i 
 

BOTANICAL MOTIFS IN THE ROCK ART OF ZIMBABWE 

 

 

Stephen van den Heever 

681984 

School of Geography, Archaeology and Environmental Studies (GAES), 

Faculty of Science 

University of Witwatersrand 

Supervisors: 

Doctor Sam Challis & Doctor Jeremy Hollmann 

 

 

A Dissertation submitted to the Faculty of Science, University of the Witwatersrand, 

Johannesburg, in fulfilment of the requirements for the degree of Master of Science. 

Johannesburg, September 2021 

 

 

 

 

 

 



i 
 

 

 

DECLARATION 

 

I declare that this dissertation is my own, unaided work. It is being submitted for the Master of 

Science at the University of the Witwatersrand, Johannesburg. It has not been submitted before 

for any degree or examination at any other University. 

 

 

 

Stephen Wepener van den Heever 

Johannesburg 10th September 2021 

 

 

 

 

 

 



iii 
 

 

ABSTRACT 
  

Botanical motifs are found in hunter-gatherer rock paintings throughout Zimbabwe, and to a 

lesser extent in other parts of southern Africa. However, even with plants being frequently 

mentioned in ethnographic material collected from San and hunter-gatherer societies in South 

Africa, Botswana, Lesotho, and Namibia, there has been a lack of engagement with plants in 

rock art research. In order to redress this I conduct the first thorough review to date of botanical 

forms and species in the rock art of Zimbabwe. This is achieved by a thorough analysis of 

literature on plants in southern African rock art, followed by the classifications and typologies 

used to identify and categorise botanical motifs recorded in Zimbabwe. These are informed by 

an extensive examination of ethnographic material concerning plants, and analysed through the 

lenses of the shamanistic model and the ‘New Animisms’. Botanical motifs are modelled on 

natural and supernatural plants including trees, tubers, fruit, leaves, and phytomorphs (plant-

human/animal amalgamations). They are found in numerous contexts, frequently painted with 

specific motif types such as termites and formlings, antelope (tsessebe, kudu, roan and sable), 

birds (swifts and potentially red-billed queleas), ‘non-real’ creatures, and humans. These 

patterns of co-occurrence, I suggest, illustrate numerous and varied associations between these 

motif subjects and plants. I focus on five central themes: trees and their ability to transgress the 

boundary between the physical and spirit worlds; roots and tubers as potent underground 

metaphors; as well as rain; femaleness; and hunter-gatherer animistic ontologies relating to 

botanical forms with personhood and as a means to mediate relationships between human and 

other-than-human persons. In this way I hope to offer a more accurate understanding of the 

lives and beliefs of the peoples behind this treasure-trove of rock art. 

  



iv 
 

ACKNOWLEDGEMENTS 
 

First, I would like to thank my supervisors, Dr Sam Challis and Dr Jeremy Hollmann, for their 

unwavering support, guidance, and enthusiasm. They both went above and beyond. I am 

grateful to the National Research Foundation for their funding through Dr Challis, without 

which this dissertation would not have been possible; and for all of photographs and knowledge 

of the sites with which Dr Hollmann provided me, and for bringing this topic to my attention. 

Second, I express my gratitude all those whom helped me with my archival work and research, 

specifically Azizo da Fonseca (SARADA), Dr Richard Kuba (Frobenius Institute), and Dr 

Helen Anderson (British Museum), as well as Steve Tatum, Ian Withnall, Nicole Calame-

Darbellay for providing me with vast amounts of photographs. 

 

Finally, to all my friends and family who helped me during the process with specific mention 

to my father, Etienne van den Heever, Amy Leibbrandt, and Kiah Johnson who read through 

the roughest of rough drafts and provided insightful advice in the editing process. 

 

  



v 
 

I dedicate this to the plants and all the things that grow.  



vi 
 

TABLE OF CONTENTS 
BOTANICAL MOTIFS IN THE ROCK ART OF ZIMBABWE ..................................................... i 

DECLARATION ................................................................................................................................. ii 

ABSTRACT .......................................................................................................................................... iii 

ACKNOWLEDGEMENTS ................................................................................................................ iv 

LIST OF FIGURES ............................................................................................................................. ix 

LIST OF TABLES ..............................................................................................................................xiv 

GLOSSARY, ORTHOGRAPHY, AND CONVENTIONS ............................................................. xv 

CHAPTER 1 – INTRODUCTION ...................................................................................................... 1 

CHAPTER 2 – A REVIEW OF LITERATURE AND THEORY .................................................... 3 

2.1 – A Brief Overview of the Rock Art of Zimbabwe ................................................................... 3 

2.1.1 – Early interpretations of botanical motifs ........................................................................... 8 

2.1.2 – Early Interpretations of associations ................................................................................. 9 

2.2 – Searching for Meaning in San Rock Art .............................................................................. 10 

2.2.1 – Theoretical framework ..................................................................................................... 12 

2.2.2 – Meaning in the rock art of Zimbabwe ............................................................................. 15 

2.3 – Trees and Arboreal Forms .................................................................................................... 16 

2.3.1 – Understanding trees in their own right ........................................................................... 17 

2.3.2 – Understanding the meanings of ‘unidentifiable’ trees’ through association ................ 18 

2.4 – Geophytes, Leaves, Flowers, Fruits & Seeds ....................................................................... 20 

2.5 – Plant Charms .......................................................................................................................... 23 

2.6 – Phytanthropes......................................................................................................................... 23 

CHAPTER 3 – METHODS ................................................................................................................ 24 

3.1 – Source Material ...................................................................................................................... 25 

3.1.1 – Historical copies ............................................................................................................... 25 

3.1.2 – Photographs ..................................................................................................................... 27 

3.1.3 – Digital redrawings ............................................................................................................ 28 

3.2 – Morphology and Natural Models of Botanical Motifs ........................................................ 28 

3.3 – Typology and Identification .................................................................................................. 29 

3.4 – Analysis of Botanical Motifs in their Painted Contexts and Associated Motifs ............... 32 

3.5 – Quantitative Analysis ............................................................................................................. 33 

3.6 – Limitations .............................................................................................................................. 35 

3.7 – Analysis of San Symbolic and Idiomatic Systems, Ontologies, Rituals, and Beliefs 

Pertaining to Plants. ........................................................................................................................ 35 

CHAPTER 4 – ETHNOGRAPHIC CONTEXT FOR BOTANICAL MOTIFS ........................... 37 

4.1 – Trees in San Mythology and Cosmology .............................................................................. 39 



vii 
 

4.2 – The Significance of Other Plant Forms and Parts for the San .......................................... 41 

4.2.1 – Underground storage organs ........................................................................................... 41 

4.2.2 – Fruit and seeds ................................................................................................................. 48 

4.2.3 – Plant sap and gum ............................................................................................................ 50 

4.2.4 – Leaves ............................................................................................................................... 53 

4.2.5 – Phytanthropes ................................................................................................................... 55 

CHAPTER 5 – BOTANICAL MOTIFS AND THEIR NATURAL MODELS ............................. 58 

5.2 - Morphology of painted trees and plants ............................................................................... 59 

5.2.1 – Trees and arboreal forms ................................................................................................. 59 

5.2.2 – Underground Storage Organs ......................................................................................... 87 

5.2.3 – Leaves ............................................................................................................................... 94 

5.2.4 – Fruit .................................................................................................................................. 99 

5.2.5 – Mushrooms ..................................................................................................................... 101 

5.2.6 – Flowers ........................................................................................................................... 105 

5.2.7 – Phytomorphs ................................................................................................................... 106 

CHAPTER 6 – PAINTED CONTEXTS AND ASSOCIATIONS ................................................ 112 

6.1 – Rain ....................................................................................................................................... 112 

6.2 – Women and ‘femaleness’ ..................................................................................................... 113 

6.3 – Ethology and Natural Modelling ........................................................................................ 115 

6.4 – Termites and termite mounds ............................................................................................. 116 

6.5 – Antelope ................................................................................................................................ 121 

6.5.1 – Tsessebe .......................................................................................................................... 121 

6.5.2 – Roan and sable ............................................................................................................... 128 

6.5.3 – Kudu ............................................................................................................................... 134 

6.6 – Giraffe ................................................................................................................................... 136 

6.7 – Birds ...................................................................................................................................... 139 

6.7.1 – Swifts and swallows ........................................................................................................ 142 

6.7.2 – Red-billed queleas .......................................................................................................... 145 

6.8 – Rain and Rain-animals ........................................................................................................ 146 

6.8.1 – Rain-quadrupeds (cows/bulls) ....................................................................................... 148 

6.8.2 – Rain-snakes .................................................................................................................... 150 

6.9 – Flecks and dots ..................................................................................................................... 154 

6.10 – Humans ............................................................................................................................... 162 

6.10.1 – Women .......................................................................................................................... 162 

6.10.2 – Men ............................................................................................................................... 163 

 



viii 
 

CHAPTER 7 – QUANTITATIVE DATA ...................................................................................... 166 

7.1 – Distribution of botanical motifs .......................................................................................... 167 

7.1.1 – Distribution of trees and arboreal motifs ...................................................................... 168 

7.1.2 – Underground Storage Organs (USOs) .......................................................................... 171 

7.2 – Plants and their associated motifs ...................................................................................... 172 

CHAPTER 8 – DISCUSSION AND CONCLUDING REMARKS .............................................. 174 

8.1 – Plants as a means for travel ................................................................................................ 174 

8.2 – Plants as metaphors for underground ............................................................................... 175 

8.3 – Rain and plants .................................................................................................................... 175 

8.3.1 – Flecks as graphic representations of rain ..................................................................... 176 

8.3.2 – Women and rainmaking ................................................................................................ 177 

8.3.3 – Tsessebe and rain ........................................................................................................... 178 

8.4 – Plant-people .......................................................................................................................... 178 

8.6 – Avenues for future research ................................................................................................ 180 

CHAPTER 9 – CONCLUSION ....................................................................................................... 181 

REFERENCE LIST .......................................................................................................................... 182 

IMAGE ATTRIBUTIONS ............................................................................................................. 2099 

APPENDIX A .................................................................................................................................. 2122 

Harare ............................................................................................................................................ 212 

Manicaland .................................................................................................................................... 218 

Mashonaland Central ................................................................................................................... 228 

Mashonaland East ......................................................................................................................... 238 

Mashonaland West ........................................................................................................................ 284 

Masvingo ........................................................................................................................................ 289 

Matabeleland South ...................................................................................................................... 297 

Zimbabwe (Exact locations not specified from sources) ........................................................... 316 

APPENDIX B .................................................................................................................................. 3199 

 

  



ix 
 

LIST OF FIGURES 

 FIG. 2.1. Schematic diagram summarizes the perceptual interconnections between 

various natural and non-natural phenomena with a notional tree as the central “bridging 

conceptual pillar” in hunter-gatherer cosmology (Mguni 2002)………….…………...5 

 FIG. 2.2. Depictions of human figures interacting with botanical motifs……………22 

 FIG. 4.1. The collection and preparation of Asparagus exuvialis additive at Eagle Pos. 

