produced volcanic activity, which helped to increase the size of the developing Cape Fold 
Mountain range to the extent that the weight of the mountains caused a sagging of the 
lithosphere to the north. As this basin, created by this sagging, drifted northwards and 
away from the polar latitudes, the ice sheets melted (Hancox and Rubidge, 2002), giving 
rise to a large inland sea referred to as the Karoo Sea. The deepest part of the basin lies to 
the south in the foredeep on the landward side of the Cape Fold Belt, where it is 12 km 
deep, gradually becoming shallower to the north (Woodford and Chevallier, 2002) 
(Figure 9b). It was in this basin that sedimentation cycles would deposit the sediments, 
which were later to form the rocks of the Karoo Supergroup (McCarthy and Rubidge, 
2005). 
 
 
 
 
 37
2.2.3 The Karoo Supergroup 
 
The Karoo Supergroup comprises a number of Groups and Formations deposited within 
the Karoo Basin through progressively changing depositional environments during the 
Late Carboniferous (310 Ma) through to the Mid Jurassic (185 Ma). In order to 
understand the position of the Ecca Group, in which Florisbad lies, within the Karoo 
Supergroup, a brief discussion on the Groups and Formations is presented. 
 
2.2.3.1 Dwyka Group 
 
Dwyka Group rocks comprise glacially derived tillites deposited in a marine environment 
and consists mainly of diamictite that grades upwards into conglomerates, mudstone and 
shale. Lithological differences occur within this Group over the Basin. The detection of 
these glacial sediments in India, and South America, provided early evidence in support 
of the Theory of Continental Drift (McCarthy and Rubidge, 2005). Sedimentation during 
the Dwyka period was so considerable that in the southern part of the Karoo Sea, the 
sediments began to form vast tracts of land. The Dwyka Group covers a time span of 300-
289 Ma, extending from the early to Late Carboniferous to the early Permian (Catuneanu 
et al., 2005). Some authors have indicated a swathe of Dwyka Group rocks along the 
western boundary of the Free State (De Bruiyn, 1971; De Waal, 1978), while others have 
included the Ecca and Dwyka Groups under a single category (Earlé and Grobler, 1987). 
By implication, the combining of the Ecca and Dwyka Groups over the Free State implies 
that Dwyka Group rocks do occur in the Free State. The references used in Figure 8 all 
indicated that no Dwyka Group rocks occur in the Free State, with the nearest location of 
these rock types being west of Kimberley, and west of Christiana in the North West 
Province. This Group forms the base of the Karoo Supergroup with a thickness of up to 
330 m, and is of little significance in terms of the geomorphology of the Free State 
landscape. 
 
 
 
 38
2.2.3.2 Ecca Group 
 
The Ecca Group comprises 16 formations of which four dominate in the Free State, 
namely the Volksrust, Vryheid, Tierberg and Prince Albert Formations. These 
sedimentary rocks consist of muds, silts and other deltatic sediments, accumulated under 
brackish and fresh water conditions from the drainage of large swampy deltas located 
along the northern shores of the basin, into the Karoo Sea (McCarthy and Rubidge, 
2005). The Ecca Group shales were formed in a shallow intracratonic depression, 
probably the result of preceding Dwyka glaciation. 
 
The rocks of the Vryheid Formation in the extreme western Free State are of a cyclic 
deltaic and fluvial origin, and are more arenaceous than the predominantly argillaceous 
Volksrust formation. This formation comprises thick mudstone and fine- to coarse- 
sandstone, with secondary shale layers with a thickness of up to 500 m. Being of marine 
origin, the Volksrust Formation comprises shale, siltstone, mudstone and fine sandstone 
at the top, and is more argillaceous than the Vryheid Formation. Loock and Grobler 
(1988) state that Florisbad is underlain by Tierberg Formations The Tierberg Formation 
largely comprises well laminated, dark grey to black shale, which has been described by 
Nolte (1995) as a lateral equivalent of the Volksrust Formation mapped further to the 
east. Catuneanu et al. (2005) indicate that the contact lies north of Florisbad, while 
Villjoen (2005) noted that the vertical division between the two formations occurs in the 
Boshfof-Hertzogville area, where the Whitehill Formation pinches out. Viljoen (2005) 
indicates that this contact zone may pass very close to, if not beneath Florisbad (Figure 
8). 
 
The Volksrust Formation, with westerly thickness of up to 380 m, extends from the 
Florisbad area, north, in an arc through to the north-eastern Free State, into Gauteng, 
Mpumalanga, and down into KwaZulu-Natal. On the other hand, the Tierberg Formation, 
with a thickness of up to 7700 m, extends from the vicinity of Florisbad, south into, and 
across, the Northern Cape Province, then southward again into the Western Cape towards 
Matjiesfontein. The Ecca Group is well represented in the western Free State, and has 
 39
weathered and eroded to produce a flat to undulating landscape, broken by flat-topped 
dolerite capped mesas, which will be discussed later. Coal is a major economic deposit 
within the Ecca Group. The Ecca Group covers a time span of 289-255 Ma, extending 
from the late to mid Permian period. 
 
2.2.3.3 Beaufort Group 
 
These rocks formed from fluvial and deltaic derived sediments which dominate the 
central and eastern Free State. As deltas to the south gradually built up, the Karoo Sea 
shrank to become a lake, with this transition to more terrestrial environments signalling 
the boundary between the Ecca and Beaufort Groups (McCarthy and Rubidge, 2005). 
Beaufort Group rocks were deposited by north-flowing, meandering rivers, in which sand 
accumulated, flanked by large floodplains where periodic flooding deposited mud 
(McCarthy and Rubidge, 2005). This environment provided an ideal habitat for the 
diversification and evolution of early reptiles, in fact, to such an extent that therapsid 
fossils have been used to biostratigraphically subdivide the rocks of the Beaufort Group 
(Kitching, 1977; Keyser and Smith, 1978; Rubidge et al. 1995). Within this Group, the 
older Adelaide Subgroup, comprising shale, siltstone, and fine sandstone, with a 
thickness of up to 5000 m, dominates in the west. In the east, and the younger Tarkastad 
Subgroup, comprising mudstone and sandstone, with a thickness of up to 2000 m, 
dominates in the east. Some areas may record a 70% predominance of sandstone. 
Weathering and erosion is similar to that of the Ecca Group shale, but due to the higher 
occurrence of dolerite (see below) in the central and eastern Free State, the relief is not as 
flat, with a higher occurrence of dolerite capped mesas. Depositional environments 
indicate a progressive desiccation up the sequence. The Beaufort Group covers a time 
span of 255-237 Ma, extending from the late Permian to the mid to early Triassic. 
 
2.2.3.4 Stormberg Group 
 
Some controversy has surrounded the classification of the Drakensberg and Stormberg 
Groups over the years, resulting in various classifications. For example, Rubidge (2005) 
 40
initially referred to the Drakensberg and Stormberg Groups, while later in the paper the 
Drakensberg Group was referred to as the Drakensberg Formation, and was included 
under the Stormberg Group. Neveling (2004) classified the Molteno, Elliot and Clarens 
Formations as independent formations, omitting the encompassing Stormberg Group 
term, while maintaining the succeeding Drakensberg Group. Catuneanu et al. (2005) 
grouped the Molteno, Elliot and Clarens sedimentary formations under the Stormberg 
Group, while also maintaining the succeeding Drakensberg Group. The reasoning for this 
latter decision was that the angular unconformity of the base-Molteno indicated a 
significant tectonic event across the region, which heralded the Stormberg sedimentation 
period. The classification by Catuneanu et al. (2005) as been adopted in this thesis. 
 
The Stormberg Group comprises three sedimentary formations and covers a time span 
from 230-183 Ma, extending from the mid Triassic to late Jurassic periods. The Molteno 
Formation (230-216 Ma), was deposited largely by shallow braided rivers (Hancox, 
2000), and hosts an abundance of insect, fish and sea-fern fossils (Anderson et al., 1999). 
As indicated by Catuneanu et al. (2005), the Molteno Formation does not lie directly on 
the Burgersdorp Formation of the upper Beaufort Group, with a ±7 Ma unconformable 
stratigraphic hiatus between the two. Molteno Formation rocks are of fluvial origin, 
dominated by sandstone (70%) and mudstone (20 %), with a thickness of up to 70 m. 
 
Rocks of the Elliot Formation (215-203 Ma) were deposited in drier conditions, with 
loess type aeolian sedimentation of mudstone and siltstone, and fluvial subordinate 
sandstone (Baran, 2003; Hancox and Rubidge, 2002). This formation, with a thickness of 
up to 150 m, preserves a large variety of reptile fossils including the oldest known 
tortoise from Africa (Gaffeny and Kitching, 1994). These deposits also include some of 
the earliest mammals (Gow, 1986). The Clarens Formation (203-183 Ma) was formed in 
arid conditions and is of aeolian origin comprising sandstone layers derived from sand 
dune deposits, which may attain a thickness of up to 230 m. These sandstones (66%) have 
weathered to produce spectacular vertical cliffs, while differential weathering has 
produced overhangs and caves. Small quantities of mudstone and siltstone also occur. 
The Stormberg Group reflects a gradual increase to more arid conditions, sequentially 
 41
recorded in the rocks, which make up this group (McCarthy and Rubidge, 2005). 
Towards the end of the deposition of the Elliot Formation, these rocks indicate desert 
conditions. The rocks of the upper Clarens Formation attests to true desert conditions and 
a situation similar to that of the Namib Desert (McCarthy and Rubidge, 2005). 
 
2.2.3.5 Drakensberg Group 
 
This Group comprises horizontally stratified basaltic lavas, including numerous flows of 
up to 50 m thick, contributing to a total thickness of 1400 m. When compression of the 
Karoo Basin was relaxed, due to a drop in sedimentation rate, the Earth’s crust erupted 
along fissures, spreading basaltic lave across the Clarens desert over most of South Africa 
(McCarthy and Rubidge, 2005). In the eastern Free State, these basalts only remain as 
remnant capping overlying the Clarens Formation at the highest elevations. This volcanic 
event signalled the end of Karoo Supergroup sedimentation and the beginning of the 
fragmentation and dispersal of Godwana into the continents, as we know them today 
(McCarthy and Rubidge, 2005). Much of this magma was injected into the horizontal 
sedimentary layers of the Karoo Supergroup, where upon cooling, crystallized to form 
dolerite sills. The Stormberg Group covers a short time span from 183-179 Ma in the late 
to mid Jurassic. 
 
2.2.4 Karoo Dolerite Suite 
 
The Karoo dolerite suite represents a post Karoo sedimentation, Jurassic period, of 
magmatic intrusions, which are younger than the basalts of the Drakensberg Group 
(Vegter, 2001). This extrusion of basalt took place over the entire African subcontinent, 
making this event one of the largest flood-basalt outpourings in the world (Chevallier et 
al., 2001). Dolerite dykes and sills are important in the hydrogeology of the Karoo 
Supergroup as their intrusion resulted in the creation of fracture zones within the host 
rocks and themselves, forming an important source of underground water (Vivier, 1996; 
Baran, 2003)). Most of the second-order geomorphological features and drainage systems 
of the main Karoo Basin are controlled by dolerite dykes, sills and ring-complexes 
 42
(Chevallier et al., 2001) Dolerite intrusions have had a number of effects on the host 
rocks. These include the metamorphosis of the host rock due to contact metamorphism, as 
well as mechanical deformation of the host rocks as in dilation and bending, which in 
turn may result in fracturing (Vivier, 1996). The suite can be divided into two basic 
categories. 
 
2.2.4.1 Dolerite Sills and Ring-Complexes 
 
These formations are one of the most are the most common type on intrusion in the 
Karoo Basin, and one of the most prominent features of the Karoo landscape (Woodford 
and Chevallier, 2002). Their distribution is the same as that of the dolerite dykes, with 
their emplacement being strongly controlled by the lithology of the country rock 
(Chevallier et al., 2001; Woodford and Chevallier, 2002). Dolerite sills are mostly 
associated with the argillaceous, lower units of the Karoo Supergroup (Baran, 2003). 
They re also preferentially associated with the contact zones of the Dwyka- Ecca Group, 
the Prince Albert- White Hill Formation, the Upper Ecca- Lower Beaufort Group, and 
other lithological boundaries within the Beaufort Group (Woodford and Chevallier, 
2002). Dolerite sills are sheet-like structures exhibiting a saucer-like morphology (Figure 
10 and 11) that follows the bedding planes of Karoo structures, with a thickness of 
between 15 and 300 m, and the outer sill having a diameter of from 10 to 50 km (Vivier, 
1996; Baran, 2003; Chevallier et al., 2001; Woodford and Chevallier, 2002). Dolerite 
sills form large coalescing circular, oval, or kidney-shaped structural units, with each unit 
being composed of several sub-units of smaller size, which are in turn comprise even 
smaller units (Chevallier et al., 2001; Chevallier and Woodford, 2002) (Figure 10). A 
number of other morpho-tectonic emplacement models have been proposed over the 
years by researchers such as Rogers and Schwarz (1902), Du Toit (1905), Du Toit (1920), 
Lombard (1952), Johnson and Pollard (1973), Meyboom and Wallace (1978), Burger et 
al. (1981), Kattenhorn (1994), Vivier et al., (1995), and Vivier, (1996). Structurally 
dolerite sills are extremely complex with an intricate connectivity, and may be composed 
of many linear to curvi-linear segments. 
 
 43
 
 44
Chevallier et al (2001) proposed three morpho-tectonic models for dolerite sill and ring-
complexes, with Woodford and Chevalier (2002) presented two mechanisms for the 
emplacement of Karoo dolerite sills: 
 
(a) Sills in the northern Karoo have a form similar to that of a laccolith, with the feeder 
to the laccolith being a central dyke, with a thickening from the outer rim towards 
the centre of the structure. The rings are seen as peripheral offshoots formed as a 
result of the warping of the overlying host rock. Vivier (1996) stated that the arms 
of some laccoliths could be interpreted as sills if the arms of a laccolith were long 
enough, or, that some sills were extensions of laccoliths. 
 
(b) In the western Karoo it was proposed a feeding system of magma along the inclined 
sheet, or the ring itself, using a coalescing ring-dyke network. The 60° inward-
dipping inclined sheet therefore changes into an upper outer sill, feeding a lower 
inner sill at the same time. 
 
The geometry of larger Karoo ring-complexes can be expanded on from (b) above in that 
there are four distinctive morpho-tectonic units associated with ring-complexes (Figure 
10). A flat inner sill forming the bottom of the saucer with a thickness of between 30 to 
60 m; a flat-lying outer sill exhibiting extensive fracturing and joining with a thickness of 
between 50 and 100 m, that can extend for hundred of kilometres; a peripheral inclined 
sheet, where dependant on the area, may dip at angles of between 30° to 80°, and attain a 
thickness of between 20 and 150 m: and abundant feeder dykes which cut through, into, 
or out, of the sills and ring-complexes. 
 
In Figure 11, the three main types of complex fracturing associated with dolerite sills and 
ring-complexes are illustrated (Chevallier and Woodford, 1999; Chevallier et al., 2001; 
Woodford and Chevallier, 2002) and are of particular importance in that they form a basis 
for the storage of ground water. 
 
 45
• Well developed vertical thermal columnar jointing is found within the flat lying 
inner and outer sills. 
• Within the inclined sheet, fractures parallel to the strike of the intrusion are 
dominant, and this is the most fractured part of the complex. 
• Well developed sub-horizontal, or oblique fracture develop within the curved 
portion of the sill, and may be filled with secondary calcite. 
 
2.2.4.2 Dolerite Dykes 
 
Dolerite dykes are associated with the younger Karoo units from the Beaufort Group 
through to the Clarens Formation, and less frequently in the Drakensberg Group (Baran, 
2003). Chevalier et al. (2001) proposed three major structural domains based on dyke 
distribution. There is an extremely complex and intricate relationship between dykes and 
sills. Dolerite dykes may feed inclined sill sheets and therefore control the shape of ring-
complexes as well as uniting with adjacent rings Chevallier et al. (2001). Dolerite dykes 
are vertical to sub-vertical discontinuities that generally represent thin, linear zones of 
relatively high permeability which act as conduits for groundwater within the aquifer 
Chevallier et al. (2001). Alternatively, they may also act as semi- to impenetrable barriers 
to groundwater Chevallier et al. (2001). Their average thickness varies from 2 to 10 m, 
with dykes as thick as 300 m having been recorded, and seldom exceed 18.5 m in width, 
usually being between 2 and 8 m wide (Woodford and Chevallier, 2002). 
 
2.2.5 Other Post-Karoo Intrusions 
 
These post Karoo intrusions are mentioned only briefly as they are not directly related to 
this study, but are a part of the geology of the Karoo Basin. 
 
2.2.5.1 Breccia Plugs 
 
Breccia plugs are mostly restricted to the Ecca Group, occurring in clusters along the 
western and northern edges of the Karoo, but not restricted to this area (Woodford and 
 46
Chevallier, 2002). They occur in clusters of up to >80 plugs, with clusters varying in size 
from a few hundred metres to 50 km in diameter, usually forming low-relief, circular hills 
50 to 80 m in diameter (Woodford and Chevallier, 2002), Alternatively they may form 
negative-relief depressions characterized by calcrete development with a white alteration 
halo around them (Woodford and Chevallier, 2002). In the western Karoo they are also 
associated with swarms of kimberlite fissures in the western Karoo. 
 
It has been proposed by Woodford and Chevallier (2002) that explosive hydrothermal 
activity often took place when the early dolerite sills intruded into the partially indurated 
Karoo Supergroup sediments. Two main facies are recognised by Woodford and 
Chevallier (2002). The first is a molten facies, domed, baked, molten, re-crystallized, and 
highly contorted sedimentary host rock. These breccias contain xenoliths from the 
underling strata, with this facies being the most common in the field, possibly due to is 
resistance to erosion. The second is a breccia facies comprising fractured, broken, 
shattered displaced, and re-cemented blocks of sedimentary rocks. Breccia plugs may 
contain extensive mineralization. 
 
2.2.5.2 Volcanic Vents 
 
Volcanic vents, or diatrems, represent the first volcanic activity prior to the extrusion of 
the lava flows at c. 180 Ma (Chevallier and Woodford, 1999; Chevallier et al., 2001; 
Woodford and Chevallier, 2002). Largely restricted to the Clarens Formation they 
occasionally occur in the Drakensberg Mountains (Woodford and Chevallier, 2002). It 
has been proposed that they were probably formed by pheratic-explosive activity where 
excessive water and steam pressure overcame that of the magma, resulting in 
fragmentation, shattering, fluidization, and mobilization of the host rock and/or 
surrounding basalt by (Woodford and Chevallier, 2002; Svensen, 2006). In diameter, 
vents may vary in shape and size from a few metres to kilometres, and comprise layered 
successions of block- and matrix supported breccia and sandstone, as well as tuffs 
(Woodford and Chevallier, 2002). The only hydrothermal mineral in vents examined by 
Svensen et al. (2006) was zeolite, with cemented sandstone clasts and breccias. 
 47
2.2.5.3 Kimberlites 
 
Kimberlites within the Karoo Basin occur in clusters of linear swarms of dykes, fissures, 
and pipes, from Sutherland in the south, through to Kroonstad in the north, as well as in 
the western Free State, Lesotho and East Griqualand (Woodford and Chevallier, 2002). 
Kimberlite fracture swarms consist of parallel fissures and associated fractures, or joints, 
often with an associated upwarping of the surrounding Karoo beds (Woodford and 
Chevallier, 2002). Kimberlites are largely composed of decomposed igneous material 
with quantities of crustal, or mantle, xenoliths and megacrysts (Woodford and Chevallier, 
2002). Fresh kimberlite may commonly be referred to as blue-ground, weathered 
kimberlite as green-ground, and decomposed kimberlite as yellow-ground (Woodford and 
Chevallier, 2002). Deposits of well developed calcrete usually make it easy to identify 
kimberlites from aerial photographs (Woodford and Chevallier, 2002). 
 
A swarm can be divided into sub-swarms of smaller size where fissures are closely 
spaced approximately 10 to 50 m apart (Woodford and Chevallier, 2002). Hypabyssal 
kimberlite can form positive-relief hills, or negative-relief, calcrete rich depressions, with 
a diameter of 10 to 400 m in the western Karoo, and from 200 to 1000 m on the Kaapvaal 
craton (Woodford and Chevallier, 2002).  
 
Much debate has surrounded the factors controlling the emplacement of kimberlites 
including hot-spot tracks, crustal tectonics, and dynamics of the mantle (Woodford and 
Chevallier, 2002). It would appear that tectonics guided the emplacement of kimberlite 
swarms, with the ability of swarm sub-division indicating a vertical hierarchy of fissures, 
dykes, parental dykes, and larger bodies, within the fracturing system, at depth 
(Woodford and Chevallier, 2002). 
 
2.2.6 Cenozoic Deposits and Soils.  
 
The 180 Ma gap between the Drakensberg Group and the Quaternary deposits in the Free 
State consists of rocks which do not occur in the province, such as those belonging to the 
 48
Kalahari Group. Quaternary deposits are normally considered to be <2 Ma, and can be 
split in to three broad categories, namely, alluvial and colluvial deposits, calcretes, and 
sands and soils. 
 
2.2.6.1 Alluvial and Colluvial Deposits 
 
There is evidence of aggrading conditions along many Free State rivers in the form of 
river terraces. Grobler et al. (1988) proposed that the change from a degrading to an 
aggrading situation was responsible for the formation of many of the pans in the area (see 
section 2.3.1 for more detail). Holmes and Barker (2006) refer to a number of Free State 
rivers which show signs of aggradation, for example, the Orange, Caledon, Vaal, 
Modder, and Vet-Sand rivers. 
 
2.2.6.2 Calcretes 
 
Particularly in the arid to semi-arid western Free State calcretes, in a nodular and hardpan 
form, are wide spread. Calcretes form from the evaporation of groundwater and the 
resulting precipitation of calcium carbonate, or may form as a weathering product of 
dolerite (Woodford and Chevallier, 2002). Because carbonates of different ages occur, 
they do not appear to be related to any specific events (Myburg, 1977). Calcretes may 
attain thickness of 30 m, but seldom remain homogeneous over depths exceeding 1 to 5m 
(Woodford and Chevallier, 2002). Holmes and Barker (2006) state that due to calcretes 
not being related to any specific event, they serve little purpose as palaeoenvironmental 
indicators. Calcretes can occur as either surficial deposits, or beneath soil, or sand cover 
(Myburg, 1977). Nodular calcrete is often associated with unconsolidated sediments and 
hardpan calcrete with little, or no, surficial cover (Myburg, 1977). 
 
2.2.6.3 Sands and Soils 
 
Soil and sand formation is influenced by four factors, namely, parent material, or 
underlying formations; climate, topography and biological factors, of which parent 
 49
material and climate have played a dominant role in the Free State (Hensley et al., 2006). 
Unconsolidated sand is a feature of large parts of the western Free State (Figure 12). 
Holmes and Baker (2006) refer to unconsolidated sands possibly being of aeolian origin, 
based on morphometric properties. They also mention the presence of lunettes and 
aeolian sand as a feature of current aeolian processes along fence lines and roads in the 
western Free State (Holmes and Baker, 2006) (see 2.1.2). Hensley et al (2006) noted that 
in the western Free State sands are of an aeolian nature, and were derived from the Vaal 
River and its tributaries over the millennia. The classification of soil and sand types will 
be dependant on the characteristics of the soil used in the classification, and researchers 
may use different criteria. Table 5 gives a comparison of soil classification after 
MacVicar (1973) and MacVicar et al. (1977), and Hensley et .al. (2006). 
 
Descriptions of the soils of the Free State presented in Figure 12 have been taken from 
Earlé and Grobler, 1987. Freely drained latosols (#1) with a moderately advanced stage 
of laterization occur in the high rainfall areas of the east. A red-yellow-grey latosol 
plinthic catena forms the dominant soil pattern west of the former (#2-#6). The term 
plinthic indicates that these soils contain a highly weathered mixture of the sesquioxides 
of iron and aluminium as red mottles, which change to hardpan during alternative wet and 
dry cycles. These plinthic latosols vary due to drainage and relief. Sub-type #5 is 
characterized by large tracks of rock outcrop in the Vredefort area, while #6 in the west 
comprises well drained, sandy, partly aeolian soils. 
 
Solonetzic soils, some with black montmorillontic clay, other with reddish clays, 
predominate in the central and southern areas, including some areas east of the Vredefort 
structure (#7-#9). These solonetzic soils have a thin porous upper layer, underlain by a 
hard clay rich, highly alkaline, columnar horizon. The high alkaline content of mainly 
sodium and magnesium causes the soil to defloculate and become impervious to water. 
Usually not suitable for crops, these lands are used for grazing sheep. These poorly 
drained, clayey, solonetzic soils do not usually support tree growth, which explains the 
endless grassy plains over the province, broken only by riverine bush along water 
 
 50
 
 51
Table 5. A comparison of Free State soil classifications after MacVicar (1973), 
MacVicar et al. (1977), and Hensley et al. (2006). 
 
 
Soil classification after MacVicar  Soil clasification after Hensley 
(1973), MacVicar et al. (1977), as per et al. (2006). 
Fig.7. 
 
 Predominant soil categories from  
Hensley et al. (2006) reflected in 
the soil categories of MacVicar 
(1973), MacVicar et al. (1977). 
 
 Latosols A, B, C. A Red-yellow, structureless freely 
1 Non-humic red and yellow with drained soils 
varying amounts of rock and 
lithosols 
     
 Red-yellow-grey latosol plinthic B, C. B Plinthic catena: upland duplex and 
 catena margalitic soils rare 
2 Acidic sands/loams yellow/grey 
dominant 
     
3 As for 2, but with many rocky B, C. C Plinthic catena: upland duplex and 
areas magalitic soils common 
     
4 Neutral sands/loams red A. D Duplex soils dominant 
dominant 
     
5 As for 4, but with many rocky B. E Dark coloured margalitic clay soils 
areas with marked swell-shrink properties 
     
6 Neutral sands/loams B, C, D. F Shallow soils on rock 
yellow/grey dominant 
     
 Black montmorillonitic clays, red E, D. I Miscellaneous soils 
 clays, solonetzic soils 
7 Black clays and solonetzic soils 
     
8 Red clays and solonetzic soils A, C, D.   
     
9 Solonetzic soils E, D, I.   
     
 Sands A.   
10 Continental red or brown on-
shifting sand with varying 
amounts of rock and lithosols 
     
 Rocks and lithosols F, I.   
11 Undifferentiated 
 
 52
courses. However where the lithosol is sufficiently well drained, such as on the slopes of 
koppies (mesas), trees and shrubs can be supported. 
 
In the south-west of the province well drained red sandy soils of aeolian origin occur 
(#10). #10 comprises patches of rock and lithosol of poorly developed morphology on 
partially weathered rock, with some outcropping. Florisbad lies within this sub-group #10 
A analogous situation occurs along the Lesotho border where similar lithosols occur with 
a high proportion of outcrop (#11). 
 
2.3 GEOLOGY OF THE FLORISBAD AREA 
 
There is no outcropping of Karoo Supergroup bedrock on the Florisbad farm (pers. obs.). 
Florisbad lies within the Tierberg Formation of the Ecca Group (Loock and Grobler, 
1988) (see 2.1.1.3). Loock and Grobler (1988) note that the Ecca beds in the vicinity of 
Florisbad are sediments of the Tierberg Formation, which comprises well bedded shales 
and thin siltstones, the deposition of which took place through suspension settling of fine 
mud and silt under reducing conditions in an inland sea (Loock and Grobler, 1988). As 
progressively shallower conditions set in, the upper Tierberg beds of coarser material 
were deposited in a prograding deltaic environment (Loock and Grobler, 1988). Beaufort 
Group rocks have been recorded 3 km south west of Florisbad, and 14 km south-east of 
Florisbad on the farm Kalkwal 17, where approximately 4 metres of Beaufort Group 
rocks were found overlying the Ecca Group rocks (Loock and Grobler, 1988) (Figure 13). 
eyer and Lyle (1931) thought that the blue basal shale was of the Dwyka Group, but were 
later described as “Blue Ground”, or kimberlite, by Dreyer (1938a). A representative of 
De Beers reportedly identified ilmenite and corundum (ruby) (Fourie, 1970), but Fourie 
(1970) noted that, with the exception of some fine ilmenite found throughout the 
Florisbad sands, no other kimbelitic minerals were identified. The surface area of 
Florisbad is composed of an unconsolidated covering of red-yellow and pale bleached  
 
 53
 
 54
aeolian sand of varying depth (Loock and Grobler, 1988). The aeolian nature of the sand 
at Florisbad is supported by the large sand dune which straddles the Florisbad spring site, 
the extensive row of sand dunes lying further to the south and south-east of the site 
(Figure 14), and. by the presence of sand dunes along the southerly margins of most 
western Free State pans. The aeolian nature of the sand at Florisbad is further supported 
by researchers such as Grobler and Loock (1988a, 1988b), Van Zinderen Bakker (1989), 
Visser and Joubert, (1991), and Kuman et al., (1999). Calcrete horizons, which are 
common in the area, have been exposed through erosion that has occurred along the 
eastern bank of the vlei (low profile drainage system with grasses and semi-aquatic plants 
growing along the course – often marshy) draining from the spring site. Dolerite 
intrusions, intermixed with Ecca Group rocks are visible to the west and through to the 
north of Soutpan (Figure 14).  
 
2.4 GEOMORPHOLOGY 
 
2.4.1 Topography 
 
The extreme planation of the regional landscape of the Highveld, on which Florisbad lies, 
was hypothesised as being the product of an erosional phase referred to as the ‘Great’ 
African planation cycle which occurred over large areas of sub-Saharan Africa (King, 
1978). The African planation and denudation cycles occurred over a prolonged 
geotectonically stable period from the mid-Cretaceous to the mid-Tertiary (80 Ma) (King, 
1978), were thought to have manifested themselves in the extensive typical grassland 
areas that stretch from the interior of southern Africa, through Zimbabwe, Zambia, and 
East Africa (King, 1978). 
 
The topography of the western to central parts of the Free State province is generally flat 
to undulating, increasing in altitude from 900 m above sea level (asl.) in the west, to 1265 
m asl at Florisbad (Figure 14). Kruger (1983) described the western Free State as slightly 
irregular plains, the north-western areas as plains and pans, and the central and southern 
areas as lowlands with hills. From the central to eastern areas of the province the  
 
 55
 
 
 
 
 56
topography increases considerably to 3282 m asl at the Sentinel, in the Drakensberg 
mountain range. Kruger (1983) described this area as slightly irregular, undulating plains, 
with occasional hills rising to mountains. Holmes and Barker (2006) note that the Free 
State is dominated by areas where slope is <5%. Drainage patterns provide a good 
reflection of the topography of the province, and this can be clearly seen by the reduction 
of primary drainage from the higher lying eastern areas to the low lying western areas 
(Figure 15). 
 
2.4.2 The Western Free State Panveld 
 
Pans are the foci of poorly developed drainage basins that develop within environments 
of high structural uniformity and low surface relief (Goudie and Wells 1995). Florisbad is 
located on the eastern boundary of the western Free State panveld, west of the 500 mm 
isohyet (Figure 15). Verster et al. (1992) classed the geomorphic surface of Soutpan as 
being Post African Phase I. Many researchers have sought to explain the origins and 
formation of this panveld (Geyser, 1950; De Bruiyn, 1971: Le Roux, 1978; Marshall, 
1987a, 1987b; Grobler et al. 1988: Marshall and Harmse, 1992). Since the classic study 
by Hutchinson et al. (1932) on the hydrology of pans and other inland waters of South 
Africa, considerable attention has been given to South African pans. Wellington (1945) 
described the western Free State panveld as having the greatest density of pans anywhere 
in South Africa. The floors of the western Free State pans show considerable variation in 
the soil and rock composition as well as vegetation cover. De Bruiyn (1971) classified 
pans as follows, with further subdivisions indicating the presence or absence of 
vegetation. However, based on observations vegetation may vary seasonally, or over the 
loner term, due to the influence of rainfall. 
 
• Salt pans – occur mainly in the central region, extending to the east of the 
panveld. 
• Calcrete pans – have the widest distribution occurring over the entire panveld 
with the exception of the extreme north of the panveld. 
• Gypsum pans – limited to small areas in the western central part of the panveld. 
 57
 
 
 
 58
• Clay pans – occur mainly along the central eastern border, and northern areas of 
the panveld. 
• Sand pans – are limited to a belt in the southern central area of the panveld, from 
the eastern border, westward to the centre of the panveld. 
• Water covered pans – permanent water covered pans are limited to one near 
Viljoenskroon and one near Brandfort. Other pans may be artificially filled for 
irrigation and evaporative purposes, the latter being related to mining. 
 
The panveld, and the influence of the panveld on faunal distribution would appear to be a 
somewhat unobtrusive and ignored as a barrier to biogeographic faunal distribution. 
Zoogeographically the panveld represents a natural western boundary to at least 
herpetofaunal distribution (Douglas, 1992), and possibly other fauna as well. 
 
Dependant on the criteria used, the classification of a deflation hollow as a pan rests with 
the particular researcher, and therefore not all deflation hollows may be classed as pans. 
Therefore, the way in which deflation hollows are defined as pans will determine the 
boundaries of the panveld as well as the number of pans within the panveld. According to 
Geldenhuys (1982) the panveld covers an area of 1227 km2, with some 8 803 pans having 
been recorded (Figure 15). Douglas (1992) gave a concentration high of 82 pans per km2. 
Geldenhuys (1982) defined the panveld as being approximately 400 km long, running 
from the Vals River in the north to the Orange River in the south. Geldenhuys (1982) 
indicated the panveld to be an average of 140 km wide (min 50 km, max 200 km), 
extending eastwards from the western Free State border eastwards towards Welkom, and 
almost to Bloemfontein. This definition is consistent with definitions given by other 
researchers (De Bruiyn, 1971; Le Roux, 1978; Earlé and Grobler, 1987; Marshall, 1987a, 
1987b; Seaman et al., 1991; Lawson and Thomas, 2002). Although not defining the 
boundaries of the panveld, Holmes and Barker (2006) showed all pans occurring within 
the Riet-Modder-Vet-Vals river drainage complex. The highest concentration of pans 
indicated by Holmes and Baker (2006) clearly fits the boundaries of the western panveld 
as defined by other researchers, with Le Roux’s (1978) delineation of the western 
panveld being similar to that of Geldenhuys (1982). Le Roux (1978) also indicated large, 
 59
but very low density, panvelds in the southern, extreme east, and north-east of the 
province, with minor low density panvelds at, or near, Bethlehem, Harrismith, Kroonstad 
and Parys. 
 
Data from Geldenhuys (1982) indicates that 54% of all pans occurring in the western 
Free State, occur in the two degrees, 2825 and 2826, representing 4719 pans (Figure 15). 
Analysis of further data indicates that size of pans varies considerably with 2561 (54.3%) 
pans being <2 ha in extent, 1855 (39%) being 2-2.5 ha, 303 (6.4%) being > 25 ha. The 
southern areas of the panveld, as defined by Geldenhuys (1982), is blanketed by 
Quaternary red and grey aeolian sand with areas of calcrete and surface limestone. 
Largely underlying the Quaternary sands, and having more exposure towards the north, 
are Tierberg and Volksrust Formation rocks of the Ecca Group, with some Adelaide 
Subgroup rocks of the Beaufort Group occurring along the eastern boundary (Figure 8). 
Rocks of the Volksrust Formation and Adelaide Subgroup again become covered by 
Quaternary deposits in the region of the Vet River to the north. Dolerite sills and dykes 
are common over the entire area.  
 
Holmes and Barker (2006) recorded 16 830 pans occurring in the Riet-Modder-Vet-Vals 
river drainage complex. If the 10 253 pans recorded by Holmes and Barker (2006) as 
occurring on the Ecca Group rocks only are considered, this number approximates the 8 
803 pans recorded by (Geldenhuys, 1982). Geldenhuys (1982) classified pans based on 
the breeding of water-fowl, where the criteria for defining pans was based on vegetation 
growth, determined two months after flooding (see Section 4.3).  
 
The derangement of drainage patterns, or the disturbance of drainage patterns, by tectonic 
and/or climatic factors, as a possible cause for the pans was suggested by Wellington 
(1945) and Geyser (1950). Le Roux (1978) disagreed with this hypothesis, stating that it 
could not account for the wide distribution of pans, and that climatic derangement in the 
form of wind was the only agent that could have been responsible for the formation of 
most pans. Grobler et al. (1988) saw pan formation being largely as a result of climatic 
causes and stated that whatever the causes, the hydrodynamic equilibrium of the streams 
 60
was altered from an actively degrading drainage to an aggrading situation. When dry 
periods occurred, prevailing winds deflated the sediments, forming hollows, which were 
the initial sites of the pans (Grobler et al., 1988). Lancaster (1978) also stated that 
deflation was the major factor contributing to the formation of pans, while Goudie and 
Wells (1995) noted that bedrock weathering was promoted by un-vegetated, dry and 
saline conditions, with the relationship between pan hydrology, sediment supply, and 
deflation, being complex. Goudie and Wells (1995) concluded that the result of deflation 
was the formation of pan-margin lunettes from deflated pan sediments. 
 
Le Roux (1978) stated that there was no evidence of river capture, or backward tilting, as 
a cause for pan formation in the then Orange Free State. Van Zinderen Bakker (1989) 
was of the opinion that the Jurassic dolerite intruded the Karoo beds forming sills and 
dykes, with pans such as Soutpan, which is one of the largest pans at 19.4 km2 or 1 940, 
ha, 1.6 km north-west of the spring site, being formed during subsequent erosion cycles 
through deflation. 
 
Marshal (1987a) examined the panveld within a morpho-tectonic framework and 
postulated that the pans were remnants of a tectonically disturbed major palaeodrainage 
system that previously drained the area between the Vaal and Modder Rivers in an east to 
west direction. This system was referred to by Marshall (1987a) as the palaeo-Kimberley 
River. Marshall (1987a) envisaged an ancient Modder River flowing directly north-west 
from near Bloemfontein, to join the Vaal River at Christiana. Through headward erosion 
and Miocene structural displacements, the palaeo-Kimberley River eroded eastwards and 
captured the middle reaches of the ancestral Modder River, which then dried out to the 
north (Marshall, 1987a). At this time the the ancestral Riet-Modder River drained parallel 
to the palaeo-Kimberley River. Later Pliocene uplift in the region of the ancestral Riet-
Modder River also resulted in headward erosion eastwards, which again resulted in the 
capture of the upper reaches of the ancestral Modder River at the present day elbow at 
Lombards Drift (Marshall, 1987a), close to Florisbad. Downwarping of the palaeo-
Kimberley River during the same period resulted in the disruption of the river, with 
 61
Quaternary climates drying out the rivers to form the beginnings of the modern day 
panveld. 
 
2.4.3 The Florisbad Sand Dune 
 
The Florisbad sand dune, a crescent shaped aeolian deposit, is something of an enigma. 
Possibly, due to the excavations being concentrated in the area of springs activity, and a 
preoccupation with the term spring mound (Brink, 1987) no research appears to have 
been carried out on the Florisbad sand dune, as an entity, in order to determine its actual 
status. The Florisbad sand dune has been previously referred to as a lunette (Brink, 1987, 
Loock and Grobler, 1988) and as a spring mound (Brink, 1987). The latter term is 
disregarded as it does not seem plausible that such large quantities of spring sand would 
be available in order to replicate a sand dune to such an extent. There also appear to be no 
other such examples of springs in South Africa producing such large deposits. While it is 
not questioned that, in some areas the base of the dune does rest on such spring deposits it 
is felt that these are minimal. Owing to the status of the Florisbad sand dune still being 
undetermined, and therefore unclassified, due to the lack of information on its 
composition, sedimentation, and structure, the dune has been referred to in this thesis 
simply as the Florisbad “sand dune”, and no attempt has been made at classification. 
 
Pan fringe lunette dunes within the western Free State panveld, and in the Kalahari, are 
formed on the lee side of deflation hollows mainly by unconsolidated material being 
blown from the pan floor by a prevailing north-west wind (De Bruiyn, 1971; Goudie and 
Thomas, 1986; Loock and Grobler, 1988; Van Zinderen Bakker, 1989; Lawson and 
Thomas, 2000; Holmes et al., 2008). Such dunes are easily recognizable on aerial 
photographs, and in the field, from their peculiar vegetation cover, colour, shape, erosion 
degradation, and topographic expression (Loock and Grobler, 1988; Holmes et al. (2008). 
Another characteristic of lunette dues in the southern African region is their often 
extensive gully erosion on the windward side of the lunettes, with Holmes et al. (2008) 
noting that all lunettes examined by him in the western Free State panveld showed signs 
of erosion degradation. These gullies allows for the recycling of sediments, where dune 
 62
sediments are washed back into the pan, and then re-deposited on the windward slope of 
the dune by aeolian action during drier periods (De Bruiyn, 1971; Telfer and Thomas, 
2006; Holmes et al 2008). 
 