………………………………………………………………………………………..44 

 FIG. 4.2. Preparing arrow poison at Dou Pos. ………………………………………45 

 FIG. 4.3. Tapping and collecting sap from palm trees. ……………………………….53 

 FIG. 5.1. Breakdown of botanical motif classification………………………………58 

 FIG. 5.2. Hunter-gatherer rock paintings of bare trees in four different contexts..……61 

 FIG. 5.3. The roots of bare trees connected to a formling with animals…………..…61 

 FIG. 5.4. The roots of bare trees creating ‘groundlines’…………...…………………62 

 FIG. 5.5. A bare tree near Nanke 1 (Matabeleland South)………..…………………63 

 FIG. 5.6. Female kudu and a bare tree (Shumbashaba 1, Matabeleland South. Mguni 

2009: 141 figure 2) ……………………………...……………………………………64 

 FIG. 5.7. Two female kudu browsing on a bare tree (Amadzimba 1, Matabeleland 

South. Mguni 2009: 141 figure 2). ………………………………………...…………64 

 FIG. 5.8. A rain-snake and a bare tree, Gulubahwe 1 (Matabeleland South. RARI)…65 

 FIG. 5.9. Human figures climbing bare trees or animals associated juxtaposed with 

them………………………………………..…………………………………………66 

 FIG. 5.10. A human figure climbing a line-like tree…………………….……………67 

 FIG. 5.11. Tree motifs with solid canopies……………………………..……………69 

 FIG. 5.12. Trees with lobed canopies…………………………………...……………70 

 FIG. 5.13. Trees motifs with fronds. …………………………………………………71 

 FIG. 5.14. Palm morphology diagrams. ………………………………..……………72 

 FIG. 5.15. Hyphaene petersiana ………………………………………………..……73 

 FIG. 5.16. Borassus aethiopum………………………………………………………73 

 FIG. 5.17. Examples of fan-palm motifs ………………………………..……………74 

 FIG. 5.18. Feather-palm motifs………………………………………………………77 

 FIG. 5.19. Paintings interpreted as the common date palm Phoenix dactylifera, 

Libya………………………………………………………………………………….78 



x 
 

 FIG. 5.20. Painting of a tree with fan-shaped fronds tentatively identified as Hyphaene 

petersiana……………………………………………….……………………………80 

 FIG. 5.21. Two copies of a tree motif at Mucheka 1 (Mashonaland East)…………..81 

 FIG. 5.22. Five examples of trees with fruit……………………………….…………83 

 FIG. 5.23. A human figure with a distended stomach eating fruits from a Ficus spp. 

tree……………………………………………………………………………………84 

 FIG. 5.24. Uapaca Kirkiana fruit growing directly from the branches………………85 

 FIG. 5.25. Tree motifs at Ngomakurira 1 (Mashonaland East)………………..……..86 

 FIG. 5.26. Flowering tree with bees (Efifi 1, Matabeleland South) ………..…………87 

 FIG. 5.27. Root motifs………..………………………………………………………88 

 FIG. 5.28. Root and bulb/corm motifs (Thetford, Mashonaland Central)…………….89 

 FIG. 5.29. USOs at Surtic Farm (Mashonaland Central)…………………………….89 

 FIG. 5.30. Bulb and corm motifs………………..……………………………………91 

 FIG. 5.31. Human figures interacting with USOs………………...………………..…92 

 FIG. 5.32. Mucheka 1 (Mashonaland East)………………………..…………………93 

 FIG. 5.33. Two humans holding fan-palm fronds at Gwangwadza 1, Manicaland…..94 

 FIG. 5.34. Sindjanja (Mashonaland East)……...……………………………………..96 

 FIG. 5.35. Two examples of female figures with distended stomachs and oval 

‘charms’…..…………………………………………………………………………..97 

 FIG. 5.36. Fan ‘charms’…….………………………………………………………..98 

 FIG. 5.37. Gourd-like motifs with white dots in a frieze with various gathering bags 

and baskets (Diana’s Vow 1, Makoni, Manicaland)……….…….…………………..99 

 FIG. 5.38. Men carry hunting equipment and, possibly, gourds (Chehara, Masvingo) 

………………………………………………………………………………………100  

 FIG. 5.39. Possible depictions of fruit and a fantastical creature (Ngomakurira 1, 

Mashonaland East)………….………………………………………………………101  

 FIG. 5.40. Schematic cross-section of a termite mound……………………….……102 

 FIG. 5.41. Mushroom motifs at varying stages of maturation, as indicated by the 

difference in cap shape (Gambarimwe 1, Mashonaland East.)………………………103 

 FIG. 5.42. Mushrooms held by human and therianthropic figures…………………104 

 FIG. 5.43. Flower motifs. …………………………………..………………………105 

 FIG. 5.44. Three phytanthrope figures isolated from the upper section of a panel at 

Thetford with numerous roots, and tubers (Thetford, Mashonaland Central)…..…..107 



xi 
 

 FIG. 5.45. Geophytic phytanthropes. ………………………………………………108 

 FIG 5.46. Human figure with plant-like projections emanating from his head (Murehwa 

1, Mashonaland East).……………………………………………………….………109 

 FIG. 5.47. A procession of human figures with plant-like projections emanating from 

their head (Murehwa 1, Mashonaland East)…………………………………………109 

 FIG. 5.48. A creature with elephant, ant bear, and plant features (Ngomakurira 1, 

Mashonaland East)………….………………………………………………………110 

 FIG. 5.49. A line of human figures leading to a creature with elephant, ant bear, and 

plant features (Ngomakurira 1, Mashonaland East)…………………………………111 

 FIG. 6.1. Schematic diagram summarizes the perceptual interconnections between 

various natural and non-natural phenomena with a notional tree as the central “bridging 

conceptual pillar” in hunter-gatherer cosmology……………………………………118 

 FIG. 6.2. Man holding up plant material at the opening of a formling, out of which 

winged insects fly (Toghwana Dam 1, Matabeleland South)……………………..…119 

 FIG. 6.3. A small section at the bottom of the large panel at Nanke 1 (Matabeleland 

South) ………………………………………………….……………………………120 

 Fig. 6.4. A) kneeling figures holds out a fan ‘charm’ (Amadzimba 1, Matabeleland 

South)…………………………………….…………………………………………121 

 FIG. 6.5. Kneeling behaviour in tsessebe and analogous postures in ritual specialists 

when in altered states of consciousness…………………………..…………………122 

 FIG. 6.6. Tsessebe and arboreal motifs…………………………..…………………124 

 FIG. 6.7. A tsessebe therianthrope reaches out an arm to a bare tree (Murombedzi, 

Mashonaland West)………...………………….……………………………………125 

 FIG. 6.8. A panel USOs, tsessebe, roan, sable, warthogs, and women (Thetford, 

Mashonaland Central. Redrawn from image courtesy of J. Hollmann)……………..126 

 FIG. 6.9. Tsessebe and formlings. …………………….……………………………127 

 Fig. 6.10. Sable and Roan……………………………………………...……………128 

 FIG. 6.11 Upside-down sable, formlings and baboons (Zombepata, Mashonaland 

Central) ……………………..………………………………………………………129 

 FIG. 6.12. Sable, roan, USOs, and other underground pictorial tropes…………….130 

 FIG. 6.13. Antelope, trees and formlings (Nswatugi 1, Matabeleland South)…..…131 

 FIG. 6.14. Sable standing above an ant bear burrow (Cairnsmore Ranch, Mashonaland 

Central)…...…………………………………………………………………………131 



xii 
 

 FIG. 6.15. Sable therianthrope (Diana’s Vow 1, Manicaland)…...…………………132 

 FIG. 6.16. A man touching the horn of a sable (Ngomakurira 1, Mashonaland 

East)…………………………………………………………………………………133 

 FIG. 6.17. Kudu and trees…………………………...………………………………135 

 FIG. 6.18. Kudu bulls painted with various USOs and warthogs or bush pigs (Surtic 

Farm, Mashonaland Central) ………………….……………………………………135 

 Fig. 6.19. Historical copy of a section of the panel of Nanke 1 (Matabeleland South) 

………………………………………………………………………………………136 

 FIG. 6.20. Frieze at Gulumbezhe 1 with bare trees, giraffe, formlings, various antelope 

and a possible rain-animal (Matabeleland South)…..………………………………137 

 FIG. 6.21. Giraffes and arboreal motifs………………………………………….…138 

 FIG. 6.22. Rain-snake, giraffe, and a tree, Gulubahwe 1 (Matabeleland South……139 

 FIG. 6.23. Avemorphs, trees, flecks, formlings, and mushrooms (Gambarimwe 1, 

Mashonaland East) …………………………………………………………………141 

 FIG. 6.24. Swallows flying into a nest………………………………………………142 

 FIG. 6.25. Winged forms and a formling (Zombepata, Mashonaland Central)…….143 

 FIG. 6.26. Lobed trees painted behind and below a distended female figure 

(Goromonzi, Mashonaland East) ………………………………...…………………145 

 FIG. 6.27. A swirling flocks of red-billed Q. quelea………………………..………146 

 FIG. 6.28. Sehonghong redrawing Vinnicombe (RARI) ……………...……………147 

 FIG. 6.29. Tracing of the rain snake at Rain Snake Shelter, Lesotho…………….….148 

 FIG. 6.30. Rain-quadrupeds.………………...…………………...…………………150 

 FIG. 6.31. Rain-snake.………………… ……………………………………...……152 

 FIG. 6.32. Rain-snake comprised of flecks (Tsindi, Mashonaland East)……………154 

 FIG. 6.33. Flecks and trees. ……………………………………………………...…155 

 FIG. 6.34. Flecks forming snake-like patterns. …………………………………..…157 

 FIG. 6.35. Giraffe, trees, and flecks (Hazelside, Matobo, Matabeleland South)158–159 

 FIG. 6.36. A large man and two bare trees in a pool of flecks (Nanke 1, Matabeleland 

South) ……………………….………………………………………………………159 

 FIG. 6.37. Amadzimba 1 (Matabeleland South) ……………………………………160 

 FIG. 6.38. Flecks , tsessebe, and a formling (Gulumbezhe 1a, Matabeleland South)161  

 FIG. 6.39. Palms and flecks at Zombepata…………….……………………………161 



xiii 
 

 FIG. 6.40. Women and USOs.………………… …………………………………..162 

 FIG. 6.41. Three examples of human figures cutting into trees………..……………165 

 FIG. 7.1. A map of Zimbabwe, showing sites with botanical motifs and the extent of 

granite outcrops with surfaces suitable for painting…………………………………166 

 FIG. 7.2. Geographic distribution of various botanical motifs by province………....167 

 FIG. 7.3. Distribution of sites with botanical motifs…………………………………168 

 FIG. 7.4. Number of individual motifs of each tree type.……………….……………169 

 FIG. 7.5.  Geographic distribution of tree motifs by province…………...……………170 

 FIG. 7.6. Distribution of sites with tree motifs by province. …………………………170 

 FIG. 7.7. Underground storage organ motif site distribution in Zimbabwe by province. 