In many instances, pan-margin sand dunes are inextricably related to pan formation in the 
drier areas of southern Africa, as indicated by Lancaster (1978, 1986, 1989) in the 
southern Kalahari of Botswana, by Lawson and Thomas (2002) west of the Molopo River 
valley in the south-west Kalahari, and De Bruiyn (1971) in the Free State. Of the ten pans 
examined by De Bruiyn (1971), with the exception of one, all had sand dunes lying to the 
south, south-east, and east. De Bruiyn (1971) also noted that these sand dunes had been 
eroded, in some cases rather severely, by small courses on the windward side. Pan margin 
related sand dunes therefore form an integral part of the pan system in the drier regions of 
southern Africa. Telfer and Thomas (2006) used optically simulated luminescence dating 
to show that lunette features are spatially complex and that their formation may be 
relatively rapid. A 5 m accumulation of sediment over 570±40 years, and a 6 m 
accumulation over 660±40 were recorded by Telfer and Thomas (2006) at Witpan. 
Studies on the Witpan lunette showed that formation may have taken place within 2 ka, 
and that primary deflation from the pan had not contributed significantly to the formation 
of these lunettes (Telfer and Thomas, 2006). In fact, Telfer and Thomas (2006) noted that 
most of the lunette material was derived from the recycling of older lunette sediments 
from neighbouring dunes, with some contribution from the linear dunes. 
 
It is also recognized that the southern African Kalahari dune system, and that of the Sahel 
of West Africa, were both established during multiple arid phases since the last 
interglacial (Thomas et al. 2005). Therefore, climate plays a critical role in dunefield 
dynamics and the interplay between dune surface erodibility and atmospheric erosivity 
(Bullard et al. 1997; Thomas et al. 2005) and thus the mobility of sand dunes. Lawson 
and Thomas (2002) state that elsewhere lunettes have been derived from deflated 
sediments transported downwind to the pan margin by wave action. Thomas et al. (1993) 
state that the relatively small size of most southern African pans mitigates against 
inundation related lunette development, particularly in the Kalahari. In the case of 
 63
Florisbad, the considerable size of the adjacent Soutpan does not place it in this category, 
and its extensive surface area would possibly have made it more conducive to deflation 
factors and the formation of the Florisbad sand dune. 
 
Lawson and Thomas (2002) saw pan-margin lunettes as palaeoenvironmental indicators, 
and their view supports the theory put forward in this thesis that the Florisbad sand dune 
may hold the key to the history of the site. Pans with more than one lunette could 
possibly preserve evidence of episodic, or multiple, episodes of aridity with dune 
orientation indicating variations in the prevailing wind direction and thus the direction of 
aeolian deposition (Lancaster, 1978). Young and Evans (1986) note that a large portion of 
pan deflated sediments are deposited within a short distance of the pan on the downwind 
side. Pan deflation occurs only to a limited degree today in the Kalahari and significant 
lunette construction is not evident, therefore pan margin lunettes may be regarded as 
palaeomorphs, representing periods which indicated favourable deflation conditions 
(Lawson and Thomas, 2002). 
 
Lawson and Thomas (2002) note that the Kalahari pans are the most westerly expressions 
of closed deflated basins and display characteristics of pan basins operating under 
relatively arid conditions. Therefore, pan and lunette processes which operate in the 
Kalahari may be similar to those which operate to the east, and consequently may have 
implications, and shed greater insight, into contemporary and recent environmental 
processes over much of southern Africa and other dryland regions. 
 
The Florisbad dune rises 27 m above the pan floor and is located 1.65 km from the pan 
margin. South and south-east of Florisbad aeolian deposition rises to 37 m above the pan 
floor on the crest of the south-east dune belt, while south of the Modder River, this 
aeolian deposition continues to rise to 93 m above the pan floor. It would appear that the 
lands on the farm directly west of Florisbad were once used for agricultural purposes and 
reached almost to the pan shore, indicating fairly deep aeolian deposits. 
 
 64
It is evident from aerial photographs and satellite images that vast areas of the western 
Free State were blanketed with a thick layer of aeolian deposits. It is suggested that, if 
Florisbad and surrounding areas were blanketed with a thick layer of aeolian sand during 
the Holocene, the pans would also have been covered. Considering the extensive area of 
Soutpan, this accumulation on the pan floor would have been quite considerable. 
Therefore, it is postulated that at various times phases of heavy aeolian deposition 
covered Soutpan with sand from outside the area. Subsequent dry and windy periods then 
deflated the recently deposited aeolian sands from the pan floor, until the harder pan floor 
was again exposed. 
 
It is suggested that it was these more recent aeolian deposits that were responsible for the 
formation of the sand dune which, as it grew, migrated towards the Florisbad spring site. 
These deflated deposits then further contributed to the formation of the dunes south and 
east of Florisbad. Alternatively, it is possible that an outer lunette may have formed near 
the Florisbad spring site, to be later covered by aeolian deposition. This would imply the 
presence of a fossil lunette beneath the more recent aeolian deposits. The four metres of 
red sand recorded by Rubidge and Brink (1985) on the western flank of the site would 
tend to support both these hypotheses. De Briuyn (1971) has recorded individual sand 
dunes forming on the floor of the Kalgat (421) pan. Holmes et al. (2008) noted that the 
upper levels of the lunettes at Morgenzon and Sunnyside Pan both contained from 80- 
95% sand. This could possibly be related to periods of Holocene aeolian sand deposition 
and deflation. 
 
What is significant about Soutpan is that, although it is not entirely devoid of fringing 
lunettes, such lunettes are far less common and pronounced than at other pans in the area 
(Figure 16), particularly when considering the vast expanse of the pan. Fringe lunette 
formation at most pans in the area, including pans such as Morgenzon, Sunnyside, and 
Geluk Pan, are very distinct, well defined, and clearly merge with the pan margins 
(Figure 16). As mentioned in subsection 2.4.2, De Bruiyn (1971) recorded nine out of ten 
pans examined by him as having distinctive lunette margins. 
 
 65
 
 
 66
De Bruiyn (1971) also noted that pan fringe lunettes did not attain great heights. Holmes 
et al. (2008) recorded a prominent fringing lunette with a height of 5 m, 60 m from the 
edge of Morgenzon Pan, while similar sized lunettes were recorded at Sunnyside Pan, 
Deelpan and Salpeterpan.. Lawson and Thomas (2002) and Holmes et al., (2008) mention 
outer lunettes which can occur at far greater distances from the pan margins than pan 
fringing lunettes themselves. Lawson and Thomas (2002) recorded the height of an inner 
lunette at Koopan Suid at 9 m, while an outer lunette was recorded 1,25 km from the pan 
with a height of 80 m. This would beg the question as to whether or not outer lunettes 
occur at the base of the Florisbad and south east dune belt dunes. Holmes et al., (2008) 
recorded an outer lunette 660 m from the primary lunette at Morgenzon Pan. 
 
Rubidge and Brink (1985) drilled 31 auger holes at the site, and although the thickness of 
individual lithostratigraphic units, sand size and colour were recorded, little information 
was produced regarding the composition and status of the dune. That up to 4 metres of 
red to brown sand were recorded on the western side of the springs and towards the top of 
the dune (Rubidge and Brink, 1985), would seen to confirm a thick layers of more recent 
aeolian deposition. By examining sand size of the top 200 mm of the dune over a four km 
distance, from the Soutpan shore in a south-easterly direction, Loock and Grobler (1988) 
conclude that the true aeolian nature and origin of the dune had been established. As 
previously mentioned, Mucina and Rutherford (2006) noted that aeolian dust may be 
transported several thousand metres into the air by strong winds. 
 
It is put forward here that many of the questions relating to inconsistencies in the 
stratigraphy of the site, particularly through auger drilling, (see subsections 2.3.3 and 
2.3.4), are directly related to the formation of the dune and its internal stratification. In 
this instance, a combination of factors such as variations and t Rubidge and Brink (1985) 
drilled 31 auger holes at the site, and although the thickness of individual 
lithostratigraphic units, sand size and colour were recorded, little information was 
produced regarding the composition and status of the dune. That up to 4 metres of red to 
brown sand were recorded on the western side of the springs and towards the top of the 
dune (Rubidge and Brink, 1985), would seen to confirm a thick layers of more recent 
 67
aeolian deposition. By examining sand size of the top 200 mm of the dune over a four km 
distance, from the Soutpan shore in a south-easterly direction, Loock and Grobler (1988) 
conclude that the true aeolian nature and origin of the dune had been established. As 
previously mentioned, Mucina and Rutherford (2006) noted that aeolian dust may be 
transported several thousand metres into the air by strong winds.he composition of 
aeolian deposition, the aquatic environment which existed when the dune initially moved 
over the site, and most importantly, the angle of internal cross-stratification of the dune, 
where the angle of repose on the leeward slope is usually between 30 and 35 degrees, but 
as the dune migrates, the angle of internal stratification decreased down the windward 
slope (Appendix IV), are responsible. 
 
It is suggested that the vegetation cover and the low angle of repose of the Florisbad 
dune, as well as the dunes making up the south-east dune belt, indicate that the 
transportation of aeolian material has been inactive for some considerable time. 
According to Loock and Grobler (1988), the Florisbad site varies from other dunes in the 
area in that it has a higher brackish soil content which has been derived from Soutpan. 
 
In summary, the Florisbad sand dune shares no similarities or characteristics with other 
lunette dunes referenced in this subsection, nor to those observed from aerial photographs 
and satellite images. Further to this, characteristics such as size, height above the pan 
floor, distance from the pan, vegetation, extensive gully erosion and possibly shape, all 
preclude the Florisbad sand dune from being classified as a lunette in the sense that the 
term has previously been used. It is however acknowledged that an as yet undetected 
fossil outer lunette may exist under the more recent aeolian deposition. As no detailed 
study of the Florisbad san dune has been undertaken in order to establish its true status, it 
is proposed that the Florisbad sand dune is more closely related to a barchanoid dune, 
rather than a lunette. Only further analysis and dating at greater depths will determine the 
Florisbad sand dunes true nature and status. 
 
In that the Florisbad dune is considered here as being more barchanoid related, having 
being formed from aeolian sand deposits originally from outside the area, rather than a 
 68
lunette comprising deflated pan floor and recycled lunette material, periods of dune 
activity in the south-west Kalahari may be pertinent. In data compiled by O’Conner and 
Thomas (1999), from Thomas et al. (1997), Stokes et al. (1997) and Eifel and Blumel 
(1998), it is shown that primary phases of linear dune activity, which is thought to have 
had an influence on the Florisbad sand dune, occurred at 30-23 ka and 16-10 ka, with the 
probability of the 16-10 ka high activity phase being extended to 17-8 ka. Holmes et al. 
(2008) recorded lunette building periods in the western Free State panveld at 12-10 ka, 
5.5-3 ka, 2-1 ka, and 0.3-0.07 ka, indicating the more recent nature of lunette 
development. 
 
2.5 SYNOPSIS 
 
By relating the regional geology of the Free State to the Karoo Basin, a foundation has 
been established for understanding the geology of the Florisbad area, as well as the many 
aspects that will emanate from the geology in future chapters. The importance of 
geomorphological features, such as the western Free State panveld and the Florisbad sand 
dune, that have played such an important role in the morphology, formation, 
sedimentation, and chemistry of the Florisbad spring site, have also been brought to the 
fore as background, and as a prelude to their contributions towards the site. A new 
hypothesis has been proposed regarding the formation of the Florisbad sand dune. 
 
Two other critical components, namely, climate and vegetation, which have also played 
major contributing roles in the depositional environment at the Florisbad spring site, will 
be dealt with in the following chapter. 
 
 69
 
 
 
 
Chapter 
 
3 
 
 
 
 
 
The Physical 
Environment 
 
Part II 
 
Climate and 
Vegetation 
 
 
 70
CHAPTER 3 
 
PART II 
 
CLIMATE AND VEGETATION 
 
 
3.1 INTRODUCTION 
 
The various facets of climate are responsible for a number of processes that have 
contributed to the depositional and geohydrological environment at Florisbad. These 
would include weathering, diagenesis, chemical reactions, the formation of aquifers, 
including related aspects such as recharge and flow, aeolian deposition, and 
vegetation. Climate would have influenced the type and production of vegetation, 
which in turn would have determined the species composition of the area. Vegetation 
would also have influenced the morphology of the area by varying the degree of sand 
dune mobility during long-term wet and dry periods. 
 
The climate of the Free State is partly influenced by the relief of the province and is 
associated with a west to east increase in altitude from 900 to 3282 metres asl. 
Florisbad Research Station is located on a plateau almost in the centre of the southern 
African region (Brink, 1987). This effectively prevents the free circulation of air from 
the costal regions from reaching the inland plateau, resulting in Florisbad having a 
climate of extremes (Brink, 1987). Aspects of climate are discussed below. 
 
3.2 PALAEOCLIMATE 
 
Palaeoclimate is important in that it would have determined aspects such as 
temperature, wet and dry cycles, vegetation, and thus the environment in which the 
Florisbad spring site was formed, and faunal remains fossilized. Many different 
factors have been used in attempts to project palaeoclimate locally and on a world 
wide scale. A few of these factors such as pollen, stalagmites, dust, rainfall and 
temperature, are presented in Table 6. In many instances, the projection of  
 
 71
Table 6. An interpretation of examples of global and local palaeoclimates as determined by various 
               authors (Log scale ).
Palaeoclimates as indicated by:
A B C D E
Dust content from the Interior rainfall Global oceanic Pollen data for Florisbad Scott Stalagmite based 
Vostok ice cores, record for southern temperature records & Nykale (2002), Scott & δ18O data for 
Antartica. (after Petit et Africa. (after Tyson, (after Shackelton, Russouw (2005). ~ = High lake Makapansgat (after 
al., 1990; Joubert & 1986; Scott, 1989; 1969; Joubert & 
Visser, 1991; McCarthy Lancaster, 1990; Visser, 1991) levels, spring activity (after 
Holmgren et al ., 
1999; 2003) 
& Rubidge, 2005) Joubert & Visser, Butzer 1994).
1991)
ka High Dry Wet Dry Wet Dry Wet Semi Sub Semi Sub Dry Wet
dust and and arid humid arid humid and and
peaks cold warm cool warm
0.0
0.4
0.4
0.5
0.8
0.9
1.0
1.2
1.3 ~ ~
1.7 ~ ~
1.8 ~ ~
2.0 ~ ~
2.1 ~ ~
2.5 ~ ~
3.0 ~ ~
35 ~ ~
4.0 ~ ~
4.2 ~ ~
4.5 ~ ~
5.0
5.5
6.0
6.3 ~ ~
6.5 ~ ~
7.0 ~ ~
8.0
9.0
10.0
11.0
12.0
15.5
16.0
17.5
18.0
19.0
20.0
21.0
23.0 C
24.0
25.0
30.0
32.0
40.0
45.0
46.0
50.0
52.0
60.0
65.0
70.0
76.0
80.0
90.0
91.0
100.0
105.0
110.0
120.0       MSA Human Occupation Horizon
129.0
130.0
135.0
140.0
150.0
160.0
170.0
180.0
190.0
200.0
210.0
220.0
250.0       Florisbab Skull
260.0
275.0
280.0
300.0
345.0
350.0
355.0       Lower Sedimentary Deposits
370.0
400.0
435.0
450.0  
 
 
 72
Scott & Nykale (2002)
Scott & Russouw (2005)
P             e              a             t           IV
palaeoclimate is uncertain and speculative. Table 6 gives examples of palaeoclimatic 
interpretations based on a various factors and reflect the differences in the results 
between broader surveys covering extended periods, and the more detailed results 
obtained from shorter term individual site climatic predictions.. It will be noted in 
Table 6 that there is often little correlation of the broader interpretation of 
palaeoclimatic conditions between studies. It is also apparent that as more intense and 
detailed research is carried out for a specific region or site, although be it over shorter 
periods, the accuracy of palaeoclimatic projections may become more reliable. For 
example, on a broader scale the hydrogen isotope ratio, temperature, oxygen isotope 
ratio, and sea surface temperature of the Vostok ice core analysis all follow the highs 
and lows of the core dust content closely (McCarthy and Rubidge, 2005). This would 
indicate that as these highs and lows are comparable, they are also fairly accurate in 
supporting the occurrence of events at a particular time but reveal little regarding 
short-term palaeoclimate. Variability at a regional level may be judge from data 
compiled by O’Conner and Thomas (1999) which shows the differences in phases of 
linear dune activity (arid periods) for western Zambia, western Zimbabwe and the 
south-west Kalahari, but all within the Kalahari region. 
 
Even although palaeoclimate is an all encompassing term covering millennia, 
palaeoclimate as such must be seen in a very site specific context. This is largely 
because of the many interrelated climatic factors, as well as factors such as the 
geology, vegetation, deposition and erosion, all on both a micro- and macro-scale, 
will differ considerably between sites, even within a region. This is illustrated in 
studies such as those by Scott and Nykale (2002) and Scott and Russouw (2005) on 
the Florisbad pollen, and Holmgren et al. (1999; 2003) on the Makapansgat 
stalagmite. It will be noted from columns “D” and “E” in Table 6, that there is little 
correlation between the two sites, and attempts to correlate different sites in different 
regions are unlikely to correspond due to climatic variability and local conditions. 
 
It can be deduced from evidence presented by previous researchers that, unlike the 
present day relatively dry climate, the Florisbad area previously experienced more 
humid periods (Butzer, 1984; Brink and Lee-Thorpe, 1992; Scott and Nyakale, 2002). 
Joubert and Visser (1991) suggested that lacustrine deposits had formed at the 
Florisbad site due to the water level from the nearby Soutpan rising and expanding to 
 73
an area of twice its current size during cyclical humid periods. Scott and Brink (1992) 
stated that the pollen levels could not be precisely linked to layers cited in the 
sedimentological study by Visser and Joubert (1991). Scott and Russouw (2005) 
noted that the high water levels proposed by Visser and Joubert (1991) were in 
conflict with the pollen evidence where Chenopodiaceae pollen was an indication of 
dry conditions. Scott and Brink (1992) also contended that the findings of Van 
Zinderen Bakker (1957), who interpreted the ratio between Compositeae pollen and 
grass pollen as an index of dry periods, was contradictory to the rises of the 
palaeolake as proposed by Visser and Joubert (1991). Rises and falls in the lake 
complex were correlated by Joubert and Visser (1991) to past southern African and 
global climatic data. Scott and Brink (1992) also rejected Van Zinderen Bakker’s 
(1957) interpretation and proposed that an increase in the ratio between 
Chenopoiaceae, grasses, and other pollens signified moist periods, while a decrease in 
the ratio signified dry periods.  
 
The presence of fresh water gastropods in the clay facies of the Florisbad site also 
indicates the previous presence of large bodies of fresh water (Fourie, 1970; Visser 
and Joubert, 1991). Other indications of a more humid past is the fossil evidence of 
species which might rely on an aquatic environment, such as hippopotamus 
(Hippopotamus amphibius), lechwe (Kobus leche), clawless otter (Aonyx capensis) 
and water mongoose (Atilax paludinosus) as well as the possibility of waterbuck 
(Kobus ellipsiprymnus) (Brink, 1987; Henderson, 2001a; Appendix II). 
 
Brink and Lee-Thorpe (1992) suggested that the area was grassland with a grazing 
succession similar to that of the Serengeti in east Africa. This suggestion was based 
on the presence and grazing habits of Antidorcus bondi, an extinct springbok, and a 
relative of the extant springbok A. marsupialis (Brink and Lee-Thorpe, 1992). A. 
bondi, unlike A. marsupialis, was a specialized grazer and is thought to have fed 
almost entirely on newly sprouted grass shoots, which could only be ensured through 
regular mowing by larger herbivores and sufficient soil moisture (Brink and Lee-
Thorpe, 1992). This specialized diet meant that A. bondi could not shift its diet in 
winter, and therefore the veld must have had a high primary production to allow for 
year-round production of new shoots, which could only have been attained with year-
 74
round high soil moisture content (Brink and Lee-Thorpe, 1992), and a frost-free 
climate. 
 
It must therefore be assumed that localized sedimentary deposits, such as the peat 
layers, were formed during wet periods, and that the sand layers were accumulated 
during dry periods. This would be accordance with the stratigraphy given in Figure 4 
where major wet and dry periods are reflected in the cross-section. There does seem to 
be consensus between the hypothesis proposed in this thesis that the water level in the 
Florisbad dam rose to the top of the arms of the san dune, and palaeoclimate. Both 
Scott and Nykale (2002) and Scott and Russouw (2005) concur that the development 
of Peat IV occurred during a predominantly wet period. Scott and Russouw (2005) 
also concur that Peat IV almost reached to the top of the site during this wet period, 
with its width indicating that the wet period must have been a considerably extended 
one. From the results of Scott and Nykale (2002) and Scott and Russouw (2005) who 
base their analysis on pollen derived from the surrounding area, it can also be deduced 
that the Florisbad dam was not filled entirely by spring-water during this wet period. 
 
As the stratigraphy of the sand dune and sedimentary deposits vary considerably, this 
would reflect a multitude of different environmental influences, with each section 
having to be considered independently. Greater detail on the environmental history of 
the site is given under section 5.6.1 and many of the other sections. Another drawback 
with the current state of palaeoclimate analysis and interpretation is that it is very 
much limited to the near surface sediments. As will be seen from Table 6, these 
interpretations come no where near to resolving issues in important layers such as the 
MSA Human Occupation Horizon, the Florisbad skull, or the lower sedimentary 
deposits where the Old Collection is to be found. However, much more research is 
required in order to associate the sedimentary developmental stages at Florisbad 
directly to the palaeoclimate. 
 
 
 
 
 
 
 75
3.3 PRESENT CLIMATE 
 
3.3.1 Rainfall 
 
Evidence presented in the appendices (Appendix I, II, III, IV) indicates that rainfall 
was a critical factor in the mobilization and transport of ions for the salinization of the 
ground water, the salinization of the clay and organic material, and the fossilization of 
faunal remains. Rainfall was also a critical factor to the formation of the Florisbad 
spring site, with alternating layers of sand and organic clay, reflecting alternating dry 
and wet periods (Figure 4). 
 
The South African climate is almost entirely under the influence of circulation to the 
west of the country, with changes being dominated by disturbances in the southern 
hemispheres westerly circulation, which in turn appear as cyclones or anticyclones, 
moving across, or around the coast. (Schulze, 1994). Climatic conditions are further 
influenced by factors such as latitude and solar radiation, altitude, position relative to 
land and sea, and ocean currents and temperature (Schulze, 1994). Rainfall in the Free 
State can be brought about by three basic weather systems (Louw, 1979; Schulze, 
1994; Kruger, 2007). 
 
3.3.1.1 Convergent Rainfall 
 
Convergent rainfall is the primary cause of rainfall over the interior of South Africa, 
including the Free State. This rainfall is largely as a result of a cut off low pressure 
cell forming over southern Africa, and drawing in a broad stream of warm, moist, 
equatorial air, from the north and north-west. This influx of warm moist air, and 
concentration of air in the lower atmosphere, causes a simultaneous outflow, or 
divergence, in the upper layers. A well developed cell of low pressure over southern 
Africa favours strong convergence over the Free State, resulting in convective storms 
and heavy rainfall. 
 
 
 
 
 76
3.3.1.2 Orographic Rainfall 
 
A low pressure system over southern Africa is usually followed by an anticyclone 
moving from the west, to the north east along the coast, gathering moisture from 
above the warm Auglhus and Mozambique ocean currents. Easterly and south-easterly 
air is then forced against the Drakensberg escarpment, resulting in orographic rain 
along the eastern side of the country, which affects the eastern and north-eastern Free 
State. At the same time this anticyclonic system also promotes uplifting and 
convergence of the equatorial air flowing southwards. Orographic rainfall is also 
important over the south-western and southern Cape.  
 
3.3.1.3 Frontal Systems 
 
These cold systems, which originate in the higher latitudes of the Atlantic Ocean to 
the west of South Africa, where warm air is forced to rise, cool, and moisture to 
condense. Frontal systems play an important role in both summer and winter rainfall, 
but particularly in winter rainfall over the south-western Cape and along the Cape 
coast, while strong frontal systems also bring some winter rainfall to the Free State. 
Contrary to the above three systems, the development of a high pressure cell over 
southern Africa results in dry periods. 
 
3.3.1.4  Rainfall of the Florisbad Area 
 
Rainfall in the Free State province increases from west to east, as illustrated by the 
mean annual rainfall for Jacobsdal (west) of 349 mm, and the mean annual rainfall for 
Witzieshoek (east), 1016 mm (Douglas, 1992). Florisbad Research Station falls within 
a summer rainfall region where most rainfall occurs between October and March, with 
approximately 78% of rainfall occurring in summer (Earle and Grobler, 1987). 
 
Florisbad lies just west of the 500 mm isohyet within the 400 mm to 500 mm mean 
annual rainfall zone of Van der Wal (1977). The 500 mm isohyet is of particular 
importance because, as a meteorological feature, areas receiving rainfall above 500 
mm are regarded as being suitable for dryland agriculture. (Petja et al., 2004). As with 
all isohyets, the position of the 500 mm isohyet is variable, with its position being 
 77
determined by fluctuating seasonal and annual rainfall. The mean annual rainfall for 
the 78 year period 1922 to 1998 was 496 mm, with a minimum rainfall of 271 mm in 
1965 and a maximum rainfall of 957 in 1988. When expressed as a percentage of the 
mean annual rainfall, between 1961 and 1999, Florisbad experienced a maximum of 
175% of the mean, and a minimum of 46% of the mean. Long dry periods are 
characteristic of the area with mean annual rainfall for the period 1964 to 1970 being 
406 mm and for the period 1982 to 1986 being 379 mm (Douglas, 1992). Figure 17 
provides a 23 year rainfall record for Florisbad showing the considerable annual 
variation in rainfall. More details on the rainfall at Florisbad can be found in 
Appendix I. 
 
3.3.2 Temperature 
 
Figure 18 gives an example of mean monthly maximum and minimum temperatures 
for the Florisbad area. Temperatures in the Free State province decrease from west to 
east. This is illustrated by the mean January (summer) temperature for Boshof (west) 
of 24 °C and the mean January temperature for Harrismith (east), 18 °C (Douglas, 
1992). The mean July (winter) temperature for Boshof is in the order of 10 °C and the 
mean July temperature for Harrismith, 7 °C (Douglas, 1992). In the Florisbad area, 
temperatures can vary considerably between summer and winter, with a maximum of 
38.5 °C (January) and a minimum of –7 °C (June) having been recorded at Glen 
Agricultural College, 32 km east of Florisbad. 
 
3.3.3 Evaporation 
 
The most important requirements for the deposition of salts in a semi-arid 
environment such as Florisbad, and the salinization of sands, soils, and organic 
material, are evaporation and capillarity. Capillarity is reliant on the conductivity of 
the soil in relation to the surface evaporation rate (Hillel, 1971), and will increase in 
fine textured soils of high clay and humus content. Bohn et al. (1985) illustrated that 
where the water table was 900 mm below the surface, the effects of salt deposition 
began at about 400 mm, where the electrical conductivity (EC) was 0.2 ms/m, with 
EC increasing to EC 70 ms/m at 50 mm. In order to illustrate the effect of evaporation 
 
 78
1000
900
800
700
600
500
400
300
200
100
0
Year
 
Figure 17. Twenty-three year annual rainfall at Florisbad showing the variation in 
annual rainfall between the lowest (1965) and highest (1988) recorded 
rainfall periods. 
 
 
 
35
30
Max
25
20
15
Min
10
5
0
-5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Months
Figure 18. An example of the monthly minimum and maximum temperatures for the 
Florisbad area (taken at Glen Agricultural College 1987). 
 
 79
Rainfall mm
Temperature °C
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
and matrix suction at Florisbad, rainfall and evaporation figures were taken from Glen 
Agricultural College, 32 km east of Florisbad. Rainfall for the period January, 
February and March 1992 was 66.5 mm, while the mean evaporation rate for the same 
period was 315.0 mm. This resulted in a rainfall deficit of 248.5 mm for the three-
month period, which is an indication of the amount of moisture and salts being 
brought to the surface by capillarity and matrix suction. 
 
Van Zinderen Bakkker (1989) gave an average annual rainfall deficit for eight 
stations, with records varying from 18 to 47 years, of 1260 mm, while the mean 
annual evaporation rate given by Baran (2003) for the area is 1 600 – 1 700 mm. 
Evaporation in the Free State is converse to rainfall decreasing from west to east, with 
higher rainfall areas experiencing less evaporation, and lower rainfall areas 
experiencing higher evaporation. This is largely due to the west to east increase in 
altitude and cooler temperatures. Particularly in semi-arid environments where 
evaporation exceeds precipitation, calcium carbonate forms as a result of the high 
evaporation rate (Hillel, 1971). The effects of capillarity and salinization have both 
played a major role in the high mineralization of the upper Peat IV horizon (Appendix 
II). 
 
3.3.4 Wind 
 
Winds at Florisbad (Figure 19), as with rainfall, are regulated by disturbances in the 
circulation systems and patterns to the west and south-west of the country, and their 
influence on the intensity of low and high pressure cells over the land. The latter 
factors will determine the intensity with which warm, moist, tropical air is drawn in 
from the tropics during summer, and the intensity with which cold fronts reach inland 
during winter. Wind has played a major role in the formation of the Florisbad spring 
site (Appendix IV). Wind is critically important in the Florisbad context in that it not 
only determines rainfall, but also the aeolian transport of sand, and the resultant 
formation of sand dunes. Mucina and Rutherford (2006) have noted that aeolian dust 
may be carried several thousand metres into the air by strong winds. Florisbad may 
never have existed as such an important archaeozoological and archaeological site had 
 
 
 80
 
 
 
 81
it not been for the formation of the sand dunes (Appendix IV). In Figure 19, wind 
roses show the predominantly north-easterly, through to south-westerly, and even 
southerly (anticlockwise), wind flow throughout most of the year. 
 
Despite some authors (Loock and Grobler, 1988; Van Zinderen Bakker, 1989) 
referring to a prevailing north-west wind, it is evident from Figure 19, that currently, 
it is difficult to assign any prevailing direction to local winds, as was confirmed by 
Kruger (2002). However, data on wind direction by Schulze (1994) does show a 
north-west predominance for both January and July. Current wind directions in Figure 
19 may therefore not be a reflection of historical wind direction. Judging from the 
extent of the south-east dune belt (Figure 14), which continues for kilometres south 
and south-east from the southern banks of the Modder River, it is apparent that 
historically, there must have been extended periods with a very predominant 
prevailing north to north-west wind in order to have moved such large quantities of 
sand. 
 
3.4 VEGETATION 
 
The Grassland Biome of Rutherford and Westfall (1986) is defined as having two 
categories of grasses. The first category is the sweet grasses, which have a low fibre 
content and maintain their nutrients in their leaves throughout winter. The second 
category is the sour grasses, which have a high fibre content and tend to withdraw 
their nutrients from the leaves during winter. Sour grasses prevail with a higher 
rainfall and more acidic soil, with 625 mm being the level at which sour grasses 
prevail. This then places Florisbad in the sweet grass category. Acocks (1988) veld 
type (AVT) 50 (Figure 20) has since been classified by Low and Rebelo (1996) as 
Dry Sandy Highveld Grassland. Mucina and Rutherford (2006) describe the 
vegetation unit as Dry Highveld Grassland and the Florisbad sub-units as Western 
Free State Clay Grassland (Gh 9), partly surrounded by Vaal-Vet Sandy Grassland 
(Gh 10), which extends from the north-west of Florisbad, around to the south east 
(anticlockwise). 
 
 
 
 82
 
 
 
 
 
 
 83
The Florisbad spring site is located in Dry Cymbopogen- Themeda Grassveld, Acocks 
Veld Type 50 (Acocks, 1988) (Figure 20), and in close proximity to an easterly 
protrusion of False Upper Karoo (AVT 36). The site also lies within, and close too, 
the western boundary of the Grassland Biome of Rutherford and Westfall (1986), with 
the Nama-Karoo Biome lying directly to the west. Karoo and Desert vegetation types 
which are encroaching from the south-west (Acocks, 1988), along with the close 
proximity of the False Upper Karoo and Nama-Karoo Biomes, may well see the area 
falling into a considerably drier vegetation zone. Vegetation is a reflection, and 
integral aspect, of past and present climates and environments. Vegetation is not only 
a reflection and record of present climatic and soil conditions, but also a record of 
vegetation and climatic changes over the millennia. This history is recorded in the 
palynology (study of pollens) of the area, in the form of pollen (organic) analysis from 
the sediments and hyaena coprolites. The study of phytoliths (inorganic biogenetic 
silica formed in the intercellular and intracellular spaces in plants), particularly from 
grasses, is another valuable analytical tool in recording the vegetation and climatic 
history of an area. Phytoloths also occur in hyaena coprolites (Scott and Russouw, 
2005), and phytoliths from the dental cavities of Middle Stone Age land surface bovid 
species were examined by Russouw (1996), providing an indication of the species 
diet. Another illustration of the importance of vegetation is that the only piece of 
wood recovered from the sediments was identified as the tree Xanthoxylim 
chalybeum, which could imply a major climatic shift to warmer, wetter inter glacial 
conditions (Bamford and Henderson, 2003; Scott and Russouw, 2005). 
 
Highveld salt pans comprise a large variety of low shrubs, succulent shrubs, 
megagraminoids, graminoids, herbs, and succulent herbs (Mucina and Rutherford, 
2006). Douglas (1992) recorded the following dominant vegetation types for the study 
area. The vegetation of the site largely comprises taller Themeda triandra, Eragrostis 
lehmanniana and Hetrodon contortus grasses, with low growing grass being 
represented by Tragus koeleroides. Bushes are scattered, and represented by Salsola 
glabrescens and Filicia muricata. Along the banks of the vlei area Cynodon dactylon 
grass is predominant, with dense stands of Salsola glabrescens bushes and Conyza 
bonariensis annuals. In the wetter regions of the vlei some areas are densely vegetated 
with Schoenoplectus (Scirpus) triqueter and Typha latifolius with 
Mesembryanthemum sp. growing along the drier edges. 
 84
The western Free State panveld vegetation unit is classified as Inland Azonal 
Vegetation, with the sub-unit of the pan vegetation being classified as Highveld Salt 
Pans (Azi 10) (Mucina and Rutherford, 2006). Based on vegetation growth, 
determined two months after flooding, Geldenhuys (1982) classified pans as bare 
pans, sedge pans, scrub pans, mixed grass pans, closed Diplachne (Leptochloa) pans, 
and open Diplachne (Leptochloa) pans. Highveld salt pan vegetation differs 
considerably from pans in other areas due to their predominance of cyperoids (Mucina 
and Rutherford, 2006). 
 
The Florisbad area is an area of natural transitional zones for a number of factors, and 
these factors have been grouped together here because vegetation and climatic 
boundaries play a dominant role (Table 7). The actual influences of these transitional 
zones on the Florisbad area were not examined in this study. It is however proposed 
that, in conjunction with morphological features such as the Western Free State pan 
veld, they contribute in making the Florisbad area distinctive in its own right.  
 
3.5 SYNOPSIS 
 
The introduction of the regional climate of the Free State province establishes a 
further background for understanding the local climate of the Florisbad area. It will be 
noted from this chapter how various aspects of the climate are interrelated, and how in 
future chapters, they will contribute and influence, in their own individual way, to the 
depositional and fossilization environment at Florisbad mentioned in the Introduction 
to this chapter. 
 
In the following chapter, yet another feature having a direct influence on the Florisbad 
spring site, geohydrology, is examined. As with geology in Chapter 1, geohydrology 
is discussed on a regional basis in order to bring perspective and understanding to the 
geohydrology of the Florisbad area. 
 85
Table 7. Natural zones which have their boundaries passing through the Florisbad 
area, as defined by various authors. 
 
 
Zone    Detail of natural zone   Reference 
 
   
Grasslands Boundary between the Karoo and Sweet Grassveld. Acocks 1953 
   
Grasslands Boundary between the Karoo (South-West Aarid) and Acocks 1975 
Grassveld (Southern Savanna Grassland). 
   
Biotic zones Boundary between the South-West Arid and Southern Keay, 1959; 
Savanna Grassland Biotic Zones. Davis, 1962; 
Meester 1965; 
Rautenbach 1978 
   
Biotic zones  Boundary between the South-West Arid and Southern Wellington, 1955; 
Grassland Biotic Zone based on the rainfall map of Davis 1962 
Wellington (1955). 
   
Biotic zones Boundary between the Grassland and Nama-Karoo Rutherford & 
Biome based on vegetation. Westfall 1986 
   
Vegetation Boundary between the Highveld Grassland (58) and the White 1981 
Karoo Grassy Shrubland Transition from Karoo 
Shrubland to Highveld (57b). 
   
Vegetation Boundary between the Eastern Mixed Nama Karoo (52) Rebelo & Low, 
and the Dry Sandy Highveld Grassland (37) 1996 
   
Rainfall The 500 mm isohyet. Poynton, 1964; 
Van Der Wal 
1977 
   
Humidity Boundary between the Sub-Humid Moisture Region Poynton, 1971 
(C) [moisture index –20 to 0] and the Semi arid 
Moisture Region (D) [moisture index –40 to –20 based 
on the silvicultural map of Poynton (1971). 
   
Frost Boundary of the Cooler –Temperate (Mesothermal) Poynton, 1971 
Thermal Region (5) [Thermal Efficiency Index – severe 
frost] and the Cooler-Temperate (Mesothermal) 
Thermal Region  (6) [Thermal Efficiency Index – 
moderate frost]. 
   
Panveld The eastern boundary of the western Free State Geldenhuys, 1982 
panveld. 
   
Herpetofauna The eastern boundary of the T2 Squamata Transitional De Waal, 1978 
Zone and B2 Squamata Province for the Free State. 
   
 
 86
 
 
 
 
Chapter 
 
4 
 
 
 
 
 
 
The Physical 
Environment 
 
Part III 
 
Geohydrology 
 
 
 
 87
CHAPTER 4 
 
THE PHYSICAL ENVIRONMENT 
 
PART III 
 
GEOHYDROLOGY 
 
 
4.1 INTRODUCTION 
 
Water is a scarce resource in South Africa, with Vivier (1996) and Chevallier et al. 
(2001) stating that groundwater contributes to only 10% to South Africa’s water 
needs, while surface waters have been virtually exploited to their limit. However, Free 
State towns such as Ficksberg, Jagersfontein, and Fauresmith, may towns derive from 
50 to 100% of their water requirements from groundwater (Vivier, 1996). Vegter 
(1995) lists 74 municipal supply schemes with populations >2500, whose water 
supply is based solely on groundwater. By virtue of secondary openings, over 90% of 
South Africa’s rocks are water bearing (Vegter, 2001) This means that groundwater is 
becoming a more important and valuable resource, with the Karoo Supergroup 
formations, which underlie approximately 50% of South Africa, potentially becoming 
South Africa’s largest and most important source of water (Vivier, 1996). Chevallier 
et al. (2001) note that, within the Karoo Basin, dolerite dykes represent the most 
commonly exploited targets due to their well known water yielding capacity. 
Although having considerable water bearing potential, it was noted by Chevallier et 
al. (2001) that dolerite ring and sill complexes, were on the other hand, overlooked as 
a source of water, and extensive research is currently being carried out in this area. 
 
That the Florisbad springs have a perennial-, as opposed to a sporadic or cyclical 
flow, requires greater emphasis to be placed on understanding the geohydrological 
properties and processes of the area. With Florisbad lying within the Karoo Basin, and 
the Florisbad springs being controlled by dolerite intrusions, a background to the 
 88
regional geohydrology, and factors and processes affecting this hydrology, are 
presented in this chapter. 
 
The geohydrological chapter has been compiled from Vivier (1996), Vegter (1995), 
Chevallier and Woodford (1999), Chevallier et al. (2001), Vegter, (2001), Woodford 
and Chevalier (2002), and Baran (2003) 
 
4.2 BACKGROUND 
 
South Africa is divided into 64 groundwater regions based on primary, or secondary 
pores, physiography, and climate, while conformity between lithostratigraphic units 
and phsiographic features strengthens the case for lithostratigraphy as a primary basis 
for subdivision. Vegter (2001) suggests that the following criteria govern the 
availability and occurrence of ground water: 
 
• the storage and transmissive properties of the geological formation; 
• topographic relief, which in turn determines the hydraulic gradient  
• the volume and frequency of discharge; 
• the rate of groundwater movement to discharge points/areas; 
• the rate of groundwater discharge as springs and effluent seepage in streams; 
and 
• loss through evapotranspiration. 
 
The transmissive properties of a formation may be restricted by: 
 
• fractures being under compressional stress and consequently tight; 
• recrystalization in fractures by hydrothermal fluids or a drop in the 
groundwater level; 
• fractures being filled with swelling, or non-swelling clays as a result of 
weathering, or past hydrothermal activity. 
 
As opposed to the availability and occurrence of groundwater, Vegter (2001) proposes 
the following criteria for recharge: 
 89
 
• factors in relation to rainfall such as, volume, intensity, frequency, and 
temporal distribution, 
• availability of surface water, 
• land surface configuration, 
• soil and vegetation cover; and 
• subsurface moisture retention and evapotranspiration. 
 
Lithification occurs in the form of, compaction and cementation desiccation, 
crystallization, recrystallization, and compression, or a combination of processes. 
Compression of the Karoo sediments, caused largely by the Drakensberg lavas 
covering the area, caused a considerable decrease of porosity in the sediments, 
particularly the porosity and elasticity of the clays (Vivier, 1996). A relatively rapid 
erosion of the Drakensberg lavas then removed the pressures being exerted on the 
underlying Karoo sediments, resulting in an isostatic uplifting of the sediments, and 
the formation of fractures and interstices, which in turn improved the permeability and 
porosity of the strata (Vivier, 1996). Subsequent weathering and diagenesis of the 
strata further improved permeability and porosity (Vivier, 1996). 
 