………………………………………………………………………………………171 

 FIG. 7.8. Sites with USOs in Zimbabwe…………………………………...………...172 

 FIG. 7.9. Percentage of the frequency of association between botanical motifs and 

motifs discussed in Chapter 6……………………………………..…………………173 

  



xiv 
 

LIST OF TABLES 

Table 1. Frequency of association between motifs discussed in Chapter 6 and botanical 

motifs…………………………………………………………………………………..…317 

  



xv 
 

GLOSSARY, ORTHOGRAPHY, AND CONVENTIONS 

 

Glossary 

Cauliflory – The production of flowers [and fruit] on the trunk and branches of trees rather than at the 

ends of twigs (Hickey & King 2000: 7. My parentheses) 

Fasciculated – arranged in clusters or bundles (Hickey & King 2000: 16)  

Fusiform – spindle-shaped (Hickey & King 2000: 17) 

Geophyte – An herbaceous plant that perennates by means of underground buds, e.g., bulbs, corms, etc. 

(Hickey & King 2000: 18) 

Imparipinnate – Pinnate, with a terminal leaflet (Hickey & King 2000: 22) 

Inflorescence – The arrangement of flowers on floral axis; a flower cluster (Hickey & King 2000: 22)  

Palmate – Divided to the base into separate leaflets, all the leaflets arising from the end of the leaf stalk 

(Hickey & King 2000: 30) 

Pellucid – not quite transparent but with some light passing through when held up to the light, e.g., of 

gland dots in the leaves of Rutaceae (Hickey & King 2000: 61) 

Pinnate – Having separate leaflets along each side of a common stalk (Hickey & King 2000: 32) 

Underground Storage Organ (USO) – includes bulbs, corms, rhizomes, tubers 

 

 

 

Orthography of ǀXam and Juǀ’hoãn words used in text (McGranaghan 2012: 377–378. For a 

more complete orthography see Hollmann 2004) 

ǀ - the dental click. A dental or alveolar affricate (fricative), produced by sucking the tip of the 

tongue back from the teeth, as in English ‘Tsk, tsk’ 

ǁ - the lateral click. A lateral affricate (fricative), produced with the tongue on the roof of the 

mouth, releasing air between the tongue and one or other cheek, approximating the sound made 

to urge on a horse. 



xvi 
 

ǃ - the palatal (cerebral, retroflex) click. An alveopalatal or palatal stop, produced by sharply 

pulling the tip of the tongue down from the hard palate, sounding much like an imitation of a 

cork pulled from a bottle. 

ǂ - the alveolar click. An alveolar stop, produced by sharply pulling the tongue away from the 

alveolar ridge, behind the teeth. 

ʘ - the bilabial click. A bilabial stop or affricate, produced by releasing air between the lips, 

somewhat like a ‘kissing’ noise but without rounded lips. 

X - is used to represent a guttural sound similar to the ch in ‘loch’, 

ʃ - sh 

 

Conventions used to cite the various Bleek & Lloyd archive 

(http://lloydbleekcollection.cs.uct.ac.za/) are the same as those laid out by Skinner (2017):  

‘Citations to Wilhelm Bleek’s and Lucy Lloyds’ original notebooks take the following 

form: LL.VIII.26.7826-7827. The first component of the entry represents the initials of 

the collector; either Wilhelm Bleek (WB) of Lucy Lloyd (LL). This is followed by a 

Roman numeral, which represents either a notebook number (in the case of Bleek’s 

entries), or refers to a certain narrator (in the case of Lloyd’s entries) in which case the 

next number corresponds to the notebook. The final part of the citation refers to the 

consecutively-numbered notebook pages, with ranges indicated as above. Reverso 

notes are indicated with an apostrophe (e.g. LL.VIII.10.6892’).’ 

All of these notebooks are online and reference the Digital Bleek & Lloyd: Digital Bleek and 

Lloyd n.d. Centre for Curating the Archive, Michaelis School of Fine Art, University of Cape 

Town. http://lloydbleekcollection.cs.uct.ac.za, site accessed 2019–2021. References for images 

from the Bleek and Lloyd Archive will be given as the either page number from Bleek and 

Lloyd 1911) or page link and image number from the online archive. 

Redrawings by author unless otherwise stated. All rights Reserved for images by original 

photographers.



1 
 

CHAPTER 1 – INTRODUCTION 
 

San and hunter-gatherer rock art1 is comprised of a multitude of meaningful motifs, some of 

which are well understood and others understudied. Rock art research endeavours to understand 

and interpret these motifs to gain a better understanding of the peoples and cultures that 

produced them. In order to interpret these with any degree of accuracy and in a meaningful 

way, interpretations must align with their indigenous knowledge systems and world-views. 

This is primarily informed by ethnographic material recorded in the Kalahari, Northern Cape 

Province in South Africa, and the southern Maloti-Drakensberg as an analogy which is largely 

accepted in the field.  

This dissertation concerns botanical motifs, image-types depicting plants or those with features 

of plants, in Zimbabwe. Early interpretations of plant motifs, conducted by colonialists, include 

paintings as decoration of rock shelters, paintings of landscapes, and art for art’s sake. These 

interpretations were based off of a ‘gaze-and-guess’ approach, which proved to be not only 

incorrect but also Eurocentric and denying any symbolic and nuanced meanings. Later 

interpretations by Peter Garlake (e.g., (1987a,b, 1990, 1995, 1997) and Nicholas Walker (1987, 

1996) moved away from this approach with success, but the scope of their research did not add 

much to understanding of plants and the roles they played in the image-makers’ society. 

Siyakha Mguni’s research on trees and formlings is the first to yield meaningful results on the 

topic by applying the concept of cable-like reasoning (Bernstein 1983; Wylie 1993) to the rock 

art contexts in Zimbabwe. He considers three separate ‘strands’ of evidence: a formal analysis 

of botanical and formling motifs, analysis of associated motifs, and the use of ethnographic 

material collected from San and other hunter-gatherer groups in southern Africa as analogies 

to inform his interpretations and typologies. This method proves to be extremely successful in 

the understanding of these understudied motifs from a region with a negligible amount of 

ethnographic material. His focus solely on trees and their ability to transcend the three-tiered 

San cosmos demonstrates how important the understanding of botanical motifs is, and 

highlights the need for further research on the subject of plant motifs and to increase the scope 

beyond just trees.  

                                                           
1 From here on rock art will only refer to rock paintings and not rock engravings 



2 
 

This dissertation aims to drastically increase our understanding of the botanical world for past 

hunter-gatherer societies in what is now Zimbabwe, following the methods used by Mguni and 

by creating the first detailed review of all plant types in Zimbabwean rock art. This is achieved 

by conducting a formal analysis of botanical forms based on likely emic typologies, informed 

by the corpus of botanical rock art in Zimbabwe and tested against the relevant ethnographic 

material, and by applying the same approach to the identification of other image-types 

repeatedly found in association with them. Determining associations in this manner has 

limitations in that it assumes a relationship between motifs because of patterns of co-occurrence 

and potential interaction of plants and these motifs. I do so understanding that these paintings 

likely were produced over a substantial time-frame with significant periods of time between 

painting episodes. I argue that the possibility of association is plausible, as it is likely that each 

placement of subsequent painting was often intentional even if they were painted centuries 

apart (Lewis-Williams 1998: 88). Although the bulk of this dissertation focuses on 

identification and classification, it is primarily used as a solid foundation for interpretation, 

providing a body of knowledge with which further research can take place, and bringing to the 

fore to usefulness of interpreting botanical motifs through the widely accepted shamanistic 

theory and New Animisms. 

 

 

  



3 
 

CHAPTER 2 – A REVIEW OF LITERATURE AND THEORY 

 

Hunter-gatherer rock art proliferates across almost the entire southern African subcontinent. 

The primary emphasis on botanical motifs (and secondary emphasis on associated animal 

motifs) has of necessity brought Zimbabwe into sharp focus as the ideal geographical location 

in which to explore the role played by the botanical world within the rich history, ethnography 

and cosmology of southern African hunter-gatherers.  

 

2.1 – A Brief Overview of the Rock Art of Zimbabwe 

 

Rock art is found throughout Zimbabwe. Paintings have been formally recorded since at least 

the 1890s and the discovery of new sites continues through to today (Nhamo 2018: 58, 65). 

The rock art has been recorded, studied, and interpreted through various lenses, ranging from 

a gaze-and-guess approach, to one informed by ethnographic information from hunter-

gatherers with similar world views. Below, I outline a brief history of rock art studies in the 

country, then focus on studies and interpretations of botanical motifs. 

Most of the early studies were empirical, with minimal attempt at an informed interpretation 

(Mguni 2005: 34). It is only in the past few decades that the rock art of Zimbabwe has been 

interpreted with the aid of ethnography collected from hunter-gatherer and San groups 

(Huffman 1983: 49; Garlake 1987a, 1990; Walker 1995, 1996; Eastwood 1999; Eastwood & 

Eastwood 2006; Hubbard & Mabrey 2007; Mguni 2002, 2004, 2005, 2006a,b, 2009, Pearce 

2009). It is now generally agreed by archaeologists that the art displays themes, motifs, and 

metaphors pertaining to the acquisition and use of supernatural potency (Huffman 1983: 49; 

Garlake 1987, 1990; cf. Bahn 2010). 

 

‘There is a significant range of differences between the paintings of Zimbabwe and 

South Africa though both undoubtedly belong to the same artistic tradition and 

derive from the same basic system of belief’ (Garlake 1990: 17). 

San and other hunter-gatherer groups are highly diverse, yet there are several commonalities 

or similarities in their practices, ontologies, and colloquial metaphors (Guenther 1989: 33; 



4 
 

Lewis-Williams 1981; Lewis-Williams & Biesele 1978). This does not mean that all the San 

have a homogeneous culture, but rather that there are strong overall commonalities, evidenced 

by both the ethnographic records and the rock art itself (e.g., Lewis-Williams & Biesele 1978; 

Lewis-Williams 1981; Hollmann 2007: 85). 

The most notable of these commonalities are the centrality of the Great Dance (trance 

dance), supernatural potency, tiered cosmos, creation mythology, the roles of ritual 

specialists, and the belief that personhood is not limited to humans. It is vital to understand 

these in order to understand and interpret rock art, especially in Zimbabwe, where there is 

no ethnographic evidence from the descendants of the painters. These complex and nuanced 

beliefs are briefly outlined below, and the concepts, and how they link to beliefs about plants, 

are discussed in depth in Chapter 4. 