In categorizing different interstices, Vegter (2001) differentiated between four types 
of saturated interstices: 
 
• pores in disintegrated/decomposed, partly decomposed rock and fractures 
restricted primarily to a zone below groundwater level; 
• fractures restricted primarily to a zone directly below groundwater level; 
• openings varying in size from fissures to caves; also pores in dissolution 
residuum and collapsed unconsolidated deposits; and 
• pores in semi- and unconsolidated deposits. 
 
• Vegter (2001) expanded on this by examining interrelated factors contributing 
to the complex interactions of weathered thickness, porosity, permeability and 
groundwater flow (Table 8). 
 
 90
Vegter (2001) classified groundwater regions into two subdivisions: 
 
• Regions of primarily water-bearing formations comprising hard-rock 
formations underlying primary water-bearing formations which may act as 
aquifers. 
• Regions of secondary water-bearing formations, including fluvial deposits, 
while being further classified on principal rock type and erathem/system: 
• crystalline metamorphic and igneous; 
• intrusive 
• extrusive 
• sedimentary 
• composite. 
 
Composite regions were further classified as: 
• having no dominant major rock type; 
• having several major lithostraphic units involved; and 
• having more extensive primary aquifers than alluvial deposits 
present. 
 
Two basic types of flow are present in Karoo aquifers, namely, matrix flow and 
mesofractural flow. Fractures in Karoo aquifers have a limited storage capacity, and 
therefore storage must be in the matrix, which is usually composed of fine grained 
mudstones, siltstones and shale, interbedded with coarser grained sandstone (Vivier, 
1996). Karoo rocks originally had a very high porosity, but as previously mentioned, 
through compaction and cementation due to the overburden, the primary porosity of 
the rocks was severely reduced, resulting in pores and microfractures being very 
small. Thus, the modern hydraulic conductivity of these rocks is also very low. The 
flow of water in a fractured medium is largely controlled by the dimension of the 
fractures, connectivity, and orientation, with the former two being either limited, 
weak, or non-existent in Karoo aquifers. Vivier (1966) concluded that borehole yields 
in Karoo aquifers was largely controlled by the horizontal fractures which may 
deform considerably due to pumping and extraction. 
 
 91
Table 8. Interrelated factors contributing to the complex interactions of weathered 
thickness, porosity, permeability, and groundwater flow (after Vegter, 
2001). 
 
 
The degree of tectonic deformation and fracturing. 
 
The degree of non-tectonic fracturing such as thermal shrinkage and sheet formation. 
 
The climate, as in past and present rainfall and temperature. 
 
The age of the land surface. 
 
The relief. 
 
The mineralogical composition and texture of the host rock 
 
Chemical weathering is the dominant process in the development of a weathering 
profile on crystalline basement rocks. 
 
Groundwater is the principal weathering agent in the saturated lower part of the 
weathering profile. 
 
The amount of dissolved oxygen and carbon dioxide in the percolating water, where 
groundwater flow rates control the rate of chemical reaction by the extent to which it 
supplies hydrogen ions and dissolved oxygen. 
 
The flow rate through the weathering system is dependant on the availability of 
recharge, the permeability of the weathered material, and the hydraulic gradient 
between the recharge and discharge areas. 
 
The hydraulic conductivity of the weathered material is therefore a function of the 
extent of chemical weathering, and is inherently linked to the history of the 
groundwater flow through the system. 
 
 
 
4.2.1 Aquifers and Aquitards 
 
Vegter (1995) defined an aquifer as a stratum which contains intergranular interstices, 
or a fissure/fracture (as such), or a system of interconnected fissures/fractures (as 
such) capable of transmitting groundwater rapidly enough to directly supply a 
borehole or spring”. Groundwater fills intergranular open spaces in the unconsolidated 
sediment and in weathered previously consolidated rock, as well as fractures in the 
hard rock. Saturation alone does not imply the existence of an aquifer, as a 
 92
precondition in identifying the formation of an aquifer is the gravitational mobility of 
water in a saturated rock formation as a result of its permeability. 
 
An aquitard was defined by Vegter (1995) as “a body of poorly permeable rock that is 
capable of slowly absorbing water from and releasing water to an aquifer. Aquitards 
do not transmit groundwater rapidly enough by themselves to directly supply a 
borehole or spring as they usually comprise layers of clay or non-porous rock with a 
low hydraulic conductivity. A formation that has no interconnected openings, and 
therefore cannot absorb or transmit water (impenetrable), is referred to as an 
aquiclude, or an aquifuge.  
 
Baran (2003) recognized four basic types of aquifer. 
 
4.2.1.1 Intergranular Aquifers 
 
Intergranular aquifers are comprised of unconsolidated sediment, such as sand and 
gravel, where water is stored in the intergranular pores. In the Free State, intergranular 
aquifers are poorly represented and only occur as narrow strips of alluvium along 
some major river valleys such as the Vaal, Wilge, Sand, Vet Modder and Vals Rivers. 
Some towns, such as Ficksberg in the eastern Free Sate, obtain up to 50 % of their 
water from alluvial beds along the Caledon River (Vivier, 1996). 
 
4.2.1.2 Fractured Aquifers 
 
Fractured aquifers occur in hard, mainly quartzitic, rock formations where water 
occurs in fissures, joints, fractures, and faults. Tectonic forces, and to a much lesser 
degree, weathering processes, have produced these fractures, resulting in a limited and 
lower storage capacity than other aquifers. 
 
4.2.1.3 Karstic Aquifers 
 
Karstic aquifers are associated with carbonate rocks such as dolomite and limestone, 
where water is stored, and transmitted through solution cavities, channels, and 
fractures. A characteristic of these aquifers are their solution chambers, formed by the 
 93
enlargement of cracks and fissures through circulating groundwater containing 
carbonic acid, which in turn dissolves the carbonate rock. These aquifers are 
characterized by a high storage capacity and high yields. 
 
4.2.1.4 Intergranular  and Fractured Aquifers 
 
Intergranular and fractured aquifers occur in tectonically altered and subsequently 
weathered and fractured rock. They are found in Precambrian rocks and Karoo Basin 
sediments. Water occurs in a vertical profile where the upper weathered and 
decomposed rock zone is in hydraulic contact with fractured zone down to the solid 
and fresh, rock formations below. These aquifers occur throughout most of the Free 
State with the Florisbad spring aquifer falling into this category. It will be noted from 
Figure 21 that there is little difference between the yield of boreholes located in Ecca 
Group rocks and those located in dolerite sills. Borehole yields from the Ecca and 
other Groups, as well as the dolerites, area are considered low. 
 
4.3 GEOHYDROLOGY OF THE KAROO SUPERGROUP 
 
Usher et al. (2006) state that the behaviour of Karoo fractured aquifers is ultimately 
determined by their unusual geometry, particularly where horizontal, bedding-parallel 
fractures are present. These features provide not only the conduits for water to 
boreholes in Karoo aquifers, but also play a prominent role in the interactions 
responsible for the behaviour of these aquifers. One very important consequence of 
these interactions is that flow in Karoo aquifers is not only radial and horizontal, but 
also linear and vertical. This property differs so much from that of the theoretical 
media usually presented in the literature on aquifer mechanics that the existing 
conceptual models are useless for the analysis of hydraulic tests performed on Karoo 
aquifers. 
 
4.3.1 Dwyka Group 
 
Dwyka Group rocks have a very low hydraulic conductivity and virtually no primary 
voids, with the Group constituting of a low-yielding, fractured, aquifers where water 
 
 94
 
50
45
40
35
30
25
20
15
10
5
0
0-0.1 0.1-0.5 0.5-2.0 2.0-5.0 > 5.0
Yield (l /s)
Ecca Group Beaufort Group Molteno Form. Elliot Form. Dolerite sills
 
 
Figure 21.  Some comparative examples of borehole yields between the Karoo Group Formations and dolerite sills (after Baran. 
    2003). 
 
 95
P e r c e n t a g e
is confined to narrow discontinuities like jointing and fracturing (Woodford and 
Chevallier, 2002). This has resulted in the formation of aquitards, rather than 
aquifers (Woodford and Chevallier, 2002). Because of the marine environment 
under which the Dwyka sediments were deposited water from these aquifers 
tends to be saline. Exploitable aquifers therefore only occur where sand and 
gravel were deposited on beaches, or where the Dwyka group was significantly 
fractured (Woodford and Chevallier, 2002). Little information is available in the 
geohydrology of the Group. Despite this, the artesian borehole at Glen, produced 
a yield of 2.53 l/s in Dwyka tillites at ± 760 m (Baran, 2003). 
 
4.3.2 Ecca Group 
 
Ecca Group rocks do not form high yielding aquifers due to the prevalence of 
dense shales, with the Volksrust shales being considered totally impenetrable 
(Woodford and Chevallier, 2002); Baran, 2003). Shale thickness varies from 
1500 m in the south to 600 m in the north (Woodford and Chevallier, 2002). 
Ecca sandstones have a low permeability as a result of them having been poorly 
sorted, with porosity being further decreased through diagenesis (Woodford and 
Chevallier, 2002). Porosity decreases from 0.10% in the north to >0.02% in the 
south (Woodford and Chevallier, 2002). Being formed in a fluvial environment, 
sand-grains tend to align their longest axis parallel to the flow direction to form 
an anisotropic matrix (Vivier, 1996). This suggests that the hydraulic 
conductivity of the matrix will be greater in the direction of the alignment of 
sand grains, as opposed to perpendicular to the alignment (Vivier, 1996). Baran 
(2003) expected the weathering and fracturing of the arenaceous Vryheid 
Formation to be more developed than that of the argillaceous Volksrust 
Formation. Ecca Group waters are of a sodium bicarbonate nature, with 45% of 
boreholes having an EC value above the recommended limit for drinking water 
(70 mS/m), 2% above 300 mS/m, with an average of 78.7 mS/m (Baran, 2003). 
Yields from boreholes are generally considered as low, with comparative 
borehole yields being given in Figure 21. 
 
 
 
 96
4.3.3 Beaufort Group 
 
The older Adelaide Subgroup consists predominantly of argillaceous sediments, 
while the younger Tarkastad Subgroup is more arenaceous and described as 
being argillaceous and arenaceous (Baran, 2003). Since the depositional 
environment of the Beaufort Group has many similarities to that of the Ecca 
Group, it can be expected that Beaufort aquifers will also be of an anisotropic 
nature (Vivier, 1996). The geometry of Beaufort Group aquifers is further 
complicated by the migration of braided and meandering streams over a flood 
plain, resulting in multi-layered, multi-porous, aquifers of variable thickness 
(Vivier, 1996; Woodford and Chevallier, 2002). The dominance of mudstone, 
shale and fine-grained sandstone render the sedimentary units of the Beaufort 
Group, low in permeability. The complexity of Beaufort Group aquifers is 
enhanced by many of the coarser and more permeable sediments being lens 
shaped (Woodford and Chevallier, 2002). Borehole yields of the entire Beaufort 
Group are considered low (Figure 21). 
 
4.3.4 Stormberg Group 
 
4.3.4.1 Molteno Formation 
 
Molteno Formation rocks were formed by braided and meandering streams with 
basal layers being comprised of conglomerates and coarse sandstone followed by 
fine-grained sandstones and shales (Vivier, 1996; Woodford and Chevallier, 
2002). A characteristic feature of the Molteno Formation is that the sediments 
are laterally uniform over large areas (Vivier, 1996). However, despite borehole 
yields being considered low to moderate yielding (Baran, 2003), the persistence 
of the sedimentary bodies makes for a more favourable aquifer geometry in so 
far as groundwater storage is concerned (Woodford and Chevallier, 2002). 
Pebble conglomerates and coarse- grained sandstone at the base of the formation 
should also allow for ideal aquifers (Woodford and Chevallier, 2002). There 
appears to be little information on the geohydrology of this formation. (Figure 
21). 
 
 97
4.3.4.2 Elliot Formation 
 
Rocks of the Elliot Formation were deposited by meandering streams under 
highly oxidizing conditions, and comprise red mudstone, succeeded by a red 
fine-grained sandstone (Vivier, 1996; Woodford and Chevallier, 2002). The red 
mudstone is more conducive to aquitards than aquifers (Woodford and 
Chevallier, 2002). Boreholes are considered moderate to high yielding (Figure 
21). 
 
4.3.4.3 Clarens Formation 
 
The sediments of the Clarens Formation are related to aeolian sand, playa lakes, 
and sheet-flood deposits (Woodford and Chevallier, 2002). Owing to the aeolian 
nature of the deposits, the rocks comprise homogeneous fine to very fine-grained 
sandstone, with a high clay content being derived from silt deposited in the playa 
lakes (Woodford and Chevallier, 2002). The Clarens Formation is dominated by 
red mudstones, which again are more conducive to aquitards, rather than 
aquifers. This Formation is considered to have the most homogeneous geometry 
in the Karoo Supergroup, and is poorly fractured with a low permeability, but 
with a relatively uniform and a relatively high porosity average of 8.5% 
(Woodford and Chevallier, 2002). Due to the formation being poorly fractured, 
hydraulic conductivity is low. Boreholes are considered moderate to low 
yielding 
 
4.4 GEOHYDROLOGY OF THE KAROO DOLERITE SUITE 
 
4.4.1 Dolerite Dykes 
 
The intrusion of dolerite bodies in the host rock formations created zones of 
fracturing in both the host rock and the dolerite itself, with these fracture zones 
becoming natural underground drainage systems for groundwater (refer to 
previous discussion in section 2.2.4.2). Fracturing usually occurs on both sides 
of the dolerite intrusion, with subsequent weathering enhancing their 
permeability, but at greater depths (>30-40 m) fractures become fewer, and 
 98
borehole yields decrease considerably. Dolerite dykes are, to a large extent, 
impenetrable and restrict the flow of groundwater (Vivier, 1996; Woodford and 
Chevallier, 2002). Vivier (1996) states that a solid dolerite dyke’s state of 
weathering is an important indication as to yields that can be expected from 
boreholes drilled near to dolerite dykes, with highly weathered dykes producing 
higher yielding boreholes. 
 
Dolerite dykes are the preferred drilling targets for groundwater in the Karoo 
Basin, with 25% of boreholes in the Victoria West district having being sited 
into, or along side, dolerite dykes (Woodford and Chevallier, 2002). In northern 
Kwazulu-Natal, 80% of successful boreholes are directly, or indirectly, related to 
dolerite intrusions (Woodford and Chevallier, 2002). The state of a dyke may be 
critical in the success of a borehole. For example, weathered dykes may 
themselves be fractured, while unweathered dykes may have few, or no, 
fractures. The hydrogeology of intrusive dykes is complex. It has been stated 
that the high permeability of dyke contact zones was not only as a result of the 
fracturing of the host rock, but also as a result of shrinkage joints developed 
during the cooling of the intrusion (Woodford and Chevallier, 2002). The degree 
to which the host rock is metamorphosed and fractured away from the dyke is 
also critical, with the zone of metamorphosis commonly being less-than, or equal 
to the width/ thickness of the intrusive body (Woodford and Chevallier, 2002). 
This distance varies in the literature from 1 to 20 m, depending on the area, 
while the dip of the dyke will also play an important role in determinations 
(Woodford and Chevallier, 2002). Where younger dolerite dykes cut across older 
sills, particularly in valley-bottom situations where sill material is highly 
weathered, these provide good locations for groundwater (Woodford and 
Chevallier, 2002). Transgressive fracturing is often well developed in such areas 
(Woodford and Chevallier, 2002). 
 
Horizontal and oblique transgressive fractures within the dolerite and country 
rock are the dominant water bearing features in the Queenstown area, with 
fractures in some areas extending for up to 90 m away from the dyke contact 
(Woodford and Chevallier, 2002). The dip of a dyke per se does not appear to 
have an influence on the borehole yield, but is important in the sitting of the 
 99
borehole. Conversely, porosity of the host rock may decline in the host rock 
contact area with the metamorphic aureole intrusive, as a result of re-
crystallisation of the host rock silicates and infiltration/cementation by magmatic 
silica (Woodford and Chevallier, 2002). It has also been found that in the eastern 
Free State the best borehole yields originated from dykes which were from 7 to 
11 m wide (Woodford and Chevallier, 2002), but again, this will vary from area 
to area. Although groundwater is often associated with dolerite dykes, this is not 
always the case. In the drilling of the Fish River tunnel between the Gariep Dam 
and the upper reaches of the Fish River, a distance of 83 km, 55 dykes were 
intersected, of which 49 were dry (Chevallier et al. (2001). 
 
4.4.2. Dolerite Sills and Ring-Complexes 
 
Because of factors such as size, hardness, thickness, and structural complexity, 
dolerite sill and ring-complexes have largely been overlooked in a hydrological 
and water exploration sense (Vivier, 1996; Woodford and Chevallier, 2002). 
This is often due to the easier way in which the multitude of linear dykes and 
fractures can be detected and located (Woodford and Chevallier, 2002). 
 
Baran (2003) provided almost no information on the geohydrological 
characteristics of dolerite sills and ring-complexes, while due to the above 
mentioned factors, Woodford and Chevallier (2002) cite specific site 
investigations. At their Victoria West dolerite sill and ring-complex site 
Woodford and Chevallier (2002) noted that very few water bearing fractures 
were encountered between the top of the inner sill and the host rock. High yields 
were encountered in horizontal water bearing dolerite offshoots, or fractures, 
found in the host-rock on the foot wall contact of the inclined sheet, as well as at 
the base of the of the sill. Moderate to high yields were encountered in the 
sediments above the inner sill, most probably due to weathering and the near 
surface release of residual compressive stress developed inside the saucer during 
intrusion. It has been shown that water bearing open fractures occur at specific 
locations within the dolerite and surrounding host-rock (Woodford and 
Chevallier, 2002). These are to be found in the sediment above an up-stepping 
sill or at the base of an inner-sill, and at the junction between a feeder dyke / 
 100
inclined sheet and sill (Woodford and Chevallier, 2002). Generally dolerite sill 
borehole yields are considered low to moderate. 
 
4.5 GEOHYDROLOGY OF OTHER POST KAROO INTRUSIONS 
 
4.5.1 Breccia Plugs and Volcanic Vents 
 
These structures are both pipe-like and filled with brecciated fractured material, 
providing highly permeable targets for groundwater (Woodford and Chevallier, 
2002). This is due to their geometry and large vertical extent of uniform 
material. Due to their limited size, it has been proposed that they are only of 
significance as conduits for rapidly extracting water from the subsurface, and 
therefore their yield sustainability is dependant on recharge and storativity of the 
host-rock (Woodford and Chevallier, 2002). The high success rate of boreholes 
(70% yield in excess of 3 l/s) at Calvinia, Luiperdkop, Doorn-Laagte, Carnarvon, 
and Bitter Poort, can be considered excellent when compared to yields 
encountered in Ecca shale of the western Karoo (Woodford and Chevallier, 
2002). Yields of >8 l/s are almost always encountered in intensely brecciated 
sections of a plug (Woodford and Chevallier, 2002). 
 
Volcanic vents have a similar occurrence of groundwater to breccia plugs, due 
them having similar shapes, sizes, and degree of brecciation (Woodford and 
Chevallier, 2002). However yields may be higher due to their eastern location in 
a higher rainfall and recharge area. Boreholes in Ladybrand and Matatiele 
produced yields of 38 l/s and 20 l/s respectively (Woodford and Chevallier, 
2002). 
 
4.5.2 Kimberlites 
 
Unlike the dolerites and volcanic plugs, kimberlite intrusion did not result in an 
intense thermal metamorphism of the Karoo sediments, and therefore did not 
significantly alter the hydrological properties of the host sediments (Woodford 
and Chevallier, 2002). On a larger scale strong regional jointing and reactivation 
of existing structures that accompanies the emplacement of kimberlite swarms 
 101
may be important for the occurrence of groundwater (Woodford and Chevallier, 
2002). Larger kimberlite blows may represent more permeable zones as they are 
more heterogeneous, brecciated, and more highly weathered. Jagersfontein and 
Fauresmith in the Free State both obtain water from the now abandoned 
Jagersfontein diamond mine (Vivier, 1996; Woodford and Chevallier, 2002). 
Overall, kimberlite fissures are seen to yield small amounts of water due to the 
clogging of near surface joints by clays derived from the weathering of the 
kimberlite (Woodford and Chevallier, 2002). Transgressive water bearing 
fractures seen in dykes do not appear to have been developed along kimberlite 
fissures (Woodford and Chevallier, 2002). 
 
4.6 GEOHYDROLOGY OF RECENT DEPOSITS 
 
Intergranular aquifers in alluvium are discussed in Section 4.2.1.1. Alluvial beds, 
although being limited to narrow strips along main river courses, supply large 
volumes of groundwater to towns in the Karoo Basin. The Caledon River, for 
example supplies Ficksburg with approximately 50% of its water (Vivier, 1996). 
Calcrete deposits are yet another source of underground water: Vivier (2003) and 
Woodford and Chevallier (2002) note that recharge to these aquifers is 2 to 5% 
higher than in average Karoo aquifers. Farmers in the Petrusburg district of the 
central Free State tap into calcrete aquifers, up to 30 m thick, for irrigation 
purposes (Vivier, 2003; Woodford and Chevallier, 2002). 
 
4.7 NON-INTRUSIVE TECTONIC FEATURES 
 
A number of non-intrusive tectonic features may also influence the hydrology of 
areas of the Karoo Basin, but these are not within the scope of this study, and are 
not dealt with detail here. These include regional lineaments, folding, vertical 
jointing and faulting, bedding-plane fracturing, and sesmotectonic / neotectonic / 
uploading features (Woodford and Chevallier, 2002). 
 
 
 
 
 102
4.8 SYNOPSIS 
 
As has been emphasised in the foregoing sections, the rocks of the Karoo Super 
Group and Karoo dolerite suite are extremely complex, while research on the 
geohydrology of dolerite ring and sill complexes can be considered to be in its 
infancy. It is only in recent years that hydrocensuses have been carried out on 
previously existing data. However researchers have had to be extremely careful 
in compiling this information as often no proper records were kept, or the 
records are not of a suitable standard for such an exercise. 
 
The geohydrology of the main Karoo Basin is as complex as the formations it 
examines and, interested parties wishing to obtain more comprehensive 
information on the subject, are advised to obtain this from any of the references 
used here, but in particular, from Woodford and Chevallier (2002) 
 
 
 103
 
 
 
 
 
 
 
 
 
 
 
Chapter 
 
5 
 
 
 
 
 
 
 
The 
Florisbad 
Springs 
 
 
 104
CHAPTER 5 
 
THE FLORISBAD SPRINGS 
 
 
5.1 INTRODUCTION 
 
The previous chapters have provided information on the many components which 
have contributed to the formation of the Florisbad spring site and the fossilization of 
faunal remains. Had the merger of all of these components, in conjunction with their 
interrelated processes, not occurred in the way, and at the times that they did, the 
Florisbad spring site, as we know it today, may very well never have existed. It is 
therefore imperative that as many components, and sub-components, pertaining to the 
formation and sedimentation of the Florisbad spring site are examined in order to 
obtain a holistic representation of how the site developed into a unique and important 
archaeozoological site. This chapter synthesises these components in developing and 
presenting a comprehensive picture of the geohydrological and surface processes 
controlling the depositional and fossilization environment at Florisbad. 
 
5.2 CLASSIFICATION OF SOUTH AFRICAN SPRINGS 
 
As the Florisbad springs have played such an important role in the history of the site it 
is prudent to briefly examine the classification of South African springs (Table 9) in 
order to be able to compare the Florisbad springs to the various types of springs that 
occur in South Africa. 
 
Springs often form at the site of faults and fissures, but they may also form at dykes, 
when such a dyke impedes the subterranean water flow. A number of springs, such as 
those at Aliwal North, Cradock, Rooiwal, Badsfontein, and Knegha Drift, have been 
recorded as forming at dolerite dykes (Kent, 1949). It is at such a dolerite intrusion  
 
 
 
 105
Table 9. Classification of South African spring waters and the criteria used in their classification. 
 
 
Reference Basis of Classification Classification Criteria for Classification 
 
    
Rindl (1916) Chemical composition Indifferent springs <1000 mg/l dissolved solids; <1000 mg/l CO2. Water 
temperature well above ambient temperature 
    
  Sulphur springs Based on an odour of sulphurated hydrogen. Water temperature 
usually above ambient temperature 
    
  Chalybeate springs >700 mg/l of ferrous iron. South African chalybeate springs 
thermal, other countries cold 
    
  Saline springs* High sodium chloride content and appreciable potassium (values 
not given) (K at Florisbad low) 
    
  Alkaline springs >1000 mg/l of solids with carbonate or hydrocarbonate ions 
predominating 
    
  Earthy springs >1000 mg/l of solids with bicarbonates of calcium and 
magnesium predominating 
    
  Table water springs Form a heterogeneous group comprising alkaline waters 
    
Bond (1946) Industrial and power Highly mineralised chloride- Total solids >1 000 mg/l:- Cl >27%; SO4 >5% 
production use sulphate waters 
    
  Slightly saline chloride Total solids >300 - <500 mg/l:- Cl >27%; SO4 >3% 
waters 
    
  Temporary hard carbonate Total solids <800 mg/l:- pH 7.6 
waters 
    
  Alkaline soda carbonate Total solids <1 000 mg/l:- Na2Co3 or NaHCO3 >15%. 
waters Permanent hardness nil 
(continued…..) 
 106
Table 9 (continued). 
 
 
Reference Basis of Classification Classification Criteria for Classification 
 
 
Bond (1946)  “Pure” waters Total solids <150 mg/l:- pH <7.1 
(cont.) 
    
Kent (1949) Allocated temperature Warm* 25°-37°C 
ranges to spring waters 
based on the air temperature 
of the region 
  Hot or hyperthermal 37°-50°C 
    
  Scalding >50°C 
    
Kent (1949) Hydrogen sulphide content Sulphuretted waters Over 10 mg/l dissolved H2S (including HS`) 
as a measure of therapeutic 
value 
    
  Moderately sulphuretted Between 5 and 10 mg/l dissolved H2S (including HS`) 
waters 
    
  Slightly suphuretted waters Between 1 and 5 mg/l dissolved H2S (including HS`) 
    
  Sulphurous waters Springs containing dissolved sulphur dioxide – not known from 
South Africa 
    
Mazor and Total dissolved ions Fresher water springs Total dissolved ions 90-432 mg/l (no intermediate group 
Verhagen classifications are given) 
(1983) 
  More saline springs* Total dissolved ions 936-2364 mg/l 
    
 
 107
 
that the Florisbad springs have formed (see Section 5.4.1). Most springs, within the 
Karoo Supergroup rocks, rise along dolerite dykes of an early Jurassic (Karoo) age and 
can mostly be classified as warm water springs, as opposed to hot water springs which 
originate from older and deeper archaean rocks (Kent, 1949). Vegter (1995) lists 59 
thermal springs and a further 57 cold water springs yielding >1000 cm3 per day. 
 
South African spring-waters have been classified by a number of researchers, dependant 
on their particular interest in spring-waters. From the categorization of spring-waters in 
Table 9, and marked with an apteryx, the Florisbad spring can be described a highly 
salinized, highly saline, predominantly methane gas, warm water spring. 
 
5.3 GEOHYDROLOGY OF THE FLORISBAD AREA 
 
Florisbad falls into a region composed of mid Palaeozoic to early Mesoic strata, namely 
groundwater region 31, the Central Pan Belt (Vegter, 2001). The principal rock types are 
shales of the Tierberg Formation (Ecca Group) and dolerite dykes (see Sections 2.2.4 and 
4.4). The surface lithology is predominantly argillaceous rocks comprising shale, 
mudstone and subordinate sandstone. Florisbad also lies in a zone of inter-granular and 
fractured aquifers in to which dolerite dykes and sills have intruded (see Section 5.4.1). 
With a groundwater yield of 0.5 –2.0 l/s (Vegter, 2001), the region could be described as 
having a very low, to low, development potential (pers. com. Bertram, E. Department of 
Water Affairs and Forestry, 20 May 2008). 
 
From information supplied by Vegter (1995, 2001) the hydrology of the Florisbad area 
can be described as follows: 
 
• Florisbad falls into the sedimentary groundwater region. 
• The depth to groundwater level is 10-20 m with a range depth of 8-15 m. 
 108
• The storage coefficient for the compact (i.e. lacking significant primary porosity) 
argillaceous strata is <0.001. This is represents 1 cm3 of rock containing less than 
one thousandth part of a cubic metre of water i.e. one litre 
• The recommended drilling depth below groundwater level is <20 m. 
• The mean annual recharge in mm per annum is 15-25 mm. 
• The total dissolved solids (mg/l) range in terms of geometric standard deviation (it 
is noted that fluoride concentrations exceed 1 mg/l in more than 20% of the 
analysed samples) is 1000-1500 mg/l. 
• The ground water quality of the area will have an EC range of from 70 to >1000 
mS/m. 
• There are two principal hydrochemical facies, based on a Piper diagram, with 
each comprising between 30% and 40% of analysed samples: (a) 
calcium/magnesium chloride/sulphate facies, and (b) calcium/magnesium 
bicarbonate/carbonate facies. 
• The contribution of groundwater to river flow is negligible, or non-existent. 
• The probability of a successful borehole yielding greater than 2.0 l/s is 20-30%. 
• The probability of drilling a successful borehole yielding more than 0.1 l/s is 40-
60%. 
 
5.4 THE FLORISBAD SPRING AQUIFER 
 
The Florisbad spring aquifer, as related to the consistency of the spring flow is of 
considerable importance in that it has never been reported that the springs have ceased to 
flow. It is therefore surprising that no specific research has been carried out on the 
Florisbad spring aquifer, and therefore information regarding its location, size, recharge, 
storage capacity, flow rates, travel times, and abstraction, is either unavailable, or 
unreliable. The research reported on in Appendix I determined that short-term rainfall had 
a considerable effect on the chemistry of the groundwater. It was further concluded that 
neither short-term, nor longer-term rainfall had any significant effect on the properties of 
the spring-water over the study period. It is proposed that any probable fluctuations in the 
quality of the spring water, due to recharge from long-term rainfall, will be smoothed out 
 109
by the suggested large size of the aquifer, its probably considerable distance from the 
spring eyes, and its depth. (Appendix I). On this basis it was concluded that if long-term 
rainfall did have any affect on the spring-water, this would be over a considerably longer 
period than the 10 year period examined in this study, or the 84 year period over which 
the spring-water had been analysed. This, however, does not alter the opinion put forward 
in this thesis that the spring-water was never historically responsible for the fossilization 
of faunal remains (see Section 5.6.4). 
 
Grobler and Loock (1988a) postulated that the intake of the Florisbad aquifer was located 
30 km north of Florisbad at Basberg where permeable Beaufort Group sandstone 
occurred at an elevation 150 m above Florisbad. The possibility that the intake area of the 
spring may lie equidistant to the south-east of Florisbad, in the hills north of 
Bloemfontein, was also suggested by Grobler and Loock (1988a). Further more, there 
appears to be uncertainty as to the travel time that recharge water could take to travel 
from the intake area, through the aquifer, to the spring eyes. Grobler and Loock (1988a) 
calculated a travel time of anything from 160 to 16 000 years, with a probable 1 600 
years. 
 
It has been shown that warmer and deeper circulating water has a higher fluoride content 
(up to 5 mg/l) than cold-water springs and shallow aquifers (0.2 mg/l), indicating that a 
high fluoride content may be an indication of deep-seated thermal groundwater 
(Woodford and Chevallier, 2002). This would also tend to give support to Grobler and 
Loock’s (1988a) estimates of the Florisbad aquifer being 500 metres below ground level 
(m. bgl), with water issuing at 29 °C with a fluoride content of 5.5 mg/l. In an attempt to 
locate oil within the Karoo sediments, a deep exploration borehole (1 495m) was sunk at 
Glen railway station, 32 km east of Florisbad. This borehole encountered warm artesian 
water in Dwyka tillite at between 747 and 775 m. bgl, with water issuing at 35 °C, and a 
fluoride content of 5.3 mg/l (Baran, 2003). The properties of the Glen borehole (Table 
10) are in many ways similar to those of the Florisbad springs. Fluoride concentrations at 
 
 110
 
Table 10. A comparison of water quality between the Glen artesian borehole and the 
Florisbad spring (after Baran, 2003 [B03]); Kent, 1971 [K71]); Fourie, 1970 
[F70]; App. I [Appendix 1]. 
 
 
Property Glen Borehole Florisbad spring (App. I) Florisbad 
 (B03) 1988 1999 Other analyses 
 
 
Temperature °C 35 °C issuing - 29 °C issuing - 
from between 747 from 500 m 
and 775 m 
Flow rate l/s 2.5 - - 2.5  (K71)
TDS mg/l 2 034 2202.9 2362.7 2277   (F70)
EC mS/m 325 388 398 -
pH 7.9 8.91 9.43 8.3   (F70)
Flouride mg/l 5.3 - 5.5 6   (F70)
NaCl mg/l 1 895 2086 2226 2099   (F70)
Ca mg/l 6 (K71) 98.4 100 87   (F70)
Bicarbonate 
(HCO3) mg/l 43 - - 49  (F70))
Temporary 
hardness (CaCO3) 
mg/l 35 - - 177   (F70)
Permanent 
hardness (CaCO3) 
mg/l 215 - - 40   (F70)
     
 
 
 
 111
other warm water springs in the Karoo basin range from 4.8 mg/l at Aliwal North (36.9 
°C) to 13.2 mg/l at Fort Beaufort (27.0-29.0 °C) (Woodford and Chevallier, 2002). 
 
5.4.1 Control of the Springs by Geological Features 
 
The intrusion of dolerite bodies into host rock formations created zones of fracturing in 
the host rock and in the dolerite itself, with the fracture zones becoming a natural 
underground drainage system for groundwater stored in the fractured and weathered, rock 
(Baran, 2003). In dolerite related zones weathering is important as it enhances the 
permeability of the rock, and thus the degree and depth of weathering could be critical in 
water yields (Baran, 2003). No specific data is available for the yield of water from 
dolerite dykes in the area due to the poor quality of borehole logs over the years, by, in 
many instances, unqualified persons (Baran, 2003). 
 
Dreyer (1938a) uncovered a section of a dolerite dyke in the excavations at Florisbad, 
which led him to divide the site into eastern and western sections (Figure 22). It was 
suggested that the unconsolidated deposits on the eastern side of the dyke were older than 
those on the western side of the dyke (Dreyer, 1938a). This was based on the difference 
in the way in which the artefacts he had found had been prepared: The western artefacts 
were described as being parallel-sided flakes, while the eastern artefacts exhibited 
convergent flakes and points (Dreyer, 1938a; Oakley, 1954; Sampson 1974). 
 
The spring eyes rise along an east to west strike and in an attempt to establish what was 
causing the spring eyes to surface along this strike, Fourie (1970) carried out a 
magnetometer survey. The results showed no anomalies outside of the excavations, 
possibly due to the thickness of the sand (Fourie, 1970). In the excavation, where there 
was minimal sand cover, anomalies were recorded, but these were attributed to noise 
from nearby buildings and water tanks, and were considered unreliable (Fourie, 1970). 
Grobler and Loock (1988b) also carried out a magnetometer survey of the site, stating 
that all the data, including borehole results, defined a dolerite sill with a plunge to the 
 
 112
 
 
 
 
 
 113
north-west (Figure 22). 
 
In the excavations, just west of the dyke, Dreyer (1938a) unearthed what he described as, 
a boulder layer in a deep blue clay matrix. A 1.2 m excavation into this layer revealed a 
zone of homogeneous, highly weathered material. The dyke delineated by Grobler and 
Loock (1988b) was thought by them to be an appropriate structure for the channelling 
and emission of underground water in the form of springs at Florisbad. 
 
There is no information for groundwater yields of boreholes located on dolerite dykes in 
the immediate Florisbad area. However, considerable success has been recorded for 
borehole yields in dolerite dykes located in other areas of the Karoo Supergroup. At 
Phuthaditjhaba in the eastern Free State, for example, 87% of boreholes sunk into dolerite 
dykes were successful with an average yield of 3.2 l/s (Kruger and Kok, 1976), where 
0.125 l/s is considered successful (Baran, 2003). Where boreholes were sunk into the 
fractured sediments along the dykes, only 40% of boreholes were successful, with an 
average yield of 1.4 l/s (Kruger and Kok, 1976). These results can be compared to 
boreholes sunk into Karoo sediments not affected by dolerite dykes, which were 16% 
successful, with a yield of 0.5 l/s (Baran, 2003). 
 
5.4.2 Temperature 
 
The temperature of the Florisbad spring-water has remained constant at ± 29 °C over the 
past 84 years, having only fluctuated by 0.17 °C over this period, with a maximum of 
29.05 °C and a minimum of 28.88 °C (Appendix I). Beside the temperatures mentioned 
in Appendix I, Kent (1964) measured 29.8 °C, while Kent (1948) recorded a temperature 
of 34.4 °C for the main eye, which is considered here to be in error. It was calculated by 
Grobler and Loock (1988a) that, if the water intake recharge area were located at 
Basberg, north of Florisbad, water would have to descend to the contact zone between the 
Karoo Supergroup sediments and Upper Ventersdorp Group basement rocks at 
approximately 500 m in order to reach a temperature of 32 °C to 33 °C, so as to issue at 
29 °C. 
 114
It is interesting to note that of the many boreholes drilled in the immediate Florisbad area, 
none encountered warm water above 25 °C (Kent, 1964) (see Kent, 1949, Table 9). At 
Vlakkraal, 5 km south of Florisbad, two springs existed in a vlei which was later covered 
by the waters of the Krugersdrift Dam. These springs occur in shales and sandstones of 
the Ecca Group, and according to Kent (1971), their chemical composition was similar to 
that of Florisbad springs, although the temperature of the water was only 22 °C, implying 
a much shallower aquifer. 
 
The temperature of spring water is largely determined by the depth at which it originates. 
It is apparent that if the hypothesis put forward by Grobler and Loock (1988a) is correct, 
and water must descend to approximately 500 m in order to reach a temperature of 32 °C 
to 33 °C in order to issue at 29 °C, then the constant temperatures measured at the 
Florisbad springs would suggest that the temperature of the spring water must have 
remained fairly constant throughout its history, with no evidence suggesting that it ever 
had deeper origins, or higher temperatures. Thermal springs with higher temperatures 
usually exhibit characteristics reflecting these higher temperatures, such as deposits of 
travertine, sinters, silicified or mineralised micro-organisms, stromatolites, and other 
chemical, or biochemical precipitation. (Kuman et al., 1999; Farmer, 2000; Allen et al. 
2000). At Florisbad there is no evidence of these, supporting a hypotheses for low 
palaeotemperatures. Allen et al. (2000) noted that organic material is rare in carbonate 
sinters deposited at above 30 °C, due to the high decomposition rates in high thermal 
environments. Therefore, the large quantities of organic material in the Florisbad 
sediments would also strongly contradict a higher palaeotemperature. Kent (1949) noted 
that, possibly due to their low temperatures, very few South African thermal springs have 
given rise to mineral deposits. Kuman et al., (1999), and research reported on in 
Appendix III, concluded that the palaeotemperature of the spring-water was similar to the 
present. 
 
 
 
 115
5.4.3 Discharge Rate 
 
Information on the discharge rate of the Florisbad springs, such as the methods used and 
location of the samples, appear to be mostly vague and unreliable. The sampling locations 
could have a considerable influence on the discharge rate measured, and it appears that 
discharge rates in the literature are confined to the area of the swimming baths. Fourie 
(1970) recorded 21 spring eyes of varying sizes on the floors of the three swimming 
baths, and another five eyes mainly to the west-north-west of the indoor pool. Due to the 
lack of detail in the literature, these latter springs appear to have been largely disregarded 
in any discharge evaluations, which would mean that any flow rate given for the springs 
as an entirety, would be a minimum flow rate. 
 
For example, Rindl (1915) gave a flow rate of 31.25 l/s (112.5 m3/h), while Kent (1948) 
has provided the highest flow rate at Florisbad of 44.72 l/s (161 m3/h). However, neither 
of the aforementioned provided information regarding where the flow test was taken, how 
the flow rate was determined, or who carried out the determination. In 1961, Mr N. 
Viljoen of the Provincial Administration measured flow rates of between 1.01 and 3.28 
l/s, with an average of 2.27 l/s (Kent, 1971). On a visit to Florisbad in 1971, Kent (1971) 
estimated a flow rate of 2.53 l/s. Fourie (1970), who himself carried out flow 
determinations in 1953 on two of the pools (Kent, 1971), gave a combined flow rate of 
13.86 l/s (49.4) – 15.53 l/s (55.9 m3/h) with an average yield of 14.65 l/s. Grobler and 
Loock (1988a) gave an increase in the flow rate from 1.25 l/s (4.5 m3/h) to 5.22 l/s (18.8 
m3/h), after a light earthquake was felt by the residents at the spring site in 1912 (Anon 
[1], 1980) (see Section 1.1). The accuracy of this latter flow rate may be in question. 
 