Supernatural potency2 (or just potency see Low 2011: 297), is a powerful force, central to San 

beliefs, (Lewis-Williams 1981: 76–77; Katz 1982) and pervades every part of San culture and 

their daily lives (Low 2014: 164). Potency is used by ritual specialists to perform various 

supernatural tasks (healing the sick, changing the weather, and influencing the movements and 

behaviour game animals) on behalf of their community (Lewis-Williams 1992: 56–59; Lewis-

Williams & Pearce 2004b: 205–207; McGranaghan 2012: 454; McGranaghan & Challis 2016: 

590). Potency is concentrated in various substances (e.g., honey, as well as the fat, meat, and 

blood of specific animals), as well as medicine songs (of specific animals) and specific places 

in both the ‘physical’ world and the ‘spirit’ world (Mguni, 2006: 62). It is acquired and used 

during the trance dance (see Marshall 1999: 63–90; Katz 1982; Biesele 1993; Katz et al. 1997) 

in which ritual specialists and other initiates go into altered states of consciousness (Schapera 

1930: 180, 198; Barnard 1992: 57; Walker 1996: 66‒7; Lewis-Williams 1998: 92‒3). When in 

this state, a person ‘dies’ and leaves their body, entering the spirit-world in order to obtain 

potency from gods and supernatural beings (Katz 1976; Barnard 1979; Lewis-Williams & 

Pearce 2004a: 59). The spirit-world is located underground, underwater, and in the sky, but the 

San do not see these realms as distinct, as they are held to interdigitate (Lewis-Williams 1992: 

57). This tiered cosmos is graphically encapsulated by Siyakha Mguni (2006: 62; see figure 

2.1) to show how the spirit world can be entered via the trance dance, and through certain areas 

or features which act as portals, such as rock-faces, termite mounds, waterholes, and bodies of 

water.  

                                                           
2 Nǀom in Juǀ’hoãn, !gi in ǀXam: will be referred to as potency unless specifically referring to one group  



5 
 

 

FIG. 2.1. Schematic diagram summarizes the perceptual interconnections between various natural and non-natural 

phenomena with a notional tree as the central “bridging conceptual pillar” in hunter-gatherer cosmology (Mguni 

2006a: 66 figure 7). 

 

A widely-held belief amongst San groups is that creation occurred in two orders (Guenther 

1989: 31; Hewitt 2008: 48). During the first order all the people and animals were created, and 

they all lived in ambiguous forms where animals, people (Guenther 1989: 31; Keeney 2003: 

29), as well as plants (Mguni 2009: 145) appeared to be one, but could easily shift according 

to need. It was only after an event, the exact nature of which varies from group to group, that 

all things were settled into their current forms (LL.II.33.394-406, 416-428, 442-448, 466, 

475; Biesele 1975: 202‒204; Guenther 1989: 31‒33, 84‒85, 1999: 146, 152; Eastwood & 

Eastwood 2006: 93‒94; Mguni 2009: 145). It is because of this that the San believe animals 

and humans are in many ways equal, in that they all have some degree of personhood (Guenther 



6 
 

1989: 31–4). These beliefs extend to include natural forces such as rain as persons with agency 

(Low 2014: 171).  

 

The content of the rock art has many similarities to that of the rest of southern Africa, with 

human and animal figures forming the bulk of the corpus. Unlike much of South Africa, 

however, formling (motifs display elements and characteristics of termite mounds) and 

botanical motifs abound (Cooke 1969; Willcox 1984; Walker 1987, 1996; Garlake 1995; 

Mguni 2002: 1). Eland, the most frequently-painted animal in the rock art of the well-known 

Maloti-Drakensberg region (Lewis-Williams 1981) are far less common in Zimbabwean rock 

art (Nhamo 2014: 331), and animals such as kudu, giraffe, tsessebe, elephant, buffalo, 

sable/roan, and smaller antelope are far more common (Garlake 1995: 123–128; Walker 1996: 

89; Mguni 2002: 1). 

There are significantly fewer examples of domesticated animals such as fat-tailed sheep, dogs, 

or cattle, found in the hunter-gatherer rock art of Zimbabwe, when compared to that of South 

Africa (Walker 1996: 13). Nicholas Walker (1995: 14, 247, 255, 1996: 13) argues, therefore, 

that this means most of the art was created before the arrival of African farmers and pastoralists 

from approximately 2100 BP (Walker 1983: 88, 90). 

Peter Garlake (1987a: 190–191) noted that there were fewer depictions of nasal bleeding, 

therianthropes, and dancers in Zimbabwe, especially in the north of Zimbabwe, compared to 

South Africa. According to Royden Yates et al. (1985: 73–74) and Ghilraen Laue (2017: 182–

183) these are rare in South Africa outside of the Maloti-Drakensberg. He took this to mean 

that there were fewer depictions pertaining specifically to trance. However, this is quite 

possibly not because there are no trance metaphors, but rather that there are different 

conventions in the representation of altered states of consciousness (ASC) and out-of- body 

travel. 

Garlake (1987a: 191) also highlighted that the art of Zimbabwe is significantly older than the 

art of South Africa. Using indirect dates of painted spalls found in dateable deposits, Walker 

(1996: 13) surmises that these paintings were made starting from between 14 000 and 10 000 

BP in Pomongwe Cave (Walker 1996: 14), with most paintings dated between 10 000 to 8000 

BP (Walker 1996: 14). In comparison, rock art was still being created in the Maloti-

Drakensberg in the 19th century (McGranaghan et al. 2013). It has been shown that South 

African art in the sandstone of the Maloti-Drakensberg is as old as 2998 – 2381 cal. BP 



7 
 

(Bonneau et al. 2017; see also van der Merwe, Sealy and Yates 1987; Mazel & Watchman 

1997). The difference in ages is likely, at least in part, to be a function of the different materials 

of the rock faces in Zimbabwe (granite) (Cooke 1969: 27; Garlake 1995: 8; Walker 1996: 9), 

and South Africa (sandstone) which weather at significantly different rates (Yates & Manhire 

1991: 3; Garlake 1995: 18; Walker 1996: 9, 14; Meiklejohn et al. 2009: 973; Mol & Viles 

2010: 280). Rock art painted on portable stones (art mobilier) has been found on burial stones 

in South Africa dating to 6430 BP (Pta-4612) (Binneman & Hall 1993: 93–94; Pearce 2002: 

29) and on stone slabs in Namibia at Apollo 11 Cave (27 000 BP) (Rifkin et al. 2015: 115) 

The hunter-gatherer rock art found in Zimbabwe has a high level of diversity in both technique 

and content (Mguni 2006b: 538; Nhamo 2014: 49–82), with variations occurring temporally 

and regionally. Walker (1996: 31) and Mguni (2002: 1) argue that kudu, tsessebe, giraffe, 

formlings and plants are the most commonly depicted in the Matopos in Matabeleland South, 

and as one moves further north, elephant, buffalo, and waterbuck are found in greater numbers. 

This may be true for these animal motifs but myself (pers. obs. 2019), and Jeremy Hollmann 

(pers. comm. 2019) note that plant motifs are more common in the North-East of the country. 

According to Ancila Nhamo (2014: 354), formlings are rare in Central and Eastern Zimbabwe 

but occur more frequently in the north in Mashonaland. 

Most of the formal archaeological studies conducted in this area have focused on lithics, 

pottery, and other excavated material culture (Jones 1926; Schofield & Stapleton 1938; Cooke 

1969, 1983; Wilcox 1984; Walker 1987, 1996; Garlake 1995). Over 4000 rock art sites have 

been recorded across Zimbabwe, many of which were recorded during the colonial period by 

rock art enthusiasts and not trained archaeologists (Nhamo 2018: 65). It would be wasteful not 

to use the empirical data of rock art (historical copies, tracings, and other quantitative data) 

collected by these early writers because of their mostly erroneous interpretations. 

 

Early recording of botanical motifs 

 

The first Westerners to pay any attention to botanical motifs were Richard Hall (1912: 593–

596 in Mguni 2002: 31), Samuel Dornan (1925: 186), Leo Frobenius (1930: 338 in Mguni 

2009: 139), and Margaret Taylor (1927: 1085 in Mguni 2002: 153). These early writers 

interpreted the art as a recording of daily events, landscapes, plants used, or ‘art for art’s sake’ 

—the decoration of rock shelters (e.g., Breuil & van Riet Lowe 1944; Goodall 1959; Cooke 



8 
 

1959, 1969, 1971: 19; Woodhouse 1979). These interpretations are deeply Eurocentric, 

denying any deeper meaning or cultural significance attributable to what has since been shown 

to be a highly sophisticated, nuanced, and complex art tradition. It is unfortunate that 

uninformed approaches are still employed by some today (e.g., Henning & le Clus 2011: 4).  

Despite their erroneous interpretations of the content of the art, and the reasons for its creation, 

these early writers did make note of, and sometimes recorded, many sites; and the prevalence 

of botanical motifs became apparent. Frobenius (1930: 338, cited in Mguni 2009: 139) notes 

that in Zimbabwe, sites with paintings of trees appear almost as frequently as those with 

animals do in South Africa. However, further surveys have revealed that botanical motifs 

appear far less frequently, and that botanical figures make up a small percentage of the total 

number of figures across Zimbabwe (Breuil & van Riet Lowe 1944: 4; Garlake 1987a: 60; 

Mguni 2009: 139). Walker (1996: 80) found that only, 1.9% of the Motobo sites contained 

plants.  

 

2.1.1 – Early interpretations of botanical motifs 

 

Trees and arboreal forms (motifs or figures that incorporate elements of trees) are found in the 

highest concentration in Zimbabwe, to the extent that Frobenius (1929; 1931 in Mguni 2002: 

7) believed tree motifs, along with formlings, to be the defining feature of what was then 

referred to as Southern Rhodesian art. Botanical motifs are not limited to Zimbabwe, however. 

They are found in Namibia (Mason 1958, 357–368; Pager 1989: 109, 152, 250, 276; Lenssen-

Erz 2001: 96; Mguni 2009: 142), South Africa (van Riet Lowe 1949: 37; Vinnicombe 1976: 

280; Lewis-Williams 1980: 469; Mguni 2002: 155; Hollmann 2015: 11), and Lesotho (Orpen 

1874; Challis et al. 2013: 343), although they are few and far between. This scarcity of 

botanical motifs is not necessarily indicative of a lack of importance (Mguni 2002: 9), as plants 

feature in the ethnographic record from these regions and are discussed in the following 

chapters. 

Mguni (2002: 153) states that a proportion of the botanical motifs are obviously arboreal or 

tree-like. Their ease of identification as trees, coupled with their high frequency, is the reason 

that they have been the focus of almost all studies of these motifs, with other plant forms being 

largely ignored (e.g., Frobenius 1929, 1931; Taylor 1929; Cooke 1969, 1971; Rudner & Rudner 

1970; Woodhouse 1979; Walker 1996; Mguni 2009).  



9 
 

2.1.2 – Early Interpretations of associations  

 

‘I draw attention to the repeated association of formlings and botanical motifs. 

Their coexistence is crucial to understanding these motifs’ (Mguni 2002: 5).  

 

From the very first recordings it was evident that formling motifs and botanical motifs are 

linked (Frobenius 1931 in Mguni 2002: 7; Goodall 1959 in Mguni 2002: 4; Mguni 2002: 1, 

2006b: 583, 2009: 142, 2015: 156; Eastwood & Eastwood 2006: 128). In reviewing botanical 

motifs one must, to a certain extent, review formlings. The connections and associations 

between these motifs are clearly evident in both the frequency in which they are painted 

together, and in the interactions between them. This connection led early writers to interpret 

these motifs as a dyad. In this dissertation they are examined together. 