Why there should be such a discrepancy between the values given by different 
researchers is not known. However, Kent (1971), strongly suggests that pumping and 
dewatering by the mines of the Free State goldfields, centred on Welkom, 90 km north-
north-east of Florisbad, and 60 km north of Basberg, the assumed recharge area for 
Florisbad, has affected the discharge rate of not only the Florisbad springs, but also of 
other springs and boreholes in the area. Table 11 gives the chronological order of flow 
 116
rate determinations at Florisbad and lends strong support for Kent’s (1971) hypothesis. 
This would strongly suggest an extensive system of inter-linked aquifers, of varying 
depths, underlying the region. Fourie (1970) also recorded a daily increase in flow rate 
and gas discharge between 09h00 and 11h00, and again at 15h00, with a corresponding 
slight increase in atmospheric pressure, although this was not considered significant. The 
flow rates for Florisbad (Table 11) can be compared to the Brandvlei spring, which is the 
strongest recorded spring in South Africa, at 128.1 l/s (461 m3/h), the Aliwal North spring 
at 28.61 \/s (103 m3/h), and the Warmbaths spring at 8.33 l/s (30 m3/h) (Kent, 1949). 
 
It has been mentioned (section 5.3) that Vegter (1995, 2001) gave a recharge value of 15-
25 mm per annum for the Florisbad area. Kok (1992) estimated that Karoo springs 
recharged, on average, 8% of the annual rainfall. This would suggest a 40 mm recharge 
 
Table 11. The chronological order of flow rate determinations at the Florisbad spring site 
showing the decrease in flow rate as a possible result of the dewatering of the 
Free State goldfields. 
 
 
Year Flow Rate Average flow Reference 
 l/s rate l/s  
 
 
1912             1.25 – 5.22            -  Grobler and Loock (1988a) 
1915                      31.25            -             Rindl (1915) 
1946    First deep mine shaft sunk            -              Kent (1971) 
1948                      44.72            -              Kent (1948) 
1953         13.86 – 15.53    14.65           Fourie (1970) 
1961             1.01 – 3.28    2.27              Kent (1971) 
1971                        2.53           -              Kent (1971) 
    
 
rate per annum for the Florisbad area, with an average rainfall of slightly less than 500 
mm per annum . Kok (1992) noted that areas of mean low rainfall could also experience 
 117
extreme events, which would increase the recharge rate Kok (1992) also noted that 
thermal springs, with temperature >25 °C may well be related to deeper regional 
groundwater systems that extend beyond the localized rainfall area. This extension might 
imply that groundwater systems may extend into areas with a much higher recharge rate 
due to a combination of factors such as higher rainfall, a more porous and/or fractured 
local geology, and the effects of geomorphology on the catchment area. In particular, this 
may be relevant to permanent springs. 
 
5.5 CHEMISTRY OF THE SPRING- AND GROUNDWATER 
 
As previously mentioned in Section 1.7, the 1988 water sampling was in light of the 
water bodies being suitable as habitats and breeding environments for amphibians 
(Douglas, 1992, 1995). Water samples were taken from the spring, vlei, an exploration pit 
that amphibians were inhabiting, the farm dam, and Soutpan (Figure 2). Water sampling 
was carried out under the supervision of the Institute of Groundwater Studies, University 
of the Free State. The Institute of Groundwater Studies analysed the water samples using 
Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) analysis. 
Samples from the vlei and the farm dam were excluded from this study as the water 
originated directly from the swimming pools, and thus the springs. The Soutpan water 
sample was also excluded, as it was the first time in living memory that the pan had filled 
with water, and the piles of evaporated salt would have had an artificial influence on the 
results.  
 
Of the water samples taken during the 1988 herpetofaunal study only the results of those 
from the spring eye and one of the exploration pits (Figure 23, E88, P188) were deemed 
relevant to this study. In 1999, the 1988 localities were re-sampled (E99, P199) and 
sampling expanded to the three other excavation pits (P299, P399, P499). Water sampling 
was carried out under the supervision of the Institute of Groundwater Studies, University 
of the Free State. The Institute of Groundwater Studies analysed the water samples using 
Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) analysis. All 
 
 118
 
 
 
 119
previous water sampling records (Rindl, 1915; Fourie, 1970; Mazor and Verhagen, 1983) 
were incorporated as part of the data. Rainfall for the year preceding the 1988 water 
sampling was exceptionally high at 957 mm, while rainfall for the year prior to the 1999 
water and peat sampling was slightly above average at 545 mm. This reflected a 42% 
drop in rainfall between the two periods. It was hypothesised that if the spring-water was 
the source of the water in the exploration pit, then the quality of the water at the two 
localities should be comparable. 
 
Rainfall for the year prior to the 1988 sampling period was nearly double the mean 
annual rainfall of 500 mm, at 957 mm, while rainfall for the year prior to the 1999 
sampling period was slightly higher than the mean annual rainfall at 545 mm. The 1998/9 
rainfall could therefore be seen as being much lower than the 1987/8 rainfall, but was in 
fact average rainfall relative to the mean annual rainfall. Mean annual rainfall over the 10 
year period prior to the 1988 sampling period was 459 mm, while the mean annual 
rainfall over the 10 year period prior to the 1999 sampling period was 514 mm, despite 
the high 1987/8 rainfall. Therefore, despite short-term punctuated high and low extremes 
in rainfall a mean annual rainfall of around 500 mm appears to have been maintained 
over the short to medium term. This resulted in a 42% decrease in rainfall between the 
1988 and 1999 sampling periods. When comparing the 1988 spring-and exploration pit 
water sample results to the 1999 results, it was found that the TDS of the exploration pit 
water was now 58% lower than that of the spring-water (Table 12). This was contrary to 
the hypothesis mentioned above that, if the spring-water was responsible for the water in 
the exploration pit, then the quality of the water at the two localities should be 
comparable. This was also contrary to evidence in the literature, which stated that if 
concentrations of dissolved substances are high, then the rate of groundwater renewal is 
low (concentration factor), whereas low concentrations indicate regular recharge (dilution 
factor) (Bredenkamp, 2000). 
 
Although there was only a slight increase between the 1988 spring-water results (2203 
mg/l TDS, high rainfall period) and 1999 results (2 363 mg/l TDS (average rainfall 
period) (Appendix I, Table 2), these results conformed with Bredenkamp’s (2000) 
 120
statement. However, results from Appendix I (Table 1, 2) also showed that, not only was 
the TDS of the pit waters generally higher than that of the spring-water, but that the TDS 
of the water in pit 1 was higher (2799 mg/l TDS) during a high rainfall period (1988) and 
lower (1174 mg/l TDS) during an average rainfall period (1999) (Table 12), this being 
contrary to Bredenkamp’s (2000) statement. 
 
Results between exploration pits 1 and 4 (Appendix 1, Figure 1) showed a west to east 
TDS increased ion concentration of 666%, while the TDS increase between the spring-
water and pit 4 was 287% (Appendix I, Table 1). Groundwater, per se, has not been 
mentioned in the literature, with the general assumption appearing to be that all 
 
Table 12. The difference in water quality (TDS, mg/l) between the spring-water and 
exploration pit 1 water during high and average rainfall periods (after 
Appendix I). 
 
 
Locality TDS mg/l TDS mg/l % increase/ 
 1988 wet period  1999 average period decrease 
 
 
Spring-water 2203 2364 +7 
    
Exploration pit 1 2799 1174 -58 
   
Spring-water/Exploration 
pit 1 2203/2799 - +27 
   
Spring-water/Exploration 
pit 1 - 2364/1174 -50 
  
 
subterranean water at Florisbad originated from the spring eyes. This strongly supported 
the argument for two separate hydrological entities. Table 13 provides some examples of  
the variations in ion concentrations, between the spring-water and exploration pit 4. It 
will be noted from Table 13 that higher ion concentrations occurred at all individual 
 121
levels of concentration, over a wide range of ions. Further to this, 73% of ions in 
 
Table 13. Some examples of variations in the increase in anion and cation concentrations 
between the spring-water (1999) and the water of exploration pit 4 (1999) 
(after Appendix 1). 
 
 
Anion/Cation Spring-water Exploration pit 4 Relative ion 
 mg/l water mg/l increase in Pit 4 
   as a % 
 
 
Na 784.00 2 920.00 272.45 
Bromide 19.31 71.64 273.00 
Cl 1 442.00 5 648.00 291.68 
B 2.59 10.16 292.28 
Ca 100.00 400.00 300.00 
Sr 2.97 11.20 301.43 
K 10.10 42.54 321.19 
Cu 0.004 0.017 325.00 
Nitrite 0.73 4.81 566.21 
Li 0.35 2.31 560.00 
Mg 0.44 9.04 1 954.56 
Sulphate 1.50 81.60 5 340.00 
 
exploration Pit 2 showed higher concentrations over the spring-water, while ions in 
exploration Pits 3 and 4 were 86% higher. This evidence strongly indicated that the 
exploration pit water could not be directly related to the spring-water, and should be seen 
as a separate entity, namely groundwater. The organic-clay layers, and in particular Peat 
II, has contributed to an elevated, or perched, water table within areas of the spring site. 
Greater detail of the chemistry of the spring- and groundwater is given in Appendices I 
and II. 
 
 122
The ground water is not thought to be of meteoric origin, as with the TDS of the 
groundwater already being that much higher than the spring-water, a higher undiluted 
groundwater TDS would only provide even stronger support for the groundwater 
hypothesis. The level of the water in the excavation pits did not drop from evaporation or 
seepage even during hot dry spell, which further supported the hypothesis of a perched 
water table. 
 
Figure 24 is a Piper diagram plot of all the Florisbad water samples in this study. It 
should be noted that some of the results were so similar in Figure 24, that a number of the 
points have been concealed behind other points. The Piper plot confirms the high Na, Cl 
content of both the spring- and groundwater, representing a strong sodium chloride 
facies. This does not agree with the deduction made by Vegter (1995, 2001) that there are 
two principal groundwater hydrochemical facies for the Florisbad area, namely, a 
calcium/magnesium chloride/sulphate facies, and a calcium/magnesium 
bicarbonate/carbonate facies. The Florisbad waters analysed reflect a very low Ca, Mg 
and S04 content. A Piper diagram presented by Grobler and Loock (1988a) supports these 
results, with Grobler and Loock (1988a) drawing the conclusion that the Florisbad spring 
water is relatively old. 
 
For purposes of comparing the TDS of the Florisbad water, Figure 25 represents Stiff 
diagrams of the water samples from Figure 24, based on the same scale. Samples E88 and 
E99 reflect the chemical stability of the spring eye water over the high (E88) and average 
(E99) rainfall periods respectively, while samples P188 and P199 show the TDS variation 
of the water in exploration pit 1 over the high (P188) and average (P199) rainfall periods. 
Samples P299, P399, and P499 show the increase in salinity in a west to east direction 
between the exploration pits, and in particular the increase in Ca in pit 4 (P499). Stiff 
diagram, R15, based on Rindl’s (1915) spring water analysis, reflects the stability of the 
spring-water over a longer period, while the rainwater sample of Litthauer (2007) 
provides further comparison. 
 
 
 123
 
 124
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  E88 E99 P188
Na+K Cl Na+K Cl Na+K Cl
Ca Alk Ca Alk Ca Alk
Mg SO4 Mg SO4 Mg SO4
180 meq/l 180 180 meq/l 180 180 meq/l 180
 
 P199 
 P299  P399
Na+K Cl Na+K Cl Na+K Cl
Ca Alk Ca Alk Ca Alk
Mg SO4 Mg SO4 Mg SO4
180 meq/l 180 180 meq/l 180 180 meq/l 180
   
 P499 R15 Spring eye water(Rindl 1915) Rain water Bloemfontein (Litthauer 2007)Na+K Cl Na+K Cl Na+K Cl
Ca Alk Ca Alk Ca Alk
Mg SO4 Mg SO4 Mg SO4
180 meq/l 180 180 meq/l 180 180 meq/l 180
Figure 25. Stiff diagrams for the Florisbad spring eye water (E88, E99), exploration pit 1 waters (P188, P199), and the 
waters of exploration pits P299, P399, P499. The spring eye water analysis of Rindl 1915 and Bloemfontein 
rainwater (Litthauer, 2007) are presented for comparison. For further comparison purposes all samples have 
been plotted on the maximum scale. For further detail on the sample codes see Figure 23. 
125
As the surface elevation of exploration pit mouths (P299 to P499) fell from west to east, 
so did the depth of the pits to groundwater level, with P299 being 6.70 m deep, P399 
being 4.20 m deep and P499 being 2.95 m deep, suggesting that the top of the water table 
was more or less at the same elevation between pits. The walls of P399 and P499 had 
both slumped to some extent, with P499 having slumped significantly more than P399, 
providing a larger groundwater surface for evaporation. The depth of P299 helped reduce 
the effects of evaporation, as the groundwater was shaded from the sun for most of the 
day and was also relatively sheltered from the effects of wind. P399, although being 
shallower, was heavily shaded by large eucalyptus trees, and was later partially covered 
with iron sheeting. P499, on the other hand, was totally exposed to the sun and wind. In 
examining these factors it is suggested that this increase in TDS from west to east is 
largely as a result of evaporation resulting in a concentration of salts as suggested in 
Appendix I. The lower TDS of samples P188 and P199 (450 mm deep), in relation to the 
other exploration pits, is thought to be due to it being located in the open excavation area 
where the elevation of the excavation was below the level and influence of Peat II, with 
the dilution effect of the nearby spring eyes having a significant effect. 
 
Loock and Grobler (1988) and Grobler and Loock (1988a) suggested that the chemistry 
of the spring-water was directly related to the organic content of the Tierberg Formation 
sediments of the Ecca Group. The Tierberg Formation sediments were deposited in a 
marine environment containing well defined zones of carbonate concretions and 
fossiliferous zones, which also contained carbonate concretions (Loock and Grobler, 
1988). Kent (1949) felt that such a high NaCl content was uncharacteristic of the Ecca 
Group rocks, and was more related to the underlying Dwyka Group rocks, through which 
he suggested the water had permeated. Loock and Grobler (1988) credited salinization of 
the spring water to the Ecca Group rocks, but mentioned that the water would have to 
have descended to the contact zone between the Karoo Supergroup sediments and Upper 
Ventersdorp Group basement rocks, upon which the Dwyka Group rocks rests. With the 
possible exception of Na, and Cl, and work reported on in Appendices II and III, it is 
concluded that, due to the low water temperature and alkalinity of the spring water, 
dissolution in the sub-surface geology had contributed little in the form of salinization to 
 126
the spring water. These conclusions are further supported by the low Ca content of the 
spring water, particularly in light of the carbonate concretions mentioned by Loock and 
Grobler (1988). 
 
The fluoride content of the Florisbad spring-water is high at 5.5 to 6.0 mg/l, this being in 
relation to the recommended limit for drinking water quality of 1.0 mg/l, and the 
maximum allowable limit for drinking water quality of 1.5 mg/l. This fluoride level 
compares to that of the artesian borehole at Glen Agricultural College of 5.3 mg/l (see 
section 4.4). Flourine is the most abundant halogen in sedimentary rocks and must be 
considered as a major source of fluoride in groundwater (Woodford and Chevallier, 
2002). Woodford and Chevallier (2002) note that fluoride concentrations cannot only be 
controlled by near surface lithology, because this does not explain the marked fluoride 
variability found, on a local scale, in shallow aquifers in the upper 150 m bgl, where the 
host-rocks are similar. 
 
Analysis of borehole water indicates that the chemical composition of groundwater varies 
considerably within Ecca Group rocks. Generally, the salinization of borehole water 
within Ecca Group rocks are considerably lower than that of the Florisbad spring water. 
Chemical values taken from Baran (2003 for Ecca Group rocks (Florisbad spring-water 
value are given in brackets) show that TDS ranged from 130.0 to 778 mg/l (2363 mg/l), 
NaCl 11 to 208.9 mg/l (2226 mg/l), and Ca 5.9 to 15.5 mg/l (100 mg/l). Woodford and 
Chevallier (2002) give the following groundwater ranges for Karoo sediments: TDS 450 
–1000 mg/l; pH 8.0-8.5, dropping to 7.5 in the east and north; Ca 30- 80 mg/l, dropping 
to 10-30 mg/l in the east; Na <100 mg/l, Cl >1000 mg/l; K 10-20 mg/l; total alkalinity 
200-300 mg/l; Si 15-25 mg/. 
 
While there is a faint hydrogen sulphide (H2S) smell emanating from the springs, the 
quantity of dissolved H2S is very low at 0.003 mg/l (Fourie, 1970). The main constituent 
of the free gas component is methane (CH4) at 70%, with the balance being comprised of 
nitrogen (N2) (27.4%) and carbon dioxide (CO2) (3.6%) (Rindl, 1916; Fourie, 1970). 
 
 127
With total dissolved ions of 2363 mg/l, Mazor and Verhagen (1983) noted that the 
Florisbad spring was the most saline of all South African springs. While the Florisbad 
water may be the most saline of South African spring-waters, even at current day levels, 
it is still relatively sweet and potable for most modern animals, and was most probably 
just as potable for the palaeofauna, (Douglas, 1992). Smit (1977) noted that Kalahari 
animals could tolerate water with a TDS of 6 000 mg/l. Kruger and Lubczenko (1994) 
gave the salinity tolerances for pigs and milking cows as <3 200 mg/l, dry dairy cows and 
horses as <4 500 mg/l, beef cattle as <5 760 mg/l, and sheep as 6 400 mg/l. Kempster et 
al. (1980) set an upper limit for livestock water quality at TDS 14 000 mg/l. This would 
also imply that, even if spring-water salinity levels had risen in the past, for whatever 
reasons, the substantial presence of the fossil fauna indicates that the spring-water was 
potable to the vast herds of game inhabiting the area over time. Based on total dissolved 
solids of 2201 mg/l, Florisbad has a salinity 82 % higher than that of the Aliwal North 
spring, and 886 % higher than the Warmbaths spring (Douglas, 1992). Based on saline 
content, Florisbad has a salinity 112% higher than the Aliwal North spring and 2617% 
higher than the Warmbaths spring (Douglas, 1992). 
 
Not only are the Florisbad TDS levels within the potable drinking water range for 
mammals, but also well within the tolerance levels for habitation by a number of 
amphibian species, which are used as modern water quality bio-indicators. With TDS 
levels of up to 2798 mg/l, the outdoor swimming baths, vlei area and exploration pits 
were inhabited at times by permanent water breeding species, such as Xenopus laevis 
laevis, Amietia fuscigula, and Amietia angolensis, although these species were never 
observed breeding (Douglas, 1992). Munsey (1972) gave an NaCl tolerance level for X. l. 
laevis of 7582 mg/l, while the pH of 9.43 for the spring-water (Douglas, 1992) was 
higher than the maximum natural pH of 8.86 given by Picker (1985) for this species. 
 
Temporary pond breeders, on the other hand, where breeding sites are used only in 
conjunction with seasonal rainfall, and where the mineral content of the water is much 
lower, appeared to avoid waters associated with the spring and groundwater i.e. 
swimming pools, the vlei, and exploration pits. These species included Pyxicephalus 
 128
adspersus, Cacosternum boettgeri, Kassina senegalensis and Tomopterna cryptotis, who 
were all observed breeding at the only fresh water site on the farm, namely, the area 
above the dam site (Figure 2) (Douglas, 1992). The use of amphibians as bio-indicators 
of water quality was particularly evident at this dam where the actual dam was avoided as 
a breeding and habitation site due to water being pumped from the swimming baths into 
the dam. The saline Florisbad spring- and groundwater therefore appear to be totally 
unsuitable for the permanent habitation of temporary pond breeders. The terrapin, 
Pelomedusa subrufa, was also often observed in the outdoor swimming pools (Douglas, 
1992).  
 
5.6 THE DEPOSITIONAL ENVIRONMENT AND SEDIMENTATION 
 
5.6.1 Theories on the Deposition of the Florisbad Sediments 
 
Numerous researchers have put forward theories on the formation of the spring site. 
Gardner (1932) and Deacon (1970) considered spring mounds to be derived from 
underlying bedrock through elutriation. Brink (1987) saw the spring site developing 
through sand being brought to the surface by spring vents, which became enlarged due to 
mechanical and chemical actions, with vegetation developing around the margins of the 
spring pools. The vents then became blocked, with the size of the mound increasing due 
to a combination of factors such as choking vegetation, the deposition of windblown 
sediments, and a diminishing supply of groundwater (spring-water) (Brink, 1987). Brink 
(1987) further hypothesised that after this closure, the spring eye moved laterally along a 
bedrock fissure to form a new passage, which cut through the existing strata to form 
another sand unit, partly overlapping the mound created by the original vent. It was 
contended that these processes would have repeated themselves to form a build-up of 
sediments, which includes alternative layers of organic and non-organic deposits (Brink, 
1987). 
 
There appears to be little contention in the literature that “peat” (Figure 4 and 23) has 
been used as a convenient term for describing the four organic rich horizons at Florisbad. 
 129
These horizons were described by Van Zinderen Bakker (1989) as so-called ‘peat’ layers 
of carbonaceous clay and silt, and by Butzer (1988) as organic horizons where the ‘peat’ 
lacked definition in terms of inter-woven vegetation structure. The quantities of sand 
deposited within these layers would also tend to make the term “organic rich” more 
appropriate. The lack of inter-woven vegetation structure, as noted by Butzer (1988), was 
particularly noticeable in Peat II, but vegetation structure was very high in Peat IV. The 
author supports the point of view that the “peat” layers should rather be seen as layers of 
organic rich material, clay, silt, and sand, and these layers are therefore referred to here as 
organic-clay layers, based on their principal components. 
 
Butzer (1988) described the site as a 7 m mound of spring beds inter-bedded with organic 
layers, and only partly covered by aeolian material. The site was seen as having 
developed through spring flow, which was determined by a deep-seated regional aquifer 
and fluctuations in recharge, with the bulk of the quartz grains comprised of detrial sand 
released from the underlying Ecca, through which the spring-waters had passed (Butzer, 
1988). Sandy pools then developed at the spring site, with peaty organic horizons 
developing as vegetation encroached on less active springs during periods of low 
discharge (Butzer, 1988). Vegetation was later submerged by periods of more active 
spring discharge, and subsequently buried by spring sediments (Butzer, 1988). Dreyer 
(1938a) also suggested that the Florisbad deposits represented sand output from the 
spring and considered the Florisbad “mound” as being formed by the sand from a huge 
(unknown) eye beneath the highest point of the “mound”. This would suggest that the 
spring pan extended well to the north of the current eyes. 
 
It has been contended that a close relationship and correlation existed between the spring 
sedimentation and the shoreline position of the adjoining palaeolake complex (Soutpan) 
and that this relationship played a significant role in the modification of the spring mound 
sediments (Joubert and Visser, 1991; Visser and Joubert, 1991; Kuman et al. 1999; 
Henderson, 2001a). It was proposed that deposition at Florisbad was directly related to 
the palaeolake levels (Soutpan), which reflected climatic conditions at the time (Joubert 
and Visser, 1991; Visser and Joubert, 1991). This involved cyclic sedimentation with soil 
 130
horizons forming during arid stages when palaeolake levels were low, while deposition of 
palaeolake bottom silts occurred during wet periods when the spring area was flooded by 
the palaeolake (Joubert and Visser, 1991; Visser and Joubert, 1991). These cyclic 
transgression and regression sequences of the palaeolake shore line were translated into 
four low water-level phases and three high phases (Joubert and Visser, 1991). It was 
noted that waterlogged conditions had existed at levels higher than the existing water 
table, and that load structures had been identified as high as the top of Peat II (Brink, 
1987). These load structures, as well as faunal evidence, led to the suggestion of the 
existence of a large water body in the past, but whether or not this was related to the 
water levels in the palaeopan still needed to be established (Brink, 1987). 
 
Sedimentary deposits to the east of the spring site have been referred to as the “lacustrine 
sequence” (Joubert and Visser, 1991; Visser and Joubert, 1991) because of their 
presumed association with the Soutpan complex. These sediments were interpreted as 
being palaeolake bottom deposits, directly related to Soutpan, being based largely on the 
presence of freshwater gastropods in the clay facies,  (Joubert and Visser, 1991). It was 
also suggested that during humid phases the Soutpan palaeolake complex enlarged to 
more than twice its present size (Visser and Joubert, 1991). Despite their “flooding of the 
palaeopan” hypothesis, Visser and Joubert (1991) noted that the distribution and lateral 
variation of the organic-rich deposits at the spring site reflected waterlogged, or bog 
conditions, on a poorly drained flood plain, reflecting the local influence of constant 
freshwater discharge at the spring. Fourie (1970), who originally described these deposits 
and the freshwater gastropods, did not see them as being related to the palaeolake, but 
rather as an integral part of the spring site itself. As mentioned in section 3.2, Scott and 
Rusouw (2005) noted that the high water levels proposed by Visser and Joubert (1991) 
were in conflict with the pollen evidence where Chenopodiaceae pollen was an indication 
of dry conditions. 
 
Butzer (1984) postulated that sedimentation had occurred in a flood plain environment 
and suggested that Peat I had a similar semi-aquatic accumulation as Peat II. Van 
Zinderen Bakker (1989) agreed with this, stating that the intervals in the organic-clay 
 131
layers were due to rises in the level of water in the pan (Soutpan) and water table, and 
that during periods of stable spring flow, vegetation around the spring produced organic 
deposits. 
 
It was noted by Grobler et al. (1988) that prior to the formation of the pans, the area was 
already in a state of low topographic relief. In effect, there was most probably not much 
difference between the elevation of the Soutpan floor, which came under the influence of 
deflation, and the spring site, which came under the influence of aeolian deposition. As 
the processes of deflation and deposition continued, the surrounding area deflated, 
lowering the elevation, while aeolian deposition increased the elevation to the south and 
south-east of Soutpan. The top of the Florisbad dune is presently 25 m (approximately) 
above the Soutpan floor (Appendix IV) while there is only a 19 m decrease in elevation 
across the panveld from Florisbad westward, over a distance of some 120 km (Douglas, 
1992). Should the flooding of the palaeopan hypothesis be a consideration (Joubert and 
Visser, 1991; Visser and Joubert, 1991), including the suggestion that the palaeolake may 
have enlarged to twice its size (Visser and Joubert, 1991), this would have implied that 
the water in Soutpan would have had to have risen to at least the level of the top of Peat 
IV (±21 m) to ensure its formation. 
 
Any such significant increase in the relatively fresh water level in Soutpan would 
therefore have resulted in a sheet of water spreading far beyond the Florisbad spring site, 
covering a large part of the western Free State. There appears to be no evidence of this, 
and with so many pans in the area, any such flooding would have filled these pans and 
other low lying areas with relatively fresh water as well. Judging by the large number of 
fossil remains recovered from such a small area at Florisbad (Brink, 1987), and with the 
availability of so much standing fresh water over such a large area, it may be asked why 
game would have concentrated to the extent they did, at the Florisbad spring site. There 
also appears to be no evidence of sedimentary deposits similar to those at Florisbad 
occurring at any other localities in the area, particularly the presence of the organic-clay 
layers. The pollen content of the argillaceous green sands between Peats I and II, which 
corresponded to one of the proposed high lake levels by Visser and Joubert (1991), was 
 132
seen as a horizon dominated by Chenopodiaceae-type pollen, indicating a period of dry 
conditions (Van Zinderen Bakker, 1989; Scott and Brink, 1992). 
 
Kuman et al. (1999) suggested that a number of micro-environments existed in response 
to the waxing and waning of the spring, as well as the expansion and contraction of the 
nearby palaeolake. Based on grain sample size and distribution, it was concluded that 
most of the sediments not only accumulated under uniform, low-energy, subaqueous 
environmental conditions, but also under several composite geomorphic regimes, and had 
not been reworked after deposition (Kuman et al., 1999). It was further noted that the 
sand layers were usually more than 10 cm thick, with an absence of thin layers, which 
would denote annual, or shorter, cyclical deposition systems (Kuman et al., 1999). The 
sharp bedding contacts of the Pleistocene levels were seen as indicating sudden changes 
in environmental conditions due to the possible increase and decrease in the spring-water 
discharge, with the stratigraphy of the layers reflecting changes in the environmental 
water regime, from open dam conditions to vegetated marshland (Kuman et al., 1999). 
 
5.6.2 Current State of the Florisbad Sediments 
 
The sediments and stratigraphy of the Florisbad spring site are important in that they hold 
the key to the history of the site. Lawson and Thomas (2002) noted that geomorphic 
features can be seen as an environmental archives. This archival record holds the key to 
aspects such as the formation of the site, palaeoclimate, and the preservation and 
fossilization of both the Old Collection and the MSA human occupation horizon. 
 
Rubidge and Brink (1985) described investigations into the lithostratigraphy and 
depositional history of the Florisbad sediments as still being in their initial stages, and 
suggested that several models could be proposed. Kuman and Clarke (1986) noted that 
the spring cycles, which are recorded in a repeated alternation of sands, silts and organic 
deposits, are poorly understood and controversial. The stratigraphy and sedimentology 
for a relatively small part of the Florisbad spring site has been documented, and it is 
hoped that the hypotheses presented in this thesis will go some way to resolving some of 
 133
the issues. The stratigraphy of the Florisbad site is complex (Rubidge and Brink, 1985), 
which has been compounded by a yet unresolved depositional history, internal 
stratification of the sand dune, and erosional forces on the leeward and eastern side of the 
sand dune. These aspects are addressed in Section 5.6.5 and Appendix IV. The 
stratigraphy on the face of the excavations may vary considerably from one section to the 
next, over relatively short distances, while being unbroken and horizontal on other faces. 
The generalized stratigraphy is reflected in Figure 26. 
 
The sedimentary deposits underlying Florisbad are made up of alternating layers of sand, 
and what are generally referred to as peat layers (Figure 26) (Dreyer, 1938a; Kuman and 
Clarke, 1986; Brink, 1987; Henderson, 2001a). Four of these organic clay layers have 
been identified at Florisbad, but not all are in evidence on the face of current excavations, 
with only Peat II and Peat IV being clearly discernable during the study. Various authors 
have described the state of the present day organic-clay layers. Fourie (1970) mentioned a 
thin basal peat layer (Peat I) resting directly on the dolerite. Peat I, which is located 
below the water table was described by Dreyer (1938a) as being bituminous, and by 
Meiring (1956) as being waxy and free of modern tree roots. Fourie (1970) described 
Peat I as being waxy with scattered pieces of decomposed wood, where the fibrous 
structure could still be identified. 
 
Peat II is currently at the same level as the water table in pits 2, 3 and 4. Meiring (1956) 
described Peat II and III as being heavily contaminated with modern tree roots. The Peat 
II sampled in this study did not however reflect any heavy contamination of modern tree 
roots, vegetation or fibrous structure. Butzer (1984, 1988) saw Peat II as an organic, very 
dark grey sandy loam layer, inter-bedded with dark grey loam and pockets of light grey 
loam. Organic material comprised what Butzer (1984, 1988) referred to as vegetation 
structures, carbonised with fine vertical roots, and also noted the lack of inter-woven 
vegetation structure. The Peat II examined in this study was of a grey sandy/clay nature 
with a very fine black clay fraction which had the consistency of light grease. The effect 
 
 134
 
 135
of compression due to the vertical weight of the sand is evident in Peat II in that the fine 
black clay fraction is being squeezed out from between the sand, and onto the walls of the 
pits. However, this oozing of the fine clay fraction may also partly be due to swelling 
pressure caused through the absorption of water by the clay. Load structures, mentioned 
by Brink (1987), are yet another aspect and indication of this compaction. 
 
Meiring (1956) noted that in a part of the 1952 excavation (Figure 27), none of the layers 
were disturbed, but were continuous throughout the exposed face. This is further 
illustrated in Figure 28 which shows the upper section of excavation pit 2 (Figure 23). It 
was noted that because of the apparently relatively low spring-water temperature (29° C) 
over time, diagenesis of the underlying Ecca and Dwyka formation rocks had contributed 
little to the salinization of the site (Appendix I; Appendix II). Based on sand grain shape 
and surface features, Kuman et al. (1999) and Van Zinderen Bakker (1989) concurred 
that the spring sediments appear to be derived predominantly from an aeolian source, 
while sands from the lower levels, although showing signs of water transportation, were 
also originally of an aeolian nature. Grobler and Loock (1988a, 1988b) also stated that 
deposition was largely as a result of aeolian processes, and this was supported by Joubert 
and Visser (1991). 
 
Dreyer (1938a) and Fourie (1970) both noted that the organic-clay layers became 
progressively more sandy towards the surface. Peat IV was described by Butzer (1984, 
1988) as organic, grey to dark grey loam to loam clay, comprising abundant organic 
voids and abundant root casts, as well as black organic muck of semi-aquatic facies. 
Joubert and Visser (1991) described Peat IV as clayey silt, which contained the remains 
of reeds, disseminated carbonate and calcareous nodules. Peat IV, on the western wall of 
the excavation, is located just 650 mm below the soil surface and is 2.8 m thick. This can 
be compared to the relatively narrow thickness of the Peat IV layer reflected in the 
section in Figure 26. 
 
 136
 
 
 137
It should be born in mind that the degree of deposition and sedimentation was highly 
variable, being punctuated by high and low varying peaks. Variability would have been 
due to factors such as rainfall and temperature, which would have in turn related to wet 
and dry periods, wind, aeolian deposition, and the development of the lower organic 
layers, their decomposition and subsequent compression and compaction. Compression 
and compaction would also have been variable, dependant on the aforementioned factors, 
but they would also have been a varying constant in that they were always a present 
force, dependant on the accumulation of sediments and water. The lower deposits would 
always have been under the increasing effect of compression and compaction as the 
accumulation of the sediments and water increased. It would, at this time, be difficult to 
determine whether narrower layers of organic matter actually formed as narrow layers, or 
whether they were much wider layers that had been compacted over time. However, 
judging by the loose structure of Peat IV, I would appear that the lower organic layers 
were originally similar in their formation, and became compacted at a later stage.  
 
As with the stratification of the sand dune, there are many variables influencing 
descriptions of the organic-clay layers. These include the composition and quantities of 
organic matter, clay, silt, and sand; the thickness of the overburden; the thickness of the 
section being examined; the location of the section being examined, the water content of 
the layer; as well as many other unknown historical factors. These factors can in turn be 
applied to variations in the salinization of the organic-clay layers as well as that of the 
groundwater. It is therefore apparent that from the above mentioned variables affecting 
the formation of the organic-clay layers, that the present state and composition of the 
organic-clay layers may vary considerably for samples from the same layer, taken at 
different localities. In relation to European peat, the Florisbad organic-clay layers in their 
current state, and in particularly the lower organic-clay layers, can hardly be classified as 
peat. 
 
 
 
 
 138
 
 
 139
5.6.3 Salinization of the Organic-Clay Layers 
 
Once water quality criteria for the spring- and groundwater were established (Appendix 
I), a number of questions arose. Firstly, why was the salinization of the groundwater 
generally so much higher than the spring-water? Secondly, how did the groundwater 
manage to maintain such a high salinity? Thirdly, what were the origins of the higher 
concentrations? Fourthly, why were there variations in TDS between high and average 
rainfall periods, and lastly, why was there a variation, contrary to norm, of lower ion 
concentrations with high rainfall, and higher ion concentrations with low rainfall? 
 
For this part of the study, the Peat II organic/clay layer was sampled from the walls of the 
exploration pits (P2, X1; P3, X2; P4, X3) by first removing ±80 mm of the outer layer 
before taking the sample. In a similar manner, the north wall of the MSA excavation was 
sampled for Peat II (X1), while Peat IV was sampled from the west wall of the open 
excavation (X4; X5). In November 199 an additional Peat II sample was taken from Pit 2 
(X6) for water extraction analysis by All Peat samples were analysed by the Department 
of Soil Science, University of the Free State, using Atomic Absorption and Spectrometric 
analysis. Sample X6 was analysed as a water extraction sample by the Institute of 
Groundwater Studies, University of the Free State, using ICP-OES analysis (Douglas, 
2009). X-ray difractometric analysis was carried out by the Department of Geology, 
University of the Free State, on Peat II samples from Pit 2 as well as sand expelled from 
the spring eye. Sedimentation was used to separate and determine the clay fractions of 
Peat II. Rainfall was also examined over a 78 year period in order to determine whether 
rainfall had any effect on the long- and short-term quality of the spring-and groundwater. 
 
Results of the organic-clay layers analyses are given in Appendix II, where Tables 1 to 4 
show that the organic-clay layers were considerably more highly salinized than either the 
spring-, or the groundwater. Peat II results showed a higher ion concentration of 945% 
over the 1999 spring-water results, with Peat IV showing a maximum higher ion 
concentration of 902%. It was therefore determined that salinization was at its highest 
levels within the organic-clay layers (Table 14). 
 140
 
Table 14.  The difference in ion concentrations (mg/l) between the groundwater and the  
Peat II organic-clay layer in exploration pits 2 and 4, during 1999, with a 
comparison to water from the spring eye (after Appendix I; Appendix II). 
 
 
Ions Spring Ground- Peat II Ground- Peat II 
(mg/l) eye water Pit 2 water Pit 4 
  Pit 2  Pit4  
 
 
Cl 1 442.00 1 104.00 7 727.40 5 648.00 12 513.00 
      
Na 784.00 554.00 4 240.00 2 920.00 6 800.00 
      
Ca 100.00 136.00 2 768.00 400.00 5 000.00 
      
K 10.10 12.90 640.00 42.54 352.00 
      
P 1.93 0.01 26.00 0.01 34.00 
      
Mg 0.44 5.15 24.80 9.04 70.00 
      
Zn 0.02 0.03 2.40 0.03 6.00 
      
Total 2 338.49 1 812.09 15 428.60 9 019.62 24 772 
ions 
  
 
Solutions to the above questions were initially investigated by referring to references 
relating to previous stratigraphic and sedimentary research at Florisbad (Dreyer, 1938a; 
Meiring, 1956; Fourie, 1970; Rubidge and Brink, 1985; Brink, 1987; Butzer, 1988; Van 
 Zinderen Bakker. 1989; Joubert and Visser, 1991; Visser and Joubert, 1991; Kuman et 
al., 1999). This literature was reviewed for processes and scenarios which could possibly 
explain and support the results of a higher salinity for both the groundwater and organic-
clay layers. The processes provided in the above references dealt fundamentally with the 
formation of the site, and were gradually disregarded, as none of these theories were 
considered capable of explaining the fact that the organic-clay layers and groundwater 
were so highly salinized. This was due to the fact that no previous chemical analysis was 
 141
carried out by previous researchers on the organic-clay layers. Where previous spring-
water analysis was available (Rindl, 1915, 1916, 1928; Fourie, 1970; Mazor and 
Verhagen, 1983, Appendix I), these were only been briefly mentioned by researchers 
such as Brink (1987) and Grobler and Loock (1988a), or ignored. Further to this, 
groundwater has rarely been mentioned in the literature (Brink, 1987), and it appears that 
the assumption has been made that all subterranean water at Florisbad originated from the 
springs. 
 
A further literature review was focussed on specific chemical and physical processes 
which could possibly explain and support the results of a higher salinization in the 
groundwater and organic-clay layers (Hillel, 1971; Taylor and Ashcroft, 1972; Tóth, 
1972; Blatt et al., 1972; Dykyjová, 1978; Tildon and Kadlec, 1979; Lakshman, 1979; 
Larcher, 1983;Wetzel, 1983; Brady, 1984; Bohn et al., 1985; Hatano et al., 1994; St-Cyr 
et al., 1997). From the above references, it was concluded that the process of salinization 
was the only process that could account for the high ion concentrations of the 
groundwater and organic-clay layers.  
 
Processes and conditions which may result in the accumulation of salts in a semi-arid 
environment and therefore influence salinization include low rainfall and semi-arid 
conditions; strong winds with associated aeolian sand deposition; adsorption of ions by 
clay and organic material; evaporation, capillarity and matrix suction associated with 
high temperatures; and the accumulation of minerals through drainage impediment. The 
organic-clay layers at Florisbad also specifically accommodated two major components, 
namely, the organic matter and the clay. Owing to the many factors involved in the 
salinization process, and the complexity of the mechanisms involved, the process was 
examined in some considerable detail. This aspect of the study was also of prime 
importance in that the results would either support or disprove future hypotheses on the 
fossilization of faunal remains and formation of the site. 
 
Organic matter, or humus, has a structure, or colloidal organization, of submicroscopic 
particles, referred to as colloids, which are somewhat larger than ordinary molecules and 
 142
ions (Taylor and Ashcroft, 1972). The structure is similar to that of clay, with highly 
charged anions being surrounded by a large number of adsorbed cations (Brady, 1984). 
However, the negative charge on humus colloids are very much pH dependant with their 
absorptive capacity decreasing with a decrease in pH (Brady, 1984). On the other hand, 
under more alkaline conditions, the adsorptive properties of humus far exceed that of 
clays. With the Florisbad organic-clay layers having a pH range of 8.8 to 4.2 (Appendix 
II), it is apparent that the degree of salinization of the organic material will vary over the 
site, and will further vary with the quantities of organic material and clay present. Even 
before humus is formed through the decomposition of vegetable matter, hydrophytes 
(aquatic plants living in, or on, water), and in this instance particularly halophytes (plants 
capable of living in salt impregnated soils) and helophytes (plants growing in water 
saturated soil), have the ability to remove large quantities of salts from both the soil and 
water, as well as the ability to accumulate them at far greater concentrations than those in 
external solution (Larcher, 1983). 
 
It is therefore apparent that the accumulated minerals absorbed and stored by halophytes 
and helophytes would contribute significantly to the salinization of the organic-clay 
layers when decomposition took place. These processes are not only applicable to 
halophytes and helophytes, as even micro-organisms such as phytoplankton and 
periphyton have the ability to absorb and accumulate salts (St-Cyr, 1997; Wetzel, 1983). 
This can be illustrated by the use of hydrophytes and halophytes in eutrophication control 
and bioaccumulation: examples are: the treatment of eutrophication caused by fertilizers 
and phosphorus-based insecticides in natural lake waters (Tóth, 1972); effluent 
wastewater treatment (Lakshman, 1979, Tildon and Kadlec, 1979); swine wastewater 
treatment (Hunt et al., 1993); dairy wastewater treatment (Davis et al., 1992); pulp 
wastewater treatment (Hatano et al., 1994); and mine wastewater treatment (Noller et al., 
1994). 
 
Clay particles are often negatively charged and capable of attracting positively charged 
ions to each colloid, therefore, having the ability to adsorb on their outer and inner 
surfaces large quantities of ions supplied by percolating water (Blatt et al., 1972; Brady, 
 143
1984). In many instances, salinization in clay beds is capable of concentrating minerals to 
such an extent, for example, through percolating water, that economic deposits of clay 
minerals such as gypsum, magnesite, halite, and mirabilite may form (Hillel, 1971; Blatt 
et al., 1972; Brady 1984). On the other hand, with the relative weakness of the structural 
unit bonds in clay, this may lead to the inter-layer ions being easily removed by 
percolating waters during diagenesis and weathering (Blatt et al., 1972). This further 
supports the hypothesis that recharge water flushes elements from the organic-clay layers 
into the groundwater, rather than rainfall having a diluting effect (Appendix II). 
Montmorillonite clay found in the soil at Florisbad (Appendix II), characteristically 
absorbs water between successive layers, causing it to expand (Taylor and Ashcroft, 
1972), and thus has a large specific surface area due to its lattice expansion and exposure 
of internal surfaces (Hillel, 1971). The potential and ability for both the organic matter 
and clay to attract, adsorb, concentrate and store ions in considerable excess to the 
groundwater, further supporting the choice of the salinization process. 
 
Now that salinization processes had been resolved, the question arose as to the origin of 
the elements. Examination of the literature revealed that there were a number of 
possibilities for the origin of ions in the spring-, groundwater, clay and organic matter. 
Kent (1949) suggested that the high Na, Cl content of the spring water was due to the 
water having percolated through the underlying Dwyka Group rocks, as opposed to the 
Ecca Group rocks. Butzer (1988) on the other hand suggested that, due to the steady 
release of methane gas combined with the Na, Cl content of the water, the chemical had 
been derived from saline and carbonaceous facies of the Ecca Group rocks. Dolerite, 
which contains plagioclase, feldspar, and ferromagnesian silicate minerals may have 
contributed ions to the spring-water in the form of Fe, Si, Mg and Ca, but due to the low 
levels of these minerals in the spring water, this was not thought to be significant 
(Appendix I). 
 
The influence of aeolian deposition from Soutpan on the salinization at Florisbad during 
average raingall periods can be gauged from water quality results obtained during the 
extremely high rainfall in February/March 1988. Soutpan was completely full for the first 
 144
time in living memory, dissolving the salt deposits on the pan floor to produce very 
diluted standing pan water with an Na, Cl content of 9 135 mg/l (Douglas, 1992). This 
equated to the Na, Cl content of the pan water being 30.96% that of sea water, where the 
Na, Cl content of sea water was taken at 29 500 mg/l (Douglas, 1992). Seaman et al. 
(1991) recorded an Na, Cl level of 188 000 mg/l, with a TDS of 197 295 mg/l for 
Soutpan water, which may suggest that this sample was taken close to, or at, an 
evaporative salt dam. In this respect the result was converse, with the Na, Cl content of 
sea water being only 15.69% that of the pan water. This illustrates the large potentially 
large reserve of salts that are available from the pan floor sands for aeolian deposition by 
the north-west prevailing winds at Florisbad. 
 
Salinization will also occur at the contact zone between the spring-water and 
groundwater, where there will be an exchange of ions through the process of miscible-
displacement, through the intrinsic movement of molecules from areas of high to low 
concentration (Taylor and Ashcroft, 1972). Hydrodynamic dispersion resulting from the 
continually changing direction of the spring- and ground water through collision with 
pore walls will also either increase or decrease salinization (Taylor and Ashcroft, 1972). 
It is thought that the largest contribution of ions came from factors related to the 
environment of the area, particularly during semi-arid phases. In semi-arid environments, 
metallic cations do not to leach from the soil, and tend to dominate when pH values 
exceed 7.0. (Brady, 1984).  
 
Figure 29 summarizes the processes and stages involved in the movement of ions and the 
salinization and fossilization of the Old Collection at Florisbad. The question of 
groundwater ion concentrations being higher during wet periods and lower during 
average rainfall periods, is explained by the organic-clay layers adsorbing ions during 
periods of low recharge, and the ions being flushed from these layers by the groundwater 
during periods of high recharge (Appendix II). 
 
 145
 
 146
In support of the above statements, Table 13 and Table 14 clearly shows the higher 
concentrations in TDS between the spring- groundwater, and organic–peat layers. 
Notable is the unexplained west to east increase in the TDS of both the groundwater and 
Peat II organic–clay layer. As there is a predominance of spring eyes in the area of the 
swimming pools, it is suspected that the spring water could be having a diluting effect on 
the TDS of groundwater in the area of the pits. This influence would appear to decrease 
from west to east, which in turn would affect the ion concentrations of organic-clay layer, 
Peat II. 
 
5.6.4 A New Perspective on the Fossilization of Faunal Remains 
 
It has now been established that the salinization of the organic-clay layers and 
groundwater was considerably higher than that of the spring-water, and that processes 
were in place to produce this higher salinization. The question then arose as to how these 
results could be applied to the physical environment at Florisbad and the fossilization of 
faunal remains. Because the large majority of fossils in the Old Collection have been 
recovered from what has been referred to as areas of spring activity (Brink, 1987), Brink 
(1987) stated, and it has generally been accepted, that the saline spring-water was 
responsible for fossilization. Brink (1987) also stated that, since the spring water was 
carbonate rich, the evidence of calcium carbonate deposition in the bones further 
suggested that they must have been in contact with the spring water for some time. It was 
further noted by Brink (1987) that, because fossilized remains of the Old Collection were 
found in areas of spring activity, post-depositional mechanical and chemical weathering 
clearly showed that the remains had become fossilized in a spring context. From a 
chemical perspective, Brink (1987) reported that the spring water has caused the remains 
of the Old Collection to be well preserved in a characteristic way. 
 
Despite Brink’s (1987) strong defence of fossilization in a spring context, he also 
admitted to some doubts in that, if the presence of calcite was an indication of contact 
with the spring-water in the original sedimentary environment, then the relatively low 
incidence of calcite concretions might contradict the assumption that the Old Collection 
 147
was entirely derived from spring deposits. It was noted that only 48.2% of Bovidae and 
Hippopotamus bones showed signs of calcium carbonate deposition (Brink, 1987).  
 
Further to this it was also suggested by Brink (1987) that acid groundwater mobilized 
carbonate cement from underlying geological strata, which may have become re-
deposited, despite being of an alkaline nature. Due to already existing water analyses 
indicating the low Ca level of the spring-water, it is somewhat surprising that the spring-
water had been credited with the fossilization of the faunal remains over the years. 
 
The first spring-water analyses by Rindl (1915) clearly showed that the calcium level of 
the spring-water was very low at 93.42 mg/l. Subsequent water analysis by Fourie (1970), 
Mazor and Verhagen (1983) and Douglas (1992), all confirmed this level of calcium at 
<100 mg/l. Slight variations could be expected through technological advances in water 
analysis. Grobler and Loock (1988a) also mentioned the high sodium content associated 
with the low calcium content; if these authors are correct in their assumption that the low 
calcium content is due to cation exchange as the water moves through the underlying 
Ecca shales, then there is a strong probability that the spring-water has historically 
maintained a low calcium content. This would then tend to lend further credibility to the 
hypothesis put forward in this thesis that the spring-water could never have been 
responsible for fossilization. Dreyer and Lyle (1931) appear to have summed up the 
situation by stating that animal remains found at the base of such pipes (spring vents) 
seem to have been accumulated and concentrated by the sorting action of the spring 
water, with coarser materials finding the lower levels, and finer material filling the upper 
ends of the vents, but with no mention of any possible fossilization. Detailed chemistry of 
the spring-water, groundwater, and organic-clay layers is given in Appendices I and II, 
and summarized in Tables 12, 13 and 14. 
 
Fourie (1970) found a lack of carbonate in the sands of the eyes, as well in the lower 
consolidated green sediment, although some evidence of calcite cementation was evident 
in the basal zone of Peat I. Butzer (1988) found no evidence of calcium carbonate in his 
1982 explorations, noting that the spring water was under saturated in both calcium and 
 148
bicarbonate. This led Butzer (1988) to conclude that any such carbonate enrichment must 
have originated from either direct, or indirect, external sources, other than the spring 
water. When the spring-water Ca concentration is compared to the Ca concentrations of 
the ground water, 82–400 mg/l, Peat II, 2 008–5 000 mg/l, and Peat IV, 6 720–7 360 
mg/l, it is evident which environment is more suited to fossilization processes. Therefore, 
even early water analysis by Rindl (1915) with a value of 93.4 mg/l Ca, indicated that 
there was a primary under saturation of Ca in the spring-water impeding any fossilization 
processes. 
 
There are a number of other important environmentally related factors, which are 
required individually, or as a combination of factors, in order to contribute to the 
effective precipitation of CaCO3 and other authigenic minerals for fossilization. These 
include pH (alkalinity/acidity), Eh (oxidizing potential), PCO2 (partial pressure of carbon 
dioxide) PO2 (partial pressure of oxygen), DOC (dissolved organic carbon/hydrophobic 
acid), CaCO3 saturation, the decomposition of aquatic plants including phytoplankton and 
periphyton, phosphates, biomineralization, capillarity (Appendix III). Krauskopf (1967) 
mentions that the solubility of CaCO3 calcium carbonate is controlled by the pH of the 
environment, changes in temperature and pressure, as well as organic matter activity and 
decay. Freeze and Cherry (1979) stated that the solubility of carbonate minerals is largely 
dependent on the partial pressure of carbon dioxide (PCO2), giving the range for the 
solubility of calcite and dolomite for natural groundwater’s as P  10-3 CO2 bar and 10-1 bar. 
The solubility of calcite in water at 25° C, pH 7, 1 bar total pressure, and a P  of 10-3CO2  
bar is 100 mg/l, while the solubility at a PCO2 of 10-1 bar is 500 mg/l (Freeze and Cherry, 
1979). Most of the above factors can be related to the clay, organic matter and 
groundwater environment at Florisbad in providing a suitable fossilization process, while 
few of the above factors can be related, or applied, to the spring-water, or its immediate 
environment. The absence of factors such as the under-saturation of CaCO3, Eh, and a 
restricted water flow in a spring context, may well be responsible for the de-fossilization 
of material (Appendix III). 
 
 149
In considering that a suitable environment had been established within the groundwater 
and organic-clay layers for fossilization, it was now important to ascertain why 
fossilization should be precluded from the effects of the spring water. It was concluded 
that there were considerably more reasons for precluding fossilization in a spring context 
than reasons for initiating fossilization. This conclusion was significant in that it would 
also indicate that the spring-water, which had developed over thousands of years, would 
never historically have been in a position to initiate fossilization. It was further concluded 
that the only way in which fossilization could have been initiated in a spring context 
would have been if the spring-water had have been supersaturated in CaCO3, had an 
alkaline pH, with a considerably reduced flow rate. Although the spring-water has an 
alkaline pH, this is irrelevant on its own An important factor governing the presence of 
carbonates in natural waters is the presence of organic matter within a low Eh 
environment with a restricted water circulation (Blatt et al., 1972). A low Eh environment 
with a restricted water flow is necessary for the preservation of organic matter in order to 
prevent the complete oxidation of organic compounds to CO2 and H2O (Blatt et al., 
1972).  
 
An integral part of organic matter preservation is the formation of DOC, which is used as 
a food source by carbonate precipitating organisms, Conversely, DOC may act as a 
kinetic growth inhibitor, inhibiting calcite crystal growth (Hoch et al., 1998). This has 
been established in the Everglades, where the kinetic inhibition of DOC prevents any 
calcite precipitation in calcite supersaturated water (Hoch et al., 1998). A characteristic of 
the Florisbad spring vents are their white quartz sands (Brink 1987), which appear as 
such because all clay and organic material has been washed out of the vents by the spring 
flow. This was confirmed by spring eye sand analysis, which resulted in 0 % clay and 
organic matter content (Appendix II). Another aspect of carbonate precipitation which 
would have been negatively affected by the spring flow is Biomineralization. This is the 
process by which carbonate precipitating organisms sequester CO2 into a solid carbonate 
mineral phase as a result of biomass degradation, also referred too as microbial carbonate 
precipitation (MCP) (Roh et al., 2001; Hammer and Verstraete, 2002). this processes 
could not have taken place in the spring vents in the absence of organic matter, and it is 
 150
also questionable as to whether carbonate sequestrating organisms could survive in the 
flow from the eyes.  
 
An important constituent in the formation of authigenic carbonate apatite is the presence 
of phosphates. Phosphates in the spring-water were very low, recording 1.93 mg/l in 1999 
and 0 mg/l in 1988, while the groundwater registered between 0.010 and 0.100 mg/l 
(Appendix I). Water analysis by previous researches indicated no phosphates in the 
spring-water, further indicating that fossilization could not have taken place in the spring 
vents. The organic Peat II layers showed a large increase in phosphates to between 17 and 
34 mg/l (Appendix II). Phosphates in the organic-clay layers peat were derived from 
three main sources. 
 
• Dissolved oxygen in the water would have reacted with the organic material to 
release CO2, which in turn would have lowered the pH and liberated phosphates 
(Karkanas et al., 2000). 
 
• Carcass remains, which the Old Collection indicates were abundant, would have 
released phosphates before and after burial, resulting in a build up of phosphates 
within the sediments for fossilization. Brink (1987) and Kuman and Clarke (1986) 
both noted that fossilized bones showed signs of pre-burial weathering such as 
sun cracking, and horn cores riddled with grooves made by the horn moth Family 
Tineidae. 
 
• Karkanas et al., 2000 noted that phosphates in caves were largely due to the 
oxidation of bird and bat guano, so the decomposition of waste material from the 
vast herds of game coming to drink at the site would have further released 
phosphates into the sediments during oxidation. See Chapter 6 for further 
discussion on this point. 
 
The recorded pitting and hollowing, and the low incidence of calcium carbonate in, and 
on, Hippopotamus amphius and bovid bones (Brink, 1987), strongly suggests 
 151
demineralisation. This would occur through the under saturation of calcium carbonate 
and phosphates in the spring-water, the dissolution of calcite, and abrasion weathering 
through spring-flow action. Although the incidence of pitting and hollowing is reportedly 
low, Brink, (1987) noted that many bones, which showed no evidence of calcite 
deposition on their outer surface, did have calcite development in the internal cavities. 
Brink (1987) considered pitting to have occurred through solution while the bones were 
still under-mineralised, and that pitting preceded calcite deposition. On the other hand, 
Brink (1987) believed that calcite deposition preceded hollowing, and that hollowing was 
the result of chemical solution, or weathering due to mechanical erosion through water 
action. Therefore, Brink (1987) seems to suggest that there was both calcite deposition 
and demineralisation within the spring vents. Dissolution is also expected to increase in 
sandy soils with flowing water (Hedges and Millard, 1995), and the relatively large 
surface area of bone apatite would account for its correspondingly high rate of dissolution 
in natural waters (Trueman and Tuross 2002). 
 
The salinization and fossilization processes described here may have major implications 
at other archaeological sites in determining the fossilization of faunal remains, and 
history of the site. Butzer (1974) recorded eight distinct layers at the Cornelia fossil beds 
in the North-eastern Free State, of which six comprised clay and/or black organic soil. 
Brink and Rossouw (2000) recorded a massive yellow clay horizon beneath the Unit 1 
described by Butzer (1974) both of which contained fossil remains and a salinization 
process similar to that at Florisbad cannot be ignored. It was noted from the literature that 
no other fossil sites in South Africa appear to have been examined in detail in order to 
determine the fossilization processes involved. It was suggested by Brouwer (1967) that 
many detailed palaeontological descriptions make no mention at all of the conditions 
under which fossilisation has taken place. Brink (1987) stated that, to fully understand a 
fossil fauna, it is important that the palaeoenvironment from which the fossil remains 
were extracted is better known. 
 
 
 
 152
5.6.5 A New Perspective on the Formation of the Florisbad Spring Site 
 
There were two primary reasons for revisiting theories on the formation of the Florisbad 
spring site. 
 
Firstly, on reviewing the literature, it became apparent that none of the previous theories 
on the formation of the spring site could explain the stratigraphy of the sediments in light 
of the salinization process. Three basic developmental theories have been put forward in 
the literature for the formation of the Florisbad spring site: 
• The site developed from sand output from the springs, which was most 
probably derived from the underlying bedrock (Dreyer, 1938a; Brink, 1987; 
Butzer, 1988; Van Zinderen Bakker, 1989). 
• Deposition at the site was directly related, or influenced by, water levels in the 
adjacent palaeolake (Soutpan), or the flooding of the palaeolake in four low 
water level phase and three high level water phases. (Visser and Joubert, 
1991; Joubert and Visser, 1991; Henderson 2001a). 
• Deposition was almost entirely derived from aeolian sources (Appendix I; 
Appendix II; Appendix III; Grobler and Loock, 1988a, 1988b; Van Zinderen 
Bakker, 1989; Kuman et al., 1999). 
 
Some researchers have suggested that a larger body of water might have existed at 
Florisbad in the past (Brink, 1987; Kuman et al., 1999), with little detail, or evidence, 
being given regarding actual developmental processes of this larger water body. 
 
Secondly, a number of unexplained, and debatable points were found in the literature to 
which it was hoped that answers could be provided. These included: 
 
• The lack of upper red-brown sand units on the eastern side of the “mound” 
appeared to have either not been deposited, or had been eroded by spring 
discharge (Rubidge and Brink, 1985; Brink, 1987). 
 153
• A seven metres thick layer of clay deposits in the modified vlei area (Butzer, 
1984), which could not have been deposited by the pre-existing weak ephemeral 
drainage line. 
• Older aeolian deposits, evident over the rest of the site, are absent from the vlei 
area, while more recent aeolian deposits were present (Fourie, 1970). This 
represents an anomaly in relation to the area in general, where the erosion of the 
older aeolian sands along the vlei being seen as an uncharacteristic erosional 
character of the weak ephemeral drainage line. 
• The extent of the dogleg in the ephemeral drainage line from the inter-dune 
valley, cutting across the natural drainage at almost right angles from NW to 
NNE, and back to NW. This could also be seen as being contrary to any erosional 
forces that could have been exerted by the weak ephemeral drainage line. 
• The disproportionate width and depth of the vlei area in relation to the ephemeral 
drainage line. 
• The erosion of the east bank of the vlei area. 
• The pinching out of the Peat IV layer on the western wall of the excavation area. 
 
Rubidge and Brink (1985) stated that the lithostratigraphy and depositional history of the 
Florisbad sediments were still in their initial stages of investigation, and that several 
models could be proposed. Rubidge and Brink (1985) also noted that, in their opinion, 
because the deposits were lithologically variable, they were the product of an unusual 
depositional environment. Kuman and Clarke (1986) noted that the sand, silts, and 
organic deposits were poorly understood and controversial. The above questions and 
statements therefore opened the way for further interpretations on the formation of the 
site. Unless otherwise referenced, all references are from Appendix IV, or are updated 
versions of the hypothesis put forward in Appendix IV. 
 
It is hypothesized that the genesis of the Florisbad site occurred in a slightly undulating 
topographical environment, with the spring site evolving prior to the formation of the 
panveld. At the time a spring pan had probably already formed, but was covered by the 
migrating sand dunes, referred to as the south-east dune belt, resulting in a fossil pan 
 154
being buried beneath the sand dunes. As the dune belt continued migrating in a south-
easterly direction, so the fossil pan became uncovered and the springs again became 
active on the surface. 
 
Figure 26, a simplified schematic cross-section of the sediments, has been replicated for 
convenience purposes from Figure 4, to be read in conjunction with Figures 30, 31 and 
32. Figure 30 is the legend for the interpretation of the five developmental stages of the 
formation of the Florisbad spring site reflected as a plans in Figure 31, and as a cross-
section in Figure 32. 
 
Stage I: Growth of the modern spring pan would have initially developed along the form 
of the fossil pan, from when it would have developed its own character. This character 
would have been based on factors such as the large number of animals coming to drink at 
the pan, deflation during dry periods, aeolian deposition, growth and decay of aquatic 
vegetation, and the contribution of waste from the herds of animals. In the initial 
developmental stages it is suggested that the inter-dune ephemeral drainage line drained 
directly into Soutpan, and that it was later captured by the spring pan. This ephemeral 
drainage would also have contributed water to the pan during wet periods. Around about 
this time a sand dune started to develop on the south-eastern bank of Soutpan, derived 
from the deflation of the Soutpan floor and aeolian deposition from the surrounding area 
and further a field from the west and north-west. 
 
Much has been made throughout this thesis regarding the influence of the vast herds of 
game roaming the Free State plains in the past, and the contribution they may have made 
to the erosional and chemical processes at the spring site. The contribution to erosion by 
the herds of game coming to drink at the spring site, in search of minerals such as salt, 
resulted in the breaking down of sand (weathering and deflation) and aquatic vegetation 
(decomposition) by trampling. These vast herds of game also contributed to the 
mineralization of the site through their waste contribution (decomposition, phosphates), 
and the contribution of carcases, either from carnivore killing, or natural causes 
(phosphates). But just how significant were these herds of game? 
 155
 
 156
 
 157
 
 158
The size of the herds inhabiting these grass planes can be judged by the results of a hunt 
carried out by Prince Albert on the 24 August 1860. Some 1000 Baralong tribesman 
rounded up between 20 000 and 30 000 head of game at Bain’s Vlei, 50 km south of 
Florisbad, and within an hour, the Prince and his hunting party managed to slaughter an 
estimated 5 000 animals (Anon [5], 1860; Anon [6], 1977). At the time, it was reported 
that the number of animals rounded up for the hunt was disappointing, and that the 
original number of animals driven onto the plain for the hunt should have been far greater 
(Anon [5]; 1860). 
 
Stage II: The spring pan had grown considerably in size, and was now supporting aquatic 
dependant species such as hippopotamus (Hippopotamus amphibious), lechwe (Kobus 
leche), clawless otter (Aonyx capensis), and water mongoose (Atilax paludinosus) (Brink, 
1987: Henderson, 1996; Henderson, 2001a). During wet periods, water was still being 
supplied by the ephemeral drainage. The now referred to Florisbad dune continued to 
grew in size from the deflation of Soutpan and other aeolian deposits, and started 
migrating in a south easterly direction towards the spring site. 
 
Stage III: The Florisbad dune now started migrating across the spring pan and blanketing 
it with sand, thus halting any further northward expansion of the pan. The arms of the 
sand dune now also begun to encompass the spring pan from the sides, and thus a 
damming effect was began, with further expansion of the pan being restricted to areas 
east, west and south of the advancing dune. Fresh water was still being supplied by the 
ephemeral drainage during wet periods. At this time is it is thought that the distribution of 
organic matter, as indicated by Grobler and Loock (1988b) (Fig 9), was far more 
extensive to the east and west, but particularly so to the south. It is suggested that the 
Basal Peat, Peat I and Peat II were formed during the developmental stage of the spring 
pan 
 
Stage IV: The arms of the sand dune now met and overrode the windward base of the 
south-eastern dune belt, cutting off any possible outlet for the spring pan water. The 
spring pan, which had always been seen as a drainage-impeded pan by the author, now 
 159
became a dam, completely enclosed by sand dunes. As the dune migrated it continued to 
cover the spring pan and ride further up the windward slope of the south-eastern dune 
belt, allowing for a steady increase in the depth of the water through spring flow and the 
ephemeral drainage. Considering the thickness, friability, and relatively low degree of 
decomposition of the Peat IV layer (2.8 m wide, and to within 650 mm of the current 
surface) on the west wall of the excavations, the rise in the water level must have been 
relatively slow in order for these organic beds to have developed and attain this height 
(see Section 3.2 for aspects of palaeoclimate supporting this). The low degree of 
decomposition in Peat IV would also suggest that these reed beds had been, at various 
stages of their formation, returned to a dry state during their growth, which inhibited any 
further decomposition. 
 
It is proposed that Peat III and Peat IV developed during this damming-up period, with 
Peat III possibly being isolated remnants of Peat IV which were left behind after Stage 5. 
The proposed area of the Florisbad dam is given in Figure 31, 4a and 5a. 
 
Stage V: With the increase in sediments and height of the water, the pressure against the 
arms of the sand dune increased. The eastern arm of the sand dune was then breached, 
taking most of the eastern arm with it, with the contents (sediments and water) of the 
spring dam evacuating the site in a flash flood. This flash flood gouged out a new 
drainage line from the soft dune sands to form the vlei, with the force of the flash flood 
severely modifying the new drainage line into a deep, wide, flood plain. The contents of 
the spring dam were deposited along the length of the vlei, giving it its current form. As 
the force and volume of water decreased along the newly formed drainage line, the vlei 
began to narrow, and the flow took an almost right angle turn, resuming a natural north-
westerly flow into Varspan. The large flat amphitheatre like area to the east and south- 
east of the springs is thought to be a part of the original base of the spring pan where it 
dammed up against the windward slope of the south-eastern dune belt. 
 
Three possible scenarios are envisaged, all of which could have led to the final breach. 
One was that the water rose to a height where it began to flow over the top of the eastern 
 160
arm of the sand dune and eroding the wall until it breached. A second alternative was that 
the weakest point in dune wall was the angle of repose of the eastern arm up against the 
south-eastern dune belt, which, as the height of the water increased, gave way under the 
pressure. The third alternative is that the herds of game coming to drink at the site, eroded 
the eastern arm of the dune to a stage where it began to overflow, and then as the arm 
further eroded, breached. Whichever of the fore mentioned scenarios was the cause of the 
breach, considerable erosion would have taken place on either side of the breach. 
 
Perhaps one of the most compelling pieces of evidence supporting this formation 
hypothesis is the high salinization of the organic clay in the vlei. Table 15 gives a 
comparison of the degree of salinization between the Peat II samples from the north wall 
of the excavations (X1), excavation pit 2 (X2), excavation pit 4 (X3) (Figure 23), and the 
organic-clay from the vlei area, approximately 400 metres from the residence. This high 
salinization of the organic-clay from the vlei further supports the hypothesis that the 
deposits in the vlei originated from the dam site, and not Soutpan. Further support is 
provided by the total ion count for the vlei clay being close to the average (14 897 mg/l) 
for the three Peat II samples in Table 15, strongly suggesting that a homogeneous mixing 
of the sediments from the dam site had occurred on entering the vlei area. It is regrettable 
that samples could not be obtained from the Peat I layer as it is presumed that the TDS of 
this layer would be even higher than the Peat II layer. 
 
In conclusion it can be stated that this research provides an alternative hypothesis on the 
formation of the Florisbad spring site, which accommodates all aspects of sedimentation 
and fossilization, as well as seriously questioning previous theories. Further to this, this 
hypothesis explains the eroding, or non-deposition, of the red-brown units on the eastern 
side of the site, the 7 m of clay deposits along the vlei area, the disproportionate size of 
the vlei relative to the ephemeral drainage line, and the changes in direction of the 
ephemeral drainage flow. Further evidence in support of this hypothesis is the deep 
cutting into the base of the windward side of the south-east dune belt on the eastern bank 
of the vlei, which would accommodate the directional flow hypothesis of a flash flood 
from the eastern wall of the dam site. The pinching out of the Peat IV layer on the 
 161
western side of the site would be explained by the greatest erosion having taking place on 
the opposite, or eastern side of the dam site. This is because the western portion was 
furthest from these erosional forces, and the Peat IV layer remained largely intact on the 
western face. In Figure 32, 4b, it is not proposed that Peat IV completely filled the dam,  
 
 
Table 15. A comparison between the salinization of the various Peat II samples and the 
organic-clay from the vlei area showing the high salinization of the vlei area 
organic-clay 
 
 Sample X1 Sample X2 Sample X3 Vlei 
 Peat II Peat II Peat II Organic-clay 
 
 
PH 8.30 8.00 6.65 6.75
 
Resistance (Ohm) 26 38 19 57
 
NaCl (mg/l) 2 427 11 967 19 193 8 499
     
IONS (mg/l)     
     
P 22 26 34 66
 
Cl 1 567 7 727 12 393 5 499
 
Ca 2 008 2 768 5 000 4 540
 
Mg 23 25 70 48
 
K 129 640 352 638
 
Na 860 4 240 6 800 3 000
 
Zn 1.2 2.4 6.0 5.5
 
Total ions 4 610.2 15 428.4 24 655.0 13 796.5
 
 
 
 162
but reflects the Peat IV layer against the western wall of the dam. It was also proposed 
that much of the leeward side of the dune face slumped into the spring dam as it drained, 
having previously been supported by the roots of aquatic vegetation and the pressure of 
the water. 
 
5.7 MEDICINAL PROPERTIES 
 
Because Florisbad built up an extensive reputation for its healing properties over the 
years, and because Floris Venter became known as the “healer”, a brief hypothesis on this 
aspect of the springs is provided. No scientific studies have been carried out on any 
possible medicinal properties of the Florisbad spring-waters, but as far back as the early 
1900s, the Florisbad spring was included in papers and articles on the medicinal springs 
of South Africa (Rindl, 1915, 1916; Kent, 1948, 1964; 1971). In some instances the 
Florisbad spring water was reported to be particularly effective in the treatment of 
sciatica, and muscular and articular rheumatism (Rindl, 1916)  
 
An examination of the water analysis tables in the appendices reveals that, with the 
exception of two elements, there appear to be no other elements indicating that the spring 
water could be medicinally beneficial. The two elements, about which much has been 
written and researched over the past 90 years, are strontium (Sr) and boron (B). Strontium 
had the fourth highest spring-water cation concentration of 2.8 mg/l after Na, Ca, and K, 
and B the fifth highest concentration of 2.6 mg/l (Appendix 1). In exploration pit 4, 
groundwater values for B increased by 288%, over that of the spring-water, to 10.1 mg/l, 
and Sr by 300% to 11.2 mg/l (Appendix I). Today Sr and B are credited in relieving the 
symptoms of inflammatory processes, immune function, calcium metabolism and 
absorption, osteoarthritis, rheumatoid arthritis, osteoporosis, and ruptured discs (Herbert, 
1921; Hall, 1976; Skoryna, 1981; Marie et al., 1985: Marie and Holt, 1986; Gaby, 1994; 
Newham, 1994a, 1994b; Travers et al., 1990; Nielsen, 1992: Hunt and Idso, 1999; 
Henrotin et al., 2001; Meunier et al., 2002; Meunier et al., 2004). 
 
 163
It is postulated that, in line with the large ion concentration increases between the 
groundwater and organic-clay layers (Appendix II), Sr and B will exhibit a similar 
increases within the organic clay layers. There are a number of reasons for this 
supposition. Results from Boon and McIntyre (1968) gave evidence supporting the 
hypothesis of B incorporation in fine-grained sediments in amounts proportional to the 
salinity of the depositional environment, which was supported by Brooks and De Wall 
(1976). Shimp et al. (1969) noted that the B content was correlated with the content of 
clay finer than 2µ, and that this gave the best discrimination between marine and fresh 
water clays. Mortvedt et al. (1972) noted that the amount of B in the soil was directly 
proportional to the amount of organic matter, while Villumsen and Nielsen (1976) found 
variations in B being related to the content of montmorillonite in the clay fraction. 
Frederickson and Reynolds (1960) summarized the situation by stating that the B content 
in saline waters increased with the salinity of the water, finding that B in the clay mineral 
fraction of sedimentary rocks was associated with the clay mineral illite. 
 
Large quantities of montmorillonite, illite, and organic matter are recorded from the 
Florisbad organic-clay layers, adding further support for salinization (Appendix II), and 
making them ideal depositories for Sr and B. The Florisbad organic-clay layers have an 
NaCl content of up to 65% that of sea water, compared to the 7.5% of the spring-water 
(Appendix II). It therefore seems feasible that the Sr and B content of the Florisbad clays 
would be considerably higher than that of the spring-water. Grobler and Loock (1988a) 
considered that, as the salts Ca2+, Na+, SO2 -4  and Cl-, were derived from the underlying 
Ecca Group rocks, which were deposited in a marine environment, this may be an 
explanation for the presence and levels of Sr and B. 
 
Newham (1994a) noted that in areas of the world where B intake was <1-2 mg per day, 
the incidence of arthritis varied between 20-70%, whereas, where B intake ranged from 
3-10 mg per day, the incidence of arthritis varied between 0-10%. Although the effect of 
percutaneous absorption through the low levels of Sr and B in the spring-water is not 
known, the effects of percutaneous absorption through the higher organic-clay medium 
might be considerably higher. Two treatment methods were used at Florisbad. The first 
 164
was to pack heated mud, taken from the vlei area below the swimming baths 
(Hendersom, Z. pers com. 20/09/2007) between two layers of newsprint, and pack these 
onto the patient’s bodies (Henderson, 1995). The second was to submerge ones self in a 
mud pool located on the east side of the baths (Henderson, 1995). It is contended that it 
was contact with the organic-clay mud, which contained elevated levels of Sr and B, 
which brought relief from the arthritis. Although research has shown that Sr and B have 
definite anti-inflammatory benefits, the effects of these elements in the Florisbad context 
still need to be proven. 
 
5.8 SYNOPSIS 
 
In this chapter, as in previous chapters, a logical progression has been followed in dealing 
with the numerous components, or elements, contributing towards the formation and 
fossilization of faunal remains at the Florisbad spring site. These include providing a 
comprehensive background in order to bring together these elements in presenting a 
holistic picture of the spring site. 
 
For example, in this chapter, the chemistry of the spring- and groundwater was examined 
in order to make a distinction between the two entities. This then led to an examination of 
the sedimentation, chemistry, properties, and processes, of the organic-clay layers in 
order to be able to relate the results to mineralization and salinization. The chemistry of 
the spring-, groundwater, organic-clay layers were then used to determine where the most 
likely areas for fossilization were likely to occur. From these results it was concluded that 
fossilization could not have occurred in a spring-water context, and that fossilization must 
have occurred in conjunction with the organic clay layers. As no available theories on the 
formation of the site could accommodate the various hypotheses that had been formulated 
from these results, an hypothesis was developed which would accommodate all the 
components. This hypothesis further resulted in a number of previously unanswered 
questions in the literature, regarding the formation and morphology of the site, being 
resolved, and new evidence being established in support of these answers. It is therefore 
 165
concluded that a comprehensive and holistic picture on the formation and fossilization of 
faunal remains at Florisbad has been achieved. 
 
Problems with the dating of the Florisbad deposits has been emphasised in these chapters. 
However, published data on the most recent dating of fossils from one of the excavation 
pits, indicates equal compaction and compression of the sediments over time and depth. 
This was questioned, in that logically, the deeper sediments should be far more 
compacted than the shallower sediments, with time and decomposition increasing the 
compaction. This issue is examined in some detail and discussed in Chapter 6. 
 
 166
 
 
 
 
 
Chapter 
 
6 
 
 
 
 
The use of 
Mathematical 
Trend Lines in 
Evaluating ESR and 
OSL Dating at 
Florisbad 
 
 
 
 
 
 
 
 167
CHAPTER 6 
 
THE USE OF MATHEMATICAL TREND LINES IN EVALUATING ESR AND 
OSL DATING AT FLORISBAD  
 
 
6.1 INTRODUCTION 
 
Florisbad is a significant archaeozoological and archaeological site situated in the Free 
State Province of South Africa (28° 46’S 26° 04’E). The importance of Florisbad lies in 
three main areas. Firstly, the discovery of the Florisbad hominid cranium by Prof. T.F. 
Dreyer in 1932 and the description thereof in 1935 (Dreyer, 1935), brought recognition to 
Florisbad. Currently the cranium is electron spin resonance (ESR) dated at 259 ±35 ka, 
(Grün et al., 1996). Secondly, a vast collection of faunal fossil remains, and artefacts, 
representing some 26 species, referred to as the Old Collection, (Brink, 1987), which 
represents the Florisian-Cornelian faunal boundary with a possible age of c. 400 ka 
(Klein, 1984). Thirdly, a Middle Stone Age Human Occupation horizon, which reflects a 
temporary butchering site, and which has delivered butchering tools as well as fossils 
(Brink, 1987; Henderson, 2001a, 2001b; Brink and Henderson, 2001). 
 
While researching various aspects of the Florisbad spring site, such as water quality 
(Appendix I), salinization (Appendix II), and the formation of the site (Appendix IV), a 
paper on the most recent dating of the spring sediments by electron spin resonance (ESR) 
and optically simulated luminescence (OSL) was examined. It became apparent that the 
ages given for the sediments at descending depths in profile (a) Grün et al. (1996), did 
not conform to a logical progressive compaction of the sediments with increasing depth, 
and increased overburden, over time. That variation in factors such as, climate, spring 
flow, vegetation growth, aeolian deposition, and decomposition, would all equal out to 
produce two depositional zones of almost equal thickness and compaction, over almost 
equal time frames, at different depths, was considered highly improbable. Due to the high 
water table, which fluctuates between three and four metres above basement, it was not 
possible to obtain more detailed compaction data. A major influence on the compaction 
 168
of the basal sediments would have been the decay and compaction of the organic 
material, with decomposition also contributing to greater compaction. Before the build up 
of aeolian sand deposits, the basal organic layers were in all probability wide porous 
bodies similar to the present day extremely porous Peat IV organic layer, which extends 
from just below the surface down to 2.8 m on the western wall of the excavations 
(Appendix IV). Despite its depth, the Peat IV layer is only covered by a few centimetres 
of overburden, which was not enough to cause any compaction. The basal sediments 
would have gone through processes of vegetation build-up during wet periods, being 
covered and compressed by the mass of aeolian deposits during dry periods, and 
decomposition during further wet periods. These cyclic events would have compacted 
and compressed the basal organic layers into the narrow bands they are today. It is 
therefore evident that the basal sediments have not only been in existence longer due to 
their position in the sequence, but have also been exposed to a longer and greater period 
of compaction and compression than the upper sediments. 
 
The effect of compaction and compression is further illustrated in the Unit F horizon 
where the mass of sediments was so great that the compacted skull and pelvis of 
Hippopotamus amphibius have been ascribed to this process (Henderson, 1996; 2001a). 
These remains were located approximately 3.0 m from the surface, and are less 
mineralised than remains from the Old Collection (Brink, 1987). The compaction of 
bones has not yet been recorded from the Old Collection, which in some instances is 
nearly twice the depth of the Middle Stone Age (MSA ) Occupation Horizon from 
surface. This would indicate that the Old Collection remains possibly became fossilised 
before any significant build up of sediments occurred. It is obvious from the excavations 
on the west wall that there has been an insufficient build-up of aeolian deposits above the 
porous Peat IV layer, or a prolonged aquatic environment which lasted long enough to 
decompose this organic layer (Appendix IV). If the compaction of H. amphibius bones 
occurred at a depth of just over 3.0 m, then the compaction of vegetation and sands at 
more than double this depth must have been far greater. 
 
Grün et al. (1996) noted several inconsistencies in the dating and gave possible reasons 
for these inconsistencies in their paper. After examination of the available data, it was 
 169
decided to find a method with which to test these results. It was decided that the 
application of mathematical trend lines would reflect increased, or linear, compaction of 
sediments with depth, over time. This paper is not an attempt to establish new ages for 
the Florisbad sediments and faunal remains, but rather a novel hypothetical exercise to 
examine the factors that may have influenced the most recent ESR and OSL ages, and to 
test the validity of these ages against the use of mathematical trend lines. The use of 
mathematical trend lines in evaluating sedimentary compaction and compression with 
depth does not appear to have been previously examined. 
 
6.1.1 Stratigraphy 
 
The stratigraphy of the Florisbad spring site is important because it holds the key to the 
history of the site, including aspects such as its formation and the preservation and 
fossilization of both the Old Collection and the MSA human occupation horizon. For 
example, it is here contended that the spring water was never responsible for the 
fossilization of the faunal remains, but that a multifaceted salinization process within the 
sediments, particularly in the clay and peat layers, ionically enriched the groundwater to 
the extent that the latter became responsible for fossilization (Appendix III). Further 
more, it has been contended that the spring water is responsible for the demineralisation 
of previously fossilized material (Appendix III). 
 