 

Formlings, interpreted by early researchers as rocks or kopjes, have previously been labelled 

as part of landscape paintings into which trees were also placed. This is because it was believed 

that the paintings were depictions of the natural world (e.g., Breuil & van Riet Lowe 1944: 4; 

Goodall 1959: 60–66; Cooke 1959: 42, 1969; Woodhouse 1979 in Mguni 2002: 18). Hall 

(1912: 594) and Henri Breuil (1966: 24) argued that since the painted plants in the Matopos 

did not look like any plant species found in the region that they must be depictions of a lush 

and fertile utopia (Mguni 2002: 32).  

 

Frobenius (1931 in Mguni 2002: 27) and Elizabeth Goodall (1959 in Mguni 2015: 24) believed 

formlings to be the tombs of ancient monarchs and interpreted the large therianthropic figure 

at Dianna’s Vow as an embalmed king. Following this supposition they interpreted some 

botanical motifs, which are often in association with formlings, as sources of tannins to embalm 

bodies (Frobenius 1931 in Mguni 2002: 32). There is only one example of archaeological 

evidence of hunter-gatherers potentially preserving their dead, with the use of Boophone 

disticha leaves, in Kouga Mountains, Eastern Cape, South Africa (Binneman 1999: 1–2). There 

is no ethnographic evidence of southern African hunter-gatherers preserving their dead, and 

still less evidence for ‘kings’.  

 

Eric Holm (1957: 68 in Mguni 2002: 32) interpreted formlings as succulents, not rocks—as 

was then the current theory. Cranmer Cooke (1971: 19 in Mguni 2002: 32), in his attempt at a 



10 
 

more pragmatic interpretation, argued that the botanical motifs were depictions of plants used 

by the people for their practical and economic importance, sustenance, or medicine. This is not 

necessarily the case, however. For example, we know from analyses of excavated faunal 

assemblages that the animals painted were not those most often utilised for food (Walker 1996: 

31).  

Hunter-gatherers showed a preference for hunting smaller game animals (Walker 1996: 99–8). 

The species depicted in the art are mainly larger herbivores, and to a lesser extent predators 

(Eastwood & Eastwood 2006: 98–102), and although eaten, do not align with the faunal 

assemblages (see Davies 1952; Cooke 1963; Deacon 1965: 199; Vinnicombe 1972: 194; 

Walker 1996: 31; Challis et al. 2007: 468). Therefore, it is clear that the depictions of animals 

in rock art are not a menu of what hunter-gatherers ate. This is quite possibly also the case for 

botanical motifs. Only in the late 1970s and early 1980s did archaeologists start to see the rock 

art as exemplifications of spiritual and religious beliefs—though Wilhelm Bleek (in Orpen 

1874: 11–13) had stated this as early as 1874. It was Patricia Vinnicombe (1976: 280) who first 

observed that the botanical motifs likely held a deeper spiritual or religious meaning. 

 

2.2 – Searching for Meaning in San Rock Art  

 

It has long been argued that, in order to gain any meaningful understanding of southern African 

rock art, one must abandon trying to interpret it from a Western point of view (Lewis-Williams 

1998: 87; Lewis-Williams & Pearce 2009: 42). The interpretation of the rock art must be 

informed by ethnographic material collected from southern African people who share similar 

lifeways and worldviews i.e., the San (Lewis-Williams 1998: 87; Lewis-Williams & Pearce 

2009: 42). This approach must be critically evaluated before it can be employed. I therefore 

give a brief outline of what has become known as the shamanistic model, as well the 

complimentary New Animisms before reviewing how the model has been used in the 

interpretation of the rock art of Zimbabwe, and more specifically the botanical motifs. 

It was only with the pivotal and pioneering work of Vinnicombe (e.g., 1976), and David Lewis-

Williams (e.g., 1981) that archaeologists began to examine emic, or insider, meanings of the 

art. Informed by the ethnographic accounts collected from various San groups such as the 

Juǀ’hoãnsi, Gǀwi, Naro, and the ǀXam, the art is seen as highly symbolic and fundamentally 

religious, with the religion being centred on the trance dance, or Great Dance, and the 



11 
 

acquisition of supernatural potency for healing and other tasks (Vinnicombe 1976; Lewis-

Williams 1980, 1981, 1998, 2002, 2015, inter alia; Huffman 1983; Lewis-Williams & Dowson 

1990, 1999; Garlake 1995; Walker 1996; Lewis-Williams & Pearce 2008). Supernatural 

potency is acquired by shamans in order to perform specialised duties on behalf of their group; 

namely healing the sick (Lewis-Williams 1992: 56–58; Hollmann 2002: 563), controlling the 

game (Hollmann 2002: 563; McGranaghan 2012: 454; McGranaghan & Challis 2016: 590), 

and controlling the weather (Dowson 1988; Lewis-Williams 1992: 59; Lewis-Williams & 

Pearce 2004b: 205–207; Skinner 2017: 25).  

Much like the rest of southern Africa, hunter-gatherer groups in Zimbabwe adapted to, 

assimilated into, or removed themselves from early ‘Iron Age’ or Bantu-speaking farmer 

society, and, later, European colonial and modern-day society (Jolly 1996: 287; see also: 

Gordon 2000: 49‒56; Köhler & Lewis 2002: 277; Sadr 2005: 206; Grauer 2007: 12), so that 

now ‘only snippets of their beliefs have survived’ (Walker 1996: 66). Although the 

ethnographic sources are temporally and spatially distinct, a common underlying core belief 

system has been shown to exist within these societies, which focuses on altered states of 

consciousness (Schapera 1930: 34; Garlake 1987: 178, 1990: 17; Yates et al. 1990: 55; Barnard 

1992: 57; Walker 1996: 66‒67; Lewis-Williams 1998: 92‒93; Eastwood 1999: 16; Mguni 

2006: 52, 2009: 143; Skinner 2017: 27). Nhamo (2007a) argues that because there is negligible 

ethnographic information from hunter-gatherers from Zimbabwe and that the rock art in in the 

country differs from that of South Africa in both content and conventions, the art of Zimbabwe 

is therefore distinct and must be interpreted in its own right. However, Nhamo (2007b: 13, 

2014: 67–68) does acknowledge that the use of some ethnographic material is unavoidable; 

that there are numerous similarities between the hunter-gatherer rock art across southern 

Africa, and that each links to the ethnographic material. The independent sources of 

ethnographic material each demonstrate striking similarities to the rock art in southern Africa. 

These similarities strengthen the use of these sources as analogies for the interpretation of 

Zimbabwean rock art. Although there are many similarities, she argues that the differences (in 

both content and style) are equally important, and that these changes are indicative of different 

cultures, and expressions, rather than just a change over time (Nhamo 2014). 

 

 



12 
 

2.2.1 – Theoretical framework 

 

The shamanistic model uses ethnographic material collected from San and hunter-gatherer 

societies in southern Africa as a means by which we can investigate and interpret rock art. 

There are correspondences between each of the sources of ethnography and, more importantly, 

between each source of ethnography and the rock art (Lewis-Williams & Biesele 1978: 130; 

Walker 1996: 69; Lewis-Williams 1998; Skinner 2017: 30). The contexts in which the 

ethnography was collected and the context of the people at the time gives a much more localised 

and specific view of those people (Low 2014: 173). Using all of these sources of ethnography 

together produces a much broader framework (Jolly 1983; Lewis-Williams 1998; Smith 2010: 

352; Skinner 2017: 130) with which to interpret to rock art and, more specifically, the botanical 

motifs of Zimbabwe. It must, however, be used with caution (Mguni 2009: 126). The validity 

of the ethnography and the context in which it was taken should be taken into account - 

although, by the very nature of San storytelling, there are many idiosyncrasies (Biesele 1975, 

1976, 1993; Guenther 1999: 227; Keeney 2003: 19, 24). These idiosyncrasies are the result of 

a culture that prioritises subjective physical experiences over exact objective narratives 

(Biesele 1975, 1976; Guenther 1989, 1999; Keeney 2003: 19). San people have been described 

as tolerant of ambiguity and ‘seem untroubled mentally and emotionally by such cosmological 

and logical incongruities as humans merging identities with the animals of myth and veld’ 

(Guenther 1999: 227 in Parkington 2003: 137). For example, the details of what is experienced 

by a person in trance may differ between groups or individuals, but nevertheless the core belief 

system remains. 

In light of the numerous similarities in beliefs and ontologies found in the ethnographic 

material and the rock art of South Africa, I have tested my data against the shamanistic model 

to understand the meanings of the botanical motifs in the rock art of Zimbabwe. The 

shamanistic approach is most succinctly explained by Lewis-Williams (1998: 87) thus: 

‘The making of San rock paintings was essentially (or principally) associated with a 

range of shamanistic beliefs, rituals and experiences and was situated within a tiered 

shamanistic cosmology and complex social relations. The images comprise symbols 

(or, more emically, concentrations) of supernatural potency (e.g., paintings of eland), 

images of trance dances, fragments' of trance dances (e.g., single figures in the arms-

back posture), processed' (recollected and formalised) visions (e.g., the capture of a 



13 
 

rain-animal), transformed shamans (including the so-called therianthropes), monsters 

and beings encountered in the spirit-world, activities performed in the spirit-world 

(e.g., fighting off malevolent spirits of the dead), and 'scenic' groups (loosely called 

'compositions) made by one or more painters, and complex groupings, including 

superimpositions, of many images that, in a range of ways, show the interdigitating of 

the spirit realm with the material world The spirit-world was, in some conceptual 

circumstances, believed to lie behind the walls of rock shelters.’ 

Below I outline the basic postures and representations of experiences of the humans in the 

trance dance based on ethnographic material collected from San groups in the Northern Cape 

and Kalahari, to identify them in the rock art. The Great Dance entails a procession of people 

who dance in a circle in short staggering rhythmic steps, usually around a fire. Outside this 

circle women sit nearby, and clap and sing complex, polyrhythmic medicine songs, which help 

activate the potency in the dancers and are themselves full of potency (Marshall 1969; 352; 

Lewis-Williams 1987: 166–167; Keeney 1999: 38, 41, 47, 60). The dancers have specific 

physical responses and experience various visual, auditory, and somatic hallucinations as a 

result of entering altered states of consciousness in this way. Common responses are abdominal 

pain (which causes many to bend over), shortness of breath, hyperventilating, sweating, nasal 

bleeding, trembling, and finally they collapse to their knees with their heads to the ground and 

sink into complete unconsciousness where they ‘die’ and can leave their body (Katz 1976; 

Barnard 1979; Huffman 1983: 49; Lewis-Williams 1987: 167; Walker 1996: 67). In this state, 

dancers can enter very deep trance states and experience paroxysms, which are sudden, violent 

and energetic outbursts (Marshall 1962, 1969: 360; Huffman 1983: 49). People adopt specific 

postures during the dance, which include bending over, holding their arms behind the back, 

placing their hands on their heads or hips, squatting, and kneeling on all fours (Woodhouse 

1971: 128–129; Huffman 1983: 50). Accompanied by these are sensory experiences which 

include auditory hallucinations: damped hearing, as if underwater (Lewis-Williams 1984, 

1988: 8; Walker 1996: 70), visual hallucinations: phosphenes, trails of light, and hallucinations 

(Lewis-Williams 1981, 1984, 1988: 8; Walker 1996: 70). Somatosensory hallucinations are 

also experienced, such as elongation of the body, weightlessness (as if flying or underwater), 

experiencing extreme heat or cold, boiling in the stomach, and tingling of the hair roots, which 

can make one feel as if they are growing hair (Lewis-Williams 1987: 172; Hollmann 2002: 

567). Many of these same behaviours are physiological responses are seen in the animals (most 

notably eland) depicted in the rock art in Southern Africa either naturally, or when they are 



14 
 

killed, especially via poison from poisoned arrows (Lewis-Williams 1981: 237, 1987: 172; 

Yates et al. 1985; Hollmann 2005b: 84). It must be noted that although death is a metaphor for 

entering trance, the San truly believe they die and leave their bodies when they enter a deep 

state of trance (Katz 1976: 144–145; Barnard 1979: 73, 75). Thus, when certain animals die, 

they are seen as analogous to humans going into trance (Lewis-Williams 1987: 172). The most 

notable and most studied is the eland in the rock art of the Maloti-Drakensberg (see 

Vinnicombe 1976; Lewis-Williams 1981: 91 for a more detailed discussion). 