The stratigraphy of the site is complex, poorly understood and controversial, while being 
the product of an unusual depositional environment, which can be described by several 
depositional models (Rubidge and Brink, 1985). Hypotheses on the formation of the site 
include the accumulation of spring sand, vegetation and aeolian deposits (Brink, 1987), 
the transgression and regression of waters from the nearby palaeolake (Soutpan) in a 
number of high and low cycles (Visser and Joubert, 1991), and the release of detrial sand 
from the underlying Ecca shale and dolerite (Dreyer, 1938a; Butzer, 1988). Other 
researchers have proposed that sand at the site is almost entirely of an aeolian nature 
(Grobler and Loock, 1988b; Loock and Grobler, 1988; Joubert and Visser, 1991; Van 
Zinderen Bakker 1989; Kuman et al. 1999). It is proposed here that the spring mound is a 
sand dune which moved in a south-easterly direction from Soutpan over the site, resulting 
 170
in a damming of the site by the dune, and that the spring pan and sand dune were 
inextricably interlinked in the later stages of the site’s formation (Appendix IV). It is also 
contended that the spring itself contributed little to the depositional process (Appendix 
IV). See Chapter 5, Figure 26, for detail on the stratigraphy discussed in this chapter.  
 
Adding to the complexity of the site is the considerable variation in both horizontal and 
lateral stratigraphy. This is confirmed by the results of an auger-drilling programme 
where a low degree of correlation was noted even between adjacent boreholes (Rubidge 
and Brink, 1985). From previous experience, large diameter auger drilling is considered 
an imprecise method for determining stratigraphic sequences in soft sediments, as 
opposed to core sampling. This is because the details of narrow layers may become 
distorted through the action of the auger bit, as well as through the feed rate of the auger 
bit, which in turn may result in compaction and distortion. Due to this, the accuracy of 
depth between boreholes may also come into question. Variations in the stratigraphy of 
the site can be attributed to several factors. These include the non-uniform deposition of 
aeolian sands and the uniform deposition of sediments in the aquatic environment of the 
spring pan, with further variations being due to alternating wet and dry periods. The 
reworking and remixing of sediments through the eruption and migration of spring eyes 
(Dreyer, 1938a; Brink, 1987) would have also played a significant role in the 
redistribution of material within the spring deposits and sand dune. 
 
Other potential influencing factors are earthquake-induced liquefication (Visser and 
Joubert, 1990) and the stratification of the sand dune through its growth and migration, 
with the further effect of stabilization by plants during wet periods and aeolian deposition 
during dry periods (Appendix IV). It is suggested that erosion was a major force during 
the early history of the site when deflation was dominant for much of the time, but as 
from the time of the formation of the dam, erosion decreased, and was minimal. 
Investigations by the author suggest that the effects of erosion have generally been 
relatively minor since the formation of the dam, with the site being in a largely aggrading 
state (Appendix I). This was largely due to aeolian sand deposition and the sand dune 
moving over the site (Appendix IV). With the exception of severe erosion when the 
dammed site drained itself (Appendix IV), other erosion factors may have included the 
 171
early periods of deflation of the spring pan by wind during dry periods, which in turn may 
have been offset by a relatively constant spring flow, and some erosion by the large herds 
of animals coming to drink at the pan (Appendix I; Appendix IV). 
 
It should be born in mind that the degree of deposition and sedimentation was highly 
variable, being punctuated by high and low varying peaks. Because sediments are 
deposited on top of one another over time, this presents a continual stratigraphic unit in 
situ, so there is no way of determining whether non depositional periods (events) actually 
occurred. It must therefore be assumed that deposition was more or less continuous to 
varying degrees. Variability would have been due to factors such as rainfall and 
temperature, which would have in turn related to wet and dry periods, wind, aeolian 
deposition, and the development of the lower organic layers, their decomposition and 
subsequent compression and compaction. Compression and compaction would also have 
been variable, dependant on the aforementioned factors, but they would also have been a 
varying constant in that they were always a present force, dependant on the accumulation 
of sediments and water. The lower deposits would always have been under the increasing 
effect of compression and compaction as the accumulation of the sediments and water 
increased. It would, at this time, be difficult to determine whether narrower layers of 
organic matter actually formed as narrow layers, or whether they were much wider layers 
that had been compacted over time. The latter is thought to have been the case. 
 
6.1.2 Archaeozoological and Archaeological Deposits 
 
Two distinct archaeozoological and archaeological deposits, which are separated from 
each other both horizontally and vertically, have been excavated at Florisbad. The first is 
referred to as the Old Collection, and comprises a basal accumulation of fossilized faunal 
remains in the areas of spring activity, representing a death assemblage resulting from 
largely carnivore killings. Bones in this assemblage are characterized by evidence of 
hyena damage, the unbroken state of the bones, and to a much lesser extent, porcupine 
gnawing, with no indication of cut marks (Brink, 1987; Brink, 1988). The Old Collection, 
or spring assemblage, is of particular importance in that it represents the type collection 
of the Florisian Land Mammal Age (Klein 1984). The second collection is the Middle 
 172
Stone Age (MSA) occupation horizon which occurs approximately 3.5 metres above 
basement, where species diversity and numbers are far less than in the basal Old 
Collection, and represents a butchering site (Brink, 1987; Henderson, 2001a; Brink and 
Henderson, 2001). Artefact and faunal remains were found in horizontal deposits on a 
sandy substrate, which appear to have been deposited in an aqueous environment with 
little disturbance (Meiring, 1956; Henderson, 2001a; Brink and Henderson, 2001; Kuman 
and Clarke, 1986; Kuman et al., 1999). In this assemblage, signature marks on bones 
indicate slicing, scraping and cutting, as well as bone-breaking, with no signs of 
carnivore damage (Brink, 1987; Henderson 2001). MSA faunal remains are also very 
friable in relation to remains from the Old Collection (Brink, 1987). 
 
6.1.3 Previous Dating 
 
Since the early 1950s, numerous attempts have been made to determine the age of the 
Florisbad sediments by 14C radiocarbon dating methods (Libby, 1954; Broecker, 1956; 
Barendsen, 1957; Beaumont and Vogel, 1972; Beaumont et al, 1978; Kuman and Clarke, 
1986), as well as by Amino Acid Racemization (Protsch, 1974), uranium series (Clarke, 
1985), and thermoluminescence (Joubert and Visser, 1991) methods. Earlier 14C dating of 
the more recent deposits, from Peat II to the surface, has provided ages for these upper 
sediments, although the method has also provided considerable variation in the ages for 
these horizons. 14C dating of the basal deposits and Peat II layer provided ages greater 
than the limit of the 14C method. Previous 14C dating of the Unit F horizon was >43.7 ka 
whereas the MSA marker bed, below Unit F, was dated at >47.2 ka (Kuman and Clarke, 
1986), with these ages being well below recent ESR and OSL ages This paper is based on 
ESR and OSL dating results only. 
 
ESR dating on tooth enamel and OSL dating on quartz separates from sediment samples 
have recently been employed (Grün et al., 1996) in an attempt to obtain more accurate 
dating. ESR results placed the Florisbad hominid tooth at 259 ±35 ka, Peat I from near 
the base of the formation at 281 ±73 ka, and the sandy MSA Unit F horizon near the 
middle of the formation at 121 ± 6 ka (Grün et al., 1996). 
 
 173
6.2 METHODS 
 
6.2.1 The Application of Trend Lines to the ESR and OSL Ages  
 
The ESR and OSL age estimates used in Figures 33, 34, and 35 were taken from Grün et 
al (1996), and from the following profiles: (a) Age estimates for the third test pit, and (c) 
Schematic diagram of the basal part of the sediment sequence at Florisbad with a 
reconstruction of the spring vent yielding the hominid fossil (not to scale). These results 
have been tabulated in Table 16 where the disjunction of ages and degree of age spreads 
in (a) are evident. The ages and the depth of the samples were plotted on graphs with the 
y axis representing age, and the x axis representing the depth. Linear (Lin), logarithmic 
(Log) and exponential (Exp) trend lines were then applied to the resultant graphs. It was 
hypothesised that the trend lines would reflect the varying degrees of compaction and 
compression over depth and time. 
Hypothetical  
It was further hypothesised that, over depth and time Lin would reflect equal compaction 
at all levels; that Log would reflect greater compaction in the basal levels, with a gradual 
decrease towards the surface; and that Exp would reflect an increasingly greater 
compaction with depth, with a gradual decrease towards the surface. A best fit to data 
between the original data (graph) and the trend line was tested against R2 values, where 
R2= 1.0 indicates a perfect fit. Although 3 decimal places may not be justifiable, they are 
given as an indication of variance where results are very close. Should this hypothesis be 
valid then the application of trend lines to ESR and OSL dating at Florisbad could be 
used as a tool for validating and evaluating these ages. 
 
Two approaches were adopted in this study in order to validate ESR and OSL dating 
against depth and time. In the first approach, the ESR MSA dating was used as a variable 
with the basal age remaining as a constant, and in the second approach, the basal ESR 
dating was used as a variable, with the ESR MSA dating remaining as a constant. In both 
instances, the near-surface ESR dating remained as a constant. The ESR dating used at 
1.75 m from surface was obtained from the mean of the two samples taken from (a) as 25 
ka. However, this age, was considered far too old for this shallow depth, and should 
 
 174
Table 16.  A tabular interpretation of age estimates and depths for the third test pit of 
Grün et al. (1996) (illustration (a)) (refer to Pit 2 for this study) and the lower 
spring sediments, Grün et al. (1996) (illustration (c)). 
 
 
  Illustration (a) Illustration (c) 
Depth Method Ages and age Stratigraphic Method Ages and age 
Metres  spreads in ka layer  spreads in ka 
 
 
0      
0.2 14C 1    
0.5 14C 1    
1.3 ESR 15 ±3    
2.2 ESR 93    
   White sand layers OSL 146 ±15 
   White sand layers OSL 128 ±22 
3.4 OSL 143 ±7    
4.0 ESR & OSL 145 ±40    
   Above Peat II -MSA OSL 133 ±31 
   Above Peat II –MSA ESR 121 ±6 
4.2 ESR & OSL 127 ±33 MSA    
4.4 ESR 181 ±44 Olive-green sand OSL 157 ±21 
4.7 ESR 159 ±21    
5.5 ESR 250 ±25 Peat I –OC OSL 281 ±73 
   Florisbad Skull -OC ESR 259 ±35 
6.2 ESR 140 ±10 OC   
6.2 ESR 202 ±28    
6.4 ESR 129 ±9 Brown sand –OC OSR 279 ±47 
7.0 ESR 152 ±8 OC   
7.0 ESR 177 ±5    
7.4 ESR 198 ±46 OC   
 
MSA = Middle Stone Age Occupation Horizon 
OC =Old Collection 
 175
 
possibly be in the range of c 8.8 to 11.7 ka, as indicated by the trend lines. This was 
confirmed by 14C radiocarbon dating taken from Kuman & Clark (1986), who recorded 
ages for the sand above Peat III (Table 4) as ranging from 4.37 ±.07 ka to 11.7 ±.1 ka, 
and Peat IV (Table 4) as 3.55 ±.6 ka to 5.53 ±.8 ka. With the MSA horizon in (a) having 
an ESR dating spread of 125 ka, the mean of the ESR dating of 121 ka in (a) and (c), and 
the OSL dating of 133 ka (c) at the same level, was used. This resulted in an MSA age of 
127 ka. Although the oldest age of c. 255 ka in (a) is situated 2 m above the base, it has 
been presumed that this sample was repositioned by some physical force and that it 
represents the basal sediments, based on the oldest age coming from the greatest depth. 
The mean age of the two samples in (a) was taken as representing the basal sediments at 
250 ka. Although this is not the oldest age recorded in the paper, the age fits within the 
ranges for older ages from the spring sequence (c) as well as the 259 ± 35 ka of the 
hominid tooth (Grün et al., 1996). 
 
While ages from Grün et al., (1996) are related to tooth samples, it is assumed that these 
tooth samples are directly related to the sediments in which they were found within the 
profiles. Therefore, in this context, tooth sample and sediment are synonymous, as in 
basal and lower sediments. 
 
There is no available stratigraphic cross-section for test pit 3, and in any event, because 
the test pit was located relatively high up on the dune face, a cross-section would not 
reflect many of the stratigraphic sequences referred to in the paper. This is due to the 
considerable horizontal variation in the stratigraphy, and may largely reflect the internal 
structure of the dune. Any cross-section reflecting a specific section of the site would also 
only reflect that particular sequence and would be of relatively limited application in the 
context of the site as a whole. Figure 28 provides a simplified cross-section illustrating 
the major stratigraphic sequences discussed in this paper. The deposits are composed 
entirely of sands, clay, and peat-like layers, with no rock formations, thus making 
compaction and compression major contributing factors to the formation of the sequence. 
 
 
 176
6.3 RESULTS 
 
The ESR dating of the deposits from ref. 2 (a) were plotted against depth (Figure 33). It 
was clear from the resulting graph that the MSA and basal deposits were almost 
equidistant apart, possibly indicating that compaction and compression in the upper 
layers was almost equal to that in the basal layers over two separate time periods. In order 
to test this assumption, Lin, which would corresponded to a uniform rate of compaction 
throughout the sequence, was applied to the data (Figure 33. An R2 value of 0.991 was 
obtained (Figure 33; Table 17, (a)), clearly showing that the data taken from 2 (a) 
represents equal compaction of both the upper and basal sediments, as well as suggesting 
an equal formation time. 
 
In order to test the statements on the age of the deposits being related to greater 
compaction with increased depth, Log applied to the above data (Figure 33; Table 17, 
(a)). As predicted, the trend line indicated greater compaction occurring in the basal 
levels, and gradually decreasing towards the surface. Because of the close fit of data to 
Lin, Log did not fit the data at R2= 0.95. Log provided an age for the basal deposits 
similar to that of the original data of c. 260 ka, and a basal MSA occupation horizon age 
of c. 90 ka. When the Log MSA age of c. 90 ka was substituted for the ESR MSA age of 
c. 127 ka, the fit to data improved to R2= 0.997, with an even lower Log MSA age of c. 
82 ka, and a Log basal deposit age of c. 253 ka (Table 17, (b)). The best Log fit to data 
was obtained by regressing the ESR MSA age to 78 ka (R2= 0.999). with the basal 
deposits remaining at 250 ka (Figure 33; Table 17, (c)). As the ESR MSA age was 
regressed, so the R2 value of Lin decreased away from the original R2= 0.991, to R2= 
0.919, while at the same time the Log MSA age increased towards R2= 1.0. These results 
tend to contradict the supposition of equal compaction times for the upper and basal 
sediments, by confirming a higher degree of compaction and compression in the basal 
sediments. To further test the statements on the age of the deposits being related to 
greater compaction with increased depth, Exp, which would indicate an increasingly 
greater degree of compaction at depth, was applied to the above data (Figure 33). As was 
expected, because of the aforementioned close fit of data of Lin, Exp did not fit the data 
either, with a best fit to data of R2= 0.790 (Figure 33; Table 17, (a)). Exp also suggested 
that the Unit F, horizon was considerably younger at c. 60 ka, with the basal deposits 
 177
being older at c. 465 ka. This Exp MSA age is somewhat older than the previously 
mentioned >43 to 47 ka 14C ages. When the Log MSA age of c. 90 ka was substituted for 
the ESR MSA age of c. 127 ka, Exp provided a much improved R2 value of 0.803 with an 
MSA age of c. 55 ka and a basal deposit age of c. 409 ka (Table 17 (b)). The basal 
deposit age is in partial agreement with the suggestion that the Florisian-Cornelian faunal 
boundary could be as old as c. 400 ka (Klein, 1984). There were only small changes in 
Exp ages when the MSA age of 78 ka was substituted for the Log 127 ka (Figure 34; 
Table 17, (c)). The R2 value improved slightly to 0.807 with an Exp MSA age of c. 52 ka 
and a Exp basal deposit age of c. 386 ka. 
 
In addition to the above, if it were assumed that the ESR MSA age of c 127 ka was 
correct, and held as a constant, the application of Log extends the basal deposits to c 420 
ka (R2= 0.999) (Figure 35; Table 17, (d)). The application of an Exp to the ESR MSA age 
of c 127 resulted in the age of the basal deposits being extended to c 690 ka (R2= 0.843) 
(Table 17, (d)). It was found to be pointless applying further basal deposit age variables 
to the exponential trend line, as MSA and basal deposit ages advanced rapidly to the 
point of being meaningless (Table 17, (e)), while the logarithmic trend line had already 
provided a best fit to data. By substituting the deepest average age of 200 ka from (a) for 
the basal deposit age, all the MSA ages decreased from 127 ka (Lin 120 ka, Log 82 ka, 
Exp 55 ka), with an increase in the R2 values (Lin R2= 0.994, Log R2= 0.885, Exp R2= 
0.7617). When using the oldest age of 281 ka from (c) for the basal deposit age, all MSA 
ages increased relative to (a) (Lin 150 ka, Log 100 ka, Exp 63 ka), with a decrease in the 
Lin R2 value (R2= 0.979), and an increase in the Log and Exp R2 values (Log R2= 0.976, 
Exp R2= 0.803). These increased R2 are an indication that both the MSA and basal 
deposit ages should be higher than either the Log or Exp ages in Table 17, (g), with a 
further indication that Figure 35; Table 17, (d), is most probably the more correct. A 
trend line takes into account all data, and smoothes this out over the entire data range. 
Trend lines do not take into account variables such as time frames, aeolian deposition, 
aquatic growth, decomposition and compaction, as well as wet and dry periods. This 
would imply that in Figures 33 and 34, compaction might have tracked neither Log nor 
Exp exactly. It is therefore proposed that compaction occurred somewhere between Log 
 
 178
 
 
0 Recent deposits
Dating at this level 15(a)
2
Exponential trend line R2= 0.7895
Logarithmic trend line R2= 0.9543
4 MSA human occupation horizon15(a)
Suggested compaction trend line
6 Linear trend line R
2= 0.9914
Lower deposits15(a)
8
0 50 100 150 200 250 300 350 400 450 500
Years (kyr)
 
 
Figure 33. A plot of linear, logarithmic, and exponential trend lines to recent Florisbad ESR and OSL ages from Test Pit 3 over a depth       
of 8 metres. The best fit to data of a trend line is represented by R2= 1 0. 15(a) = Grün et al. (1996) (a) represents the illustration 
identifier where data has been tabulated in Table 16. 
 
 179
D e p t h  ( m )
0 Recent deposits
Dating at this level 15(a)
2 Suggested compaction trend line
Logarithmic trend line R2= 0.9991
Regressed MSA human occupation horizon age
4
Linear trend line R2= 0.9210
Exponential trend line R2= 0.8063
6
Lower deposits 15(a)
8
0 50 100 150 200 250 300 350 400 450 500
Years (kyr)
 
 
Figure 34.  A plot showing the effect on linear, logarithmic, and exponential trend lines by regressing the MSA age to 78 ka 
and holding the basal sediment age at 250 ka. The best fit to data of a trend line is represented by R2= 1 0. 15(a) 
= Grün et al. (1996) (a) represents the illustration identifier where data has been tabulated in Table 16.
 180
D e p t h  ( m )
 
0 Recent deposits
Dating at this level15(a)
2 Exponential trend line R2= 0.8425
Logarithmic trend line R2= 0.9999
4 MSA human occupation horizon 15(a)
Linear trend line R2= 0.9159 
6 Advanced lower deposit age
8
0 50 100 150 200 250 300 350 400 450 500
Years (kyr)
 
 
Figure 35.  A plot showing the effect on linear, logarithmic, and exponential trend lines by extending the age of the basal 
sediments to 420 ka and holding the MSA age at 127 ka. The best fit to data of a trend line is represented by R2= 1 
0. 15(a) = Grün et al. (1996) (a) represents the illustration identifier where data has been tabulated in Table 16. 
 
 181
D e p t h  ( m )
 
Table 17.  A summary of the various ages and depths discussed in text and the affect of applying linear, logarithmic and
exponential trend lines to this data 
                        
   Linear trend line Logarithmic trend line Exponential trend line 
Text Ref, MSA age B. Dep.  MSA age B. Dep age R2 MSA age B. Dep age R2 MSA age B. Dep age R2 
 x plot (ka) x plot (ka) (ka) (ka)  (ka) (ka)  (ka) (ka)  
            
                        
a (Fig. 33) 127 250 137 242 0.9914 93 260 0.9543 60 465 0.7895
            
B 90 250 128 229 0.9446 82 253 0.9967 55 409 0.8029
            
c (Fig. 34) 78 250 125 222 0.9187 80 250 0.9991 52 386 0.8065
            
D (Fig. 35) 127 420 205 375 0.9159 127 420 0.9999 75 690 0.8425
            
E 127 1825 770 1460 0.7547 430 1740 0.9500 127 2170 0.9240
            
F 127 200 120 200 0.9939 82 200 0.8850 55 200 0.7617
            
G 127 281 150 281 0.9793 100 281 0.9756 63 282 0.8028
                        
          
B. Dep. Age = Basal deposit age          
 
 182
and Exp, more or less following the suggested compaction trend lines. It is proposed that 
compaction increases from the surface downwards, initially following Exp, but tending 
towards Log near the middle of the sequence, although at the same time mirroring the 
Exp gradient. With increasing depth, the suggested compaction trend line then tends back 
towards Exp, but then mirrors the Log trend line gradient, taking a path between the two. 
This would suggest that, by maintaining the age of the basal deposits at 250 ka, the age of 
the Unit F, horizon could then fall between the Log MSA age (c. 84 ka) and the Exp age 
(c. 56 ka), at 70 ± 12 ka (Table 17). The crossover point between Log and Exp, which 
remains fairly constant at about 2 m above basement in Figures 33, 34, and 35, is 
significant. From current investigations, it is contended that from the genesis of the spring 
to c. 200 ka (Figure 35), the site existed as a large drainage impeded spring pan 
(Appendix IV; Kuman et al., 1999). From c. 200 ka to the present, it is further contended 
that there was a considerable increase in the rate of aeolian deposition and a 
corresponding increase in the rate of compaction in the basal sediments due to an 
increased mass of aeolian sand moving across the site in the form of a sand dune 
(Appendix IV). 
 
6.4 DISCUSSION 
 
The ESR age of a sample normally lies somewhere between the estimates of two 
hypothetical models based on the rate and manner of the early, and linear (continuous) 
uranium uptake by an object (Grün and Thorne, 1997). These models are however only 
relevant on U absorbing material such as teeth and bone. There are inherent problems 
with particularly the ESR dating method used, and the results may reflect a lack of 
information regarding aspects such as the formation, history and chemistry of a site. 
Chemical reactions, for example, may influence the distribution of radioactive elements 
and subsequently the ages obtained by ESR and thermoluminescence dating (Mercier et 
al., 1995), with the degree of chemical reaction varying over the site, and even within 
samples. ESR and uranium-series dating also depend upon knowledge of the manner of 
uranium uptake into the bone (Millard and Hedges, 1999) and, more specifically, that the 
sample must be from a closed system with uranium being present at the time of 
deposition (Trueman and Tuross, 2002). A number of factors may have influenced the 
Florisbad uranium-dependent dating methods as well as indicating historically low 
 183
radioactivity levels. These include very low scintillation counts of between 9 and 18 
counts per minute for the area in general, including calcrete exposures, pit-waters and 
spring-water, as well as extremely low levels of stable and radioactive isotopes 
(Appendix I; Mazor and Verhagen, 1983). Uranium levels for tooth enamel (3–101 ppb) 
and dentine (10–268 ppb) in the ESR analysis were also low (Grün et al., 1996). 
 
Libby (1954) provided an age of >41 ka for the Peat I layer at Florisbad but apparently 
this greater-than-age was not based on the limitations of the methods used. According to 
Oakley (1955), the Florisbad deposits were accumulated by spring waters possibly arising 
from an underlying Palaeozoic formation, which included Ecca coal measures.  The peat 
unquestionably contains Pleistocene plant material with the possibility that it also 
contains Palaeozoic carbon (Oakley, 1955). Should this `dead` carbon have a high 
percentage, the point of no measurable radio-activity would be reached within 41 ka 
(Oakley, 1955). This was interpreted by Oakley (1955) that Libby (1954) found no radio-
activity in the Peat I layer. 
 
The accumulation of uranium in enamel and dentine, over time, can result in uncertainties 
in estimating age (Grün et al., 1990a), and it is evident that age uncertainty in ESR results 
is strongly dependent upon the circumstances of burial (Rink, 2001). Relatively high 
uranium concentrations in both enamel and dentine may result in considerable differences 
between linear and early uptake results, making it difficult to estimate ages (Grün et al., 
1990b) due to the lack of differentiation between the two uptake modes. For example, 
uranium can show steep concentration gradients in bone and teeth with uranium gradients 
decreasing towards the cortex, as can be illustrated by the difference in uranium 
concentrations between enamel and dentine (Williams, 1988; Grün et al., 1990a; Grün et 
al., 1990b; Janssens et al., 1999; Rink, 2001). The low U and Th content at Florisbad, and 
the use of the saturated content for the duration of the burial period by Grün et al. (1996), 
may have contributed to a greater source of scatter in their data. It has been proposed that 
the U content at Florisbad has always been historically low, and therefore there would 
have been little variability over time, and this would have been reflected in all the 
sediments through out the sites depositional history. 
 184
Results may be further influenced by the degree of bone preservation, with bones in 
sandy soils usually remaining more porous and attaining a more rapid equilibrium with 
the environment in which they were buried, but with a shallower concentration gradient 
(Millard and Hedges, 1995; Hedges and Millard, 1995). For example, no firm ages could 
be established for ESR dating on tooth enamel from Kromdraai B, due to high uranium 
concentrations in the enamel and dentine, neither could the reliability of determining 
isochron age estimates (Cunroe et al., 2002). Good bone preservation also usually results 
in low uranium levels while poor preservation usually leads to high uranium levels 
(Janssens et al., 1999). Bone preservation will of course largely be dependent on the 
hydrological environment and history of the site, as well as the location of bones within 
the sediments. While much attention has been given to the Florisbad fossils, no detailed 
studies or information on the historical chemical and physical processes of fossilization 
have been published. 
 
Post-depositional sample enrichment by uranium may also affect the magnitude of ESR 
signals. Uranium concentrations for the Florisbad tooth enamel and dentine can be 
considered low at 3–101 ppb and 10-268 ppb, respectively (Grün et al., 1996). Post-
depositional sample enrichment by uranium was not regarded as a problem at Border 
Cave because of what were considered negligible amounts of U in the teeth (Grün et al., 
1990b) (enamel <10–460 ppb and dentine <10–440 ppb), but which were still higher than 
the Florisbad values. These low uranium values for the Border Cave teeth indicated that 
there had been no mobilization of uranium to influence the ESR result, this probably 
being due to the cave floor deposits having been in a dry state throughout their history 
(Grün et al., 1990b). 
 
Environmental conditions may have implications at Florisbad because, despite long dry 
periods, the site was also inundated with water for long periods due to the possible 
uninterrupted spring flow from an apparently large aquifer, about which little is known 
(Appendix I). This could have had a major effect and influence on the mobilization of 
any uranium present. Furthermore, uranium may well have been flushed at a considerably 
higher rate from bones that were in close proximity to the spring vents (Appendix III). 
Variations in uranium concentrations over the site were possibly further influenced by the 
 185
Eh of the peat deposits, which tend to bind uranium, producing low solution 
concentrations (Janssens et al., 1999), and good bone preservation (Millard and Hedges, 
1995). Despite the many ideal factors involved at Border Cave, such as low uranium 
concentrations and dry conditions, it was still concluded that ESR age assessments of 
certain significant occurrences, such as the basal part of the SAS member, and the 
Howiesons Poort artefact occurrence, could not yet be resolved (Grün et al., 1990b). In 
contrast to this, at Klasies River Mouth it was noted that the high uranium values might 
have influenced ESR results, making it difficult to estimate age (Grün et al., 1990b). This 
is of interest because the Florisbad enamel and dentine uranium values are the lowest of 
the three sites, suggesting that the Florisbad ages should be the more accurate. 
Notwithstanding apparent inconsistencies and other possible problems associated with the 
dating and methods employed at Florisbad, the authors still had confidence in the 
reliability of the methods (Grün et al., 1996). 
 
In ref. 2 (c) OSR results give the Unit F horizon as 133 ±31 ka and the sands above Peat 
II as ascending in age by a probable 55 ka, from 106 ka to 161 ka. This is questioned 
when considering that spring activity probably never occurred in these upper levels, as is 
indicated by the unbroken and uniform stratigraphy of the sediments in the MSA horizon 
(Kuman and Clarke, 1986; Kuman et al., 1999; Henderson, 2001a). Another issue is the 
100–350 ka age spread for hippopotamus (Hippopotamus amphibious) and wildebeest 
(Connochaetes gnou) teeth from the spring material (Grün et al., 1996; Brink, 1997), 
which makes the Unit F horizon hippopotamus remains older than the youngest of the 
spring collection specimens. However, this spread could well be attributable to spring 
action. Contrary to the confidence expressed in the reliability of the methods, it was noted 
that ESR and OSL ages from the third test pit only confirmed the general age span of the 
site, and were considered too imprecise to provide any further insight (Brink, 1997). 
Reasons given for some of these anomalies were large errors in the OSL results due to 
saturation problems, and that the large age spread of ESR ages, in both the third test pit 
and spring material were due to re-working of the sediments by spring action (Grün et al., 
1996). Columns of sand resulting from the spring vents are distinguished by their 
whiteness (Dreyer, 1938; Brink, 1987), because they have been flushed clean of clay and 
organic material (Appendix III). There appears to be no evidence of such columns having 
 186
been recorded in the third test pit, and therefore no real evidence of spring action. This 
would imply that if any disturbance has occurred within the sediments, this must have 
had other origins. It is suggested that these large ESR and OSL age spreads, of up to 125 
ka at the MSA level levels, are more likely to be as a result of the internal stratification of 
the sand dune and debris it accumulated as it migrated across the site (Appendix IV), and 
not spring action. 
 
The location and sampling of the third test pit may also have been partially responsible 
for the perception of skewed data and erroneous dating. For example, it is possible that 
test pit 3 was not located over an area where early remains had been deposited. Had there 
been remains of earlier specimens in the area of the test pit, then the previously suggested 
spring activity (Brink 1987; Grün et al., 1996) should have made it even more likely that 
earlier remains would have been sampled. Furthermore, the size of the spatial sampling 
area of the test pit, in both horizontal and vertical dimensions, may have been too limited, 
and presented too small a window for the sampling of early remains. 
 
The application and comparison of other age-depth models was well beyond the scope of 
this study, and this study should be seen as an alternative method. There are a number of 
age-depth models based on linear regression, splines, and linear interpolation (Bennett, 
1994), fuzzy regression (Boreux et al. 1997), Bézier curves (Bennett and Fuller, 2002), 
mixed-effect models (Heegaard, 2003), polynomial regression or cubic splines, and 
monotonic smoothing splines (Enters et al., 2007), which can be used. These can be 
integrated into programs such as INTCAL98 and INTCAL04 (Reimer et al. 2004), and 
the Bayesian model, OXCAL 4.0 (Ramsey, 1995; Ramsey, 2001). Age–depth models are 
only meaningful when using calibrated radiocarbon dates where different approaches 
may produce very different results for an age at a particular depth, a difference in their 
respective sedimentation rates, as well as their uncertainties attached to the estimates 
(Bennett, 1994; Bennett and Fuller, 2002; Telford et al., 2007; Enters et al., 2007). It 
should be noted that different calibration curves are separately applicable to the northern 
and southern hemispheres due to differences in the reservoirs of 14C in the two 
hemispheres (McCormac et al., 2004). Cal ka BP dating of from 0-12.4 cal ka BP are 
based primarily on dendrochronological dating by 14C tree ring data measurements, as in 
 187
the SHCAL04 calibration set of McCormac et al. (2004) (0-11 cal ka BP) for the 
southern hemisphere. Marine dating, which covers the calendar time span of 10.5- 26 cal 
ka BP, is derived is primarily on data derived from U/Th dated corals and foraminifera as 
in the MARINE04 calibration curves (Hughen et al., 2004). 
 
BP is the abbreviation for “Before Present” and represents uncalibrated data, while cal 
BP represents data calibrated from calibration curves onto a calendar time scale 
expressed as BC/AD. “Present” is taken as 1950, being the year in which calibration 
curves for radiocarbon dating were established, and also precedes large scale testing of 
nuclear weapons which altered the global ratio of 14C to 12C. The BP scale may typically 
have uncertainties high enough that the difference between 1950 and the actual present 
year may be insignificant (Anon [7] 2008). From the literature it would appear that there 
are two distinct schools of though on the use of calibrated (cal BP, AD/BC) and 
uncalibrated (BP) 14C data. One school of thought can be typified by the archaeological 
and geological community who would like to know the age of an item, or sediment, is by 
using uncalibrated 14C radiocarbon years (BP) data. The other school of thought is 
typified by the theoretisists who find it more logical and convenient to work with 
calibrated dates, as on a calendar time scale, or calendar age range (cal BP, AD/BC). An 
argument for the use of calibrated ages is that BP gets all sorts of uses in different 
contexts, and that age is also a variable quantity (Ramsey, C.B. pers com. 24 February 
2009). 
 
Even although theoretisists emphasise the use of calibrated 14C radiocarbon dates in age-
depth models, they also concede that calibrated dates are probabilistic and add an extra 
level of complexity (Bennett, 1994; Sewell, 1998), as the resulting probability 
distributions are not Gaussian (Telford et al, 2004). By including the stratigraphic 
relationship of the dated samples, information from short-lived isotopes, and other factors 
the uncertainty of using calibrated dates can be considerably reduced (Enters et al., 
2007). 
 
The question remains as to whether the application of calibrated correction curves to the 
Florisbad BP ages is warranted, and what might be achieved from such an exercise. 
 188
Because cal ka BP dates only provide good accuracy up to 26 cal ka BP (Reimer, et al. 
2004), this would only accommodate approximately the top two metres of the Florisbad 
deposits down to roughly the Peat III horizon. This would imply that calibrated age-depth 
models would possibly also only be accurate to this level, but this was not investigated in 
detail. To date, much of the interesting material that has been excavated at Florisbad lies 
below these dates, and at this time, there does not appear to be sufficient justification for 
testing these complex exercises, which would appear to be incapable of producing any 
meaningful results. 
 
Sedimentation is another important factor in age-depth determinations. Telford et al. 
(2007) noted that the number of dates needed to construct an age-depth model for a 
sedimentary sequence will depend on the required precision and complexity of the 
sedimentation rate, and that in order to achieve accurate, high-precision, many more dates 
may have to be used than is currently the norm.  
 
6.5 SYNOPSIS 
 
It is apparent from the results in Figures 33 and 34 that the ESR and OSL ages of the 
Florisbad sediments, as reported in ref. 2 (a), do not conform to a process of increased 
compaction and compression with depth, as has been shown by the application of Lin, 
Log and Exp. As indicated by Lin, and considering the many variables involved, it is 
highly improbable that the upper and basal sediments accumulated and became equally 
compacted and compressed within almost identical time frames, at different depths. This 
leads to the conclusion that either one, or all three of the ages drawn from ref. 2 (a), and 
used in Figure 33, may be in error. Errors in the dating may have been further aggravated 
by factors such as inherent problems with the dating methods, and a lack of information 
regarding the formation, history and chemistry of the site. As regards chemistry, there 
could be grounds for the possible low, or even non-existent, occurrence of uranium for 
early uptake, with the effects of spring flow and the peat deposits influencing results 
negatively. The degree and type of bone preservation along with environmental factors, 
and about which little detail appears to be known, may also have had a considerable 
influence on results. These aspects emphasise the importance and necessity of having a 
more holistic knowledge and understanding of the many aspects associated with a 
 189
particular site, and that by simply applying a dating method to a sample, may not be 
sufficient for accurate dating. 
 
If the original ages drawn from ref. 2 (a), and used in Figure 33, are taken to be incorrect 
in relation to their depth, it must then also be presumed that although Log and Exp reflect 
a more accurate version of compaction, they may also be slightly in error due to the many 
previously mentioned variables. According to the above arguments, if the Unit F, horizon 
ESR age of c. 127 ka is found acceptable, then according to Log, the Unit F age would 
reduce to c 90 ka, with the basal deposits increasing slightly to c. 260 ka (Figure 33; 
Table 17, (a)). Exp indicates a greater increase in the age of the basal deposits to c. 465 ka 
with a conversely greater decrease in the Unit F age to c 60 ka (Figure 33; Table 17, (a)). 
Alternatively, if the basal deposits age of 250 ka is found acceptable, then the Unit F 
horizon must be much younger at c. 79 ka, as indicated by Log, or even c. 52 ka as 
indicated by Exp (Figure 34, Table 17, (c). However, both suggested compaction trend 
lines indicate a Unit F age of c. 95 ka and a basal deposit age of c. 380 ka. What Figures 
33 and 34 also show, is that compaction could not have occurred at equal rates in both the 
upper and basal sediments. 
 
Log in Figure 35 reflects the best R2 fit to data, with the Unit F age at 127 ka, and the 
basal deposit age advanced to 420 ka, indicating that the ESR 250 ka basal deposit age 
used in Figures 33 and 34 may not be a true reflection of the age of the basal deposits at 
this depth. The 420 ka age also has better agreement with the oldest possible age recorded 
for the site, namely c. 354 ka, and the minimum suggested compaction trend line ages (c. 
380 ka) in Figures 33 and 34. Additionally, the 420 ka age also corresponds closely to the 
proposed Florisian–Cornelian boundary at c. 400 ka (Klein, 1984). The ages suggested by 
the suggested compaction trend line in Figures 33 and 34 of 92 ±12 ka for the Unit F 
horizon and 400 ±20 ka for the basal deposits, seem to best reflects the ages of the 
sediments at Florisbad, but more work is needed to confirm this. 
 
In light of the above, and regardless of the Unit F horizon having the proposed smallest 
error value (Grün et al., 1996), this small error value should not necessarily be seen as 
reflecting the accuracy of the dating. Given the recent dating of the Florisbad hominid 
 190
cranium at c. 259 ka (Grün et al., 1996), the arguments advanced in this paper regarding 
the chemistry of the site and possible flaws in the dating methods may help to explain 
how such a very early age was claimed for a specimen with such an advanced 
morphology (Brink, 1997). 
 
It is concluded that mathematical trend lines could be used to evaluate and validate ESR 
and OSL ages at Florisbad, with the trend lines confirming that the ages in ref 2 (a) must 
be suspect, and in error, when related to depth. On the other hand, where it was indicated 
that ages and dating might be more correct, for example, where the oldest recorded age 
was placed at the greatest depth, the trend lines closely confirmed these ages and depths. 
It is also evident that the Unit F and basal deposit trend line ages reflect the general trend 
of the ESR and OSL ages, but with the trend lines strongly suggesting that the age of Unit 
F could possibly be younger, while the basal deposit age could possibly be older. From 
this it can be stated that the trend lines confirmed two opposing scenarios at Florisbad and 
can therefore be seen as a tool in evaluating ESR and OSL ages under these 
circumstances. 
 
 191
 
 
 
 
 
 
 
 
 
 
 
 
Chapter 
 
7 
 
 
 
 
 
 
 
Discussion 
and 
Conclusions 
 
 
 
 192
CHAPTER 7 
 
DISCUSSION AND CONCLUSIONS 
 
 
The initiation of this research was as a result of spring and excavation pit water 
samples taken during a high rainfall period in 1988. These results recorded a 
significant difference in water quality between the spring- and excavation pit water. In 
1999, during an average rainfall period, it was decided to confirm the previous results 
by re-sampling the spring and excavation pit water, and to expand the water sampling 
to other excavation pits in the area of the springs. The spring water had previous been 
analysed by Fourie (1970), so the results would not only expand the spring-water 
database, but also establish a database for the exploration pit waters. In addition to 
this, comparative results would be available for both an exceptionally high rainfall 
period as well as an average rainfall period. 
 
The TDS of the original exploration pit water showed a 58% decrease between the 
1988 high rainfall period and the 1999 average rainfall period. This was contrary to 
the norm where high rainfall would be expected to create a dilution effect on the 
excavation pit water. This incongruity provided motivation to examine other aspects 
of the sedimentation in order to resolve the issue. The increase and variability of the 
composition of the pit waters, and the stability of the spring-water, is evident in the 
Stiff diagrams (Figure 25). This evidence then also provided a strong argument for 
two separate hydrological entities, namely, the spring-water and groundwater 
(exploration pit water). This was the first time that the spring-water had been analysed 
in such detail, and the first time that the groundwater had been examined. 
 
Further to the chemical evidence, the quality of the two hydrological environments 
was also examined in relation to rainfall. It was established that neither long- nor 
short-term, rainfall had much affect on the quality of the spring-water due to the 
presumed large size, and distant location, of the spring aquifer. It is proposed that the 
spring aquifer effectively stabilizes any fluctuations in the spring-water quality. 
However, short-term rainfall had a decided affect on the quality of the groundwater by 
 193
increasing TDS during high rainfall periods and reducing TDS during average to dry 
periods. 
 
Of prime importance was the establishment of causes and processes for the higher 
salinization of the groundwater. Peat II samples were taken for analysis from the wall 
of the excavations and the excavation pits, while Peat IV samples were taken from the 
west wall of the excavations. Analysis of the organic-clay layers showed even more 
elevated TDS levels than the groundwater, providing evidence of an even higher 
degree of salinization in these layers. Causes and processes that would account for 
this higher salinization of the organic-clay layers were then sought. Primarily, the 
adsorption of ions occurs due to the presence of both clay and humus, the latter being 
derived from the decomposition of halophytes, phytoplankton, periphyton, the waste 
from the large herds of animals, which previously visited the site, and the 
decomposition of carcases. 
 
Clay and humus have a similar structure where highly charged anions attract, adsorb, 
large numbers of cations. It was concluded that the organic-clay layers at Florisbad 
acted as storage zones, with ions being adsorbed and concentrated during periods of 
average and low rainfall, and then being flushed from the organic-clay layers by 
percolating rain and groundwater during periods of high rainfall. Other factors 
contributing to the salinization of the organic layers included salts from aeolian 
deposition, particularly from Soutpan; the accumulation of salts through impeded 
drainage and a closed pan environment; capillarity; illuviation; the lack of leaching in 
a semi-arid environment; and a degree of adsorption through the miscible 
displacement with the spring-water. The organic-clay layers therefore had the ability 
to adsorb, accumulate, and retain ions at appropriate pH and Eh levels which were 
within the required dissolved organic carbon content and PCO2 conditions. These 
processes further occurred in the presence of Ca and P with a restricted water 
circulation. Salinization is a complex process with the above mentioned factors being 
found to be interrelated and often dependent on one another in creating an ideal 
equilibrium for fossilization to take place. 
 
This was the first time that the organic-clay layers at Florisbad or any other southern 
African archaeology site had been analysed from a salinization aspect, and as a 
 194
potential source of mineralization for the fossilization of faunal remains. The quality 
of the hydrological and sedimental environments in relation to one another, and 
particularly the salinization of clay and organic matter, could be important in 
understanding fossilization process when applied to investigations at other fossil sites 
where clay and/or organic matter occur. It is strongly suggested that, based on these 
results, the basal peat layer and Peat I where, at one time, appreciably more highly 
mineralised than even the Peat II layer, allowing for the fossilization of the Old 
Collection. 
 
It appears to have been generally accepted that the salinized spring-water was 
responsible for the fossilization of faunal remains at Florisbad (Brink, 1987, 1988), 
with much unsubstantiated evidence having been provided in support of this. It was 
put forward that the spring-water was carbonate rich and, with evidence of calcium 
carbonate deposition, the bones must have been in contact with the spring water for 
some time (Brink, 1987). Brink (1988) also noted that such prolonged contact with the 
spring-water had caused the remains of the Old Collection to be well preserved in a 
characteristic way. Post-depositional mechanical and chemical weathering was cited 
as further evidence that faunal remains had become fossilized in areas of spring 
activity. Butzer (1988) on the other hand stated that the spring-water was under-
saturated in both calcium and bicarbonate and that any carbonate enrichment must be 
derived from either direct, or indirect external sources, other than the spring-water. 
This further supported the salinization of the organic-clay layer hypothesis. 
 
An investigation of previous water analysis (Rindl, 1915; Fourie, 1970; Mazor and 
Verhagen, 1983; Douglas, 1992) indicated that there was insufficient Ca, CaCO3 and 
P in the spring-water for the initiation of fossilization. This substantiated Butzer’s 
(1987) statement of under-saturation, and was further verified by the analyses and 
results of this study. It was concluded that under saturation would have been 
applicable even in an historic sense because, historically, the spring vents, whether 
active or inactive, contained no clay or organic material to adsorb, accumulate, or 
retain the ions necessary for fossilization. To all intents and purposes, in the context 
of salinization and fossilization, the spring vents were found to be sterile. 
Furthermore, the high Eh environment of the spring flow would have inhibited 
precipitation and rapidly oxidized any organic material. 
 195
 
Contrary to the statement that post-depositional mechanical and chemical weathering 
clearly showed that the remains had become fossilized in a spring context, analysis of 
data indicated that post-depositional mechanical and chemical weathering, as well as 
the low incidence of calcite on, and in, bones (chemical weathering), reflected that 
bones from the spring vents have been in a state of demineralisation. At a high pH and 
low Eh, calcite and hydroxyapatite would be preserved, while at a low pH and high 
Eh, they would be dissolved and mobilized. The former conditions are also necessary 
for the production of DOC, which in turn acts as a food source for carbonate 
precipitating micro-organisms. However the level of DOC may be critical as over 
saturation may act as a kinetic growth inhibitor on calcite crystals, with a fine balance 
having to be reached between the production, and consumption of DOC by CaCO3 
precipitating micro-organisms and plant decomposition in order to prevent DOC 
acting as a CaCO3 inhibitor. It is suggested that these conditions could not have 
existed in either the active or non-active spring vents due to a lack of organic material. 
Although the spring water may have been conducive to the preservation of calcite and 
hydroxyapatite at pH 8.0, it would not have been conducive for the dissolution and 
mobilization of calcite and hydroxyapatite in order for initialise fossilization. 
 
In summary, the spring vents and their immediate environment are devoid of all 
requirements and processes for the initiation of fossilization. Brink (1987) did 
however note that if the presence of calcite was an indication of contact with the 
spring-water in the original sedimentary environment, then the relatively low 
incidence of calcite concretions might contradict the assumption that the Old 
Collection was entirely derived from spring deposits. It was noted that only 48.2% of 
Bovidae and hippopotamus bones showed signs of calcium carbonate deposition 
(Brink, 1987). 
 
Rubidge and Brink (1985) considered the lithostratigraphy and depositional history of 
the Florisbad sediments as still being in their initial stages of investigation, and felt 
that several models could be proposed. In their opinion, because the deposits were 
lithologically variable, they were seen as being the product of an unusual depositional 
environment. Kuman and Clarke (1986) also noted that the sand, silts, and organic 
deposits were poorly understood and controversial. The above questions and 
 196
statements therefore opened the way for further interpretation on the formation of the 
site. 
 
Based on contemporary theories on the formation of the Florisbad archaeological site, 
as well as the results of this study, it was concluded that aeolian deposition and 
sedimentation alone could not account for the formation and the current morphology 
of the immediate spring site and that erosion played a major role. It was further 
concluded that previous theories on the formation of the spring site could neither 
explain adequately the stratigraphy of the site, nor accommodate the fossilization 
process within the current stratigraphy. The three main theories put forward for the 
formation of the spring site were that: 
• Deposition was almost entirely derived from aeolian sources (Appendix I; 
Appendix II; Appendix III; Grobler and Loock, 1988a, 1988b; Van Zinderen 
Bakker, 1989; Kuman et al., 1999). 
• Development and formation of the spring site was from sand output from the 
springs, which most probably originated from the underlying bedrock 
(Dreyer, 1938a; Brink, 1987; Butzer, 1988; Van Zinderen Bakker, 1989). 
• That deposition was fluvial, being directly related to four low water level 
phases and three high level water phases (flooding) in the adjacent palaeolake 
(Soutpan) (Visser and Joubert, 1991; Joubert and Visser, 1991; Henderson 
200). 
 
Based on data from the water analysis and salinization processes, as well as an 
intensive study of the morphology of the region and the site, along with literature 
reviews, it was concluded that a new hypothesis was required in order to incorporate 
new evidence into the formation of the spring site. This hypothesis would have to 
account for, and accommodate, not only the unexplained morphological anomalies 
mentioned below, but at the same time be capable of accounting for the fossilization 
of faunal remains at least up to the level of the MSA Human Occupation Horizon. As 
part of this formulation considerable attention was paid to the western Free State 
panveld and dune development, with a new hypothesis on the development of the 
Florisbad sand-dune being formulated. These unexplained morphological features 
included:  
 197
 
• The lack of upper red-brown sand units on the eastern side of the “mound” 
appeared to have either not been deposited, or had been eroded by spring 
discharge (Rubidge and Brink, 1985; Brink, 1987). 
• The seven metres thick layer of clay deposits in the modified vlei area (Butzer, 
1984), which could not have been deposited by the pre-existing weak 
ephemeral drainage line. 
• The older aeolian deposits, evident over the rest of the site, were absent from 
the vlei area, while more recent aeolian deposits were present Fourie (1970).  
• The extent of the dogleg in the ephemeral drainage line from the inter-dune 
valley, cutting across the natural drainage at almost right angles from NW to 
NNE, and back to NW. 
• The disproportionate width and depth of the vlei area in relation to the 
ephemeral drainage line. 
• The erosion of the east bank of the vlei area. 
• The pinching out of the Peat IV layer on the western wall of the excavation 
area. 
 
A hypothesis was formulated whereby a dam was formed at the spring site by a 
migrating sand dune coming to rest against a previously establish dune belt. This dam 
slowly filled with sediments until the water level reached the top of the eastern arm of 
the dune. The water then broke through the eastern arm of the dune creating a flash 
flood, with the contents of the dam evacuating to create a wide vlei area. It should be 
pointed out that since the publication of Appendix IV, this hypothesis has been 
considerably revised in the introductory part of the thesis due to further investigation, 
and should be taken as contemporary 
 
This hypothesis proposes that an area of fossilized faunal remains will extend, and 
correspond, to the final size of the spring pan, before the migrating sand dune finally 
covered it. Therefore, the probability of extensive areas under the sand dune 
containing concentrations of fossil faunal remains is relatively high. It is also 
proposed that fossils located on the pan floor, within the area of the dam, would have 
 198
been washed into the vlei, suggesting that fossil faunal remains may also occur along 
the bottom of the vlei. 
 
From the results of the vlei organic-clay analysis, it can also be assumed that any 
faunal fossil remains found in this environment will be in a relatively well preserved 
state. Because these fossils would have been disseminated over a large area, they 
would have been concentrated to a far lesser degree than those on the original pan 
floor. It should be stated that the Florisbad sand dune probably had little influence on 
the fossilization of faunal remains, with the exception of any minerals it may have 
contributed to the salinization process. However, the damming of the site, the erosion 
of the eastern arm of the dune, and subsequent flash flood, severely affected, and 
changed, the local morphology of the site and surrounding area. It is believed that this 
theory offers a viable explanation for the formation of the spring site in a manner 
which accommodates the current morphology, the morphological anomalies, as well 
as providing an environment conducive for the fossilization of faunal remains. 
Induce  
There are numerous archaeozoological sites in the western and south-western Free 
State which are related to springs, pans and alluvial river terraces. It has been shown 
here that the source water of some of these springs is similar to that of Florisbad, 
which would imply that they too have a very low Ca content. This would then also 
preclude faunal remains at these sites from being fossilized in a spring or river 
context, and possibly clay and/or organic matter is the source of mineralization. It is 
therefore contended that most, if not all, faunal fossil remains of the Florisian – 
Conelian time period in the Free State were fossilized in either spring pan, or marsh 
contexts. It is also contended that the vast quantities of faunal remains at Florisbad, as 
well as at other sites in the area, were largely as a result of cyclical epidemic periods 
of bovine and equine diseases such as foot and mouth disease, anthrax, and African 
horse sickness. 
 
Dating of the Florisbad sediments and deposits has been a contentious issue over the 
years, and sometimes for valid reasons. From the many reasons given in Chapter 6 for 
discrepancies that might occur in radioactive dating methods, the extremely low 
presence of radioactive isotopes, and the lack of information on the chemistry and 
history of the Florisbad site, are most probably the most conspicuous. Grün et al., 
 199
1996 also noted that all teeth in the spring collection had low uranium concentrations. 
It was noted that the ages of the faunal remains and sediments given by Grün et al. 
(1996), were incongruous, and did not follow a logical ascending chronological 
progression from the top of the sediments down to bedrock. For example, the oldest 
age recorded was located some 2 m above the bedrock, while some of the much 
younger MSA ages were recorded well below the 2 m level. Further more, age scatter, 
or age spread, of up to 125 ka occurred in a narrow horizon of less that a metre wide 
in the area of the MSA Human Occupation Horizon, at approximately the mid-way 
point in the profile. 
 
A plot of age against depth from the excavation pit of Grün et al. (1996) suggested 
that the narrow human occupation horizon was positioned at a mid point in the profile. 
On a linear scale the implication was uniform compaction of sediments throughout the 
profile, with only c. 62.5 ka being available for the basal sediments to accumulate, 
and another c. 62.5 ka for the upper sediments to accumulate. This plot left no room 
for the possibility of greater compression and compaction with depth, particularly 
regarding the decomposing deeper organic-clay layers. Considering these factors, an 
experimental exercise was initiated whereby mathematical trend lines were applied to 
the data presented by Grün et al. (1996). Another application of this exercise was that 
data could be manipulated and varied, as in keeping one age fixed and varying another 
age to obtain different ages. 
 
The incongruous ages of Grün et al. (1996) were more acceptable in the 
reconstruction of the lower sediments around the spring vents where age distribution 
could be attributed to the redistribution of material from spring activity at different 
levels. However, although it was stated that these age spreads in the exploration pit 
were also associated with a reworking of the sediments by spring action, it is 
questioned whether spring action within the dune itself ever ascended to the height of 
the MSA horizon. The unbroken horizontal stratigraphy of the exploration pit and the 
MSA excavations, do not suggest spring activity at these levels. That the excavation 
pit was opportunely located directly over a spring eye is highly improbable as there is 
no mention of any larger fossils other than tooth fragments. It is proposed that the 
incongruity in the ages of the small tooth fragments, and their position within the 
dune, may have been as a result of changes in the internal stratigraphy of the dune as 
 200
it moved across the site. Rubidge and Brink (1985) found that over most of the site 
that rapid facies changes and the lateral wedging out of layers, over very small 
distances within the dune, made lateral correlation extremely difficult. Rubidge and 
Brink (1985) stated that this unconformity was as a result of micro-environments 
created by a number of spring eyes giving rise to ponds, channels, and over bank 
deposits which moved laterally in time. This is in agreement with the hypothesis 
proposed in Chapter 5, section 5.6.5, but this situation would only have occurred at 
the lower levels, and not have extended to the higher levels, that is, with the exception 
of the dam itself.  
 
As a linear trend line disputed the ages of the excavation pit ages derived from Grün 
et al. (1996), deliberation was given to applying logarithmic and exponential trend 
lines to the data. This, it was anticipated, would reflect increasing compaction with 
depth. The best fit to data of R2=0.9999 was obtained by the application of a 
logarithmic trend line with the MSA age fixed at 127 ka, and the basal deposit age 
extended to 420 ka.. By using a suggested trend line, a basal deposit age of 400 ±20 
ka was obtained while the MSA age was reduced to 92 ±12 ka, which may not be an 
unreasonable estimate. Application of an exponential trend line produced no best fit to 
data values higher than R2=9240, where, with the MSA age fixed at 127 ka, the basal 
deposit age was far outside any practical limit at 2170 ka. 
 
The application of trend lines confirmed the age incongruities in the excavation pit. 
By taking the youngest ages in the MSA spread of Grün et al. (1996) for the 
exploration pit at c. 115 ka, and the maximum age for the basal deposits at 354 ka, 
these ages closely approximated the ages of the suggested trend line where the 
maximum MSA age was 104 ka and the minimum basal deposit age was 380 ka. The 
suggested trend line was developed in an attempt to smooth out some of the variances 
which may have occurred in the stratigraphy 
 
In summary, the chemical database for the spring-, groundwater and organic clay 
deposits were considerably expanded on, leading to a better understanding of the 
hydrological and sedimentary environments at Florisbad. This in turn resulted in a 
scientifically supported hypothesis being formulated that states that the fossilization 
of faunal remains occurred in conjunction with the organic-clay layers and 
 201
groundwater, and not in a spring context, as previously proposed. It was further 
concluded that the spring-water is responsible for the demineralization of faunal 
remains, as opposed to their fossilization. The hypothesis on the formation of the 
Florisbad site accommodated and supported the depositional history and stratigraphy 
of the site, as well as accommodating the fossilization of faunal remains. Further to 
this, the hypothesis on the formation of the spring site also addressed, and provided, 
explanations for a number of previously unexplained morphological aspects of the 
site. While a number of hypotheses have been put forward over the years in an 
attempt to explain the history of the development and formation of the Florisbad 
spring site, and its fossilization processes, it is hoped that this research will go some 
way in contributing to a more holistic and better understanding of the site. 
 202
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
References 
 
 
 
 203
 
REFERENCES 
 
 
Acocks, J.P.H. 1953. Veld Types of South Africa. (1st ed.) Memoirs of the Botanical 
Survey South Africa 28: 1-192. 
 
Acocks, J.P.H. 1975. Veldtypes of South Africa. (2rd ed.) Memoirs of the Botanical 
Survey South Africa, No. 40.  
 
Acocks, J.P.H. 1988. Veldtypes of South Africa. (3rd ed.) Memoirs of the Botanical 
Survey South Africa 57: 1-146. 
 
Allen, C., Albert, F. Chafetz, H., Combi, J. Grahan, C. Kieft, T. Kivett, S., McKay, 
D., Steele, A. Taunton, A., Thomas-Keptra, K., and Westall, F. 2000. The search 
for signs of ancient Martian microbes: physical bookmarkers in carbonate 
thermal springs. Lunar Planetary Institute, 31st Lunar and Planetary Science 
Conference, Houston, Texas. 
 
Anderson, J.M., Berger, L., De Wit, M., Fatti, P., Holm, E., Rubidge, B., Smith, 
G.,Thackeray, F., and Van Wyk, B. 1999. Towards Gondwana Alive: Promoting 
Biodiversity and Stemming the Sixth Extinction. Godwana Alive Society, 
Pretoria. 
 
Anon [1]. 1980. Florisbad wetenskaplike vindplek by uitnemendheid. National 
Museum News 19: 3-13. 
 
Anon [2]. 2001. Stratigraphy of South Africa (Chart). South African Committee for 
Stratigraphy. Council for Geosciences, Pretoria. 
 
Anon [3], 2006). “Deep impact – The Vredefort dome” Hartebeesthoek Radio 
Astronomy Observatory. 
(Access at: http://www.hartrao.ac.za/other/vredefort/vredefort.html) 
 204
Anon [4], 2007. World Heritage –Vredefort Dome. United Nations Educational, 
Scientific and Cultural Organization. (Access at http://whc.unesco.org/en 
/list/1162) 
 
Anon [5]. 1860. The Prince’s progress: “The hunt”. The Friend of the Free State and 
Bloemfontein Gazette XII, No 515 (August 31). 
 
Anon [6]. 1977. The greatest hunt in history took place in the Orange Free State. 
National Museum News 12: 7. 
 
Anon [7] 2008. Before Present. http://www.absoluteastronomy.com/topics/Before 
_Present  
 
Ariëns Kappers, C.U.A. 1935. Note on the Endocranial Cast of the Florisbad Skull. 
In: Dreyer T.F. A Human Skull from Florisbad, Orange Free State. Koninklijke 
Akademie van Wetenschappen te Amsterdam 38: 3-12. 
 
Avenant, 2005. N.L. Barn owl pellets: a useful tool for monitoring small mammal 
communities. Belgian  Journal of. Zoology 135: 39-43. 
 
Bamford, M.K. and Henderson Z.L. 2003. A reassessment of the wooden fragment 
from Florisbad. South African Journal of Archaeological Science 30: 637-650. 
 
Baran, E. 2003. An Explanation of the 1:500 000 General Hydrogeological Map 
Kroonstad 2725. Department of Water Affairs and Forestry, Pretoria, South 
Africa. 
 
Barendsen, G.W., Deevey, E.S. and Gralenski, L.J. 1957. Yale natural radiocarbon 
measurements III. Science 126: 908-919. 
 
Bates, M.F. 1988a. Life History Note. Mabuya variegata punctulata: Avian predation. 
Journal of the Herpetological Association of Africa. 34: 48. 
 
 205
Bates, M.F. 1988b. Life History Note. Mabuya capensis: Avian predation. Journal of 
the Herpetological Association of Africa 34: 48. 
 
Bates, M.F., DE Swardt, D.H. and Louw, S. 1992.  A note on the diet of the 
Yellowbilled egret. Ostrich 63: 44. 
 
Beaumont, P.B. and Vogel, J.C. 1972. On a new radiocarbon chronology for Africa 
south of the equator. Part 2. African Studies 31: 155-182. 
 
Beaumont, P.B., DE Villiers, H. and Vogel, J.C. 1978.  Modern man in sub-Saharan 
Africa prior to 49 000 years B. P.: a review and evaluation with particular 
reference to Border Cave. South African Journal of Science 74: 409-419. 
 
Bennett, K.D. 1994. Confidence levels for age estimates and deposition times in late-
Quaternary sediment sequences. The Holocene 4: 337-348. 
 
Bennett, K.D. and Fuller J.L. 2002. Determining the age of the mid-Holocene Tsuga 
canadensis (hemlock) decline, eastern North America. The Holcene 14:421-429. 
 
Blatt, H., Middeleton, G. and Murray, R. 1972. Origin of Sedimentary Rocks. Prentice 
Hall, New Jersey. 
 
Bohn, H., MC Neal, B. and O’Conor, G. 1985. Soil Chemistry. (2nd ed.). John Wiley 
& Sons, New York. 
 
Bond, G.W. 1946. A geochemical survey of the underwater supplies of the Union of 
South Africa. Memoirs of the Geological Survey of South Africa 41: 1-208. 
 
Boon, III, J.D. and McIntyre, W.G. 1968. The boron-salinity relationship in estuarine 
sediments of the Rappahannock River, Virginia. Chesapeake Science, 9: (1) 21-
26. 
 
 206
Boreux, J.J., Pesti, G., Duckstein, L. and Nicolas, J. 1997. Age model estimation in 
palaeoclimate research: fuzzy regression and radiocarbon uncertainties. 
Palaeogeography, Palaeoclimatology, Palaeoecology 128: 29-37. 
 
Bisschoff, A. and Mayer, J.J. (Compilers) 1999. The Geology of the Vredefort Dome 
(and Geological Sheets). Explanation of Sheets 2627CA, CB, CC, CD, DA, DC. 
2727AA, AB, BA. Scale 1:50,000. Council for Geoscience, Geological Survey of 
South Africa, Pretoria. 
 
Brady, N.C. 1984. The Nature and Properties of Soils. Macmillan Publishing 
Company, New York. 
 
Bredenkamp, D.B. 2000. Groundwater Monitoring: A Critical Evaluation of 
Groundwater Monitoring in Water Resource Evaluation and Management. 
Water Research Commission Report, WRC Project No. 838/1/00. 
 
Brink, J.S. 1987. The archaeozoology of Florisbad, Orange Free State. Memoirs van 
die Nasionale Museum, Bloemfontein 24: 1-151. 
 
Brink, J.S. 1988. The taphonomy and palaeoecology of the Florisbad spring fauna. 
Palaeoecology of Africa 19: 169-179. 
 
Brink J.S. 1994. An ass, Equus (Asinus) sp. From late Quaternary mammalian 
assemblages of Florisbad and Vlakkraal, central Southern Africa. South African 
Journal of Science 90: 497. 
 
Brink, J.S. 1997. Direct dating of the Florisbad fossil hominid. Culna :52: 7-9. 
 
Brink, J.S. and Lee-Thorpe J.A. 1992. The feeding niche of an extinct springbok 
Antidorcus bondi (Antelopini: Bovidae) and its palaeoenvironmental meaning. 
South African Journal of Science 88: 227-229. 
 
 207
Brink, J.S. and Rossouw, L. 2000. New trail excavations at the Cornellia- Uitzoek 
type locality. Navorsinge van die Nationale Museum, Bloemfontein 16: 141-156. 
 
Brink J.S. and Henderson Z.L. 2001. A High Resolution Last Interglacial MSA 
Horizon at Florisbad in the Context of Other Open-air Occurrences in the 
Central Interior of Southern Africa: An interim Statement. In: Conrad, N.J. (Ed) 
Settlement dynamics of the Middle Paleolithic and Middle Stone Age. Kerns 
Verlag, Tübingen, pp 1-20. 
 
Brodrick, A.L. 1960. Man and His Ancestry. Scientific Book Club, London. 
 
Broeker, W.S., Kulp, J.L. and Tuckek, C.S. 1956. Lamont natural radiocarbon 
measurements. Science 124: 154-165. 
 
Brooks, D.J. and DeWall, A.E. 1976. Boron concentration in Chesapeake Bay 
sediments, paleosalinity and baymouth uplift. Chesapeake Science, 17: (3) 221-
224. 
 
Broom, R. 1913. Man contemperoraneous with extinct animals in South Africa. 
Annals of the South African Museum 12: 13-16. 
 
Brouwer, A. 1967. General Paleontology. Oliver and Boyd, Edinburgh. 
 
Bullard, J.E., Thomas, D.G.S., Livingstone, I. and Wiggs, G.F.S. (1997). Dunefield 
activity and interactions with climate variability in the southwest Kalahari 
Desert. Earth Surface Proceedings and Landforms 22: 165-174. 
 
Burger, C.A.J., Hodgson, F.D.I. and Van Der Linde, P.J. 1981. Hidrouliese 
Eienskappe van Akwifere in die Suid-Vrystaat. Die Ontwikkeling en Evaluering 
van Tegniek vir die Bepaling van die Ontginninspotensial van 
Grondwaterbronne in die Suid-Vrystaat en in Noord-Kaapland. Instituut vir 
Gondwaterstudies, Oranje Vrystaat Universiteit, Volume 2. (Available from the 
University of the Free State). 
 208
Butzer, K.W. 1974. Geology of the Cornelia beds. Memoirs of the National Museum, 
Bloemfontein 9: 7-32. 
 
Butzer, K.W. 1984. Archaeology and Quaternary Environment in the Interior of 
Southern Africa. In: Klein, R.G. (Ed) Southern African Prehistory and 
Paleoenvironments. A. A. Balkema, Rotterdam. pp. 1-64. 
 
Butzer, K.W. 1988. Sediment interpretation of the Florisbad spring deposits.  
Palaeoecology of Africa 19: 181-189. 
 
Catuneanu, O., Wopfner, H., Eriksson, P.G., Cairncross, B., Rubidge, B.S., Smith, 
R.M.H.. and Hancox, P.J. 2005. The Karoo basins of south-central Africa. 
Journal of African Earth Sciences 43: 211-253. 
 
Chevallier , L. and Woodford, A.C. 1999. Morpho-tectonics and mechanisms of 
emplacement of the dolerite rings and sills of the western Karoo, South Africa. 
South African Journal of Geology 102(1): 43-54. 
 
Chevallier, L., Goedhart, M. and Woodford, A.C. 2001. The Influences of Dolerite Sill 
and Ring Complexes on the Occurrence of Groundwater in Karoo Fractured 
Aquifers: A Morpho-Tectonic Approach. Water Research Comission, Pretoria. 
WRC Project 937/1/01. 
 
Churchill, S.E., Brink, J.S., Berger, L.R., Hutchinson, R.A., Rossouw, L., Stynder, D., 
Hancox, P.J., Brand, D., Woodborne, S., Loock, J.C., Scott, L. and Ungar, P. 
2000. Erfkroon: a new Florisian fossil locality from fluvial contexts in the 
western Free State, South Africa. South African Journal of Science 96: 161-163. 
 
Clark, J.D. 1955. A note on a wooden impliment from the level of Peat I at Florisbad. 
Navorsinge van die Nasionale Museum, Bloemfontein 1(6): 135-140. 
 
Clarke, R.J. 1985. A New Reconstruction of the Florisbad Cranium, with Notes on the 
Site. In: Delson, E. (Ed) Ancestors: the Hard Evidence. New York, Alan R. Liss. 
 
 209
Coetzee, L. 2001. Lohmanniidae species (Acari: Oribatida) from the Holocene 
deposits at Florisbad. Navorsinge van die Nasionale Museum, Bloemfontein 
17(5): 125-134. 
 
Coetzee, L. 2002. Preliminary Report on Fossil Oribatid Mites from Florisbad 
Quaternary Research Station. In: Bernini, R., Nannelli, R., Nuzacci, E. and De 
Lillo, E. (Eds.) Acari Phylogeny and Evolution, Adaptations in Mites and Ticks. 
pp 41-44. Netherlands, Kluwer Academic Publishers. 
 
Coetzee, L. 2003. A new genus and species Floritricus louisbothai (Acari, Oribatidar, 
Oripodoidae, Zetomotrichidae) from South Africa. Navorsinge van die 
Nasionale Museum, Bloemfontein 19(5): 93-100. 
 
Coetzee, L. and Brink, J.S. 2003. Fossil oribatid mites (Acari: Oribatida) from the 
Florisbad quaternary deposits. Quaternary Research 59(2): 246-254. 
 
Connstanz, M.T., Fulmer, I.C., Ison, Ben., Norian, C. and Cupertino, C.A. 1997. Ionic 
interactions at the hydroxyapatite/collagen interface. Proceedings of the Mineral 
Research Society, Spring Meeting, San Francisco, California. 
 
Cooke, H.B.S. 1950. A critical revision of the Quaternary Perissodactyla of southern 
Africa. Annals of the South African Museum 31: 393-479. 
 
Cooke, H.B.S. 1952.  Some fossil mammals in the South African Museum collection.  
Annals of the South African Museum 42: 161-169. 
 
Cooke, H.B.S. 1964. Pleistocene Mammal Faunas of Africa, with Particular 
Reference to Southern Africa. In: Howell, F.C. and Bourliére, F. (Eds.). African 
Ecology and Human Evolution. pp. 175-184. Metheun, London. 
 
Cunroe, D., Grün, R. and Thackeray, J.F. 2002. Electron spin resonance dating of 
tooth enamel from Kromdraai B, South Africa. South African Journal of Science 
98: 540. 
 
 210
Davis, D.H.S. 1962. The distribution patterns of some small southern African 
mammals (Mammalia: Insectivora, Rodentia). Annals of the Transvaal Museum 
29: 56-76. 
 
Davis, S. H., Ulmer, R., Strong, L., Cathcart, T., Pote, J. and Brock, W. 1992. 
Constructed wetlands for dairy wastewater treatment. Paper American Society 
of Agricultural Engineers. St. Joseph, Mich.: American Society of Agricultural 
Engineers. Paper presented at the "1992 International Winter Meeting sponsored 
by the American Society of Agricultural Engineers," December 15-18, 1992, 
Nashville, Tennessee. American Society of Agricultural Engineers (92-4525). 
 
Deacon, J. 1966. An annotated list of radio-carbon dates for sub-Saharan Africa. 
Annals of the Cape Provincial Museums 5: 5-84. 
 
Deacon, H.J. 1970. The Acheulian occupation at Amanzi Springs, Uitenhage, Cape 
Province. Annals of the Cape Provincial Museums (Natural History) 8: 89-189. 
 
Deacon, H.J. and Deacon, J.1999. Human Beginnings in South Africa. David Philip 
Publishers, Cape Town and Johannesburg. 
 
De Bruiyn, H. 1971. ’n Geologiese Studie van die Panne van die Westelike Oranje 
Vrystaat. M.Sc. thesis, University of the Orange Free State, Bloemfontein. 
(Available from the University of the Free State). 
 
De Waal, S.W.P. 1978. The Squamata (Reptilia) of the Orange Free State, South 
Africa. Memoirs of the National Museum, Bloemfontein 11: 1-160. 
 
Douglas, R.M. 1992. Investigations into the Ecology of the Herpetofauna of Florisbad 
Research Station, Orange Free State, South Africa. M.Sc. thesis, University of 
Natal, Durban. (Available from the University of Kwazulu-Natal). 
 
Douglas, R.M. 1995. The Herpetofauna of Florisbad Research Station as largely 
determined by array trapping. Navorsinge van die Nasionale Museum, 
Bloemfontein 11(6): 121-148. 
 211
Drennan. M.R. 1935. The Florisbad skull. South African Journal of Science 32: 601-
602. 
 
Drennan, M.R. 1937. The Florisbad skull and brain cast. Transactions of the Royal 
Society of South Africa. 25: 103-114. 
 
Dreyer, T.F. 1935. A human skull from Florisbad, Orange Free State, with a note on 
the endocranial cast, by C.U. Ariëns Kappers. Koninklijke Akademie van 
Wetenschappen te Amsterdam 38: 3-12. 
 
Dreyer, T.F. 1936a. The endocranial cast of the Florisbad skull – A Correction. 
Soölogiese Navorsinge van die Nasionale Museum, Bloemfontein 1(3): 21-31. 
 
Dreyer, T.F. 1936b. The Florisbad skull in light of the Steinheim discovery. 
Zeitschrift für Rassenkunde 4(3): 320-322. 
 
Dreyer, T.F. 1938a. The archaeology of the Florisbad deposits. Argeologiese 
Navorsinge van die Nasionale Museum, Bloemfontein 1: 65-77. 
 
Dreyer, T.F. 1938b. The fissuration of the frontal endocranial cast of the Florisbad 
skull compared with that of the Rhodesian skull. Zeitschrift für Rassenkunde 
8(2): 192-198. 
 
Dreyer, T.F. 1947. Further observations on the Florisbad skull. Soölogiese Navorsinge 
van die Nasionale Museum, Bloemfontein 1: 183-190. 
 
Dreyer, T.F. and Lyle, A. 1931. New Fossil Mammals and Man from South Africa.  
Nasionale Pers, Bloemfontein. 
 
Du Toit, A.L. 1905. Geological Survey of Glen Grey and Parts of Queenstown and 
Woodehouse, Including the Indwe Area. Annual Report of the Geological 
Commission of the Cape of Good Hope. 
 
 212
Du Toit, A.L. 1920. The Karoo dolerites- a study in hypabysal intrusion. Transactions 
of the Geological Society of South Africa. 23: 1-42. 
 
Dykyjová, D. 1978. Nutrient Uptake by Littoral Communities of Helophytes. In: 
Dykyjová, D. Kvet, J. (Eds.) Pond Littoral Ecosystems: Structure and 
Functioning. Springer-Verlag, Berlin, Heidelburg, New York. 
 
Earlé, R. and Grobler, N. 1987. First Atlas of Bird Distribution in the Orange Free 
State. National Museum, Bloemfontein. 
 
Eiko, H. 2001. Review of pseudotachylites in the Witwatersrand Basin, South Africa. 
Journal of Geography. 110(1): 1-16. 
 
Enters, D., Kirchner, G. and Zolitschka, B. 2007. Establishing a chronology for 
lacustrine sediments using a multiple data approach – A case study from the 
Frickenhauser See, central Germany. Quaternary Geochronology 1(4): 249-260. 
 
Ewer, R.F. 1957. The fossil pigs of Florisbad. Navorsinge van die Nasionale Museum, 
Bloemfontein 1: 275-280. 
 
Ewer, R.F. 1962. A note on some fossil otters. Navorsinge van die Nasionale 
Museum, Bloemfontein 1: 275-280. 
 
Farmer, J.D. 2000. Hydrothermal systems: doorways to early biosphere evolution. 
Geological Society of America Today 10(7): 1-9. 
 
Fernàndez, L.M. and Guzmàn, J.A. !979. Seismic history of southern Africa. 
Seismologic Survey of the Geological Survey of South Africa 9: 1-38. 
 
Foley, R.A. and Lahr, M.M. 1997. Mode 3 technologies and the evolution of modern 
humans. Cambridge Archaeological Journal 7:3-36. 
 
Ford, W.E. 1966. A Textbook of Mineralogy (4th edition). John Wiley & Sons Inc. 
New York, London, Sydney. 
 213
Fourie, G.P.1970. Die Geologie van Florisbad. Geological Survey of South Africa. 
Unpublished Report No 0144. Department of Mineral and Energy Affairs, 
Geological Survey, Pretoria. 
 
Frederickson, A.F. and Reynolds Jr., R.C. 1960. Geochemical Method for 
Determining Palaeosalinity. In: Swinefors, A (Ed) Proceedings of the Eighth 
National Conference on Clays and Clay Minerals. Pergammom Press, London. 
 
Freeze, R.A. Cherry, J.A., 1979. Groundwater. Prentice-Hall, Inc, Engelwood Cliffs, 
N.J. 
 
Gaby, A.R. 1994. Preventing and Reversing Osteoporosis. Prima Publishing, Rocklin, 
California. 
 
Gaffeny, E.S. and Kitching, J.W. 1994. The most ancient African turtle. Nature 369 
55-58. 
 
Galloway, A. 1937. The nature and status of the Florisbad skull as revealed by its 
non-metrical features. American Journal of Physical Anthropology 23: 1-16. 
 
Gardner, E.W. 1932. Some problems of the Pleistocene hydrography of Kharga Oasis. 
Geological Magazine 69: 386-421. 
 
Geldenhuys, J.N. 1982. Classification of the pans of the western Orange Free State 
according to vegetation structure, with reference to avifaunal communities. 
South African Journal of Wildlife Research. 12: 55-62. 
 
Geyser, G.W.P. 1950. Panne – Hul onstaan en die faktore wat daartoe aanleiding gee. 
South African Geographical Journal. 32: 15-31. 
 
Goudie, A. and Thomas, D.S.G. 1986. Lunette dunes in southern Africa. Journal of 
Arid Environments 10: 1-12. 
 
 214
Goudie, A.S. and Wells, G.L. 1995. The nature, distribution and formation of pans in 
arid zones. Earth Science Reviews 38: 1-69. 
 
Gow, C.E. 1986. A new skull of Megazastrodon (Mammalia: Triconodontia) from the 
Elliot Formation (Lower Jurassic) of southern Africa. Palaeontologia Africana 
26:13-23. 
 
Grandstein, F.M. and Ogg, J.G. 2004. Geological time scale 2004 – why, how, and 
where next! Lethaia 37:175-181. 
 
Grobler, N.J. and Loock, J.C. 1988a. The Florisbad mineral spring: its characteristics 
and genesis. Navorsinge van die Nasionale Museum, Bloemfontein 5(17): 473-
485. 
 
Grobler, N.J. and Loock, J.C. 1988b. Morphological development of the Florisbad 
deposit. Palaeoecology of Africa 19: 163-168. 
 
Grobler, N.J. Behounek, N.J. and Loock, J.C 1988. Development of pans in the 
palaeodrainage in the north-western Orange Free State. Palaeocology of Africa 
19: 87-96. 
 
Grün R., Beaumont, P.B. and Stringer, C.B. 1990a. ESR dating evidence for early 
modern humans at Border Cave in South Africa. Nature 344, 537-539. 
 
Grün, R., Shackelton, N.J. and Deacon, H.J. 1990b. Electron-spin resonance dating of 
tooth enamel from Klasies River Mouth Cave. Current Anthropology 31: 537-
539. 
 
Grün, R., Brink, J.S., Spooner, N.A., Taylor, L., Stringer, C.B., Franciscus, R.G. and 
Murray, A.S. 1996. Direct dating of Florisbad hominid. Nature 382: 500-501. 
 
Grün, R. and Thorne, A. 1997. Dating of the Ngandong humans. Science 276: 1575-
1576. 
 
 215
Hall, D.A. 1976. The Aging of Connective Tissue. Academic Press, San Francisco. 
 
Hammers, F. and Verstraete, W. 2002. Key roles of pH and calcium metabolism in 
microbial carbonate precipitation. Reviews in Environmental Science and 
Biotechnology. 1: 3-7. 
 
Hancox,, P.J. 2000. The continental Triassic of South Africa. Zentralblatt fur 
Geologie und Palaontologie, Teil. 1998: 1285-1324. 
 
Hancox, J. and Rubidge, B. 2002. The Karoo Supergroup: a geological and 
palaeontological perspective. Rocks & Minerals Jan-Feb 2002 pp. 54-59. 
 
Hart, R.J., Andreoli, M.A.G., Tredoux, M. and De Wit, M.J. 1990. Geochemistry 
across an exposed section of Archaean crust at Vredefort, South Africa: with 
implications for mid-crustal discontinuities. Chemical Geology 82: 21-50. 
 
Hart, R.J., Andreoli, M.A.G., Reimold, W.U., Tredoux, M. 1991. Aspects od the 
dynamic and Thermal metamorphic history of the Vredefort cryptoexplosion 
structure: implications for its origin. Tectonophysics 192: 313-331. 
 
Hart, R.J., Hargraves, R.B., Andreoli, M.A.G., Tredoux, M., Doucouré, C.M.. 1995. 
Magnetic anomaly near the centre of the Vredefort structure: implications for 
impact-related magnetic signatures. Geology 23(3): 277-280. 
 
Hatano, K., Frederick, D. J. and Moore, J. A. 1994. Microbial ecology of constructed 
wetlands used for treating pulp mill wastewater. Water Science and Technology. 
29: 233-239. 
 
Hedges, R.E.M. and Millard, A.R., 1995. Bones and groundwater: Towards modelling 
of diagenetic processes. Journal of Archaeological Science 22: 155-164. 
 
Heegaard, E. 2003. Age-depth routine for R. (Access at http://www.bio.uu. 
nl/~palaeo/Congression/Holivar/Literature_Holivar2003.htm.) 
 
 216
Henderson, Z. 1992. The Florisbad skull – a diamond jubilee. Culna 43: 18-19. 
 
Henderson, Z. 1995. The Venters of Florisbad. Culna 48: 14-16. 
 
Henderson, Z. 1995. Mud-packs and mineral baths at Florisbad. Culna 49: 38. 
 
Henderson, Z. 1996. Interesting Hippopotamus finds at Florisbad. Culna 50: 29-30. 
 
Henderson, Z. 2001a. The integrity of the Middle Stone Age horizon at Florisbad, 
South Africa. Navorsinge van die Nasionale Museum, Bloemfontein 17: 25-52. 
 
Henderson Z.L. 2001b. Spatial Patterning at Southern African Middle Pleistocene 
Open-air Sites: Florisbad, Duinfontein 2/2 and Hwanganda’s Village. Ph.D. 
thesis University of Cambridge. 
 