 

New Animism 

 

The New Animisms constitute an anthropological movement focusing on beliefs in which both 

humans and non-humans (animals, plants, natural forces e.g., wind and rain) have agency and 

personhood, away from earlier views on animistic societies that were oversimplified and 

derogatory (Low 2014: 165, 169). The degree of attributed personhood is in constant flux, and 

the manner in which it is mediated varies over time and from group to group (Low 2014: 171). 

Robert Wallis (2013: 51–53, 59, 61) puts forth the idea of ‘rhizomic personhood’,  as a means 

to describe the central tenant of New Animism in that it is fundamentally a relational ontology, 

specifically where non-human-people are extant prior to human involvement and have their 

own relationships with one another (Cruikshank 2005: 15). Wallis (2013: 61) uses the growth 

patterns of rhizomic plants as an analogy for this web of relations where the roots of a plant 

spread out in all directions with nodes where new shoots form. These nodes and shoots 

represent various persons and their interconnectedness with some only connected via another, 

with no hierarchy or centre. 

In southern Africa, New Animisms can be complementary to the shamanistic approach 

(McGranaghan & Challis 2016: 580) and is used as a framework to understand the interactions 

between specific human and non-human entities in San mythology, and to investigate whether 

these are represented the rock art (Low 2014: 171). This framework, like the shamanistic 

approach, is based on ethnography and uses indigenous epistemologies. I take a specific 

approach, focusing on how humans interacted with animals and non-real creatures that are ‘of 

the rain’ and how, in certain instances, plants or plant charms were used to facilitate these 

interactions (McGranaghan & Challis 2016: 580). For example, in the ǀXam worldview rain is 

not just a natural phenomenon but is one manifestation of !Khwa, a powerful entity who is rain 



15 
 

and in some instances water, mist, or clouds. Specific animals and plants that are associated 

with the rain, but from an etic perspective are unrelated, are !Khwa-ka-tshweng, things of the 

rain (Challis et al. 2013: 346). ‘Tameness’ and behaving ‘nicely’ are desirable qualities in both 

human and non-human persons (McGranaghan & Challis 2016: 586) that result in more 

amicable relationships. Humans behave in the right manner so that non-human persons behave 

in the same fashion, and so will be easier to hunt, or will send rain (McGranaghan & Challis 

2016: 587).  

Plants were not only used to mediate these relationships but some were persons in and of 

themselves, whether this be a plant with agency (e.g., ʃo-/õä) (Hollmann 2004: 311; 

McGranaghan 2012: 138), mythical figures from Primal Times (e.g., ǀXue) (LL.XI & XII.5. 

9362-9377, 9379, 9380v, 9381-9391, 9391v, 9401v; LL.XI & XII.8.9608-9618, 9624-9626; 

Guenther 2019: 57), or people who transform into plants (Mguni 2002: 51, 2015: 40) which I 

shall argue is to take on characteristics of plants (phytanthropes).  

 

2.2.2 – Meaning in the rock art of Zimbabwe 

 

Siyakha Mguni’s interpretation uses the ethnographic material collected from many San 

communities in the Kalahari, Northern Cape, and Eastern Cape as a central pillar to the 

interpretation of botanical, and formling, motifs (Mguni 2002: 4, 2004: 183, 2006a: 58, b: 538, 

2009: 142). While this is by no means a new approach, he utilises the ethnographic record using 

‘cable-like reasoning’, which allows for detailed and nuanced interpretations of these motifs.  

Mguni (2002: 4) makes a distinction between ‘chain-like’ and ‘cable-like’ arguments in 

scientific reasoning, as set out by Bernstein (1983) and Wylie (1989, 1993). Chain-like 

arguments work on observations that are generalised to form a ‘link’, to which the next link in 

the argument is added. If one of these links is faulty, however, the entire argument fails. Cable-

like reasoning examines ‘discrete strands of evidence’ as interconnected and intertwined 

(Bernstein 1983; Wylie 1993; Mguni 2002: 4). If one strand fails but all the other ‘strands’ 

remain intact then the argument remains strong (Bernstein 1983; Wylie 1989; Mguni 2002: 4). 

Mguni uses three strands of evidence for his interpretations: formal analysis of the rock art 

images; analysis of associated motifs; and ethnographic material collected from various San, 

and wider, hunter-gatherer groups (Mguni 2002: 4). This approach has yielded encouraging 

results in the interpretation of the rock art of Zimbabwe. 



16 
 

 

Peter Garlake is the first archaeologist to focus on the botanical images in any meaningful way. 

He sees the art as depicting trance phenomena, and the acquisition and use of supernatural 

potency, with many fundamental similarities with the rock art of South Africa. However, he 

sees several significant differences between the art of Zimbabwe and South Africa especially 

with regard to the presence of plant forms and formlings (Garlake 1987: 190–191; see also 

Vinnicombe 1976: 280). 

Garlake (1987a, b) made a major contribution to our understanding of botanical motifs, but it 

is not without its flaws. He interprets the rock art, specifically formlings and flecks, as symbols 

of potency, but later suggests that the lack of literal trance-related imagery (e.g., nasal bleeding, 

or processions of dancers) is due to shamanism being only one element—and not a significant 

one—in the art (Garlake 1987a: 190–191, 1995: 166; Mguni 2002: 51). This seems highly 

unlikely, as potency and the Great Dance are an integral part of shamanistic beliefs of San and 

southern African hunter-gatherers and the paintings thereof (Biesele 1975: 5; Huffman 1983: 

49; Jolly 1983: 47, 75; Garlake 1987: 178; Lewis-Williams 1981, 1992; Guenther 1989; 

Marshall 1999; Mguni 2015: 142; Skinner 2017: 27). This lack of trance metaphors and 

symbolism is not due to it not being shamanistic (Skinner 2017: 27), but rather to different 

conventions in the representations of altered states of consciousness (ASC) and out-of-body 

travel. 

 

2.3 – Trees and Arboreal Forms 

 

Trees and other arboreal forms are by far the most common botanical motifs (Rudner & Rudner 

1970; Walker 1996: 26; Mguni 2002: 6, 2009: 139). Like most other botanical motifs, the trees 

are highly stylised and, in some cases, have extraneous features removed and relevant features 

highlighted, most likely in order to convey a particular meaning (Garlake 1995: 23; Mguni 

2002: 157, 2004: 189, 2005: 37). Features of importance were most often the trunks, roots and 

in some cases the leaves (Walker 1996: 26; Mguni 2002: 154, 2009: 145). In some cases, these 

features are highlighted by increasing their size, and this should be kept in mind if one is using 

scale to identify plants in the botanical motifs (Mguni 2002: 157); stems, leaves, and fruit are 

not always in natural proportion to one another. When leaves are depicted, they are often 



17 
 

formed by two lobes to depict what Taylor (1927: 1058 in Mguni 2009: 140) refers to as 

‘conventional trees’. 

Highly stylised arboreal forms have almost no specifically identifiable features (Walker 1996: 

32), and so their exact type (depending on the emic classifications system used by the artists) 

might not have been important. It is likely that identity (whether a particular plant, or a group 

of plants, as classified into indigenous categories), when shown, must have been of some 

import. It is possible that some of the botanical forms that are unidentifiable to us, due to the 

lack of definable features used for Western taxonomic classification, but would have been 

easily identifiable to the hunter-gatherers of the time (Heinz & Maguire 1974). The fact that 

one can paint generalised trunks, leaves and roots is a strength and not a weakness. Botanical 

motifs were painted with as much or as little detail as was required by the image-makers to 

convey a particular set of intended meanings (Mguni 2002: 165–6). 

 

2.3.1 – Understanding trees in their own right 

The symbolisms and meanings of tree motifs must be understood in their own right, as well as 

through their associations with other linked motifs (Mguni 2002: 52, 2015: 42). It is in this 

respect that both Garlake and Walker fall short in their attempts to understand tree motifs. 

David Lewis-Williams (1987: 166) proposed interpreting rock art by looking at it as analogous 

to a written language. If we are to understand or ‘read’ what a panel is saying then we must 

understand the meanings of each word (motif) before we look at the syntax, or associations 

between motifs and the rockface (Lewis-Williams 1987: 166; Lewis-Williams & Pearce 2009: 

43). This analogy has proven to be very useful because the context and associations of images 

can alter the meaning of each motif (just as they can with words), because they are polysemic 

symbols (symbols with multiple meanings), but it is crucial that we know what these meanings 

(informed by the relevant ethnography) are beforehand (Garlake 1990: 17; Lewis-Williams 

1987: 166; Lewis-Williams & Pearce 2009: 43). That is not to say that rock art is text with a 

linear narrative, but that motifs, much like words, have specific meanings, which are used either 

in isolation or in association to covey a specific set of meanings though the form of metaphors 

and idioms (Lewis-Williams 1998; cf. Tilley 1991). 

Trees thus have the ability to transcend physical and spiritual worlds, acting as a cosmological 

axis mundi. This is because their roots go underground and their trunks and branches into the 



18 
 

sky (Rappenglück 2018), both of which are worlds of spirit and the supernatural in the tiered 

cosmos of the San (e.g., Lewis-Williams 1994: 278, 1998: 87). The trunks and boughs are 

highlighted in the paintings, probably because these are the features that allow for travel on a 

vertical axis. All other detracting structural details of the tree, such as the smaller branches and 

leaves, are removed. The act of using trees for vertical travel features frequently in San 

cosmology (Biesele 1978: 929–933; Guenther 1999: 188) and is likely a metaphor for traveling 

either underground (through holes next to the roots) or into the sky, both common during out-

of-body experiences in trance (Orpen 1874: 8–10; Biesele 1978: 929–933; Whitley 1994: 24, 

26; Guenther 1999: 188; Mguni 2009: 145). Lateral roots of trees often indicate the ground 

level and act as another way to demonstrate what is underground (Mguni 2009: 140). The roots, 

trunks, and branches of these trees are stylised and contain few details or diagnostic features. 