Hendey, Q.B. 1974. The late Cainozoic carnivora of the south-western Cape Province. 
Annals of the South African Museum 63: 1-169. 
 
Henrotin, Y., Labasse, A., Zheng, S.X., Galais, P., Tsouderos, Y., Crielaard, J.M., and 
Reginster, J.Y. 2001. Strontium ranelate increases cartilage matrix formation. 
Journal of Bone and Mineral Research 16(2): 299-308. 
 
Hensley, M., Le Roux, P. Dy Preez, C., Van Huyssteen, C. Kotze, E. and Van 
Rensburg, L. 2006. Soils: The Free State’s agricultural base. South African 
Geographical Journal 88(1): 11-21. 
 
Herbert, L.S. 1921. The Hot Springs of New Zealand. Lewis & Co., London. 
 
Hillel, D. 1971. Soil and Water: Physical Principals and Processes. Academic Press, 
New York. 
 
 
 
 217
Hoch, A., Reddy,M., Aiken, G., 1998. Inhibition of Calcite Growth by Natural 
Organic Acids from the Everglades at pH 8.5 and 25oC. American Geophysical 
Union Meeting, Boston, MA. pp. 26-29. (Access at: http://wwwbrr.cr.usgs. 
gov/projects/SW_corrosion/calcite-poster/index.htm) 
 
Hoffman, A.C. 1953. The fossil alcelaphines of South Africa - genera Peloroceras, 
Lunatocerus and Alcelaphus. Navorsinge van die Nasionale Museum, 
Bloemfontein 1: 41-56. 
 
Hoffman, A.C. 1955. Important contributions of the Orange Free State to our 
knowledge of primitive man. South African Journal of Science 51: 163-168. 
 
Holmes, P and Barker, C.H. 2006. Geological and geomorphological controls on the 
physical landscape of the Free State. South African Geographical Journal 88(1): 
3-10. 
 
Holmes, P., Bateman, M.D., Thomas, D.S.G., Telfer, M.W., Barker, C.H. and 
Lawson, M.P. 2008. A Holocene late Pleistocene aeolian record from lunette 
dunes of the western Free State panveld, South Africa. The Holocene 18: 1193-
1205. 
 
Holmgren, K., Karlén, W., Lauritzen, S.E., Lee-Thorpe, J.A., Partridge, T.C., Piketh, 
S., Repinski, P., Stevenson, C., Svanered, O. and Tyson, P.D. 1999. A 3000-year 
high resolution stalagmite record of palaeoclimate for northern South Africa. 
The Holcene 9(3): 295-309. 
 
Holmgren, K., Lee-Thorpe, J.A., Cooper, G.R.J., Lundblad, K., Partridge, T.C., Scott, 
L., Sithalden, R., Talma, A.S. and Tyson P.D. 2003. Persistent millennial-scale 
climate variability over the past 25,000 years in southern Africa. Quaternary 
Science Reviews 22: 231-2326. 
 
Hooijer, D.A. 1958. Pleistocene remains of Hippopotamus from the Orange Free 
State. Navorsinge van die Nasionale Museum, Bloemfontein  1: 259-266. 
 
 218
Hughen, K.A., Baillie, M.G.L., Bard, E., Beck, J.W., Bertrand, C.J.H., Blackwell 
P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., 
Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Kromer, B., McCormac, G., 
Manning, S., Ramsey, C.B., Reimer, P.J., Reimer, R.W., Remmele, S., Southon, 
J.R., Stuiver, M., Talamo, S., Taylor, F.W., Van Der Plicht, J. and 
Weyhenmeyer, C.E. 2004. MARINE04 radiocarbon age calibration, 0-26 cal kyr 
BP. Radiocarbon 46(3): 1059-1086. 
 
Hunt, C.D. and Idso, J.P. 1999. Dietary boron as a physiological regulator of the 
normal inflammatory response: a review and current research. Journal of Trace 
Elements in Experimental Medicine. 12(3): 221-233. 
 
Hunt, P. G.; Humenik, F. J.; Szogi, A. A.; Rice, J. M.; Stone, K. C.; Cutts, T. T. and 
Edwards, J. P. 1993. Constructed wetland treatment of swine wastewater. 
American Society of Agricultural Engineers, St. Joseph, Michigan. pp1-10 (93-
2601/93-3510). 
 
Hutchinson, G.E. Pickford, G.E. and Schuurman, J.F.M. 1932. A contribution to the 
hydrobiology of pans and other inland waters of South Africa. Archchiv für 
Hydrobiologie 24 1-136. 
 
Janssens K., Vincze L., Vekemans B., Williams C.T., Radtke M., Haller M. and 
Knochel A. (1999). The non-destructive determination of REE in fossilized bone 
using synchrotron radiation induced K-line X-ray microflourescence analysis. 
Journal of  Analytical Chemistry. 363, 413-420. 
 
Johnson, A.M. and Pollard, D.D. 1973. Mechanics of growth of some laccolithic 
intrusions in the Henry Mountains, Utah, l. Field observations, Gilberts model, 
physical properties and flow of the magma. Technophysics 18: 261-309. 
 
Joubert, A. 1990. Die Kwaternere Fontein- en Verwante Lakustriene Afsettings by 
Florisbad, Oranje-Vrystaat. M.Sc. thesis, Univerisiteit van die Oranje-Vrystaat, 
Bloemfontien. (Available at the University of the Free State). 
 
 219
Joubert, A. and Visser, J.N.J. 1991. Approximate age of the thermal spring and 
lacustrine deposits at Florisbad, Orange Free State. Navorsinge van die 
Nasionale Museum, Bloemfontein 7: 97-111. 
 
Joubert, A., Beukes, G.J., Visser, J.N.J. and De Bruin, H. 1991. A note on flourite in 
Late Pleistocene thermal spring deposits at Florisbad, Orange Free State. South 
African Journal of Geology 94: 174-171. 
 
Karkanas, P., Bar-Yosef, O., Goldberg, P. and Weiner, S. 2000. Diagenesis in 
prehistoric caves: the use of minerals that form in situ to assess the completeness 
of the archaeological record. Journal of Archaeological Science 27: 915-929. 
 
Kattenhorn, S.A. 1994. Mechanism of Sill and Dyke Intrusion. M.Sc. thesis, 
University of Natal, Durban. (Available from the University of Kwazulu-Natal). 
 
Keay, R.J.W. (Ed.) 1959. Vegetation Map of Africa South of the Tropic of Cancer. 
Oxford University Press, London. 
 
Kempster, ,P.L. and Hatting, B. Th. 1983. Summarized Water Quality Criteria. 
Hydrological Research Institute, Department of Environment, Technical Report 
TR108. 
 
Kent, L.E. 1948. Die Geneeskragtige Bronne van Suid Afrika. Publicity and Travel 
Department, South African Railways and Harbours. Stellenbosch, Pro Ecclesia 
Printers. 
 
Kent, L.E. 1949. The thermal waters of the Union of South Africa and South West 
Africa. Transactions of the Geological Society of South Africa 52: 231-264. 
Kent L.E. 1964. Warmbronne in die Oranje-Vrystaat. `n Voorlopige Verslag. 
Geological Survey, Pretoria. Report Gh 1259. 
Kent, L.E. 1971. Warmbronne in die Oranje-Vrystaat. Aanvullende Verslag. 
Geological Survey, Pretoria. Report Gh 1585 
 220
Keyser, A.W. and Smith, R.M.H. 1978. Vertebrate biozonation of the Beaufort Group 
with special reference to the Western Karoo Basin. Annals of the Geological 
Survey, South Africa 12:1-36. 
 
Keyser, N. (Compiler). 1997. Map. Geological Map of the Republic of South Africa 
and the Kingdoms of Lesotho and Swaziland. Council for Geosciences, Pretoria. 
 
King, L.C. 1978. The Geomorphology of Central and Southern Africa. In: Werger, 
M.J.A., and Van Bruggen, A.C. (Eds.) Biogeography and Ecology in Southern 
Africa. W. Junk, The Hague. 
 
Kitching, J.W. 1977. The distribution of the Karoo vertebrate fauna. Memoir of the 
Bernard Price Institute for Palaeontological Research. University of the 
Witwatersrand 1: 1-131. 
 
Klein, R.G. 1984. The Large Mammals of Southern Africa: Late Pliocene to recent. 
In: Klein, R.G. (Ed) Southern African Prehistory and Palaeoenvironments. A.A. 
Balkema, Rotterdam. 
 
Kok, T.S. 1992. Recharge of Springs in South Africa. Department of Water Affairs 
and Forestry, Pretoria. Technical Report GH 3748. 
 
Krauskopf, K. B., 1967. Introduction to Geochemistry. McGraw Hill Book Company, 
Inc., New York. 
 
Krueger, H.W. 1991. Exchange of carbon with biological apatite. Journal of 
Archaeological Science 18: 355-361. 
 
Kruger, J.C. and Kok, T.S. 1976. Die Voorkoms van Grondwater in die Dolerietgange 
in die Noorsoos-Vrystaat. Department of Water Affairs and Forestry, Pretoria. 
Directorate Geohydrology Technical Report GH 1482, Pretoria. 
 
Kruger, A.C. 2002. Climate of South Africa. Surface Winds. South African Weather 
Service ,Pretoria. 
 221
Kruger, A.C. 2007. Climate of South Africa. Precipitation. South African Weather 
Service, Pretoria. 
 
Kruger, G.P. 1983. Terrain Morphological Map of Southern Africa. Soil and 
Irrigation Research Institute, Department of Agriculture, Pretoria. 
 
Kruger, T. and Lubczenko, V. 1994. A Community Water Quality Monitoring Manual 
for Victoria. Victorian Community Water Monitoring Task Group, Victoria. 
 
Kuman, K. A. 1989. Florisbad and #GI: The Contribution of Open-air Sites to Study 
of the Middle Stone Age in Southern Africa. Ph.D. thesis, University of 
Pennsylvania. 
 
Kuman, K. and Clarke, R.J. 1986. Florisbad - New investigations at a Middle Stone 
Age hominid site in South Africa. Geoarchaeology 1: 103-125. 
 
Kuman, K., Inbar, M. and Clarke, R.J. 1999. Palaeoenvironments and cultural 
sequence of the Florisbad Stone Age hominid site, South Africa. Journal of 
Archaeological Science 26: 1409-1425. 
 
Lancaster, N. 1978. The pans of the southern Kalahari, Botswana. Geographical 
Journal 144: 81-89. 
 
Lahr, M.M. and Foley, R.A. 1988. Towards a theory of modern human origins: 
geography, demography, and diversity in recent human evolution. Yearbook of 
Physical Anthropology 41: 137-176. 
 
Lancaster, N. 1986. Pans of the south western Kalahari: a preliminary report. 
Palaeoecology of Africa 17:59-67. 
 
Lancaster, N. 1989. Late Quaternary palaeoenvironments in the southwestern 
Kalahari. Palaeogeography, Palaeoclimatology, Palaeoecology 70: 267-276. 
 
 222
Lancaster, N. 1990. Palaeoclimate evidence from sand seas. Palaeogeography 
Palaeoclimatology Palaeoecology 76: 279-290. 
 
Lakshman, G. 1979. An ecosystem approach to the treatment of waste waters. Journal 
of Environmental Quality 8: 353-361. 
 
Larcher, W. 1983. Physiological Plant Ecology. Springer-Verlag, Berlin, New York, 
Tokyo. 
 
Lawson, M.P. and Thomas, D.S.G. 2002. Late Quaternary lunette dune sedimentation 
in the southwestern Kalahari Desert, South Africa: luminescence based 
chronologies of aeolian activity. Quaternary Science Reviews 24: 825-836. 
 
Le Roux, J.S 1978. The origin and distribution of pans in the Orange Free State. South 
African Geographer 6: 167-176. 
 
Lieberman, D.E., McBratney, B.M. and Krovitz, G. 2002. The evolution and 
development of cranial form in Homo sapiens. Proceedings of the National 
Academy of Science of the United States of America (PANAS) 99(3): 1134-
1139. 
 
Libby, W.F. 1954. Chicago Radiocarbon Dates V. Science 120: 733-742. 
 
Litthauer, A. 2007. The Mobility of Uranium, Molybdenum, Arsenic, Lead, Tungsten, 
and Thorium in the Ground- and Surface Water of the Mooifontein Ore Body, 
South Africa. B.Sc. Hons. Project, University of the Free State. 
 
Lombaard, B.V. 1952. Karoo dolerites and lavas. Transactions of the Geological 
Society of South Africa 55:175-198. 
 
Loock, J.C. and Grobler, N.C. 1988. The regional geology of Florisbad. Navorsinge 
van die Nasionale Museum, Bloemfontein 7: 529-540. 
 
 223
Lotz, L.N., Seaman, M.T. and Kok, D.J. 1991. Surface active spiders (Araneae) of a 
site in semi-arid central South Africa. Navorsinge van die Nasionale Museum, 
Bloemfontein 7: 529-540. 
 
Louw, S. 1987. Species composition and seasonality of pitfall trapped Coleoptera at a 
site in the central Orange Free State, South Africa. Navorsinge van die 
Nasionale Museum, Bloemfontein  5(15): 415-453. 
 
Louw, W.J. 1979. Orange Free State Rainfall. Part 1. General Characteristics. 
Technical Paper 6. Weather Bureau, Department of Transport, Pretoria. 
 
Low, A.B. and Rebelo, A.G. (Eds.) 1996. Vegetation of South Africa, Lesotho and 
Swaziland. Department of Environmental Affairs and Tourism, Pretoria. 
 
Lynch, C.D. 1983. The Mammals of the Orange Free State. Memoirs van die 
Nationale Museum Bloemfontein. 18: 1-218. 
 
MacVicar, C.N. (Ed.) 1973. Soil Map. Republic of South Africa: An Interim 
Compilation. Soil and Irrigation Research Institute, Department of Agricultural 
and Technical Services. Government Printers, Pretoria. 
 
MacVicar, C.N., Loxton, R.F., Lambrechts, J.J.N., Le Roux J., De Villiers, JM., 
Verster, E., Merryweather, F.R., Van Rooyen, T.H. and Von M. Harmse, H.J. 
1977. Soil Classification. A Binomial System for South Africa. Soil and 
Irrigation Research Institute, Department of Agricultural Technical Services. 
Government Printers, Pretoria. 
 
Maier, W.D., Andreoli, M.A.G., McDonald, I., Higgins, M.D., Boyce, A.J., 
Shukolyukov, A., Lugmair, G.W., Ashwal, L.D. Gräser, P., Ripley, E.M. and 
Hart, R.J. 2006. Discovery of a 25-cm asteroid clast in the giant Morokweng 
impact crater, South Africa. Nature 441: 203-206. 
 
 
 224
Marie, P.J., Skoryna, S.C., Pivon, R.J., Chabot, G., Clorieux, F.H., and Stara, J.F. 
1985. Histomorphometry of Bone Changes in Stable Strontium Therapy. In: 
Hemphill, D.D. (Ed) Trace Substances in Environmental Health XIX. University 
of Missouri, Columbia, Missouri. Pp. 193-208. 
 
Marie, P.J. and Hott, M. 1986. Short-term effects of fluoride and strontium on bone 
formation and resorption in the mouse. Metabolism 35: 547-551. 
 
Marshall, T.R. 1987a. The Origins of the Pans of the Western Orange Free State: A 
Morphotectonic Study of the Palaeo-Kimberley River. Economic Geology 
Research Unit, University of the Witwatersrand, Johannesburg. Unpublished 
Report No 196. 
 
Marshall, T.R. 1987b. The Morphotectonics of the Orange Free State. M.Sc. thesis, 
University of the Witwatersrand, Johannesburg. 
 
Marshall, T.R. and Harmse, J.T. 1992. A review of the origin and propagation of 
pans. South African Geographer 19: 9-21. 
 
Martini, J.E.J. 1992. The metamorphic history of the Vredevort dome at 
approximately 2 Ga as revealed by coesite-stishovite-bearing pseudotachylites. 
Journal of Metamorphic Geology 10(4): 517-527. 
 
Mashchak, M.S. and Naumov, M.V. 1996. The Suavjarvi structure: an Early 
Proterozoic impact site on the Fennoscandian Shield. Lunar and Planetary 
Science 27: 825. 
 
Mayer, J. 2007 “The Vredefort Structure. Misconceptions and Facts.” (Access at: 
http://www.vredefortstructure.org/homepage.htm) 
 
Mazor, E. and Verhagen, B. Th. 1983. Dissolved ions, stable and radioactive isotopes 
and noble gases in thermal waters of South Africa Journal of Hydrology 63: 
315-329. 
 
 225
McCarthy, T. and Rubidge, B. 2005. The Story of Earth & Life. Struik Publishers, 
Cape Town. 
 
McConnell, D. 1952. The crystal chemistry of carbonate apatites and their 
relationship to the composition of calcified tissue. Journal of Dental Research 
31: 53-63. 
 
McCormac, F.C., Hogg, A.G., Blackwell, P.G., Buck, C.E., Higham, T.F.G. and 
Reimer, P.J. 2004. SCALo4 southern hemisphere calibration, 0-11.0 cal kyr BP. 
Radiocarbon 46(3): 1087-1092. 
 
Meester, J. 1965. The origins of the southern African mammal fauna. Zoologica 
Africana 1: 87-95. 
 
Meiring, A.J.D. 1956. The macrolithic culture of Florisbad. Navorsinge van die 
Nasionale Museum, Museum, Bloemfontein 1: 205-237. 
 
Mercier, N., Valladas, H., Joron, J.-L., Schiegel, S., Bar-Yosef, O. and Weiner, S. 
1995. Thermoluminescence dating and the problem of geochemical evolution of 
sediments. Israel Journal of Chemestry 35: 137-141. 
 
Meunier, P.J., Slosman, D.O., Delmas, P.D., Sebert, J.L., Brandi, M.L., Albanese, C., 
Lorenc, R., Pors-Nielsen, S., De Vernejoul, M.C., Roces, A. and Reginster, J.Y. 
2002. Strontium ranelate: dose-dependant effects in established postmenopausal 
vertebral osteoporosis – a 2-year randomized placebo controlled trial. Journal of 
Clinical Endocrinology and Metabolism 87(5): 2060-2066. 
 
Meunier, P.J., Roux, C., Seeman, E., Ortolani, S., Badurshi, J.E., Spector, T.D., 
Cannata, J., Balogh, A., Lemmel, E.M., Pors-Nielsen, S., Rizzoli, R., Genant, 
H.K. and Reginster, J.Y. 2004. The effects of strontium ranelate on the risk of 
vertebral fractures in women with postmenopausal osteoporosis. New England 
Journal of Medicine 350(5): 459-468. 
 
 226
Meyboom, A.F. and Wallace, R.C. 1978. Occurrence and origin of ring-shaped 
dolerite outcrops in the Eastern Cape Province and Western Transkei. 
Transactions of the Geological Society of South Africa 81: 95-99. 
 
Millard, A.R. and Hedges, R.E.M. (1995). The role of the environment in uranium 
uptake in buried bone. Journal of Archaeological Science. 22, 239-250. 
 
Millard A.R. and Hedges R.E.M. (1999). A diffusion-adsorption model of uranium 
uptake in archaeological bone. Geochimica et Cosmochimica Acta 60, 2139-
2152. 
 
Moon, B.P. and Dardis, G.F. 1988. Introduction. In: Moon, B.P.and Dardis, G.F. 
(Eds). The Geomorphology of Southern Africa. Southern ,Johannesburg. 
 
Mortveldt, J.J., Giordano, P.M. Lindsay, W.L. (Eds), 1972. Micronutrients in 
Agriculture. 3rd ed. Soil Science Society of America Inc. Madison, Wisconsin. 
 
Mucina, L. and  Rutherford, M.C. (Eds) 2006. The Vegetation of South Africa, 
Lesotho and Swaziland. Strelitzia 19. South African National Biodiversity 
Institute, Pretoria.  
 
Munsey, L.D. 1972. Salinity tolerance of the African pipid frog Xenopus laevis. 
Copeia 3:548-586. 
 
Myburg, A. 1997. The Geomorphological Evolution of the Northwestern Free State 
since the Mesozoic. Ph.D. thesis University of the Orange Free State, 
Bloemfontein. (Available at the University of the Free State). 
 
Nakamura, N. 2003. Pseudotachylite as a magnetic recorder of earth’s ambient field. 
Geological Society of America, 2003 Seattle Annual Meeting, Session 249, 
Seismogenic Friction and Pseudotachylites, Washington State Convention and 
Trade Centre. Abstracts with Programs 35(6) 628. 
 
 227
Neveling, J. 2004. Stratigraphic and Sedimentological Investigation of the Contact 
Between the Lystrosaurus and the Cynognathus Assemblage Zones (Beaufort 
Group: Karoo Supergroup). Council for Geoscience, Pretoria. Bullitin 137. 
 
Newnham, D. O. 1994a. The role of boron in human nutrition. Journal of Applied 
Nutrition. 46: 81-85. 
 
Newnham, D. O. 1994b. Essentiality of boron for healthy bones and joints. 
Environmental Health Perspectives 102: 83-85. 
 
Nielsen, F. H. 1992. Facts and fallacies about Boron. Nutrition Today. 27:6-12. 
 
Noller, B. N.,Woods, P. H. and Ross, B. J. 1994. Case studies of wetland filtration of 
mine waste water in constructed and naturally occurring systems in Northern 
Australia. Water Science and Technology 29:.257-265. 
 
Nolte, C.C. 1995. The Geology of the Winberg aAea: Explanation Sheet 2826 (1: 250 
000). Council for Geoscience, Pretoria. 
 
Oakley, K.P. 1954. Study tour of early hominid sites in southern Africa. South African 
Archaeological Bulletin 9: 75-87. 
 
Oakley, K.P. 1955. The Dating of the Broken Hill, Florisbad and Saldanha Skulls. In: 
Clark, J.D. (Ed.) Third African Congress on Prehistory. Livingstone. pp 76-79. 
Chatto & Windus, London. 
 
Pan, Y. and Fleet, M.E. 2002. Composition of the Apatite-group Minerals: 
Substitution Mechanisms and Controlling Factors. In: Kohn, M.J. Rakonan, 
J.and Hughes, J.M. (Eds.) Reviews in Mineralogy Geochemistry, Phosphates: 
Geochemical, Geobiological and Materials Importance 48: 13-49. 
 
Petit, J.R., Mounier, L., Korotkevich, Y.S., Kotlyakov, V.I. and Lorius, C. 1990. 
Palaeoclimatological and chronological implications of the Vostok core record. 
Nature 343; 56-58. 
 228
Petja, B.M., Malherbe J. and Van Zyl. J. 2004. Using satellite imagery and rainfall 
data to track climate variability in South Africa IGARSS 2004. IEEE 
International Geoscience and Remote Sensing, Ankorage, Alaska. N.J. 
Picataway IEEE. 
 
Pfeiffer, J.E. 1969. Emergence of Man. Harper & Row, New York. 
 
Picker, M.D. 1985. Hybridisation and habitat selection in Xenopus gilli and Xenopus 
laevis in the south-western Cape. Copeia 3:574-580. 
 
Poynton, J.C. 1964. The amphibia of southern Africa: a faunal study. Annals of the 
Transvaal Museum 17:1-334. 
 
Poynton, R.J. 1971. A silvicultural map of Southern Africa. South African Journal of 
Science 67:58-60. 
 
Protsch, R. 1974. Florisbad: its palaeoanthropology, chronology and archaeology.  
Homo 25: 68-78. 
 
Ramsey, C.B. 1995. Radiocarbon calibration and analysis of stratigraphy: the OxCal 
program. Radiocarbon 37: 425-430. 
 
Ramsey, C.B. 2001. Development of the Radiocarbon calibration program OxCal. 
Radiocarbon 43: 355-363. 
 
Rautenbach, I.L. 1978. Ecological distribution of mammals in the Transvaal. Bulletin 
of the Carnegie Museum of Natural History  6: 175-187. 
 
Rebelo, A.G and Low, A.B. (Compilers) 1996. Vegetation of South Africa, Lesotho 
and Swaziland (Map). Department of Environmental Affairs and Tourism, 
Pretoria, and National Botanical Institute, Pretoria. 
 
 
 229
Reimer, P.J., Baillie, M.G.L., Bard, E.; Bayliss, A., Beck, J. W., Bertrand, C.J.H., 
Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, 
R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, 
K.A., Kromer, B., McCormac, G., Manning, S., Ramsey, C.B., Reimer, R.W., 
Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., Van der 
Plicht, J., Weyhenmeyer, C.E., 2004. IntCal04 terrestrial radiocarbon age 
calibration, 0-26 cal kyr BP. Radiocarbon 46(3): 1029-1058. 
 
Rightmire, G.P. 1976. Relationships of Middle and Upper Pleistocenehominids from 
sub-Saharan Africa. Nature 260:238-240. 
 
Rightmire, G.P. 1978.  Florisbad: Its human population succession in southern Africa.  
American Journal of Physical Anthropology 48: 475 486. 
 
Rindl, M.M. 1915. The mineral spring on the farm Rietfontein, district Brandfort, 
O.F.S. South African Journal of Science 12: 561-588. 
 
Rindl, M.M. 1916. The medicinal springs of South Africa. South African Journal of 
Science 13: 528-552. 
 
Rindl, M.M. 1928. Medicinal Springs of South Africa. South African Railways and 
Harbours Administration, Johannesburg. 
 
Rink, W.J. (2001). Beyond 14-C dating: A Users Guide to Long Range Dating 
Methods in Archaeology. In: Goldberg, P, Holliday V.T., and Ferring, C.R. 
Kluver (Eds). Earth Sciences and Archaeology. Academic/Plenum, New York. 
 
Rogers, A.W. and Schwarz, E.H.L. 1902. Report on Parts of the Divisions of Beaufort 
West, Prince Albert and Sutherland. Annual Report of the Geological 
Commission of the Cape of Good Hope. 
 
 
 
 230
Roh, Y., Phelps, T.J., McMillan, A.S. and Lauf, R.J. 2001. Utilization of 
biomineralization processes with fly ash for carbon sequestration. Proceedings 
of the First National Conference on Carbon Sequestration. USA Dept. of Energy 
– National Energy and Technology Laboratory. DOE-Netl_2001-1144. 
 
Rossouw, L. 1996. The extraction of opal phytoliths from the fossilized teeth of two 
Bovid species from Florisbad. Navorsinge van die Nasionale Museum, Museum, 
Bloemfontein 12(8): 265-274. 
 
Rubidge, B.S. 2005. Re-uniting lost continents – Fossil reptiles from the ancient 
Karoo and their wanderlust. 27th Du Toit Memorial Lecture. South African 
Journal of Geology 108(1): 135- 172. 
 
Rubidge, B.S. and  Brink, J.S. 1985. Preliminary survey of the extent and nature of 
the Pleistocene sedimentary deposits at Florisbad, South Africa. Navorsinge van 
die Nasionale Museum, Museum, Bloemfontein 5: 69-76. 
 
Rubidge, B.S., Johnson, M.R., Kitching, J.W., Smith, R.M.H., Keyser, A.W. and 
Groenewalld, G.H. 1995. An Introduction to the Biozonation of the Beaufort 
Group. In: Rudidge, B.S. (Ed.) Biostratigraphy of the Beaufort Group. South 
African Council for Stratigraphy, Biostratigraphic Series, Council for 
Geoscience. 
 
Rutherford, M.C. and Westfall, R. 1986. Biomes of South Africa – Objective 
categorization. Memoirs of the Botanical Survey of South Africa 54: 1-98. 
 
Sampson, C.G. 1974. The Stone Age Archaeology of Southern Africa. New York, 
Academic Press. 
 
Schulze, B.R. 1994. Climate of South Africa. General Survey Part 8. Department of 
Environment Affairs, Weather Bureau, Pretoria. 
 
 231
Scott, L. 1989. Climatic conditions in southern Africa since the last glacial maximum, 
inferred from pollen analysis. Palaeogeography Palaeoclimatology 
Palaeoecolgy 70: 345-353. 
 
Scott, L. and Brink, J.S. 1992. Quaternary palaeoenvironments of pans in central 
South Africa: palynological and palaeontological evidence South African 
Geographer 19: 22-34. 
 
Scott, L. and Nyakale, M. 2002. Pollen indications of Holocene palaeoenvironments 
at Florisbad spring in the central Free State, South Africa. The Holocene 12(4): 
497-503. 
 
Scott, L. and Rossouw, L. 2005. Reassessment of botanical evidence for 
palaeoenvironments at Florisbad, South Africa. South African Archaeological 
Bulletin 60(182): 96-102. 
 
Scott, L., Fernández- Jalvo, Y. Carrión, J and Brink, J.S. 2003. Preservation and 
interpretation of pollen in hyaena coprolites: taphonomic observations from 
Spain and South Africa. Palaeoenvironments of Africa 39: 83-91. 
 
Seaman, M.T., Ashton P.J. and Williams, W.D. 1991. Inland salt waters of southern 
Africa. Hydrobiologia 210: 124-128. 
 
Shackelton, N.J. 1969. The last interglacial in the marine and terrestrial records. 
Proceedings of the Royal Society of London (B) 174: 135-154. 
 
Shand, S.J. 1916. The pseudotachylyte of Parijs (Orange Free State) and its relation to 
`trap-shotten gneiss` and `flinty crush rock`. Quarterly Journal of the 
Geological Society of London 72: 198-221. 
 
Sewell, D.R. 1998. A note for novices. Radiocarbon 40(3): xi 
 
Shimp, N. F., Witters, J., Potter, P. E. and Schleicher, J. A. 1969. Distinguishing 
marine and freshwater muds. Journal of Geology. 77: 566-580. 
 232
Singer, R. 1958. The Rhodesian, Florisbad and Saldanha Skulls. In: Von 
Koenigswald, G. (Ed.). Hundert Jhare Neanderthaler. Utrecht , Kemink en 
Zoon, N.V. 
 
Singer, R. and Boné, E.L. 1960. Modern giraffes and fossil giraffids of Africa. Annals 
of the South African Museum 45: 375-548.  
 
Skoryna, S.C. 1981. Effects of oral supplementation with stable strontium. Canadian 
Medical Association Journal 125: 703-712. 
 
Smit, P.J. 1977. Die Geohidrologie in die Opvangsgebied van die Moloporivier in die 
Noordlike Kalahari. Ph.D. thesis University of the Orange Free State, 
Bloemfontein (Available from the University of the Free State). 
 
St-Cyr, L., Cattaneo, A., Chassé, R. and Fraikin, C. G. J. `1997. Technical Evaluation 
of Monitoring Methods Using Macrophytes, Phytoplankton and Periphyton to 
Assess the Impacts of Mine Effluents on the Aquatic Environment. Report 
presented to: Canada Centre for Mineral and Energy Technology. 1-218. 
(Access at: http://www.nrcan.gc.ca/mets/aete/reports/2_3_2.pdf) 
 
Svensen, H., Jamtveit, B. Planke, S. and Chevallier, L. (2006). Structure and 
evolution of hydrothermal vent complexes in the Karoo Basin, South Africa. 
Journal of the Geological Society 163 (4) 671-682. 
 
Taylor, S A. and Ashcroft, G.L. 1972. Physical Edaphology. W. H. Freeman and 
Company, San Francisco. 
 
Telfer, M.W. and Thomas, D.S.G. 2006. Complex Holocene lunette dune 
development, South Africa: implications for palaeoclimate and models of pan 
development in arid regions. Geology 34(10): 853-856. 
 
Telford, R.J., Heegaard, E. and Birks, H.J.B. 2004. All Age-depth models are wrong: 
but how badly. Quaternary Science Reviews 23: 1-5. 
 
 233
Thackeray, J.F, Brink, J.S. and Plug, I. 1996. Temporal Variability in Horn-core 
Dimensions of Damaliscus niro from Olduvai, Sterkfontein, Cornelia, and 
Florisbad. In: Stewart, K.M. and Seymor, K.L. (Eds). Palaeoecology and 
Palaeoenvironments of Late Cenozoic Mammals. University of Toronto Press 
Inc. Toronto. 
 
Thomas, D.S.G., Nash, D.J., Shaw, P.A. and Van Der Post, C. 1993. Present day 
sediment cycling at Witpan in the arid southwestern Kalahari. Catena 20: 515-
527. 
 
Thomas, D.S.G., Knight, M. and Wiggs, G.F.S. 2005. Remobilization of southern 
African desert dune systems by twenty-first century global warming. Nature 
435: 1218-1221. 
 
Tildon, D. L. and Kadlec, R. H. 1979. The utilization of a fresh-water wetland for 
nutrient removal from secondary treated waste water effluent. Journal of 
Environmental Quality 8: 328-334. 
 
Tóth, L. 1972. Reeds control eutrophication of Batton Lake. Water Research. 6: 1533-
1539. 
 
Travers, R.L., Rennie, G.C. and Newnham, R.E. 1990. Boron and arthritis: the results 
of a double-blind pilot study. Journal of Nutritional Medicine 1: 127-132. 
 
Tredoux, M., Hart, R.J., Carlson, R.W. and Shirley, S.B. 1999. Ultramafic rocks at the 
center of the Vredefort structure: further evidence of the crust on edge model. 
Geology 27(10): 923-926. 
 
Trueman, C.N. Tuross, N., 2002. Trace Elements in Recent Fossil Bone Apatite. In: 
Kohn, M.J. Rakovan, J.and Hughes, J.M. (Eds.), Reviews in Mineralogy 
Geochemistry, Phosphates: Geochemical, Geobiological and Materials 
Importance 48: 489-521. 
 
 234
Tyson, P.D. 1986. Climatic Change and Variability in Southern Africa. Oxford 
University Press, Cape Town. 
 
Usher, B.H., Pretorius, J.A., and Van Tonder, G.J. 2006. Management of a Karoo 
fractured-rock aquifer system – Kalkveld Water User Association (WUA). 
Water SA 32(1) 9-19. 
 
Van der Wal, R.W.E. 1977. Die Neerslagklimaat van die Oranje-Vrystaat. M.Sc. 
thesis, University of the Orange Free State, Bloemfontein (Available from the 
University of the Free State). 
 
Van Zinderen Bakker, E.M. 1957. A Pollen Analytical Investigation of the Florisbad 
Deposits (South Africa). In: Clarke, J.D. (Ed.). Third Pan-African Congress on 
Prehistory, Livingstone 1955. pp. 56-57. London, Chatto and Windus. 
 
Van Zinderen Bakker, E.M. 1989. Middle Stone Age palaeoenvironments at Florisbad 
(South Africa). Palaeoecology of Africa  20: 133-154. 
 
Vegter, J.R. 1995. An Explanation of a Set of National Groundwater Maps. Sheets 1 
and 2, Groundwater Resources of the Republic of South Africa. WaterRresearch 
Commission, Pretoria. WRC Report TT 74/95. 
 
Vegter, J.R. 2001. Groundwater Development in South Africa and an Introduction to 
the Hydrology of Groundwater Regions. Water Research Commission, Pretoria. 
WRC Report TT 134/00. 
 
Verster, E., Van DeVenter, P.W. and Ellis, F. 1992. Soils and associated materials of 
some pan floors and margins in South Africa: a Review. South African 
Geographer 19: 35-47. 
 
Viljoen, C.C. 2005. Tierberg Formation. In: Johnson M.R. (Ed.). Catalogue of South 
African Lithostratigraphic Units. 8: 36-40. Council for Geoscience, South 
African Committee for Stratigraphy, Prtetoria. 
 
 235
Villumsen, A. and Nielsen, O.B. 1976. The influence of palaeosalinity, grain size 
distribution and clay minerals on the content of B, Li and Rb in Quaternary 
sediments from eastern Jutland, Denmark. Sedimentology 23(6): 845-855. 
 
Visser, J.N.J. and Joubert, A 1990. Possible earthquake-induced sediment liquefaction 
in the thermal spring deposits at Florisbad, Orange Free State. South African 
Journal of Geology 93: 525-530. 
 
Visser, J.N.J and Joubert, A. 1991. Cyclicity in the Late Pleistocene to Holocene 
spring and lacustrine deposits at Florisbad, Orange Free State. South African 
Journal of Geology 94: 123-131. 
 
Vivier, J.J.P., Van Der Voort I. And Botha, J.F. 1995. The Analysis and Interpretation 
of Aquifer Tests in Secondary Aquifers: The Influence of Geology on the 
Geohydrology of the Karoo Aquifers. Institute of Groundwater Studies, 
University of the Orange Free State, Unpublished Report. (Available from the 
University of the Free State). 
 
Vivier, J.J.P. 1996. The Influence of Geology on the Geohydrology of Karoo 
Aquifers. M.Sc. thesis, University of the Orange Free State, Bloemfontein. 
(Available from the University of the Free State). 
 
Weinert, H. 1936. Der Urmenschenschädel von Steinheim. Zeitschrift für 
Morphologie und anthropology 35: 463-518. 
 
Wellington, J.H. 1945. Notes on the drainage of the western Free State sandveld. 
South African Geographical Journal 27: 146-145. 
 
Wellington, J.H. 1955. Southern Africa: A Geographical Study. 2v. Cambridge 
University Press, Cambridge. 
 
Wells, L.H. 1959. The Quaternary giant hartebeests of South Africa. South African 
Journal of Science 55: 123-128. 
 
 236
Wells, L.H. 1965. Antelopes in the Pleistocene of southern Africa. Zoologica 
Africana 1: 115-120. 
 
Wells, L.H. 1967. Antelopes in the Pleistocene of Southern Africa. In: Bishop, W. W. 
and Clark, J. D. (Eds.). Background to Evolution in Africa. Chicago, University 
of Chicago Press. 
 
Wells, L.H. 1969. Homo sapiens afer Linn.. Content and earlier representatives. South 
African Archaeological Bulletin 24: 172-173. 
 
Wells, L.H. 1972. Late Stone Age and Middle Age toolmakers. South African 
Archaeological Bulletin 27: 5-9. 
 
Wetzel, R. G. 1983. Periphyton of Freshwater Ecosystems. Dr W. Junk Publishers, 
The Hague, The Netherlands. 
 
White, F. 1981. UNESCO Vegetation Map of Africa. Cook, Hammond & Kell Ltd., 
Mitcham, Surrey. 
 
White, T.D., Asfaw, B., DeGusta, D., Gilbert, H.Richards, G.D., Suwa, G. and 
Howell, F.C. 2003. Pleistocene Homo sapiens from Middle Awash, Ethiopia. 
Nature 423:742-747. 
 
Williams, C.T. 1988. Alteration of Chemical Composition of Fossil Bones by Soil 
Processes and Groundwater. In: Grupe, G. and Herman, B. (Eds) Trace 
Elements in Environmental History. Springer-Verlag. Berlin. 
 
Woodford, A.C. and Chevallier, L. 2002. Hydrology of the Main Karoo Basin: 
Current Knowledge and Future Research Needs. Water Research Commission, 
Pretoria. WRC Report TT 179/02. 
 
Young, J.A. and Evans, R.A. 1986. Erosion and deposition of fine sediments from 
playas. Journal of Arid Environments 10: 103-115. 
 
 237
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Appendices 
 
 
 
 
 238
 
 
 
 
Appendix 
 
I 
 
 
 
 
 
 
 
 
 The quality of the Florisbad spring-
water in relation to the quality of 
the groundwater and the effects of 
rainfall. 
 
Douglas, R. M. 2006. 
 
Water SA 27: 39-48. 
 
 
 239
 
 240
 
 241
 
 242
 
 243
 
 244
 
 245
 
 246
 
 247
 
 248
 
 249
 
 
 
 
Appendix 
 
II 
 
 
 
 
 
 
 
 
Salinization of the Florisbad organic 
layers, clay, and groundwater. 
 
Douglas, R. M. 2006. 
 
Navorsinge van die Nasionale Museum, 
Bloemfontein 17(1): 1-24. 
 
 
 
 
 250
 
 
 
 
 
 
 251
 
 
 
 252
 
 
 
 
 253
 
 
 
 254
 
 
 
 255
 
 
 256
 
 
 257
 
 
 
 258
 
 
 
 
 259
 
 
 
 260
 
 
 
 261
 
 
 262
 
 
 
 263
 
 
 
 
 264
 
 
 
 
 265
 
 
 266
 
 
 267
 
 
 
 
 268
 
 
 
 269
 
 
 
 
 270
 
 
 
 
 
 
 
 
 
************************************************************** 
 271
 
 
 
 272
 
 
 273
 
 
 274
 
 
 
 
Appendix 
 
III 
 
 
 
 
 
 
 
 
Is the spring-water responsible for 
the fossilization of faunal remains 
at Florisbad, South Africa. 
 
Douglas, R. M. 2006. 
 
Quaternary Research 65: 87-95. 
 
 
 
 275
 
 276
 
 277
 
 278
 
 279
 
 280
 
 281
 
 282
 
 283
 
 
 
 
 
 
 284
 
 
 
 
Appendix 
 
IV 
 
 
 
 
 
 
 
 
Formation of the Florisbad spring 
and fossil site – an alternative 
hypothesis. 
 
Douglas, R. M. 2006. 
 
Journal of Archaeological Science 33:698-
706. 
 
 
 
 285
 
 286
 
 287
 
 288
 
 289
 
 290
 
 291
 
 292
 
 293
 
 294
 
 295
 
 296