Other botanical forms such as geophytes, fruits, seeds, and flowers are much less frequently 

depicted when compared to trees (Mguni 2002: 155, 2009: 140; e.g., Mason 1958: 364; Garlake 

1987: 41, 60; Pager 1989: 276). 

 

 

2.3.2 – Understanding the meanings of ‘unidentifiable’ trees’ through association 

 

There are several motifs that are often found in association with (i.e. motifs superimposed over, 

touching or in the same cluster as another) tree motifs and can aid in the interpretation of their 

meaning in that specific context. Trees are most often found with one or more image categories, 

i.e., formlings, flecks, human figures, kudu, tsessebe, and/or giraffe (Mguni 2002: 157). Walker 

(1996: 60) states that in the Matopos, arboreal motifs are most often found in association with 

images of birds. However, this does not seem to be the case, and his lack of published 

recordings of these panels makes it difficult to verify.  

Tree motifs are very often found within or surrounded by fields of flecks (Garlake 1995: 33‒

4, 98; Mguni 2002: 15, 2009: 142‒143), which often look very similar to the leaves of those 

trees (Garlake 1995: 99). They seem to swirl, and flow around the trees, especially the crowns 

(Garlake 1990: 21, 1995: 99; Mguni 2015: 38). The claims made by early writers that these 

flecks are literal representations of rivers or other features of the landscape are easily dismissed 

as they, like all other motifs, are highly symbolic and inherently shamanistic, and so have a 



19 
 

deeper meaning beyond literal depictions (Walker 1996: 34). To Garlake (1995: 105) these 

flecks are representations of a form of potency that is specific to plants. It is likely that they are 

representations of potency, which is often the case in hunter-gatherer rock art in South Africa 

(Dowson 1989: 92; Lewis-Williams & Pearce 2004a: 98), but he ignores a fundamental feature 

of potency—that it permeates plants, animals, and people alike and that ‘… the potency that 

permeates these subjects is undifferentiated.’ (Mguni 2015: 40; cf. Low 2011: 297; Hollmann 

& Lombard 2020). This argument also ignores the association between fleck and other motifs 

such as humans (Walker 1996: 34), formlings (Mguni 2002, 2004, 2005, 2006a,b, 2009, 2015) 

and other animals, which Garlake himself later notes (1995: 46, 135‒136). 

Chris Low (2011) suggests that potency is not supernatural, as it is part of everyday KhoeSãn 

life, and that it is not substance but rather a potential to do work with. From his work with 

numerous Khoe and San groups he argues that: 

‘The idiom of work is closely tied to ownership and to taking on the characteristics and 

powers of something, or mastership. KhoeSan healers, old and new, describe how they 

own and work with particular animals or plant medicines. Similarly /Xam folklore 

reveals shamans who were owners of springbok. A Hai//om healer described the 

Ju/’hoansi as the people who ‘own the lion’ because they work with the lion. Birds that 

are strong do strong things. By taking on the potency of a strong bird a person can 

become strong and achieve the same things as the bird; they own the bird or work with 

the bird. Birds that fly before the rain are described as working with the rain or being 

owned by the rain’ (Low 2011: 300). 

Walker (1996: 71) interprets the fork-like shape of some of the bare trees as linked to or 

representative of arrows, which are commonly forked in paintings. To him, the forked trees 

have more to do with potency, due to their supposed arrow symbolism, than with trees. 

The work of Siyakha Mguni (2002, 2004, 2005, 2006a,b, 2009) changed the way we understand 

Zimbabwe’s rock art. For instance, he showed that the ubiquitous formlings contained elements 

of termite nests in their structure and, informed by ethnographic material and natural modelling, 

that they were polysemic images relating to supernatural potency, creation, and rain. Termite 

nests and mounds have deep spiritual significance because they are seen as portals to the spirit-

world, much like waterholes and rockfaces (Lewis-Williams & Pearce 2004a: 119). Trees that 

grow out of these mounds connect them to the sky and so combine to form an axis mundi 

connecting sky, ground and underground (Mguni 2006a: 63, 67). Termites are a renowned 

source of fat, which in the San idiom translates as supernatural potency; termite mounds, 



20 
 

therefore are foci for this spiritual power (Jolly 1983: 75; Walker 1996: 72; Mguni 2004: 189, 

2006: 67). Arboreal forms are regularly depicted in relation to formlings in the Matobo and 

Mguni (2002: 5, 150, 154, 2004: 183, 2006a: 58,b: 538, 2009: 142) demonstrates a convincing 

connection between the two and their associated symbolism and meaning. 

 

2.4 – Geophytes, Leaves, Flowers, Fruits & Seeds 

 

Geophytes are plants with underground storage organs (USO) which include tubers, roots, 

rhizomes, and bulbs. I draw a distinction between USOs, and tree roots, solely for the purposes 

of classification. Tree roots are used as one means to illustrate the distinction between surface 

and underground and are used to travel through and across these realms (Mguni 2002: 154, 

156, 2009: 140). In paintings, these are often simple in design in order to highlight this feature 

(Mguni 2002: 154, 2009: 140). Geophytes that are not associated with trees appear in different 

contexts, with detailed morphological features marking them as distinct from tree roots. One 

exception to this is tree roots with morphological features similar to those of USOs, and which 

could potentially be diagnostic features. Geophytic storage organs are the second most 

commonly depicted botanical motif, yet they have not been discussed in any detail by 

researchers. Little mention is made except for their presence, frequency, and on occasion their 

associations. Walker (1996: 32) observed that bulbs are found painted more frequently in 

Mashonaland than in the Matopos. Geophytes and tree roots are both found underground and 

their presence in a panel acts as an idiom relating to the ability of plants to transcend the San’s 

tiered cosmos (Mguni 2009). 

Other plant forms, particularly geophytes, feature in the ethnographic record and the rock art. 

These plants are valued for their medicinal or spiritual properties, or because they contain 

supernatural potency (Marshall 1999: 52‒53, 149‒150, 166; Mitchell & Hudson 2004: 41). 

These plant forms appear more often in the ethnography than trees (Katz 1982: 52; Marshall 

1999: 52; Mitchell & Hudson 2004: 41‒43). This could be in part due to the greater number of 

and reliance on geophytic plants when compared to trees in the arid Kalahari biome (Cowling 

et al. 1999: 3; Procheş et al. 2006: 27). That is not to say that just because geophytes are 

mentioned more frequently than trees in the Kalahari ethnography, that they were more 

important. Even if they are more important to the Kalahari San groups (whose ethnographic 



21 
 

material is used to aid interpretation of Zimbabwean rock art) one cannot say paintings of 

geophytes are of more importance than trees. An example of this:  

‘Ju/’hoansi in respect of the zao (Terminalia sericea) and /ana (Acacia giraffae, 

now Vachellia giraffae) trees, which they believe to be capable of influencing the 

rain’ (Marshall 1999, cited in Mitchell & Hudson 2004: 43). 

Fruits or seed pods are rarely painted (Mason 1958: 364; Garlake 1987: 41, 60; Pager 1989: 

276; Mguni 2002: 155, 2009: 140), and flowers or buds even less frequently (Rudner & Rudner 

1970: 155). 

Geophytes and fruits, when painted, are commonly depicted being used by people, e.g., 

pounding roots (figures 2.2A & C), and eating fruit (Garlake 1995: 60; Hubbard 2013: 17). 

There are a small number of images of people eating plants. Figure 2.2D shows a human with 

a distended stomach, eating what is most likely the fruit of Ficus sur. Further, the distended 

stomach in this example is a depiction of being full of potency (Garlake 1995: 60; Hubbard 

2013: 17) which has been recorded ethnographically (Katz 1982: 51; Yates et al. 1985: 72; 

Garlake 1995: 89). This is likely the case as a distended stomach is a widely used convention, 

both seen in the rock art of southern Africa, and from ethnographic material, for being full of 

potency (Katz 1982: 51; Yates et al. 1985: 72; Garlake 1995: 87–89; Mguni 2004: 186). Ficus 

fruit contain no psychoactive compounds, highlighting the point that psycho-activity of a plant 

does not pertain to potency or ritual significance.



22 
 

 

 

FIG. 2.2. Depictions of human figures interacting with botanical motifs described in text. A) Mucheka 1 

(Mashonaland East. Garlake 1995: 53); B) Man holding up likely plant material at the opening of a formling 

(Toghwana Dam 1, Matabeleland South. Pager 1973 fig. 5 in Mguni 2015: 27 fig. 14); C) Ngomakurira 1 

(Mashonaland East. Redrawn from photograph courtesy of the Trustees of the British Museum. Copyright David 

Coulson/TARA. Edited: cropped); D) Goromonzi (Mashonaland East. After Garlake 1995: 60). 

A plant's power can come from its medical properties, and magico-religious (either containing 

supernatural potency or having other “non-real” or perceived) healing properties, as well as 

aiding novices to enter trance (Lee 1993: 115). A plant’s psychotropic qualities could be one 

reason that they had magical or supernatural qualities ascribed to them (Sobiecki 2002: 1; 

Stafford et al. 2009: 324), but as Mguni (2009: 142) rightly points out, without a species level 

identification, one cannot use possible psychotropic qualities as the sole reasons for the trees 

being painted.  

 



23 
 

2.5 – Plant Charms 

 

Unequivocal depictions of plants used as charms (plants or plant parts that are used for their 

magico-religious properties) are rare but their depiction may be inferred with a fair degree of 

confidence from their context (e.g., Orpen 1874: 10; Pager 1976: 46; Vinnicombe 1976 fig. 

239). These charms are most likely bundles of aromatic herbs (Lewis-Williams 1980; 

McGranaghan & Challis 2016: 594) used to placate rain animals, allowing them to be captured 

and led over the land to make rain (Lewis-Williams 1977: 166; Mguni 2002: 165; Lewis-

Williams & Pearce 2004b; McGranaghan & Challis 2016: 590), as well as for other purposes. 

Plant charms are depicted as ovoid leaf-like shapes, or fan-like shapes. In the Zimbabwean 

context these plant charms appear in a greater diversity of forms (discussed in detail in 

subsequent chapters) and are used to mediate the relationships between the typical rain-

animals, along with antelope, elephants, as well as trees. 

 

2.6 – Phytanthropes  

 

One of the most interesting and, to date, understudied subsets of botanical motifs are 

phytanthropes (from the Greek phytos - pertaining to plants, and anthropos- human). 

Phytanthropes are beings that combine the features of humans and plants, similar to the way 

that therianthropes combine features of humans and animals. Depictions of phytanthropes are 

rare, and in Africa appear to be unique to the rock art of Zimbabwe (van den Heever 2020: 48). 

Hubbard (2013: 16) only briefly mentions these “plant-like eidolons” noting their association 

with formlings and flecks.  

Botanical motifs are an understudied topic in rock art research in southern Africa despite their 

prevalence in various ethnographic sources. Their meanings and associations with other motifs 

in hunter-gatherer rock art has been studied with various degrees of accuracy. It is only by 

looking at the art through the lens of the shamanistic model, informed by relevant ethnographic 

material from southern African hunter-gatherer societies, as well as New Animism, that one 

can interpret them based on likely emic perspectives. This allows for a fuller and more 

meaningful interpretation of botanical motifs and can be used to complement our understanding 

of hunter-gatherer beliefs using the methods outlined below.  



24 
 

CHAPTER 3 – METHODS 

 

The principal method I use combines three lines of evidence. Mguni (2002, 2009) successfully 

uses this approach to elucidate on the meanings and roles of trees and formlings in Zimbabwean 

hunter-gatherer rock art. The first line of evidence is a formal study of botanical motifs using 

their morphology and natural models. The second is an analysis of botanical motifs in their 

painted contexts and their associations with other motifs. Finally, there is an analysis of San 

symbolic and idiomatic systems, ontologies, rituals, and beliefs pertaining to plants. 

The research for this project began with a review of published and unpublished literature, 

tracings of rock art, and the ontologies and folkloric traditions of southern African hunter-

gatherers, particularly those pertaining to plants. It became readily apparent that botanical 

motifs, in various form, are present in seven of the ten provinces of Zimbabwe and that large-

scale analyses are necessary to do justice to this understudied area of research.  

I examine a diverse collection of graphical material taken from a wide range of sources. Due 

to the scope of the research area all credible sources were consulted. I have trawled through 

online archives (SARADA, the Frobenius Institute, the British Museum, as well as Flickr) for 

photographs, tracings, redrawings, and historical copies, as well as photographs from Jeremy 

Hollmann, Siyakha Mguni, and amateur rock art enthusiasts. I have also located tracings and 

redrawings in the RARI archives as well as published material by Garlake (1987, 1990, 1995), 

Walker (1996), Cooke (1971), Goodall (1959), and Mguni (2001, 2009). 

For practical and ethical reasons this study relies exclusively on archival sources. Fieldwork 

across the entirety of Zimbabwe is impractical because of time constraints, logistical 

difficulties, inability to enter Zimbabwe because of border closure due to SARS COVID-19, 

and a lack of resources available in Zimbabwe. The latter factor ties into (and cannot be 

separated from) the ethical concerns. At the time this research commenced, the country was 

(and is) experiencing a severe economic crisis (hyper-inflation, and a lack of currency), 

resulting in shortages of food and fuel. This dire state of affairs coupled with the tumultuous 

political situation means that research on the scale that was envisaged is not currently feasible. 

I could not justify using what little there was of these scarce resources that are available in 

Zimbabwe. 



25 
 

Fortunately, the sheer amount of material available is more than sufficient for this research. I 

have amalgamated this material into a single database. To facilitate comparison of items in the 

database, material from disparate sources and in different formats has been converted into 

modern, and standardised formats (discussed in the following section). 

 

3.1 – Source Material 

 

Multiple sources of graphic material are used in order to gain a better understanding and create 

a more complete database. The sources used are photographs (film and digital), tracings, ink 

redrawings, and historical copies. Each source has its own set of pros and cons, which I discuss 

below. It should be noted that although the level of detail and accuracy of the recorded artworks 

is extremely important, this alone is not enough; the images must be interpreted within the 

context of San hunter-gatherer ontologies. 

Originally, I used historical copies, tracing, and photographs for sites but subsequently I 

realised that using historical copies and tracing as the sole record for a panel or site was not 

reliable. I have subsequently removed all sites with only historical copies or redrawings from 

my analyses. Sites with multiple copies from separate authors e.g., historical copy and a 

redrawing are used if these seemed to corroborate one another. Redrawings, even those from 

reputable authors (e.g., Mguni, Garlake, Walker) are not used when they are the only recording 

of a panel or site because redrawings tend to make images seem clearer or they leave out motifs 

that are not pertinent to the authors’ research. They are not best used for determining 

associations. I use redrawings that I have made from photographs, or those from authors at sites 

I have photographs of, throughout most of this dissertation, because they make the images and 

motifs clearer and more distinct. The photographs of these sites and those I used to make the 

redrawings are all included, along with historical copies and all images with botanical motifs, 

in Appendix A. 

 

3.1.1 – Historical copies 

 

Historical copies are physical graphic reproductions of rock art panels, made in the late 19th 

and 20th centuries, prior to the development of other methods. The media commonly used for 



26 
 

these copies were paper, watercolour, pencil, crayon, ink, and stencils. These media were used 

because at the time photography was expensive, and lacked the quality, detail, and colour (Guy 

& Wintjes 2009: 64) needed for rock art recording.  

These copies are useful as they create a record dating back close to a century. They enable us 

to see how the paintings might have been at that time, and they provide a useful record of how 

they have deteriorated or been damaged over time providing that the copyists did not 

themselves add (or omit) details in the rock art, as is sometimes the case.  

Copies are inherently subjective (more so than other recording methods such as photography 

and direct tracing) as they are not precise, faithful reproductions, especially with regards to 

scale and content. Very few of the copies were traced directly from the rock face (Keene 2011: 

17), and in some cases, copyists omitted motifs and ‘corrected’ details in these images to what 

they thought was right. Images that are far apart on a panel were recorded closer together, or 

figures from a panel were recorded separately, and details of the rock face (steps, cracks etc.) 

were left out. This can be seen in some of the copies even by reputable copyists such as Joachim 

Lutz and Elisabeth Goodall, who Michael Raath (1971: 1–6) and Garlake (1993: 15, 1997: 47) 

said had produced some of the most detailed and precise copies. Since copyists usually had 

little knowledge of San beliefs, mythology, and idioms, these omissions and corrections often 

altered the meaning of the art (Mguni 2002: 81). For these reasons I have not used unverified 

historical copies for my analyses of botanical motifs. 

  

3.1.2 - Direct tracing and ink redrawing 

 

Direct tracings are made using semi-opaque tracing film and light pencils. Images or entire 

panels are painstakingly traced, with their colour, fading, superpositioning, and damage 

(flaking, mineral wash, anthropogenic damage etc.) marked. The outlines of the images are 

traced, and their colours are labelled. This is a time-consuming process, and although it is more 

accurate than freehand copying it is nonetheless subjective (Hollmann 2018: 167). The copies 

are the basis for ink redrawings, using standard conventions of stippling to show colour. There 

are limitations to this method: the images are removed from their context on the rock face, and 

the outlines and distinctions between images are made clearer and more distinct than they are 

on the rock face. This method makes the lines sharper and clearer than those of the original 



27 
 

image (Guy & Wintjes 2009: 75) and are therefore useful for analysis (Le Quellec et al. 2015; 

Hollmann 2018: 161). Tracings and redrawings of both published and unpublished material 

were done by Mguni, Garlake, and Walker. 

 

3.1.2 – Photographs 

 

Photographs, both historical and modern, are used in this dissertation. Historical photographs 

were taken during the Afrika Expedition, led by Frobenius, in 1929–1930. The panels were 

recorded on 16mm film (Keene 2011: 17). The low resolution and lack of colour renders these 

photographs useful only for more general analysis and for comparison with the historical 

copies. 

Digital photographs are sourced from SARADA, the British Museum, the Frobenius Institute, 

as well as from individual researchers (Jeremy Hollmann, Siyakha Mguni, and Ancila Nhamo) 

as well as private photographers and rock art enthusiasts (Ian Withnall, Nicole Calame-

Darbellay, Lyle Tatum). High resolution (Raw, Tiff) digital photographs, from the sources 

mentioned above, are used for the documentation of rock art sites. This method of recording 

and digitally enhancing images has become the standard for rock art recording (e.g., Hollmann 

2018: 157). The use of image enhancing software and algorithms such as Adobe Photoshop, 

and D-Stretch (see http://www.dstretch.com/AlgorithmDescription.html), allow certain details 

that are now difficult to see with the naked eye to be made clear. D-stretch is a plugin of ImageJ, 

developed by J. Harman, and is used to enhance pigments, especially red and orange pigments, 

and this has become an indispensable tool to aid in the depiction of rock art (J. Harman 2005; 

Hollmann 2018: 159). The use of digital photographs has many benefits over physical tracing: 

it allows the recorder to initially record more images in the same period of time, images can be 

enhanced showing details that are now invisible to the naked eye, the recorder can zoom into 

and layer images, images can be enhanced and traced back in the studio reducing the amount 

of fieldwork time, it includes features of the rock face in high detail, and most importantly it 

negates any contact with (and possible damage to) the rock face (Hollmann 2018: 159). Digital 

recording and enhancements, like other methods of recording, are still subjective, due to the 

nature of focusing on certain aspects pertinent to a research question, and the subjective nature 

of choosing outlines (Le Quellec et al. 2015; Hollmann 2018: 161). Image enhancement 

algorithms are a tool that must be used in conjunction with the relevant ethnographic materials 



28 
 

as a means of interpretation. In order to avoid or mitigate any biases and errors, I have followed 

the method laid out by Le Quellec et al. (2015), which has been used successfully in the 

southern African context (e.g., Hollmann 2018). This method involves the following: 

1. An unenhanced version of the rock painting should be published alongside the 

enhancement or digital tracing 

2. Any additions or modifications to the image should be noted 

The procedures followed to create the enhanced image or tracing should be made explicit. 

 

3.1.3 – Digital redrawings 

 

The digital redrawings I produced are made in Adobe Photoshop CS4, using a Wacom Cintiq 

22HD digital tablet with an interactive pen display. The source materials are historical copies, 

tracings, ink redrawings, scanned film photographs, and digital photographs. Digital 

photographs are enhanced using D-Stretch and contrast settings in Photoshop. Each version of 

the image, and each redrawing colour (e.g., red, white, yellow etc.) is placed on a separate 

layer. Whichever version of the image makes a specific colour, or section, of the panel clearer, 

is used. 163 photographs were digitally redrawn. I use similar conventions to ink redrawings, 

but I use grayscale in place of stippling to depict colour for the majority of panels and figures, 

especially where the original had small dots and flecks, to avoid misrepresentation of these 

features.  

 

3.2 – Morphology and Natural Models of Botanical Motifs  

 

The first step I use to gain understanding of botanical motifs, and therefore the role of plants 

in various aspects of hunter-gatherer lives, is twofold: a formal (i.e., Western taxonomy, or 

predetermined categories) and an informal (likely emic categories) identification. Botanical 

motifs are painted with varying degrees of detail and embellishment (Mguni 2002: 153). I have 

therefore classified the botanical motifs into categories based on their morphological features. 

The categories are useful for classification of the motifs, although they are unlikely to reflect 

the exact categorisation used by the hunter-gatherers of Zimbabwe. The main types of botanical 

motifs identified are trees, underground storage organs (USOs, i.e., roots, tubers, bulbs etc.), 

flowers, fruits, leaves, mushrooms, and phytomorphs.  



29 
 

Rock art motifs are not literal representations, however; they are ‘real’ things or objects which 

are representative of, and relate to, a multiplicity of culturally specific concepts (Lewis-

Williams 1998: 89), which are modelled on natural forms (Pearce 2009: 338). The concept of 

natural modelling in hunter-gatherer societies is grounded in a shamanistic worldview that sees 

the natural world as a rich source of information and inspiration (Whitley 1994). Natural 

models are identified by their morphological features (Mguni 2004, 2006a: 53; Hollmann 

2005a). These features are often not ones that are most obvious or useful to those usi